20ICSMGE Vol 1 Keynotes
Uploaded by
limara6520ICSMGE Vol 1 Keynotes
Uploaded by
limara6520th International Conference on Soil Mechanics and
Geotechnical Engineering 2022 | 1-5 May 2022 | Sydney, Australia
Proceedings of 20th International Conference on
Soil Mechanics and Geotechnical Engineering -
State of the Art and Invited Lectures
Editors: Prof. Md Mizanur Rahman and Prof. Mark Jaksa
ISBN: 978-0-9946261-6-5
Hosted by the Australian Geomechanics Society and the International
Society for Soil Mechanics and Geotechnical Engineering
ICSGME 2022 Proceedings cover Book 2.indd 1 1/4/22 11:11 am
PROCEEDINGS OF TWENTIETH INTERNATIONAL CONFERENCE ON SOIL MECHANICS AND
GEOTECHNICAL ENGINEERING (ICSMGE 2022) – STATE OF THE ART AND INVITED LECTURES,
A GEOTECHNICAL DISCOVERY DOWN UNDER, SYDNEY, NEW SOUTH WALES, AUSTRALIA,
1-5 MAY 2022
Proceedings of 20th International Conference on
Soil Mechanics and Geotechnical Engineering –
State of the Art and Invited Lectures
Editors
Professor Md Mizanur Rahman
UniSA STEM (Science, Technology, Engineering and Mathematics)
University of South Australia, Adelaide, Australia
Professor Mark Jaksa
School of Civil, Environmental and Mining Engineering
University of Adelaide, Adelaide, Australia
© 2022 Australian Geomechanics Society, Sydney, Australia
All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a
retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording
or otherwise, without written prior permission from the publishers.
Although all care is taken to ensure integrity and the quality of this publication and the information herein, no
responsibility is assumed by the publisher nor the author for any damage to the property or persons as a result of
operation or use of this publication and/or the information contained herein.
Published by: Australian Geomechanics Society
PO Box 955, St Ives, NSW 2075, Australia
e-mail: [email protected]
www.australiangeomechanics.org
ISBN 978-0-9946261-6-5
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Preface
The Conference Organising Committee (COC) are delighted to have been granted the great privilege of
organising and hosting of the 20th International Conference on Soil Mechanics and Geotechnical
Engineering (20th ICSMGE), which was held in Sydney, Australia between May 1 and 5, 2022. This,
the first of three volumes, contains all 18 of the keynote papers presented at the conference.
First, is Prof. Antonio Gens’ 10th Terzaghi Oration entitled “Geotechnical engineering in a deltaic
environment”. As outlined on the ISSMGE’s website:
“The Terzaghi Oration was instituted to recognise a person of distinction and outstanding in his
or her contribution to geotechnical engineering, for services to the International Society, or to
geotechnical engineering and is normally presented at the International Conference. The Terzaghi
Oration is delivered at each ISSMGE International Conference as a tribute to Professor Karl
Terzaghi, first President of the International Society. The Terzaghi Oration normally covers case
histories, derived from professional activities and explores the dynamic interaction between
consulting work, teaching, research and publication. The orations exemplify Prof. Terzaghi's
intellectual approach to engineering and geology and to the observational method both for
improving design and for the advancement of knowledge.”
The COC is proud to have re-established State of the Art papers, which have long been a great source
of information and benchmarks for the profession. At the 20th ICSMGE, seven State of the Art papers
were presented under the broad disciplines of Transportation Geotechnics, Energy Geotechnics, Tailing
and Mine Wastes, Environmental Geotechnics, Unsaturated Soils, Earthquake Geotechnics and
Biogeotechnics.
The State of the Art papers are followed by a paper prepared by Profs. Peter Day and Jean Louis Briaud
on “Are We Overdesigning?”, which presents the results of an ISSMGE global survey of current
geotechnical design practices worldwide.
Finally, this volume incorporates eight invited papers involving Artificial Intelligence, Planetary
Geotechnics, Engineering Geology, Rock Mechanics, and two papers by two of the ISSMGE’s young
‘Bright Sparks.’
This volume would not have been possible but for the dedication, commitment and expertise of each of
the papers’ authors. We also offer our heartfelt gratitude to Robin Lehane, for curating the papers, and
to Major Graphics Printers, who assembled the proceedings.
We have no doubt that you’ll find these proceedings, and this volume in particular, a great source of
wisdom, knowledge and inspiration for many years to come.
Professor Md Mizanur Rahman
University of South Australia, Australia
Professor Mark Jaksa
University of Adelaide, Australia
III
IV
Table of contents
Preface III
Terzaghi Oration
10th Terzaghi Oration: Geotechnical engineering in a deltaic environment 3
A. Gens
State of the Art Papers 37
State of the Art 1: Transportation Geotechnics 39
State of the Art in Transportation Geotechnics 41
A. Gens, R. Kelly, B. Indraratna, W. Powrie, C. Zapata, Y. Kikuchi, E. Tutumluer and A.G. Correia
State of the Art 2: Energy Geotechnics 93
Energy geo-engineering 95
J.C. Santamarina, A. Aftab, D.N. Espinoza, M. Dusseault, A. Gens, H. Hoteit, S. Kim, J.Y. Lee, L. Lei,
G. Narsilio, J.-M. Pereira, M. Sanchez, K. Soga, M.V. Villar and M. Violay
State of the Art 3: Tailing and Mine Wastes 119
Geotechnics of mine tailings: a 2022 State of the Art 121
A. Fourie, R. Verdugo, A. Bjelkevik, L.A. Torres-Cruz and D. Znidarcic
State of the Art 4: Environmental Geotechnics 185
Protecting the environment from contamination with barrier systems: advances and challenges 187
R. K. Rowe, S. Jefferis, F.B. Abdelaal, A. Dominijanni, K. Legge, E.M. Palmeira, J. Scheirs, C.D. Shackelford,
J. Scalia IV, T.L.T. Zhan, C.H. Benson, A. Bouazza, G.D. Emidio, J.C. Evans, E. Fratalocchi, M.G. Gardoni,
T. Katsumi, M. Manassero and D. Parra
State of the Art 5: Unsaturated Soils 295
State of the Art Report. Advances in Unsaturated Soil Mechanics: Constitutive modelling, experimental 297
investigation, and field instrumentation
N. Khalili, E. Romero and F.A.M. Marinho
V
State of the Art 6: Earthquake Geotechnics 349
Nonlinear dynamic analyses involving soil liquefaction or cyclic softening 351
R. W. Boulanger
Nonlinear dynamic analyses for evaluating seismic ground deformation—Delayed flow failure 381
R. Harran, D. Terzis & L. Laloui
State of the Art 7: Biogeotechnics 403
State of the Art: MICP soil improvement and its application to liquefaction hazard mitigation 405
J.T. DeJong, M.G. Gomez, A.C.M. San Pablo, C.M.R. Graddy, D.C. Nelson, M. Lee, K. Ziotopoulou,
M.E. Kortbawi, B. Montoya and T.-H. Kwon
Corporate Associates Presidential Group 509
Are we overdesigning: Results of ISSMGE survey 511
P. Day and J.-L. Briaud
Invited Papers 531
Artificial Intelligence 533
Deep learning for rainfall prediction: data enhancement, forecasting, and estimation 535
L. Chen, J. Wang, J. Li, C.W.W. Ng and Y. Shen
Machine learning in geotechnical engineering: opportunities and applications 543
Z. Liu and S. Lacasse
Planetary Geotechnics 559
The first ever landing and experiments on the far side of Moon & Future geotechnical considerations 561
H. Liu, G. Xie and W. Zhang
The mechanical properties of the Martian soil at the InSight landing site 567
P. Delage, E. Marteau, M.P. Golombek, K. Hurst, W.B. Banerdt, S. Piqueux, S. Kedar, M. Lemmon,
N.R. Williams, N. Brinkman, P. Edme, J. Robertsson, D. Sollberger, S. Stähler, C. Schmelbach, P. Lognonné,
K. Onodera, C. Vrettos, V. Asan-Mangold, M. Grott, N. Mueller, N. Murdoch, T. Spohn, N. Warner,
R. Widmer-Schnidrig, B. Caicedo-Hormaza, J.P. Castillo-Betancourt, L. Lange and N. Verdier
Engineering Geology 583
The impacts of climate change on chains of geological hazards on the Tibetan Plateau and its margins 585
X. Fan
VI
Rock Mechanics 595
Application of satellite technology (GPS and SAR) for monitoring ground surface displacements in rock and 597
geotechnical engineering
N. Shimizu
Bright Spark Papers 617
Prediction of earthquake-induced ground motions using simulation methods and their validation against 619
observations
B.A. Bradley
Soft peaty clay stabilization in Sri Lanka: sustainability through utilizing secondary raw materials 625
A. Ranathunga and A. Kulathilaka
Author Index 631
VII
Terzaghi Oration
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
10th Terzaghi Oration
Geotechnical engineering in a deltaic environment
10e allocution Terzaghi
La géotechnique dans un environnement deltaïque
Antonio Gens
Department of Civil and Environmental Engineering, Technical University of Catalonia, Spain, [email protected]
ABSTRACT: Two case histories involving harbour works in a deltaic environment are described in the paper. The mechanisms
underlying delta formation and development are outlined first; they involve climate changes, geological processes and human effects.
The history of the Llobregat delta, the location of the case histories, is presented. The first case concerns the design and construction
of a vertical caisson breakwater. The incorporation of storm wave loading, the description of the geological setting and the bases for
stability evaluation are discussed. Special attention is given to the stability of the structure during the construction period when
stability issues are more critical. The use of the envisaged observational method was hampered by the limited monitoring data
available resulting in a modification of the planned approach. The second case refers to the reconstruction of the Prat quay after a
major failure caused by the flow liquefaction of the hydraulically-deposited backfill. Post-failure investigation and analyses identified
the need for reducing the liquefaction potential of the backfill and enhancing the undrained shear strength of the foundation clay;
preloading was the main ground improvement method selected. The observational method was used to steer the construction activities
to ensure an adequate margin of safety. The paper closes with some general considerations of the role of geology, observational
method and judgement in the cases reported.
RÉSUMÉ : L’article présente deux études de cas historiques impliquant des travaux portuaires dans un environnement deltaïque. On y
décrit en premier lieu les mécanismes conduisant à la formation et au développement d’un delta. Ils impliquent des changements
climatiques, des processus géologiques et les effets de l’activité humaine. Ensuite, l'histoire du delta du Llobregat, site des cas historiques
étudiés, y est présentée. Le premier cas concerne la conception et la construction d'une digue verticale en caissons. On y discute la
description du cadre géologique, l’incorporation du chargement dû à la houle de tempête et les bases de l'évaluation de la stabilité. Une
attention particulière est accordée à la stabilité de la structure pendant la période de construction lorsque les problèmes de stabilité sont
les plus critiques. L’utilisation de la méthode observationnelle avait été initialement envisagée mais l’approche a dû être modifiée en
raison de la quantité limitée des données d’auscultation. Le deuxième cas concerne la reconstruction du quai Prat après une défaillance
majeure provoquée par la liquéfaction statique du remblai déposé hydrauliquement. L'enquête et les analyses réalisées après la rupture
ont identifié la nécessité de réduire le potentiel de liquéfaction du remblai et d'améliorer la résistance au cisaillement non drainé de
l'argile de fondation ; le préchargement a été la principale méthode d'amélioration du sol qui a été retenue. La méthode observationnelle
a été par ailleurs utilisée pour guider les activités de construction afin d'assurer une marge de sécurité adéquate. L'article termine avec
quelques considérations générales sur le rôle de la géologie, de la méthode observationnelle et de l’utilisation du jugement expert dans
les cas rapportés.
KEYWORDS: deltas, breakwater, harbour quay, observational method, stability.
1 INTRODUCTION testing, field observations, analysis and (in Terzaghi’s words)
“intelligently digested experience” (Terzaghi 1953).
The ISSMGE’s guidelines say that “the Terzaghi Oration Many of these threads come together in what is generally
normally covers case histories, derived from professional known as the observational method, in many ways a natural
activities and explores the dynamic interaction between development in the presence of considerable uncertainty. The
consulting work, teaching, research and publication. The orations observational method was indeed regularly advocated and used
exemplify Prof. Terzaghi's intellectual approach to engineering by Terzaghi during his career (Terzaghi 1961). In some ways, it
and geology and to the observational method both for improving can be viewed as a form of the scientific method where
design and for the advancement of knowledge”. This statement assumptions are tested against observations and, as a result,
can be viewed as a perceptive summary of Karl Terzaghi’s modifications are introduced (in a theory or in a design) in
legacy and work, in which engineering geology and the response. Finally, a keystone of geotechnical engineering is the
observational method played an increasingly important part in exercise of judgement, the necessary ingredient to deal with
his long and successful career as professional engineer (Einstein limited information, significant uncertainty and the often
1991). competing demands of theory, experience, analysis and
Geotechnical engineering practice involves some distinctive observations (Peck 1980). Undoubtedly, the profession is now
features; a major one is the fact that it is generally performed with the beneficiary of tremendous advances in theory, site
a higher level of uncertainty compared to other branches of investigation techniques, laboratory equipment, field
engineering; a consequence of dealing with intrinsically variable instrumentation, ground improvement methods and capacity for
geological materials (on which information is always limited), analysis. These advances provide a much sounder platform upon
the complexity of their behaviour and the pervasive effects of which to exercise judgement; but, at the end of the day, the basic
pore water conditions (that are often subject to changes in time). underlying principles persist.
Therefore, it is not surprising that geotechnical problems are best A way to illustrate how all those aspects interact in practice is
tackled using a multi-pronged approach (Burland 2017) by way of presentation of documented case histories, an
encompassing geological understanding, site investigation, undertaking very much encouraged by Teraghi (1948) as a major
3
avenue for progress, as explicitly recognized in the ISSMGE’s From a stratigraphic point of view, the evolution of a delta by
Terzaghi Oration guidelines. This paper presents an account of the interplay of shoreline location and sedimentation changes is
the geotechnical aspects of some major harbour works built in a better understood by distinguishing four system tracts
deltaic environment. Deltas provide both an appealing and a (Catuneanu 2002), an extension of the initial threefold scheme of
challenging setting for the practice of geotechnical engineering; Posamentier and Vail (1988): Lowstand system tract (LST),
they are the intricate product of recent geological history, climate Transgressive system tract (TST), Highstand system tract (HST)
change and human activity. In this paper, general aspects of delta and Falling stage system (FSS) tract (Figure 2). The LST is the
development and an account of the history of the Llobregat delta package of sediments deposited at the minimum sea level in the
and of the resulting ground profile are presented first. Two case available accommodation space. The TST is deposited during a
histories are then described involving the construction of a transgression episode and is located between a transgressive
breakwater and the reconstruction of a harbour quay. In the case surface or sequence boundary (SB) and the maximum flooding
of the breakwater, the level of uncertainty is increased by limited surface (mfs). The HST forms during the late or a stillstand stage
knowledge of the magnitude and effects of storm wave loading
that incorporates a major random component. In the case of the The FSST only forms during a forced regression induced by the
the quay, the reconstruction was undertaken in the aftermath of a seaward movement of the shoreline caused by the falling sea
major failure brought about by an episode of flow (static) level, irrespective of the sediment supply (Hunt & Tucker 1992,
liquefaction, a phenomenon that impacts directly on the 1995).
vulnerability of the structure. In both cases, the observational
method lies at the core of the geotechnical approach adopted.
2 DEVELOPMENT OF A DELTAIC ENVIRONMENT
Deltas are dynamical coastal systems in the interface between
land and sea. Their development is controlled by eustasy (i.e. sea
level variations), tectonics and sediment flux (Jervey 1988,
Posamentier et al. 1988, Guillocheau et al. 1989, Homewood et
al., 1992, Rabineau 2001, Rabineau at al. 2005), as illustrated in
Figure 1. Sediment supply by a river or stream provides the basic
material for building-up of the delta, a material that in turn may
be eroded and transported away by waves, tides and/or sea
currents. Those processes are very much influenced by the Figure 2. System tracts in the development of a delta during a sea level
location of the contact between the sea and the land that, in turn, drop (top) and a sea level rise (bottom): LST, TST, HST, FSST. Dashed
depends on the variation of sea level (often associated with lines indicate eroded sediments. (modified from Rabineau 2001).
climate changes) and the vertical movements of the basement
Of special interest are the postglacial materials that were
(that can be due to tectonic events, isostasy, consolidation of
deposited after the last sea level minimum as they constitute the
sediments or other processes). Some of those phenomena may be
largest part of the present deltas (Figure 3). The sea level
influenced by human actions; for instance, sediment supply is
minimum, SLM, (corresponding to the Last Glacial Maximum,
positively affected by deforestation and negatively by dam
LGM) took place about 20,000 years ago and the sea level has
construction upstream. In a sense, deltas are geo-archives that
been increasing since that time although not at a steady rate. As
provide a record of the recent history of our planet and document
Figure 4 shows. the rate of seal level rise reached a value of 10-
the transition from environments dominated by natural processes
15 mm/year about 13,000 years ago but it slowed down to about
to environments controlled by human activities. The close
1 mm/year 6,500 years ago, approximately.
relationship of delta development with sea level changes and
human activities resonates with contemporary societal concerns.
Figure 3. Conceptual model of a delta configuration showing the
postglacial sediments (Hunt & Tucker, 1992).
The lowest post glacial materials sit on the erosional surface
Figure 1. Main controlling geological factors in the development of a (SB in Figure 3) formed during the last glacial maximum or
delta.
minimum sea level. This erosional surface overlies the FSST of
Deltas have always been very important in human history; the previous sea level drop. During this low sea-level conditions
usually they can support a productive agriculture and therefore (LST) coarse materials were generally deposited initially (fluvial
they can maintain a large population. It can be argued, for non-marine deposits) followed by more fine-grained materials
instance. that the Nile, Yangtze and Ganges/Brahmaputra deltas identified as a prodelta facies. Occasionally, coarse materials
played a key role in the origin and development of the Egyptian, appear again at a higher elevation indicating a changed
Chinese and Indian early civilizations, respectively. Also, deltas hydrological regime. Subsequently, the rapid sea level rise led to
often contain areas of high ecological value. From the marine transgression with the shore line retreating landwards.
geotechnical point of view, deltas may provide a challenging This movement caused erosion and reworking of the existing
environment; the soils are generally soft and recently deposited, sediments by waves. The deposits during this period are called
the water table is high and the distribution of materials in the TST and are identified by a retrogradational stacking pattern
ground may be complex. Moreover, space constraints due to the involving non-marine to marine successions. They are typically
presence of valuable agricultural or intensely populated areas as sands and gravels (beach complex) becoming finer when moving
well as the need to preserve valuable ecosystems can provide out into the sea.
additional challenges.
4
Last glacial
maximum
(LGM)
Holocene
Figure 6. The Llobregat delta (modified from Gámez 2007).
Figure 4. Sea level evolution since the Last Glacial Maximum, LGM
(modified from Stanley & Warne 1994). Exceptionally, the Llobregat delta displays well preserved
Pleistocene onshore deposits. By correlating onshore and
The sea level rise deceleration led to a progradation of the
offshore information, it has been possible to reconstruct the
delta by HST deposits. They include the prodelta fine grained
entire sedimentary system in both the delta (onshore) and
deposits, coarser materials from delta front sediments as well as
prodelta (offshore) areas. The delta sedimentary profile is formed
fluvial channels and flood plains. The HST delta succession
by four Depositional Sequences (DS-I to DS-IV) overlying the
characteristically exhibits interfingering configurations. It is
corresponding erosion surfaces (Figure 7); Deposition Sequence
interesting to note that the large development of present deltas is
I is of Holocene age whereas Deposition Sequences II to IV
the consequence of a quasi-steady sea level over the last 6000-
belong to the Pleistocene.
7000 years. As Figure 5 indicates, carbon data shows that the age
of the softer sediments at the bottom of present deltas generally
lies in the range of 6000-8000 years Before Present (BP).
a)
DS I
DS I
DS IV
DS III
DS II
Figure 5. Dates at the base of recent world deltas (Stanley 1997). b)
Figure 7. Geological profile of the Llobregat delta. a) Profile location, b)
Depositional sequences (DSI-DSIV).
3 THE LLOBREGAT DELTA
Depositional Sequence I (named the Holocene Upper Deltaic
The delta of the Llobregat river occupies the area immediately Complex) is composed of TST and HST deposits whereas
to the SW of the city of Barcelona. The onshore delta occupies Depositional Sequences II and III (called the Pleistocene Detritic
an area of 97 km2 and it extends along 23 km of the shore line Complex) contain fine sediments interpreted as continental and
(Figure 6). Its geology has been extensively studied (e.g. Llopis marine facies, representative of HST and FSST deposits. This
1942, Solé-Sabarís 1963, Marques 1984. Manzano et al. 1986, stratigraphic sequence is closely related to the variations of sea
Checa et al. 1988, Lafuerza et al. 2005, Simó et al. 2005, Liquete levels that reflect glacial and interglacial episodes (Figure 8).
et al. 2008. Of special interest is the stratigraphic study reported Glacial periods are characterized by continental facies formed
in Gámez (2007) and Gámez et al. (2009). Abundant information during sea level falls, and by fluvial facies formed during low sea
has also been collected from site investigation works related to levels. The interglacial periods include fluvial facies that were
the numerous infrastructure projects in the area. A recent review eroded and resedimented during the sea level rise. These periods
of the data has been presented in Yubero (2021). are however mainly characterized by prodelta and delta front
facies deposited during highstand sea levels.
5
upwards succession represents transgressive deposits of the
Holocene, which is consistent with the low sedimentation rates.
Deposition of this unit occurred between 14500 and 6000 years
BP. After the transgressive period, the slowing down of the sea
level rise and the increases in sediment supply triggered
progradation of the delta. The first progradation episode started
and some human activity. A second progradational event started
around 1,500 years BP developing at quite variable rates. The
increased sediment supply has been associated with
deforestation, aridity and another cold event. Similar delta
evolutions have been observed in other Mediterranean deltas
(Goy et al. 1996, Amorosi & Milli 2001; Cattaneo et al. 2003,
Figure 8. Sea level changes for the last 400 Kyears (extracted from Boyer et al. 2005). The resulting changes of the coastline are
Rabineau 2001). Correlation with depositional sequences (DS). shown in Figure 10.
Focusing now on the Holocene period (Depositional
Sequence I), the postglacial materials overlie the gravels and
conglomerates deposited during the LGM. The oldest Holocene
materials date back to about 14,500 years BP. Figure 9 shows the
estimated deposition rate during this period; it can be observed
that it is very variable (note that the deposition rate scale is
logarithmic). Initially, the sedimentation rate is very low, about
0.65 mm/year between 14,500 years BP and 6,000 years BP but
it increases to about 5 mm/year from 6,000 years BP to 1,200 a) b)
years BP. Afterwards, there were further increases of
sedimentation rates, although very variable; they reached values
in excess of 100 mm/year on some occasions. All the
sedimentation rates quoted are maximum values corresponding
to the central area of the delta.
c) d)
e) f)
Figure 10. Evolution of the shoreline from 6,500 years BP (modified
from Gámez 2007). a) Top of post glacial gravels, b) 6500-6000 years
BP, c) 3000-5000 years BP, d) 1300-1200 years BP, e) 1050-500 yeas
BP, f) 250 years – present.
Figure 11 shows a typical geological profile of the onshore
and offshore delta indicating the various facies identified (Gámez
2007). The main body of the delta and prodelta, and the most
interesting material from the geotechnical perspective, is formed
by Prodelta deposits (P) constituted by silty clays sometimes
interbedded with thin silty sands or fine sands. That deposit is
frequently overlaid (especially onshore) by deltaic front (Df)
deposits consisting of coarse to fine sands often containing thin
layers of sandy silts or silty clays. Surface materials correspond
to flood plain deposits (Fp) and alluvial deposits (Fc) consisting
of gravels embedded in a fine-grained matrix. Finally, the
prodelta deposits are underlain by sand (Bs) and gravel (Bg)
Figure 9. Estimated sedimentation rates during the Late Pleistocene and
beach deposits. Beach deposits are part of the TST while the
Holocene.
sediments above are part of the HST. DS-II and DS-III are found
As indicated above, the deposition rate depends on the below DS-I. They have a similar structure. In the upper part there
sediment supply and on the accommodation space available that, are clays and marls and well sorted partially cemented fine to
in turn, it is controlled by basement location and sea level; there coarse sandstone (Rw). They are interpreted as a mixture of
is however some controversy about which was the major factor coastal and continental facies. Below there are beach gravels
in this process (Jalut et al.2000, Goy et al. 2003, Trincardi et al. (Bg) and alluvial deposits (Fc). In these sequences, there is more
2004, Liquete et al. 2004, Frigola et al. 2007). The sediments matrix and cement content compared to the same facies of DS-I.
immediately above the gravels and conglomerates consist of stiff The rock basement is of Pliocene age.
clays and silts interpreted as marsh deposits and are overlain by The Llobregat delta has provided the site of quite a number of
sands interpreted as beach barrier materials. The coarsening- large infrastructure projects: the airport extension (Gens & Lloret
6
2003), the new Line 9 of the Metro system (Schwarz et al. 2006, (Alonso et al. 2000), a new railway track into the airport and the
Gens et al. 2006, 2012, Yubero 2021), underground works for Barcelona Port enlargement. Only the harbour works are
the High Speed railway (Arroyo et al. 2007, 2012, Arroyo & described and discussed in this lecture.
Gens 2009, Sau et al. 2012), a large water treatment plant
Fp Fc
Df
P
Bs
Bg-Fc
Rw
Rw Bg-Fc
Bg-Fc
Figure 11. Geological profile of the onshore and offshore Llobregat delta (Gámez 2007). Bs: beach sands, Bg: beach gravels, Df: delta front deposits,
Fc: fluvial alluvial deposits, Fp: flood plain deposits, P: prodelta deposits, Rw: clays, marls and partially cemented sandstone.
4 THE BARCELONA HARBOUR
Shipping activities in the Barcelona Port area have a long history
as befits a town with a past that goes back to Roman times.
However, there were no significant harbour structures (Figure
12) prior to the 18th century when an initial breakwater and some
quays were built (Figure 13). From that time on, Barcelona
Harbour has grown continuously until the present time. Over the
last two decades or so, a major enlargement of the Barcelona Port
has been undertaken that is still ongoing. It has involved the
construction of a new breakwater 4.8 Km long and the
lengthening of an existing one by 2.2 additional Km. These
operations allow the construction of more than 10 Km of new
quays and the doubling of the operational and logistic area.
Figure 14 gives an indication of the scale of the development by Figure 13. Barcelona Harbour in the 18th Century (1739-1748).
comparing the state of the harbour in 1995 with the present one.
It can be noted (Figure 14b) that the enlargement has required the
modification of the final course of the Llobregat river. 5 CONSTRUCTION OF THE SOUTH BREAKWATER
The plan view of the 4800 m-long South breakwater is presented
in Figure 15. The breakwater is divided in three sections; rubble
mound breakwaters occupy Sections 1 and 3 whereas 46 caissons
constitute the 1700 m-long Section 2. The caissons are 24.4 m
wide, 17 m high and either 36.7 m or 49.9 m long. The seabed
depth is rather uniform throughout the caisson section, about 20
m. In this paper, attention will be mainly focused on the vertical
caisson section that occupies the central and deeper part of the
alignment. From the point of view of stability, the vertical
caisson type may not be the optimum choice for a breakwater
placed on a soft soil foundation but it was selected mainly for
logistic and environmental reasons. A cross-section of the
Figure 12. View of the Barcelona Port area painted by Jan Cornelisz vertical breakwater is presented in Figure 16. For completeness,
Vermeyen in 1535. Figure 17 shows a representative cross-section of the rubble
mound breakwater. It can be noted, that both designs involve a
As mentioned above, two specific cases, selected on the basis rather limited dredging depth, stability is achieved by the
of their geotechnical relevance and interest, have been selected provision of longer berms.
for presentation: The construction and operation of the vertical breakwater
a) The construction of the South breakwater section comprise the following Phases:
b) The reconstruction of the Prat quay - Phase I: Dredging and construction of the base mound
The locations of the two structures are shown in Figure 14b. The - Phase II: Emplacement and filling of the caissons
new harbour structures are located in the Llobregat delta and
prodelta. - Phase III: Construction of the superstructure (crown
wall) and ancillary works
- Phase IV: Construction of a quay adjacent of the
breakwater in the protected basin area.
7
Only Phases II and III are discussed in some detail here. It
should be noted that stability concerns due to wave loading are
no longer relevant in Phase IV.
1995
a)
South breakwater East breakwater
2021
b)
Figure 14. Barcelona Harbour. a) 1995, b) 2021 (the location of the East and South Breakwaters, the Prat quay and the new course of the Llobregat river
are indicated).
Figure 17. Cross-section of the rubble-mound South breakwater.
Section 2 The behaviour of a vertical breakwater is complex involving
L= 1.700 m the interaction of gravity loading, wave loading and associated
structural, hydraulic and geotechnical response. A number of key
issues regarding this structure can be identified at the outset:
Figure 15. Plan view of the South breakwater. - Definition of the storm loading for design
- Selection of an appropriate reliability index or factor of
safety
- Confirming the adequacy of the design under static and
cyclic loading
- Ensuring an adequate degree of stability during
construction
- Behaviour of the vertical breakwater under cyclic wave
loading
Figure 16. Cross-section of the vertical caisson South breakwater.
8
They are addressed in the following sections. Storm wave r = Cost of indirect and direct losses / Investment
loading estimation (5.1), site investigation and soil properties
(5.2) inform the numerical evaluation of static and cyclic stability The sea action on a vertical breakwater is much dependent on
under permanent conditions (5.4), based on rationally selected the type of wave that hits the caisson. They can be classified as
safety measures (5.3). Because of the soft nature of the (Kortenhaus & Oumeraci 1998):
foundation soil, a key geotechnical issue is the stability during - quasi standing waves observed for smaller waves heights
construction that is likely to constitute one of the critical so that the incident waves are approximately reflected
conditions over the lifetime of the structure. It was envisaged and do not break
from the start to resort to staged construction combined with the - slightly breaking waves that occur for increased wave
observational method so that the loading and duration of each height when the waves start to break in front of the
step would be decided on the bases of monitoring observations breakwater
and assisted by the performance of appropriate numerical - impact waves when the berm in front of the structure
analyses. The performance of staged construction and the induces a breaker with the breaking point just in front of
application of the observational method provides the framework the wall.
for the following sections: description of the monitoring system, Figure 18 shows some typical variations of horizontal force
(5.5): description of the construction (5.6 and 5.8), and ensuring on the breakwater with time for the different wave types.
the safety of the breakwater throughout construction (5.7).
Finally (5.9) deals with the behaviour of the vertical breakwater
under cyclic storm wave loading.
5.1 Storm wave loading
Much of the uncertainty and complexity of breakwater design a)
and performance is related to the wave loading component of the
problem. A first step for defining the wave action is to select the
design storm depending on the expected duration of each Phase
of the project. In this case, those periods were rather
conservatively estimated as 5 years for Phase II and 50 years for
Phase III. From these values, it is possible to obtain the storm
return period, R, from b)
=1 1 (1)
where L is the assigned duration of the Phase and E is the risk or
probability of occurrence of the storm during the period of the
Phase. The return period of a storm is defined as the mean time c)
interval over which the variable value (significant wave height, Figure 18. Wave forces on wall caissons (USACE, 2008). a) non-
in this case) is exceeded once. The significant wave height is breaking wave, b) impact wave with an almost vertical front, c) impact
defined as the average height of the highest third part of the wave with large air pockets.
waves in a wave field. The value of E adopted for design depends
on the potential for a failure to cause human casualties and on the The MAST III PROVERBS project (Oumeraci et al. 1999)
magnitude of the associated financial losses (Table 1). In the case developed a parameter map to predict the type of wave that
of a vertical breakwater, it is assumed that failure would involve should be expected to act on the vertical caisson depending on
the total destruction of the structure. In that case, a value of the structure geometry, water depth and wave conditions; it is a
E=0.15 has been adopted corresponding to a low probability of development from an earlier proposal by Allsop et al (1996). The
human losses and to a medium economic repercussion if the map is based on three parameters (Figure 19): relative berm
works are disabled. height (hb/hs), relative wave height (Hs/hs) and relative berm
Consequently, the return period is 31.3 years for Phase II and width (Beq/Lhs) where hb is the berm height, hs is the seabed depth,
308.2 years for Phase III. The corresponding significant wave Hs is the significant wave height, Beq is the mean berm width in
heights are 5 m for phase II and 5.91 for Phase III. It is assumed front of the structure and Lhs is the local wave length. According
that the distribution of wave heights follows a Rayleigh (scaled to the parameter map and the dimensions of the vertical South
or standard) distribution because it generally provides a good breakwater, impact waves will act on the caisson whenever the
representation of the actual variation of individual wave heights significant wave height exceeds 4 metres.
in a wave field (Longuet-Higgins 1952, 1980, Nolte & Hsu 1979, In that case, the recommended method to estimate the
Tayfun 2004, Nayak & Panchangb 2015). For the standard magnitude and distribution of horizontal forces and uplift
Rayleigh distribution, the probability that a wave selected pressures induced by wave loading is via the Goda-Takahashi
randomly has a height less than H is: formula (USACE 2008). The original Goda’s wave load formula
(Goda 1974) was based mainly on a number of flume
experiments and additionally calibrated against real breakwater
( )=1 (2) failures. It was subsequently extended by Takahashi et al. (1993,
1994) to better account for impact loads and thus widen the scope
where Hs is the significant wave height. of applicability. The resulting expressions (Voortman et al 1999,
Goda 2009) lead to a trapezoidal distribution of pressures on the
Table 1. Probability of occurrence of the design storm over the duration vertical face and a triangular distribution of uplift pressures on
of a Phase (Puertos del Estado 1990). the caisson base (Figure 19).
Possibility of human losses
Reduced Expected
Economic Low (r<5) 0.20 0.15
repercussion
Medium (5<r<20) 0.15 0.10
if work is
disabled High (r>20) 0.10 0.05
9
present in the foundation of the breakwater except in sections
close to the shore. For conciseness, the description of the soil
properties will be focused on the soft silty clay material only.
Figure 20. Location of the boreholes and CPTu tests.
Figure 19. Distribution of wave-induced pressures on the vertical caisson
face and of wave-induced uplift pressures on the base of the caisson. S-4 AS-3 S-5 AS-4 S-6 AS-5
Horizontal and vertical forces on the breakwater are
calculated using the maximum wave height. It can be shown Df
that the most probable value of the maximum wave height, Hmax,
P P
in a sea state with N waves is approximately =
Bg-Fc
0.706 ln where Hs is the significant wave height. Wave
loading records for real structures during storms are still scarce Rw
and, inevitably, there are significant uncertainties associated with Bg-Fc
Rw
these basically semi-empirical methods (Meinen et al. 2020). Bg-Fc
Consequently, a number of physical model tests were performed
for the specific configuration of the South breakwater that led to Figure 21. Geotechnical profile along the South breakwater alignment.
some modifications of the analytically-derived wave design For interpretation of the labels, see Figure 11.
loads. The final design values for Phases II and III of the project
are presented in Table 2. The soft silty clay classifies as CL-ML in the Unified Soil
Classification System (USCS); mean values of liquid and plastic
Table 2. Wave-induced horizontal and uplift forces for the different limit are 30 and 20, respectively. The site investigation revealed
project Phases, that, at a medium scale, the soft silty clay layer was very uniform
Sig. wave Wave Wave Appl. Uplift along the whole alignment of the vertical breakwater. The Soil
Crown wall Behaviour Type (Robertson & Cabal 2015, Robertson 2016)
Phase height period force point* force
elev. (m) derived from CPTu tests lies nearly entirely in the range
(m) (s) (kN/m) (m) (kN/m)
corresponding to silty clay to clay.
No crown The key parameter for stability is of course the undrained
II 5 9 1036.3 9.48 525.1
wall shear strength. Ko-consolidated compression and extension
III +6 5.91 12.7 1436.1 10.36 878.2 triaxial tests as well as direct simple shear tests were performed
on specimens of silty clay under various consolidation stresses.
* Measured from the toe of the caisson
Representative results are collected in Figure 22. Undrained
shear strength, su, appears normalisable with respect to the
5.2 Site investigation and soil properties
consolidation vertical effective stress with mean values of su/ ’v0
After a general seismic geophysical survey, the site investigation of approximately 0.33 for triaxial compression, 0.195 for triaxial
along the alignment of the breakwater was carried out from a extension and 0.25 for simple shear. The undrained shear strength
vessel fitted with a heave compensated drilling rig. It was values of su/ ’v0 from CPTu tests (using a Nk coefficient = 15)
performed in two stages, in the first one 4 boreholes were drilled tend to cluster around the 0.25 value. Results from vane tests
to a depth of 40 m (from the seabed) and 18 CPTU tests were show more scatter.
carried out. In the second stage, 9 CPTU tests were performed Other important parameters for breakwater construction and
and three deeper boreholes were drilled, one of them (S-6) was performance are the compression and swelling indices (Cc and
taken to a depth of 120 m to reach the bottom of Depositional Cs), the coefficient of secondary consolidation (C ) and the
Sequence III. Vane tests on the clayey soft soils were also carried vertical and horizontal coefficients of consolidation, (Cv and Ch).
out inside the boreholes. The location of the boreholes and CPTU A first set of parameters was determined from conventional tests.
tests are shown in Figure 20. It may be noted that some boreholes As time was available, a trial embankment loading test was
are not located precisely under the centre of the breakwater; this performed onshore, across the breakwater location, in order to
is due to slight changes in the alignment subsequent to the site obtain parameters with higher reliability. The embankment was
investigation. 160 m x 80 m in plan and 8 m high. It was built over a period of
The various ground profile layers identified in the site 5.5 months and remained at maximum height for a further 6.5
investigation are shown in Figure 21, using the same months. It was then unloaded in two stages in about three months.
nomenclature as in section 3 (Figure 11). From the geotechnical The test was instrumented by four sliding micrometres reaching
point of view, the dominant material encountered is a layer of down to a depth of 60 meters measuring vertical deformations
soft silty clay about 40-50 m thick corresponding to the prodelta every meter. Surface vertical displacements were also measured
deposit (P). It is underlain by a level of gravels and sands (Bg-Fc with settlement plates. Strings of vibrating wire piezometers
beach and alluvial deposits) that constitutes the bottom of the were installed in four locations to check the evolution of pore
Depositional Sequence I. Underneath, the materials of pressures. The observations of the loading test were
Depositional Sequences II and III appear, each of them composed backanalysed so that the values of compressibility and
of an upper layer of stiff clays (Rw deposit) and a lower layer of coefficient of consolidation parameters could be estimated on a
sands and gravels (Bg-Fc deposits) Most of the upper sand layer meter-by-meter basis. Figure 23 shows four examples of vertical
that appears in the area covered by boreholes S4 and AS-3 was strain adjustments at different depths for the construction and
in fact dredged away before construction and, therefore, it is not consolidation period. It was interesting to observe that the
10
Cc Cs C cv (cm2/s)
measurements from the settlement plates and from the sliding 0.01 0.1 1 0.01 0.1 0.001 0.01 0.001 0.1 10
micrometres were in close agreement indicating that the 0
materials below 60 m are quite stiff and do not contribute in any
significant way to the surface settlements. 20
40
Depth (m)
60
80
100
oedometer tests 1996 oedometer tests 1999 oedometer tests 2001
cptu tests 1999 triaxial 1999 cptu tests 2001
preload: SSC model preload: oedometers
Figure 24. Variation with depth (from the sea bed) of the compression
index (Cc), swelling index (Cs), coefficient of secondary consolidation
(C ) and coefficient of consolidation (Cv) derived from laboratory and
field tests.
The results of the simple shear testing programme are
Figure 22. Undrained shear strength of the soft silty clay determined from summarised in the interaction diagram shown in Figure 26; it
laboratory and field tests. indicates the number of cycles required to reach failure for
different combinations of average ( ave) and cyclic ( cyc) shear
stresses normalised by the undrained shear strength, as suggested
by Andersen & Lauritzsen (1988). It is readily apparent that the
number of cycles at failure increases rapidly as the value of cy
21.5 m reduces. In the diagram, the relationship between accumulated
and cyclic strain is also shown. For comparison, the equivalent
25.5 m contours for Drammen clay (Goulois et al 1985, Andersen 2009)
are also displayed in the Figure; they exhibit a similar shape.
12.5 m
After cyclic loading, a number of samples were loaded undrained
15.5 m
until static failure was achieved. A moderate undrained strength
reduction up to 20% was observed only for vertical consolidation
stresses larger than about 250 kPa.
Figure 23. Examples of four vertical strain model adjustments at different
depths for the construction and consolidation period of the trial 40
embankment.
The parameter values resulting from those adjustments are
presented in Figure 24 together with the values derived from 20
oedometer and triaxial tests. It can be observed that there is a
Shear stress (kPa)
large degree of agreement between the two sets of values; this is
especially significant for the vertical coefficient of consolidation,
an important parameter in staged construction. It can also be 0
noted that the horizontal coefficients of consolidation determined
from CPTu dissipation tests are nearly one order of magnitude
larger, a rather typical result in this type of materials. S511-A
-20
Because of the oscillating nature of wave loading, it was S511-B
necessary to determine also the cyclic properties of the soil. S511-C
Cyclic undrained triaxial tests were performed on silty clay
samples consolidated isotropically and anisotropically. In -40
addition, 17 undrained cyclic simple shear tests were also carried
-40 -20 0 20 40
out under different stress conditions. As an example, Figure 25
shows the development of shear strains in three simple shear Shear strain (%)
cyclic tests of different amplitudes. Failure is defined as the point Figure 25. Stress-strain behaviour observed in three cyclic simple shear
at which the accumulated shear strain reaches a value of 10%. tests under different applied cyclic stresses.
Different values of average shear stress and ( ave) and cyclic
stresses ( cyc) were used (Andersen 2009). It is recognized that For the numerical analyses, the behaviour of the soft silty clay
application of cyclic stresses may lead to pore pressure increases was represented by the Soft Soil (SS) model (Brinkgreve et al,
(cyclic and permanent), strain development (cyclic and 2017) and by the S-Clay1 model (Wheeler 1997, Wheeler et al.
permanent) and, potentially, to some undrained shear strength 2003) The first one is an isotropic model well attuned to
reduction (e.g. Andersen 2009, Patiño et al. 2013, Wang et al. reproduce the behaviour of soft fine-grained soils whereas the
2018). second one is able to yield different soil responses depending on
the mode of loading (Figure 27). Model parameters have been
calibrated to provide a value of su/ ’v0 of 0,25 for the simple
shear case because, as demonstrated below, it is the dominant
failure mode (Figure 28). As mentioned later, the UBC3D-PLM
constitutive model (Puebla et al. 1997, Beaty & Byrne 1998) has
been used to account for the cyclic behaviour of the soft clay in
11
dynamic analyses. SERI takes into consideration three different factors: possibility
and impact of loss of human life, damage to the environment
and/or to the historical heritage and potential for social
disturbance. Design probability of failure is mainly controlled by
the SERI index while ERI is taken into account via economic
optimization procedures.
Figure 29. Relationship between failure probability and reliability index.
Many of the harbour works have a low or very low SERI
index and the design probability related to geotechnical failure is
Figure 26. Interaction diagram for the silty clay of the Barcelona Port. chosen to be sufficiently low so that it does not affect the overall
The contours for Drammen clay (Goulois et al 1985, Andersen 2009) are failure probability of the works due to other more proximal
added for reference. causes. The recommended failure probability values (and
reliability indices) for the different cases are collected in Table 3.
However, when the principal agent of potential stability is wave
action, a higher probability of failure (about on order of
magnitude higher) is accepted; geotechnical collapse then
becomes the main factor (or one of the main factors) affecting
the overall probability of failure. The breakwater case described
herein has a low SERI value (5-19) and the probability of failure
over the entire lifetime of the works recommended for design is
0.01 (reliability index = 2.33).
Figure 27. Constitutive laws. a) Soft Soil model, b) S-Clay1 model.
Table 3. Maximum probability of failure and reliability index for
different values of the Social and Environmental Repercussion Index,
SERI (Puertos del Estado 2002).
Very low High / very
SERI Low 5-19
<5 high >19
Maximum failure
probability over the 10-2 10-3 10-4
works lifetime
Reliability index 2.33 3.09 3.72
It should be recognized, however, that factors of safety are
still widely used in professional practice and many designs are
Figure 28. Experimental and Soft Soil (SS) and S-clay1model results for based on them (Lacasse 2013). It must be also be said that, very
simple shear tests. The Modified Cam Clay (MCC) model results are often, the statistical information required for a rigorous
added for reference. probabilistic analysis is not available and the risk analysis has to
be performed in a qualitative way (e.g. USACE 2014). Although
5.3 Reliability and factors of safety a number of issues associated with geotechnical risk and
It has been long recognized that factors of safety are not a fully reliability (Christian & Baecher 2011) still persist, there is,
satisfactory measure of how far a geotechnical structure is from however, a definite trend towards characterizing safety in terms
failure as it does not account for uncertainty, either aleatoric or of probability and reliability. Many codes maintain the use of the
epistemic, in the design (Duncan 2000, Lacasse et al. 2019). A factor of safety concept (total or partial) but some recent
more rational metrics is the probability of failure or equivalently, developments seek to link factors of safety to reliability measures
the reliability index, in some way. Thus, the Spanish geotechnical code for Maritime
and Harbour works, ROM 0.5-05 (Puertos del Estado, 2008) has
= (3) established an approximate relationship between the reliability
index (or probability of failure) and the factor of safety using
where is the normalised standard cumulative probability the following expression:
function and Pf is the probability of failure (Figure 29).
The probability of failure to be used in design in Spanish = (4)
harbours (Puertos del Estado, 2002) depends on two indices: the
Economic Repercussion Index (ERI) and the Social and where F* is the central estimate of the factor of safety and is
Environmental Repercussion Index (SERI). ERI includes the the standard deviation of ln given by
cost of rebuilding the structure and the economic consequences
of a failure (normalized by a regularising parameter) whereas
12
material is a normally consolidated clay and, therefore,
= ln 1 + (5) undrained shear strength is the relevant strength parameter in any
stability assessment (Ladd 1991). For brevity, only Phases II and
where is the combined sensitivity of the factor of safety to the III of the vertical breakwater design are reported here. In the
variation of design parameters and other variables. If a lognormal stability analyses, the Soft Soil model was used to characterize
distribution is assumed, then: the behaviour of the soft silty clay with parameters selected to
yield a value of su/ ’v of 0.25; a prudent mean for the type of
failure surface envisaged. The factors of safety were computed
ln 1 + = ln 1 + (6) for the final situation of each phase assuming that all the excess
pore pressures had dissipated and failure would occur in
where are the individual sensitivities. undrained conditions. Adequate values are obtained, as shown in
Based on this methodology, factors of safety are Table 5. The critical failure mechanisms for the different
recommended for each type of structure, mode of failure and instability conditions are depicted in Figure 30. They are very
required probability of failure. The standard deviation of the similar, irrespective of the type of factor of safety with which
Factor of safety can be cautiously derived from the computation they are associated; the simple shear failure mode appears
of the effects of the variation of a number of parameters applied predominant. Only the failure mechanism corresponding to an
to a generic representative situation. In a specific case, dedicated increase of the storm load alone (FS3) results in a more
sensitivity analyses can be used to improve the estimation of superficial failure surface. The stability during the transient
uncertainty, following the guidelines set out in Soriano (2005), period of consolidation and excess pore pressure dissipation,
although, often, the required statistical data is scarce or potentially more critical, is discussed later.
unavailable. In the rest of the paper, the degree of stability will
be quantitatively characterized by a Factor of Safety; the values Table 5. Factors of safety for the final states of the breakwater in Phases
used, however, should be viewed in the context of their II and III.
relationship with reliability/failure probability as outlined in this FS1: 1.91 CF1: 1.77
section. Phase II FS2: 1.52 CF2: 1.42
For vertical breakwaters, the values of factors of safety for the FS3: 2.09
different geotechnical failure modes and load combinations are FS1: 2.06 CF1: 1.96
listed in Table 4. They correspond to a probability of failure of Phase III FS2: 1.51 CF2: 1.40
10-2 over the lifetime of the structure. Only global stability is FS3: 1.85
considered (the bearing capacity case is in fact akin to the global
stability case), overturning and sliding were generally not critical
in the design of the South breakwater.
Table 4. Vertical breakwaters. Factors of safety according to ROM 0.5
Probability of failure: 0.01. (Puertos del Estado 2008).
Load combination
Ultimate limit Quasi - Fundamental/ Accidental a)
state permanent Characteristic
FS1/CF1 FS2/CF2
Sliding 1.3 1.1 1.0
Bearing capacity 1.8 1.5 1.2
Overturning 1.3 1.2 1.1
Global stability 1.3 1.1 1.0
The quasi-permanent case includes the permanent actions and b)
the quasi-permanent values of variable loads. The
fundamental/characteristic situation takes into account the
variable loads potentially acting during failure; the predominant
one is taken at its characteristic value. The accidental situation
considers extraordinary loads acting during a short period (e.g.
seismic loading). In the breakwater the quasi-permanent case c)
corresponds to the application of self-weight with a horizontal Figure 30. Failure mechanisms obtained in the calculation of factors of
sea level (Factor of safety FS1/CF1) whereas the safety. a) Factors of safety FS1, CF1, b) Factors of safety FS2, CF2, c)
Fundamental/characteristic case includes the wave loading as the Factors of safety FS3, CF2.
predominant loading variable (Factor od safety FS2/CF2).
Factors of safety FS1 and FS2 indicate the increase of loading A pseudo-static evaluation of stability of the breakwater
that would be required to reach failure. In contrast Factors of taking into account the cyclic effects of wave loading was also
Safety CF1 and CF2 are defined in terms of strength reduction. performed, based on the experimentally-determined interaction
Finally, useful information is also derived from a definition of an diagram (Figure 26). The design storm is converted into a
additional factor of safety, FS3, that represents the increase of number of waves of equal size that will apply a uniform cyclic
wave action that would result in breakwater failure. The loading to the foundation soil. This conversion introduces a
consideration of this set of factors of safety instead of a single degree of uncertainty and two different cases have been
value enriches the assessment of the safety of the structure. The considered, one with a small number of waves of large magnitude
accidental case is not discussed here; it was not a critical aspect and a second one involving a large number of waves of smaller
of the design of this breakwater. size. The first conversion yields a horizontal load of 1011.5 kN/m
applied 40 times whereas the second one results in a horizontal
5.4 Design stability evaluation load of 341,6 kN/m applied 1000 times. Each wave loading cycle
is additionally associated with the application of a triangular base
The evaluation of the degree of stability of the design cross- uplift whose maximum value is 310.5 kN/m for the larger wave
sections was evaluated using numerical analyses because it was case and 223.4 kN/m for the smaller wave case.
a more reliable method to account for the variation of undrained The procedure requires the performance of a numerical
shear strength with changing effective stresses. The foundation analysis that applies the corresponding horizontal vertical load
13
and uplift water pressure on the constructed vertical breakwater
to obtain the stress increments throughout the foundation. In this
way, it is possible to obtain the values of cyc and ave at every
point of the foundation, so they can be placed in the
corresponding interaction diagram. It is assumed that the points
located beyond the interaction diagram lines may undergo cyclic
failure and, consequently, their undrained shear strength is then
reduced, based on laboratory cyclic test results. A conservative
strength reduction varying between 0.6 and 0.8 has been applied.
It turned out that the resulting reduction in Factor of safety for
both Phases II and III was very marginal due to the very limited
size of the zone affected by cyclic failure. A similar procedure
was used successfully in the design of the new Bilbao Harbour
breakwater (Alonso & Gens 1999). Figure 31. Location of the instrumented cross-sections.
5.5 Instrumentation
As noted above, the limited shear strength of the foundation soil
led to the adoption of staged construction combined with the
observational method as the procedure to build the breakwater.
The observational method naturally requires installing an
appropriate monitoring system to monitor key parameters. One
of them would be the pore pressures in the foundation soil as
stability will depend on the degree of dissipation achieved at any a)
particular time. An additional component of the procedure was
the setting up of an appropriate numerical model to assist in the
interpretation of the observations and to provide predictions to
guide future operations.
Four cross sections were instrumented (Figure 31), one in the
rubble mound section (in an area close to the caissons) and three
in the vertical breakwater section. A scheme of the instrumented
cross sections is presented in Figure 32. The monitoring system
included a string of piezometers at five different depths emplaced b)
near the centre of the cross section reaching down to a depth of Figure 32. Schematic layout of the instrumented sections. a) Rubble-
35 m below dredging level. A string of five settlement sensors, mound breakwater section, b) Vertical breakwater sections.
also 35 m long, was installed near the pore pressure measurement
points. The determination of settlements at depth was based on
the comparison of the seal level pressure measured at those
locations with the reading of a reference cell located away from
the loaded area Unfortunately, pressure insulation did not
function properly and reliable observations could only be made
when transient excesses of water pressures had dissipated. In
addition, two settlement-measuring devices were installed on the
dredged surface. Finally, horizontal movements were monitored
by two inclinometers 20 m long installed on either side of the
cross section; the locations were selected to provide indications
of impending failure. The measurement signals were carried by
cables to a buoy that transmitted the data by wireless.
Instrumentation was installed once the dredging has been
completed and before the construction of the rubble mound. The
strings of sensors were installed from a floating platform using
an innovative method that involved pushing a casing containing
the chain of sensors to the required depth (Figure 33). There was Figure 33. Instrumentation installation system.
little precedent for installing geotechnical instrumentation
beneath a breakwater and an elaborate protection system was
placed to shield the cables from the dumping of material used to
construct the foundation rubble mounds. The system was quite
successful and very few instruments were lost for this cause.
Unfortunately, there was a massive loss of operational
instruments due to maritime traffic and irregular fishing
operations (Figure 34) with the consequence, as discussed later,
that significant construction operations had to be carried out
without appropriate monitoring data. Once the slabs on top of the
caissons were constructed new strings of pore pressure sensors
were installed at similar locations as in the initial instrumentation
system. Also, the caissons were equipped with surveying targets
placed at the four corners, so caisson settlement data was
available as soon as they were emplaced.
Figure 34. Replacement of a damaged data cable.
14
5.6 Construction of the rubble mounds 300 Soft Soi
seabed (kPa)
Sclay-1
The first construction operation was the dredging of the upper 200
Load at(kPa)
part of the ground profile down to the depth specified in design. h
100
The base of the caisson section as well the submerged parts of
the rubble mound breakwater were deposited by dumping the 0
material from barges (Figure 35). The material for the emerged
(kPa)
(kPa)
part of the rubble mound was emplaced by terrestrial means. 120
Pore Pressure
80
E.ressure
40
E. Pore Pressure (kPa)
120
80
40
E. Pore Pressure (kPa)
120
80
40
E. Pore Pressure (kPa) E. Pore Pressure (kPa)
120
80
40
Figure 35. Rubble mound material loaded in a barge.
0
Construction of the rubble mound section of the South
breakwater started from the shore and proceeded seawards 120
(Figure 36). The observed evolution of pore pressures at five 80
different depths for instrumented section 1 is shown in Figure 37. 40
In the upper graph of the Figure, the estimated vertical stress at 0
the centre of the embankment is plotted; it reflects the different
0
100
300
500
700
600
800
200
400
construction stages of the embankment that lasted just over 600 Days
days. The plots of the individual measurements stop at around
750 days due to irrecoverable sensor failure; readings gaps Figure 37. Observed and computed excess pore pressures at five different
depths. The top graph shows the evolution of the vertical stress at the
correspond to repair periods. The observed evolutions of base of the embankment due to construction. Numerical model results are
settlements at the base of the breakwater and at different depths added. Instrumented section 1.
for this section 1 are shown in Figure 38. As mentioned above,
the settlement sensors readings are only valid at the end of the The excess pore pressures measured at five different depths in
various periods of water pressure dissipation, the dashed lines in instrumented section 2 during the construction of the rubble
the Figure join the observations that are considered reliable. Most mound for the caisson base are plotted in Figure 40. The plot at
sensors became lost by day 550, except one that survived until the top of the Figure indicates that construction lasted about 300
day 750. It is likely, that the settlements at dredging level in this days. Sensors failed around day 700 just before caissons were
section must have been well in excess of 1 m at the end of due to be emplaced. It can be observed that, as in instrumented
construction. Figure 39 shows the horizontal displacements section 1, pore pressures had not fully dissipated when the
measured with the inclinometer sensors placed in the port side; sensors failed. The same kind of observations were obtained for
the movements increase consistently as the construction instrumented sections 3 and 4 even though they suffered still
advances but are quite small, a few centimetres only. more extensive sensor losses. Figure 41 shows the settlements
measured at the base of the embankment (dredged level) of
instrumented section 2, no data could be obtained from the
settlement sensors at depth. Settlements of about 1 metre due to
the construction of the base rubble mound were recorded before
the sensor failed just before caisson emplacement.
Figure 36. Rubble mound breakwater construction close to completion.
15
agreement with the exception of the pore pressures at large
(kP 400 depths which are not relevant to stability concerns. Naturally
(kPa)
Soft Soi there are gaps in the instrumentation data where the comparison
seabed
300
Sclay-1 is not feasible. It is interesting to note that the, in spite of their
Loadatonseabed
200
h
significant formulation differences, the two models provide quire
100 similar predictions for the field data. Full details of constitutive
lement (m) Load
0 model calibration and of the numerical analyses performed are
0.0
given in Madrid (2021).
Settlement (m)
1.0
200
seabed (kPa)
2.0
(kPa)
100 -13.5m
3.0 -26m h
Soft Soil model
SS02 SS01
Load at
PP05
0 PP04
PP03
Sclay-1 model PP02
Settlement (m)
0.3 0 PP01
60
E. Pore Pressure (kPa)
0.6
ressure (kPa)
0.9
40
1.2
1.5 20
0
Settlement (m)
0.3 0
60
E. Pore Pressure (kPa)
0.6
0.9 40
1.2
1.5 20
0
0
Settlement (m)
0.3
60
E. Pore Pressure (kPa)
0.6
0.9 40
1.2
20
1.5
0
50
100
150
700
300
500
750
800
350
550
600
200
650
250
400
450
0
60
Time (days)
E. Pore Pressure (kPa)
Figure 38. Observed and computed settlements at the base of the rubble 40
mound breakwater and at three different depths. The top graph shows the
20
evolution of the vertical stress at the base of the embankment due to
construction. Numerical model results are added. Instrumented Section 1.
0
60
E. Pore Pressure (kPa)
40
20
0
0
100
700
300
500
600
200
400
Time (Days)
Figure 40. Observed excess pore pressures at five different depths. The
top graph shows the evolution of the vertical stress at the base of the
embankment due to construction. Numerical model results are added.
Instrumented section 2.
(kP
200
seabed(kPa)
c.s
onseabed
-13.5m
100 -26m SS02 SS01 h
PP05
Loadd at
PP04
PP03
PP02
PP01
0
0.5
Settlement (m)
Figure 39. Distribution of horizontal displacements at depth for various 1.5
times. Numerical model results are added. Instrumented section 1. 2
2.5
The interpretation of monitoring data was assisted by 3
0
100
numerical analyses that simulated the full construction sequence
300
500
700
600
200
400
of dredging and rubble mound construction. The analyses Time (days)
incorporated the duration of each construction stage so that the
resulting consolidation and undrained shear strength increase Figure 41. Evolution of settlements at the base of the breakwater and of
was taken into account by means of a proper coupled the caissons after emplacement. Numerical model results are added.
hydromechanical formulation. The behaviour of the soft silty Instrumented section 2.
clay was modelled by two constitutive laws: Soft Soil model and
S-Clay1 model. The results of the numerical analyses are 5.7 Ensuring safety during caisson emplacement and filling
compared with the field monitoring observations in Figures 37, At this point, it is of interest to examine more closely how the
38, 40 and 41. It can be observed that there is a very substantial sequence of caisson emplacement and filling was undertaken to
16
ensure a sufficient margin of safety at all times. The installation
of the rubble mound, without the emplacement of the rockfill
protection, had lasted about 350 days and a decision about the
time and sequence of caisson emplacement and filling was
required at the time. Obviously, stability would depend on the
evolution of pore pressure dissipation in the foundation soil. It
had been envisaged that the observational method would be
followed and decisions would be taken based directly on the
monitoring data, assisted by accompanying numerical analyses.
However, due to the widespread instrumentation failure, a)
available monitoring information was extremely scarce; in fact,
all pore pressure monitoring had ceased to operate just before
caisson emplacement was due. The advancement of the vertical
breakwater construction, therefore, constituted a significant
challenge under these conditions. The only observations that
were going to be available during the early construction period
would be the settlements of the caissons measured by surveying.
To proceed, it was decided to rely, with a measure of caution,
on the results of numerical analyses. There was confidence that
b)
the behaviour of the soft silty clay was reasonably well
represented in the analyses, especially regarding its undrained Figure 42. Variation of Factor of Safety with amount of caisson filling.
shear strength. In addition, the simulations had agreed rather well a) in terms of load increase, b) in terms of strength reduction. The dashed
vertical line indicted the caisson unit weight selected for partial filling.
with the monitoring data during the first stages of construction
until the sensors failed. Since the material had proved to be quite Table 6. Factors of safety for different construction stages.
uniform along the breakwater alignment, it was considered Phase Stage Factors of safety
acceptable to assume that the results of the analyses would be Caisson sinking (11.5 CF1: >1.84 (c)
II
valid throughout the breakwater alignment. kN/m3) CF2: < 1* (b)
Consequently, numerous stability evaluations using FS1: 1.57 CF1: > 1.84 (c)
Partial caisson filling
numerical analyses were performed considering a range of II FS2: 1.30 CF2: 1.28 (a)
(15.2 kN/m3)
waiting times for the installation and filling of the caissons. Due FS3: 1.33
to the limited duration of the operations, the smaller Phase II FS1:1.78 CF1: 1.45 (a)
Full caisson filling (19.1
II FS2: 1.50 CF2: 1.25 (a)
design storm was used in the stability analyses. The simulations kN/m3)
FS3: 1.88
predicted failure if the waiting time was only three months and, FS1: 1.91 CF1: 1.77 (a)
even with waiting times up to 6 months, the degree of safety was Full dissipation of pore
II FS2: 1.52 CF2: 1.42 (a)
not satisfactory yet. If full pore pressure dissipation is achieved, pressure. Phase II
FS3: 2.09
safety become acceptable but the time required was uncertain, FS1: 1.98 CF1: 1.59 (a)
Construction of the
especially in the absence of reliable pore pressure measurements. III FS2: 1.42 CF2: 1.23 (a)
superstructure
An alternative was explored consisting of emplacing the FS3: 1.71
caisson four months after the rubble mound was completed but FS1: 2.06 CF1: 1.96 (a)
Full dissipation of pore
filling it only partially. When the caisson is sunk full of water, its III FS2: 1.51 CF2: 1.40 (a)
pressure. Phase III
average unit weight is 11.5 kN/m3 and it becomes 19.8 kN/m3 FS3: 1.85
when it is totally filled. Using different unit weights, the effects * Indicates failure
of partially filling the caisson on stability can be readily (a) Global stability failure
(b) Sliding failure
examined. The factors of safety obtained are presented in Figure
(c) Failure of the rubble mound slope on the port side
42 where the failure mechanisms identified in the strength
reduction analyses are also indicated (Figure 42b). The two main
ones are a global stability failure and a sliding failure combined
with a local bearing capacity failure at the caisson edge on the
port side (Figure 43). The third mechanism involves the failure
of the rubble mound slope on the port side and is not relevant.
Based on those results, a new construction schedule was
adopted where a partial filling leading to a unit weight of 15.2
kN/m3 (corresponding to filling with sand 40% of the caisson
cells) would be implemented just after caisson sinking followed a)
by a final filling of the caisson a minimum of four months later.
The final filling was not complete but limited to an equivalent
unit weight of 19.1 kN/m3. With this strategy, the resulting
Factors of safety for all the construction stages are collected in
Table 6. Note that in Phase III the wave loading and uplift
pressures are increased with respect to Phase II because both the
design storm and the height of the breakwater structure are larger,
see Table 2. Now, all factors of safety are considered adequate
except at the moment of sinking the caisson when the factor of b)
safety considering storm loading is below 1, indicating failure by
Figure 43. Failure mechanisms identified in the stability analyses.
sliding. This result simply reflects the vulnerable situation of the
a) Global stability failure mechanism, b) Sliding failure mechanism.
caisson just after emplacement, if a storm occurs. This is the
reason why filling, even partial, proceeded immediately after the 5.8 Completing the construction of the vertical breakwater
caisson had been deposited on the base rubble mound.
Accordingly, the new construction schedule incorporating staged
construction and partial caisson filling was adopted to
accomplish the full construction of the vertical breakwater. The
17
caissons were transported floating from the shipyard and sunk
into their respective positions and 40% of the caisson cells were
(kP
400
(kPa)
immediately filled with sand (Figures 44 and 45). Full filling was
seabed
c.s
300
performed a minimum of four months later. For some caissons,
onseabed
-13.5m
construction times were in fact longer, due to delays in other parts 200 -26m SS02
PP05
SS01 h
of the works, adding an additional measure of stability. In this
Load at
PP04
100 PP03
way, construction was completed without mishaps. The final 0
PP02
PP01
stage of crown wall construction was normally undertaken about 0.5
Settlement (m)
one year later. 1
1.5
2
2.5
3
600
800
200
400
1000
1200
1400
1600
1800
3000
3200
3400
2000
2200
2400
2600
2800
Time (days)
Figure 46. Evolution of settlements at the base of the breakwater and of
the caissons after emplacement. Numerical model results are
added .Instrumented section 2.
The measurements of the piezometers installed from the top
of the caisson, about 17 months after the emplacement and filling
of the caissons are presented in Figure 47 along the observations
obtained during the construction of the rubble mound base. It was
satisfying that the measurements at the time of the second
installation were quite consistent with model predictions. It is
interesting to note that, some residual pore pressures are still
present in the foundation at that time and about 50 additional
months were required for them to reduce to negligible values.
at seabed (kPa)
400
300
-13.5m
Load (kPa)
200 -26m SS02 SS01
h
Figure 44. Vertical section of the South breakwater during construction.
PP05
PP04
100 PP03
PP02
PP01
0
(kPa)
120
(kPa)
Pore Pressure
80
E. essure
40
0
E. Pore Pressure (kPa)
120
80
40
0
E. Pore Pressure (kPa)
120
80
40
0
E. Pore Pressure (kPa)
120
80
40
0
Figure 45. Filling of the caissons.
E. Pore Pressure (kPa)
120
Caisson settlements were measured from the time of
80
emplacement, the measurements for five different caissons are
plotted in Figure 46; they are all very similar; the settlement due 40
to the rubble mound construction is about 1 m while the caissons 0
settle an additional 1.5 m. The results of the simulations have
0
600
800
200
400
1000
1200
1400
1600
1800
3000
3200
3400
2000
2200
2400
2600
2800
been added to the Figure; the fact that observations were well
Time (Days)
reproduced during the critical construction period reinforced to
some extent the confidence placed on the numerical analyses. Figure 47. Observed excess pore pressures at five different depths. The
Figure 46 also includes two long term measurements obtained top graph shows the evolution of the vertical stress at the base of the
from the global survey system of the Port Authority that reveal embankment due to construction. Numerical model results are added
that there are long-term settlements of the order of 2-3 cm per Instrumented section 2.
year eight years after the end of construction.
5.9 Post-construction behaviour under storm loading
A major issue in the design and performance of a vertical
breakwater lies in the estimation of the wave loading and its
18
effects on the structure. To contribute to the advancement of
knowledge in this area, a caisson was equipped with sensors to
measure the water pressure exerted by the waves as well the
water pressures on the base of the caisson. The layout of this
hydrodynamical monitoring system is shown in Figure 48.
Twenty sensors were placed on the face of the caisson at different
elevations along two vertical alignments whereas eight sensors
were placed in the caisson base, again in two alignments. Figure
49 shows one of the sensors about to be installed.
Figure 50. Evolution of the significant wave height and of the maximum
wave height during the December 2009 storm.
Wave pressure Wave pressure
(kPa)
-4.5 m
(kPa)
a) b)
Figure 48. Layout of the hydromechanical monitoring system. a) Sensors -9.5 m
on the caisson face, b) Sensors on the caisson base.
Increment pore
pressure (kPa)
Date
Figure 51. Water pressures on the caisson face (elevations -4.5 m and -
9.5 m) and increments of pore water pressures in the foundation recorded
in the December 2008 storm.
Hmax (m)
Wave pressure Wave pressure
(kPa)
Figure 49. Sensor to measure wave-induced water pressures on the
caisson face.
-4.5 m
A very large storm hit the Barcelona coast in December 2008
shortly after the breakwater was completed. The wave heights
(kPa)
were not recorded close to the harbour but measurements in a
nearby area provided the information shown in Figure 50; the
maximum wave height reached 9 meters with a significant wave
-9.5 m
height of 6 m. Two typical time histories of the wave-induced
Pore pressure
water pressures on the caissons is shown in Figure 51, they
correspond to observations at elevation -4.5 m and -9.5 m.
(kPa)
Unfortunately, the record is not complete because, due to the
violence of the storm, the data collection centre eventually
flooded and was damaged. It is also interesting to note that the
piezometers installed at different depths in this section showed a Date
steady increase of pore pressures at the height of the storm.
The effects of a milder storm in December of the following Figure 52. Water pressures on the caisson face (elevations -4.5 m and -
year (maximum wave height, 6 meters) were fully recorded and 9.5 m) and pore water pressures in the foundation recorded in the
December 2009 storm.
are shown in Figure 52. In this case, the wave water pressures on
the caisson at the same locations are significantly lower and no The initial, maximum, minimum and average wave-induced
pore pressures are generated in the foundation. The frequency water pressure envelopes on the face and on the base of the
content of the two December storms and an additional one caisson measured with the hydromechanical monitoring system
(January 2010, maximum wave height 6 m) are shown in Figure during the severe December 2009 storm are plotted in Figure 54.
53; they are similar with a maximum in the 0.07-0.12 Hz range. It can be noted that the observed maximum pressures are
The largest storm exhibits the lowest dominant frequency. somewhat smaller than the values proposed by Gouda (2009),
suggesting that the assumptions made in design are realistic.
19
0.5
E. Pore Pressure (kPa)
0.4 20
Amplitude
0.3
15
0.2
0.1
10
0 5
0
0.2
0.4
0.04
0.08
0.12
0.16
0.24
0.28
0.32
0.36
Frequency (Hz)
Figure 53. Frequency content for three different storms.
-5
-10
0 200 400 600
Time (s)
Figure 55. Computed pore pressure generation in the soil foundation due
to storm loading.
Figure 56. Contours of excess pore water pressure ratio ( v) at the
end of the cyclic loading.
5.10 Summary notes
The behaviour of vertical caisson breakwaters is complex as it is
the result of a combination of hydraulic, structural and
geotechnical responses to storm wave actions that, in turn,
incorporate a large degree of randomness. When the vertical
breakwater is founded on a thick layer of soft normally
consolidated clay, geotechnical considerations are dominant.
Judgement is required to introduce the inevitable simplifications
required in practice. The selection of an appropriate design storm
and wave action, the selection of suitable safety indicators
(related to probability of failure) and the site investigation
Figure 54. Initial, maximum, minimum and average wave-induced information have been combined together to evaluate the
water pressures measured on the face and on the base during the
December 2009 storm.
stability of the adopted design using numerical simulations.
A key geotechnical issue is ensuring the stability of the works
A limited dynamic finite element analysis has been carried out during construction because it is often the most critical condition.
to simulate the effects of the storm on the behaviour of the Staged construction in an observational method framework was
caisson and its foundation using appropriate boundaries and adopted as the preferred approach. There was a major setback
element sizes. The dynamic analysis was performed in the time when most of the monitoring system was lost before the
domain, using the wave pressure records measured with the installation of the caissons thus endangering the construction
instrumentation of the caisson at different elevations during the approach that had been envisaged. It was decided, however, that
large December 2008 storm. The cyclic soil behaviour was staged construction would continue based mainly on the results
modelled by the UBC3D-PLM constitutive law combined with of the numerical simulations. The judgement underlying this
Rayleigh global damping. In order to keep the calculations within decision was based on the successful performance of the
feasible bounds, the analysis has covered only 600 seconds of the numerical analyses until that point, the good knowledge of the
storm history. behaviour of the foundation soil and the high uniformity of the
Figure 55 shows that analysis predicts that pore pressures are ground profile. Based on the results of a wide-ranging set of
generated in the foundation soil as observed in the field, but the analyses, it was decided to split the construction of the vertical
rate and magnitude of increase are much higher than measured. breakwater in two steps separated by a consolidation period; the
Contours of excess pore pressures computed at the end of the first step involved filling only 40% of the caisson cells. This
cyclic load analysis are shown in Figure 56. The analysis also strategy ensured the completion of the works in a reasonable time
yields a 0.47 Hz value for the dominant vibration frequency, a frame with, presumably, an adequate margin of safety throughout.
quite typical value for vertical breakwaters, and quite similar to A view of the completed South breakwater is presented in Figure
the 0.53 Hz value calculated in Madrid (2021) using a simplified 57.
analytical method (Oumeraci & Kortenhaus 1994). This A large degree of uncertainty and risk in vertical breakwaters
fundamental frequency is quite different from that of the storm is related to the effects of storm loading on the structure and the
action. foundation. A hydromechanical monitoring system installed in
one of the caissons has provided data on wave action at full scale.
In addition, it has been observed that a large storm is able to
generate pore pressures in the soft foundation soil but, below a
certain threshold, pore pressure remained steady throughout the
storm. Preliminary analyses suggest that major advances are still
needed for a good understanding and appropriate modelling of
storm loading on vertical breakwaters.
20
the main procedure to enhance the degree of safety of the quay
by way of simultaneously reducing the liquefaction potential of
the hydraulic fill and increasing the strength of the foundation
natural ground. The evaluation of the level of safety of the Phase
2 quay and the design of the preloading sequence, to avoid an
unintended instability, are presented in (6.5). Fulfilling the
construction of Phase 2 requires the use the observational method
underpinned by instrumentation data and appropriate numerical
analyses (6.6). Finally, the reconstruction of the failed Phase 1
quay is briefly described (6.7).
6.1 Basic features of the project
Figure57. Completed South breakwater.
Using again the nomenclature of Figure 11, the ground profile
along a section perpendicular to the Prat quay is shown in Figure
6 RECONSTRUCTION OF THE PRAT QUAY 59, where the location of the quay wall is indicated. The ground
profile is very similar along the full length of the quay. The
The first quay to be built in the new harbour area was the Prat upper sand layer corresponds to alluvial (Fc) and delta front (Df)
quay intended to accommodate a new automated container deposits; it becomes thinner and finally disappears altogether as
terminal. Its total length is 1580 m and construction was divided the distance to the shore increases. Again, the main foundation
in two phases: Phase 1, 1000 m long, and Phase 2, 580 long material is the prodelta soft silty clay (P) that extends down to
(Figure 58). 25 caissons constitute the retaining wall of Phase elevation -70 m approximately. In fact, the quay wall is founded
1 of the quay whereas 12 caissons were used for Phase 2. The directly on the soft silty clay because the upper sands were
surface area of the resulting operational zone is 0.95 km2. dredged away before construction. The prodelta deposit is
A major portion of the Phase 1 quay failed catastrophically underlain by gravels and sands (Bg-Fc) that completes the
due to flow liquefaction of the backfill. Although the essential Depositional Sequence I. The lower portion of the investigated
features of the failure and its subsequent investigation are ground profile is occupied by the succession of stiff clays (Rw)
described below, this section deals mainly with the works and sands and gravels (Bg-F) corresponding to Depositional
devoted to ensuring that the completion of the rest of the quay, Sequences II and III.
including the reconstruction of the failed section, had an
adequate degree of safety. Focus will be placed on a number of Fc Quay wall location
key points: Df
- Identification of the causes and mechanism of the Phase
1 quay failure
- Determining whether the remaining backfill is
P P
susceptible to liquefaction
- Modification of the design and/or construction procedure Bg-Fc
to ensure an adequate stability for the Phase 2 quay. Rw
- The monitoring system providing the bases for the
Bg-Fc
application of the observational method
- Verifying the effectiveness of the mitigating measures to Rw
reduce liquefaction potential Bg-Fc
- Reconstruction of the failed Phase 1 quay
Figure 59. Ground profile in the quay location and surrounding areas.
Cross section perpendicular to the quay alignment. For interpretation of
the labels, see Figure 11.
Design cross sections of the two Phases of the Prat quay are
shown in Figure 60. The caissons are 17.5 m high and 40 m long.
The width is 18.5 m in Phase 1 and 11.5 m in Phase 2. Phase 1
caissons were unusually wide because they were installed before
the South breakwater was completed and they were designed to
withstand a significant storm. Another difference between the
two Phases is the presence of a rockfill wedge at the back of
Phase 2 caissons that is absent in Phase 1. In contrast to the
design that we adopted for the breakwaters of limited dredging
and wide berms, the quay designers opted for a deeper dredging
and a narrower rubble mound, partially conditioned by the need
to provide the necessary draught.
The construction sequence involved a first stage of dredging
Figure 58. Location of Prat quay with Phases 1 and 2 indicated.
from an elevation of about -8 m down to elevation -24.5 m.
Afterwards, the rubble mound was built to elevation –16 m with
This section is organized as follows. After presenting some a rockfill protection on the sea side. The caisson was then
basic features of the project (6.1), the failure of a large section of installed and filled with sand. Time for some consolidation to
the Phase 1 quay is summarily described (6.2). The information take place was allowed between successive construction stages.
obtained in the post-failure site investigation (6.3) together with Finally, over a period of 6 months, the space behind the caisson
the performance of dedicated numerical analyses (6.4) provide was filled with rainbow-deposited hydraulic fill (Figure 61)
solid evidence on the causes and trigger mechanism of the using dredged material. In some locations, embankments were
collapse. After the failure, the stability of Phase 2 quay became built on top of the backfill to create enclosures for performing
an obvious concern because the same hydraulically-deposited preloading operations in a sequential manner. The measured
material had been used for backfill. Preloading was selected as settlements of the various caissons during this stage of the works
21
are plotted in Figure 62. The caissons of phase 1 settle somewhat
more because of their larger size. The very similar settlement
behaviour confirms the quite uniform foundation conditions
along the quay wall.
a)
a)
a)
b)
Figure 60. a) Phase 1 cross-section, b) Phase 2 cross-section.
b)
Figure 62. Settlements of the caissons before failure. a) Phase 1 (dashed
lines indicate the caissons involved in the failure, b) Phase 2.
Figure 61. Hydraulic filling of the space at the back of the Prat quay
caissons.
6.2 The failure Figure 63. View of the Prat quay (Phase 1) after the failure.
600 m of Prat quay Phase 1 failed catastrophically on the
morning of the 1st of January 2007 due to flow liquefaction of
the hydraulic fill. 15 caissons were affected, some of them
travelling long distances, up to 90 m, into the sea. Figures 63 and
64 show views of the quay after the failure; the distances
travelled by the individual caissons involved are plotted in Figure
65. Phase 2 of the quay did not fail and did not show any visible
signs of distress. At the time, two embankments (South and East)
South emb.
were being constructed by dumping material on the hydraulic fill
in order to demarcate a preloading area; they are also indicated
East emb.
in Figure 64. It should be noted that parts of the two Phase 2 Phase 1
embankments have disappeared in the failure.
600 m
90 m
Figure 64. Aerial view of the failure. The locations of the South and East
embankments are indicated.
22
Borehole drilling and CPTu and DMT tests were performed
along seven cross-section alignments, as illustrated in Figure 67;
both failed and non-failed zones were investigated. In total, 32
boreholes (with SPT determinations and sample extraction) were
drilled and 120 CPTu tests (including more than 500 dissipation
tests) and 7 DMT tests were performed. A prime objective of the
investigation was to identify and confirm the liquefiable nature
of the hydraulic fill. Only a few selected results are shown here
for illustration purposes. Figure 67 also shows the extent of the
South and East embankments at the time of failure.
Figure 65. Distances travelled by the caissons involved in the failure.
The state of the hydraulic fill backfilling at the moment of
failure is depicted in Figure 66. A 2.5 m earth bund had been
emplaced on top of the caisson allowing the hydraulic fill to be
raised above the caisson top level, in order to accelerate
consolidation. At that time, the hydraulic fill had reached an
elevation of about +3.0 m and the water on top of the fill was at
elevation +3.50 m. According to the subsequent investigation, it Figure 66. State of the caisson and the fill at the time of failure.
is likely that the triggering of the liquefaction was related to the Results of the site investigation carried out in cross-section
marginal stability of the South embankment. Release of air 14A, i.e. in the zone of maximum caisson displacement, are
trapped in the backfill may have also been associated with the
shown in Figure 68. The sandy deposits in the upper part of the
failure (Hight 2007); an eyewitness reported five or six air
natural ground profile in the land-side zones can be readily
releases in the discharge area minutes before the collapse. Initial
identified in the CPTu response. It should be noted that the upper
analyses showed that the margin of safety for a sliding failure
was very similar to that of a deeper foundation failure involving sands deposit is intensely interbedded with thin layers of more
the natural ground. However, underwater investigations carried clayey materials. The lower part of the ground profile is occupied
out after the collapse proved conclusively that the mechanism of by the prodelta soft silty clay throughout. The information
failure involved the caissons sliding on the top surface of the obtained from in situ and laboratory tests on the natural ground
rubble mound. Analyses presented later have been able to silty clay was remarkably consistent with that obtained in the
reproduce this failure mechanism in a consistent manner. breakwater investigation and will not be presented again here.
Immediately after the failure a full programme of site All three boreholes drilled in the zone formerly occupied by the
investigation and analyses was undertaken, as summarised below. caisson encountered the rubble mound practically intact. The
More information on the failure is provided in Gens (2019) and hydraulic fill found around the area where the caisson had been
Tarragó (2021). previously located corresponds to the liquefied fill transported
during the failure. Therefore, this cross-section cannot be used
6.3 Post-failure site investigation for characterizing the fill material in it is initial state. Information
on the state and properties of the hydraulic fill before failure must
A large-scale site investigation was undertaken after the failure rely on data provided by the investigation of the remaining fill in
to obtain information on the causes of the collapse and to supply
Phase 1, in cross-sections not affected by the failure, and of the
data to complete the construction of the Prat quay in a safe
fill in Phase 2 that had been emplaced in the same way.
manner. Both Phase 1 and Phase 2 sections were investigated.
Borehole
CPT u
Pre-bored CPT u
DMT
PHASE 1
PHASE 2
CS – 14 A CS – 22 A
CS – 7 B CS – 3 B CS – 6 A CS – 12 A 16
21
16 20 15
16 16 13, 14
15 15 15, 14 19, 18
13 12
14 14 17
13, 12 13, 12 12 16
11 11 11 15 8
10 10 10 8, 9 14 n5
9 9 9 7 13 7, 6 8
8 8 8 6 12 n4 7
7 7 7 5 11 5 6
6 6 6 4 10 16a 5
5 5 5 3 9 3 4
4 4 4 2 8 2 3
3 3 3 7
2 2 2 1 6 1
1 1 1 5
CS – 16 A
1
Figure 67. Post-failure site investigation.
23
HYDRAULIC FILL CAISSON 14a
(after failure) SILTY CLAY SANDS
LOCATION
Figure 68. Results of the site investigation in cross-section 14A (maximum caisson displacement).
HYDRAULIC FILL
(after failure) HYDRAULIC FILL SILTY CLAY SANDS
Figure 69. Results of the site investigation in cross-section 6A (non-failed section).
Results of the site investigation in a non-failed section 6A are By virtue of its mode of deposition, the fill exhibited a degree
shown in Figure 69. Again, the different CPTu response of of heterogeneity; its grading depends on the location of the
hydraulic fill, upper sands and soft silty clay can be easily discharge point and the distance travelled by the soil before
distinguished. Regarding the hydraulic fill, it is apparent that the sedimentation. As an example, the particle size distributions of
CPTu tests yield very low cone resistance in practically all hydraulic fill samples extracted from a single borehole, 6a08, are
locations and the contrast with the natural ground is readily plotted in Figure 71 where it can be observed that the fine content
apparent. It should be noted that the hydraulic fill at the toe of is high but variable, ranging from 40% to 90%. In the same
the caisson on the sea side is part of the backfill spilled in the borehole, clay content was in the range of 20%-40%.
failure. A typical CPTu result (CPTu 6a-12) can be seen in more Considering the ensemble of results over the full quay site, it
detail in Figure 70a. By way of comparison, Figure 70b shows appears that, overall, Phase 2 backfill contained a somewhat
the CPTu test 3b-11 corresponding to the Phase 2 investigation; larger proportion of more granular material. It was not possible
a quite similar behaviour is obtained. to extract samples from the hydraulic fill undisturbed enough for
24
0 0
reliable laboratory testing and the material has been
characterized by field in situ tests (DMT and, especially, CPTu). 0.25 · '
v
0.12 · '
2 v 2
0.05 · '
v
u2 (MPa) u2 (MPa)
4 4
depth (m)
depth (m)
(m)
6 6
depth
8 8
depth (m)
depth (m)
10 10
12 12
0 10 20 30 0 0 0,5 1 1,5
ssu (kPa)
(kPa) B
u q
a) b)
Figure 72. Test CPTu-6a12. a) Distribution of undrained shear strength
with depth obtained from CPTu tests, B) Distribution of pore pressure
qc (MPa) qc (MPa) coefficient with depth.
a) b) The assessment of a potentially liquefiable material is more
Figure 70. Cone resistance and pore pressure response obtained in CPTu often performed using empirical charts based on extensive
tests. a) CPTu 6a-12, b) CPTu 3b-11. databases. In Figure 73, the results of the typical CPTu tests, 6a12
(Phase 1) and 3b11 (Phase 2), are plotted in the SBT chart
100 proposed by Robertson (2016). As expected, there is a significant
span of results ranging from sand-like to clay-like materials but
practically all points indicate contractive behaviour and therefore
Percentage passing (%)
80
with a potential for flow liquefaction, especially those points
lying in the CCS (Clay Contractive Sensitive) region. Similar
60 evidence of liquefiable fill material has been obtained (Tarragó
2021) using alternative liquefaction criteria (e.g. Plewes et al.
1992, Mayne and Sharp 2019). It was therefore confirmed that
40
liquefiable hydraulic fill was present in in the areas of Phase 1
not affected by the failure and in the Phase 2 backfill. The post-
20 liquefaction strength was estimated to be suliq = 0.06 ’v0 (Olson
& Stark 2002, Shuttle and Cunning 2007, Robertson 2021). This
0,01 0,1 1 10 100 value appears consistent with the observed sleeve fractions that
Particle size (mm) have been postulated (Lunne et al. 1997) as a potentially good
Figure 71. Particle size distributions of samples retrieved from borehole estimator for the residual undrained shear strength (Figure 74).
6a8.
1000 1000
I =32
The development of computational capabilities to simulate I =32
B
B
flow liquefaction (outlined in the next section) has allowed the I =22 I =22
B
SD B SD
detailed numerical examination of cone penetration tests in 100 CD=70 100 CD=70
brittle (and therefore liquefiable) materials. In addition to a
Qtn
QQ tn
TD SC TD
constitutive law tailored to reproduce liquefaction phenomena,
tn
SC CD
tn
CD
Q
TC TC
this advanced numerical modelling incorporates a Particle Finite 10 10
Element (PFEM) framework to deal with large displacement CC CCS CC
CCS
problems and a non-local formulation to account for softening
behaviour (Monforte, et al. 2018, Manica et al. 2018). Research 1 1
has identified some features of the cone penetration response that 0,1 1 10 0,1 1 10
F F
can be associated with brittle liquefiable material (Monforte et al. r r
2021, 2022) such as a strong pore pressure response (reflected in a) b)
a high value of the pore pressure coefficient, Bq) and in a low
Figure 73. Observations in two CPTu tests plotted in Robertson (2016)
value of undrained shear strength (or, equivalently, a low value SBT charts: a) CPTu 6a12 (Phase 1); b) CPTu 3b11 (Phase 2).
of the Nkt peak factor) where = / and = / ;
is the pore pressure increment measured at the cone shoulder, 6.4 Post-failure analysis
is the peak undrained shear strength and is the net cone
resistance. As Figure 72 shows. the typical CPTu 6a-12 response Understanding of the failure is enhanced if a numerical
exhibits the two characteristic features of brittle behaviour; simulation incorporating the actual field conditions is capable to
computed undrained shear strength (Nkt=15) is often below the capture the main features of the event. The failure spurred
usual normally consolidated range of 0.22-0.25 v0 and Bq research to try to model realistically flow liquefaction. An
important component of such capability is the adoption of
values are high, even larger than 1 on occasions. The same type
constitutive model capable to reproduce such phenomenon. A
of response was obtained across the site. Undrained shear
modified CASM constitutive law (Yu 1998, 2006, González 2011,
strength derived from DMT (Marchetti et al 2001, Marchetti Yu et al. 2019) has proved successful in this regard (Gens 2019,
2015) also yielded values below the normally consolidated range. Arroyo & Gens 2021, Manica et al 2022). The model has been
selected because i) it is relatively simple, using a single yield
surface, ii) it is based on the state parameter concept (Been &
Jefferies 1985), iii) it uses non-associated plasticity allowing for
25
instability phenomena to be simulated, and iv) it can reproduce
classical critical state models if appropriate parameters are
selected. As illustrated in Figure 75, the model is capable of
encompass a wide range of soil behaviour, including brittle
undrained softening.
0 0
CPTu 6a12 CPTu 3b11
fs fs
0.25 · 'V
0.25 · 'V
0.12 · 'V
5 5 0.12 · 'V
0.05 · 'V
0.05 · 'V
(m)
(m)
depth (m)
10
depth (m)
10
depth
depth
15 15
20 20
0 0,025 0,05 0,075 0,1 0 0,025 0,05 0,075 0,1
fs (MPa) fs (MPa)
fs (MPa) fs (MPa)
a) b)
Figure 74. Distribution of sleeve friction, ff, with depth obtained from
CPTu tests. a) CPTu 6a-12. b) CPTu 3b-11. Figure 76. Succession of computed displacements contours during
failure.
Figure 77. Caisson locations after the failure from the PFEM analysis
Figure 75. Range of soil behaviour simulated by the CASM model (Yu (Celigueta et al. 2007).
1998). Effect of state parameter on undrained triaxial stress paths.
A coupled hydromechanical analysis was performed taking 6.5 Enhancing Phase 2 quay stability after failure
into account all the stages of construction and their duration. Understandably, the evaluation of the safety of the Phase 2 quay
After reaching the situation just before failure, a small zone at was a priority after failure. Following the same methodology to
the location of the tip of the South embankment being relate reliability index (i.e. probability of failure) with factors of
constructed at the time of failure was made to liquefy by safety, the Spanish code ROM 0.5-05 (Puertos del Estado, 2008)
modifying appropriately the CASM model parameters. Figure 76 recommends, for gravity quay walls with a low SERI index (5-
illustrates how liquefaction propagated from that zone outwards 19), the factors of safety values listed in Table 7. Again, global
until a complete failure of the quay ensued. The final mechanism stability is the critical failure mode in practically all cases. In this
involves sliding on the rubble mound with a localised bearing section, all the computed factors of safety will be defined in
failure at the seaward toe, a feature also observed in the field terms of undrained strength reduction, where most of the
investigation. The subsequent displacements of the caissons have uncertainty lies in this case.
been modelled with a 3D Particle Finite Element Analysis The situation of the Phase 2 quay section at the moment of the
(PFEM) that assumes that the initial focal point of liquefaction failure is depicted in Figure 78. The hydraulic fill had reached
propagation is at the tip of the South embankment followed by a elevation +1.0 m against the caisson and the water level was at
subsequent one at the tip of the East embankment. Because of the +2.0 m elevation. If an analysis similar to that performed for
complexity of the 3D analyses, a simplified model for the Phase 1 but without a trigger event was carried out, the Phase 2
liquefaction and its propagation has been employed (Celigueta el quay sections proved stable (as observed) with a factor of safety
al. 2007). As Figure 77 shows, the movements of the different against reduction of peak undrained shear strength of about 1.25.
caissons involved in the failure are quite well reproduced by the However, the hydraulic fill was obviously liquefiable and it is
analyses. The results of this set of analyses strongly suggest that difficult to ensure completely that no triggering event will ever
the enclosure embankments were involved in the triggering of take place. Thus, common practice requires to evaluate the safety
the failure. of the structure assuming liquefaction of the hydraulic fill, i.e
considering that the hydraulic fill is at residual (liquefied)
undrained strength (Schnaid et al. 2020). In this case, the
26
computed factor of safety has a much lower value of 1.08. It is reducing significantly, the liquefaction potential in this particular
intersting to note that he failure mechanism obtained for this case fill. The results were encouraging.
is no longer by caisson sliding but a general failure involving the
natural foundation ground (Figure 79).
Table 7. Gravity quay walls with low SERI (5-19). Factors of safety
according to ROM 0.5-05 (Puertos del Estado 2008).
Load combination
Quasi - Fundamental/ Accidental
Ultimate limit state
permanent Characteristic
FS1/CF1 FS2/CF2 FS2a/CF2a
Sliding 1.5 1.3 1.1
Bearing capacity 2.5 2.0 1.8
Plastic overturning 1.5 1.3 1.1
Global stability 1.4 1.3 1.1
Figure 80. Vertical wick drain installation.
Prior to the preload installation, a 50 cm-thick drainage layer
was installed throughout to ensure an appropriate discharge of
the water produced by consolidation. The preload design
distinguished 3 different zones: deposit zone, manoeuvre zone
and caisson (Figure 81). The manoeuvre zone extends 70 m from
the back edge of the caisson landwards. The three zones could
not be preloaded simultaneously because of stability concerns;
embankment placement followed the rather intricate sequence
shown in Table 8. As a matter of fact, preloading the caisson itself
Figure 78. Cross section of Phase 2 quay at the moment of the failure of
may be considered hazardous given the precedents and the nature
Phase 1 quay. of the foundation soil but it turned out to be necessary to increase
the undrained shear strength of the soft silty clay below the
caisson (Figure 82). It was also needed to increase the safety
against sliding at some stages of preloading. As described in the
next section, the consolidation time required between preload
stages was decided, following the observational method, on the
bases of the evolution of the instrumentation data, giving special
consideration to the progress of pore pressures dissipation.
Vertical drains Terrestrial fill Gravel columns
Figure 79. Contours of incremental displacements illustrating the failure
mechanism for Phase 2 quay, assuming hydraulic fill liquefaction.
In any case, immediately after the failure, the top of the Sands
hydraulic fill was lowered by excavation to elevation -2.0 m.
Afterwards, the backfill was raised to elevation +1.5 m by Hydraulic
dumping mainly sandy material by terrestrial means; no more Silty clay fill
hydraulic fill was emplaced. However, a large amount of
Gravel
potentially liquefiable material remained on site. If no further
action was taken, the evaluation of the safety under service loads Figure 81. Phase 2 quay cross-section incorporating the measures for
revealed that the quay might not be stable (factor of safety < 1) stability enhancement.
under service loads if the hydraulic fill liquefied. The predicted A complementary stability enhancement was achieved by the
failure mechanism would be very similar to that of Figure 79 installation of 0.5 m diameter gravel (stone) columns by vibro-
involving both the backfill and the natural ground below and replacement in a 20-m wide area adjacent to the caissons (Figure
beyond the caisson. As a matter of fact, even without assuming 81). They were up to 30 m long and set in a triangular pattern
hydraulic fill liquefaction, the factor of safety under service loads with a 2 meter-spacing (Figure 83). Gravel columns provide
had a quite low 1.04 value. Therefore, it was necessary to achieve soil reinforcement and mitigate the fill liquefaction potential
a twofold soil improvement: to reduce or eliminate the (Baez 1995, Mahoney & Kupec 2014). They were installed
liquefaction potential of the remaining hydraulic fill and to
during preload stages 2 and 3. The length of some columns was
increase the undrained shear strength of the foundation natural limited to 25 m because of concerns regarding transient high
ground in the areas potentially affected by the failure. water pressures in the foundation during installation.
Preloading was selected as the primary means to achieve these
two soil improvement goals (e.g. Stamatopoulos et al. 2005, 2013, Table 8. Sequence of preloading stages for completing Phase 2 quay.
Raptakis 2013); 30-m long vertical wick drains in a 1.5 m square
Preload embankment elevation
pattern were installed to ensure an early effectiveness of the
intervention (Figure 80). Preloading required however careful Stage Deposit zone Manoeuvre zone Caisson
management because, in the short term, a preload will tend to Initial (I) + 2.0 m +2.0 m +1.5 m
destabilize the quay. Vacuum preloading (Indraratna 2017)
1 + 7.0 m +2.0 m +1.5 m
would sidestep this problem but it was discarded because of
2 +10.5 m +2.0m +4.5 m
logistic difficulties and time constraints. In this context, a key
preliminary step was to check, by means of a full-scale trial, 3 +10.5 m +2.0m +5.0 m
whether preloading would be effective in removing, or at least 4 +10.5 m +2.0 m +3.0 m
27
5 +10.5 m +5.0 m +3.0 m -32 m) because they are in vicinity of the potential sliding surface.
6 +10.5 m +6.5 m +3.0 m An additional chain of pore pressure sensors was installed in the
7 +2.0 m +6,5 m +3.0 m rockfill adjacent to the caisson to warn of possible water pressure
impacts due to the execution of gravel columns. Obviously, the
Final (F) +2.0 m +2.0 m +1.5 m
measurement of the caisson settlements by topographic means
was continued and supplemented by a clinometer attached to the
sea face of the caisson. The interpretation of the instrumentation
observations was aided by the simultaneous performance of
numerical analyses that simulated the same construction and
preloading stages as in the field. By incorporating the monitoring
data, the analyses were also used to evaluate the safety at every
stage of the works and were used to decide the time when it was
possible to proceed to the next stage.
The set of observations corresponding to the piezometer
sensors located at elevations -27 m and -32 m during the various
stages of preloading for all the instrumented caissons is shown in
Figure 85. Note that the times of the different preload stages
showed in the Figure can only be indicative because, of course,
the emplacement of the preload was progressive and not
simultaneous in all the caissons. A different view of the pore
pressure state of the foundation is presented in Figure 86 where
the distribution of excess pore pressures with depth for the
piezometer chains located in the Caisson 7b cross-section are
plotted for preload stages 6 and 7, when stability was more
critical. The results of the associated numerical analyses are also
plotted; it can be noted that there is a good correspondence at this
stage, the differences between observations and calculations are
only a few kPa. In all cases, the maximum excess pore pressures
Figure 82. Caisson preloading. appear to occur at elevation –40 m.
Figure 83. Installation of gravel columns.
6.6 Completing the construction of Phase 2 quay
Prior to the failure, only the settlements of the caissons had been
monitored showing vertical displacements in the range of 50-60 Figure 84. Instrumentation in the caisson zone.
cm (Figure 62). After the failure, an extensive monitoring system
was installed in Phase 2 quay, as required by the use of the
observational method for the completion of the works. As soon
as the esplanade became accessible, an extensive instrumentation
system was installed consisting of sliding micrometres,
piezometers and settlement plates distributed across the site.
However, for the control of stability, the relevant instrumentation
was that deployed in the caisson and adjacent areas (Figure 84);
every other caisson was monitored in this manner.
The instrumentation for one caisson included an inclinometer
reaching down to elevation -50 m to observe the horizontal
displacements at different elevations under the caisson. Given the
importance of monitoring pore pressure dissipation, three chains
of piezometers were installed, one under the caisson and two
more under the berm on the sea side, at 4 and 14 m from the
caisson sea face. One of the piezometer chains was taken to
elevation -60 m while the other two reached elevation -40 m.
Naturally, regarding stability concerns, the most important pore
pressure readings are the two upper ones (at elevations -27 m and
28
and consolidation times decided on the bases of monitoring
Gravel Stage 2
Gravel Stage 3
Stage 4
Stage 5
Stage 6
Stage 7
readings and analysis predictions. The duration times listed are
typical values as there were some minor variations between
different parts of the works. It can be observed that the factors of
columns
columns
safety for the characteristic load combination (CF2) are always
adequate although in stage 6 (full preloading of the manoeuvre
zone) a somewhat smaller factor of safety (1.24) is obtained.
Work was allowed to proceed anyway given the close watch of
the works and the limited duration of the stage. Analyses
indicated that without the gravel columns the factor of safety
would have reduced to 1.17 at this stage. Table 9 also presents
the factors of safety for accidental load combinations that, in this
case, corresponds to the possibility of hydraulic fill liquefaction.
Again, the most critical situation is stage 6 when the computed
factor of safety is 1.14. The calculations for accidental conditions
assumed, conservatively, that the hydraulic fill liquefaction
potential only disappeared at the end of the preloading, this is
why no CF2a Factors of safety are computed from stage 7
onwards. The failure mechanisms obtained for the critical stage
6 are shown in Figure 87. It can be noted that they involve the
soft silty clay in the foundation.
a)
Stage 2
Stage 3
Stage 4
Stage 5
Stage 6
Stage 7
Table 9. Factors of safety and duration of the preloading stages for
completing Phase 2 quay.
Time (days) Factors of safety
columns
columns
Gravel
Gravel
Construction /
Stage Consolidation CF2 CF2a
Removal
1 20 45 1.42 1.34
2 35 45 1.42 1.34
3 4 140 1.43 1.35
4 10 50 1.53 1.48
5 10 15 1.38 1.20
6 10 15 1.24 1.14
7 60 60 1.37 -
Final 30 1.54 -
b)
Figure 85. Evolution of excess pore pressures below all instrumented
caissons. a) elevation – 27 m, b) elevation -32 m. Approximate dates of
the start of the preload stages are indicated.
a)
b)
Figure 87. Stage 6 failure mechanisms identified in the numerical
analyses. a) characteristic case, b) accidental (liquefied) case.
Settlements continued to be monitored throughout and their
Figure 86. Observed and computed distribution of excess pore pressure
in the piezometers corresponding to Caisson 7b cross-section. Preload evolution, reflecting the various stages of preloading,
stages 6 and 7. corresponded well with the results of the analyses (Figure 88).
However, from the perspective of stability, horizontal
The computed factors of safety corresponding to the different
preload stages are listed in Table 9 together with the construction
29
movements are more relevant 1. As an example, the distribution that the reduction of liquefaction potential was especially marked
of horizontal displacements under caisson 7b for the critical in those zones where the liquefaction potential was higher before
preload stages 5 and 6 and subsequent (stage 7) consolidation are preloading. The global reduction of liquefaction potential is also
shown in Figure 89. It was reassuring that the model results demonstrated in Figure 92 where it can be noted that the
agreed well with the observations. contractive fraction of the hydraulic fill has reduced considerably
and that the points in the CCS sector have practically disappeared.
Stage 1
Stage 3
Stage 5
Stage 6
Figure 88. Evolution of the measured and computed settlement of Phase
2 caissons.
Figure 90. Evolution of horizontal and vertical displacements in caisson
7b compared with the results of the numerical simulation.
Once preloading was completed and the other soil
improvement measures were in place, construction activities
could continue by installing all the civil infrastructure and
equipment required to operate the container terminal. At the same
time, the hydraulic fill spilled by the failure and present in front
of the quay was dredged away. The computed factor of safety
under service loads fell in the range 1.28-1.36 depending on the
loading case considered; the associated failure mechanism is
shown in Figure 93. It can be observed that it affects the soft
foundation ground and that it does not extend much into the now
improved esplanade behind the caisson.
Figure 89. Distribution of horizontal displacements with depth under
caisson 7b. The results of the numerical simulation are shown for
comparison.
6.7 Reconstruction of Phase 1 quay
The reconstruction of Phase 1 quay was less challenging from a
However, it is well established in soft soil engineering that the geotechnical point of view. Taking advantage of the fact that a
relationship between horizontal and vertical movements is a large amount of the hydraulic fill had spilled out during the
more meaningful indicator of potential stability problems than failure and that some of the material still remaining was dredged
their absolute values (e.g. Tavenas et al. 1978). A sharp increase away, a new backfill with better geotechnical properties was
of horizontal displacements compared to the vertical emplaced. Typical cross-sections of the reconstructed quay are
displacements often flags an approaching failure. Such a depicted in Figure 94. A rockfill wedge against the caisson was
relationship was closely watched during the works and is plotted first installed along the entire Phase 1 caissons. In the failed zone
in Figure 90 for caisson 7b. It can be noted that no acceleration
where no hydraulic fill remained, all the backfill installed used
of horizontal movements is observed at any stage and that, in
selected material (Figure 94a). Where some hydraulic fill was
addition, there is a good correspondence with the numerical
analyses that anticipated a sufficient degree of stability still present in the non-failed zone, the selected backfill was
throughout. In fact, the model shows, if anything, a slight limited to a length of 30 m adjacent to the rockfill wedge (Figure
overestimation of the ratio between horizontal and vertical 94b). Eventually, only a very limited amount of hydraulic fill was
displacements. present in Phase 1 quay and, in addition, preloading was also
A final important check was necessary to confirm the stability applied with drainage assisted by the installation of 30 m-long
evaluation, namely, to determine whether the application of a wick drains. The CPTu results shown in Figure 95 indicate that,
preload had effectively rendered the hydraulic fill non- as expected, the new emplaced fill material was mostly sand-like
liquefiable, as assumed in the calculations. To this end an and always dilatant and, therefore, not susceptible to liquefaction.
extensive camping of 46 CPTu was performed and were The margin of safety was ample throughout.
compared with the results of a previous campaign involving 48
CPTu tests carried out before the preload. A typical result is
shown in Figure 91 where the normalised CPTu parameters are
shown for two piezocone tests performed at the same location
before and after preloading. The improvement in terms of cone
resistance and pore pressure response is apparent. It was found
1
In this context, it was interesting to learn that, in the 1940’s, Terzaghi indicator of approaching failure for the foundation of a large iron ore
had insisted on the measurement of horizontal displacements, by the pile in Cleveland (Goodman 1998)
rather rudimentary means then available, as the most discriminating
30
elevation (m)
elevation (m)
elevation (m)
Qt Fr Bq
a)
a)
elevation (m)
elevation (m)
elevation (m)
Qt Fr Bq
b)
Figure 91. Normalised CPTu parameters. a) CPTu-R16 (before
preloading), b) CPTu-67 (after preloading). b)
Figure 94. Typical cross-sections of the reconstructed Phase 1 quay. a)
1000
1000
I =32 I =32
B
Failed zone, b) Non-failed zone.
B
I =22 I =22
B
SD B SD
100 CD=70 100 CD=70
TD TD
Qtn
QQtn
CD
tn
CD
tn
SC SC
TC TC
10 10
CCS CC CCS CC
1 1
0,1 1 10 0,1 1 10
FF
r F
F
r
a) b)
Figure 92. Observations of CPTu tests performed in Phase 2 quay plotted
in Robertson (2016) SBT charts: a) before preloading; b) after preloading.
Figure 95. CPTu tests performed in Phase 1 quay plotted in Robertson
(2016) SBT charts.
6.8 Summary notes
Static (flow) liquefaction of the hydraulic fill used to backfill the
space behind the quay wall caissons caused the catastrophic
failure of a large part of the Phase 1 quay. Liquefaction was
probably triggered by the instability of an enclosure embankment
being constructed at the time. The post-failure site investigation
and associated numerical analyses have provided strong
Figure 93. Failure mechanism under service loads identified in the supporting evidence for the role of liquefaction in the failure and
numerical analyses.
for the mechanism of collapse. The failure was by sliding of the
It is interesting to record that 11 (out of 16) of the displaced caissons on the underlying rubble mound although a deeper
caisson could eventually be re-used; they were successfully re- failure involving the foundation soft silty clay had a very similar
floated and placed back in its initial location and filled again with degree of stability at the time.
sand (Figure 96). The remaining 5 caissons were demolished, The post-failure site investigation confirmed the liquefiable
removed from the site, and replaced with new caissons. A view nature of the Phase 2 backfill. Thus, the increase of the safety
of the completed Part quay before the installation of the civil margin of the Phase 2 quay posed some significant challenges in
infrastructure is shown in Figure 97. view of its presumably marginal stability; analyses showed that
it was necessary to mitigate the liquefaction potential of the
31
hydraulic fill and, at the same time, increase the strength of the
foundation soft silty clay. Preloading was the selected procedure
for soil improvement complemented with gravel columns in a
limited area. Special care had to be exercised during the
deployment of the preload because, in the short term, it is a
destabilizing measure. The observational method was a key
feature to decide the sequence and duration of the various
preloading stages. This time, a full monitoring system was
operational and analysis could return to its intended supporting
role. An extensive campaign of CPTu tests performed before and
after preloading confirmed the effectiveness of the measure to
reduce the liquefaction potential of the fill.
The reconstruction of the Phase 1 quay was more
straightforward as most of the liquefiable hydraulic fill had been
removed by the failure or dredged afterwards. The application of
a preload similar to that of Phase 2 and the fact that the newly- Figure 98. Prat quay container terminal in operation (courtesy of the
emplaced backfill was non-liquefiable provided an ample degree Autoritat Portuaria de Barcelona, APB).
of safety 2. Thus, the construction and commissioning of the Prat
quay could be accomplished and the container terminal became 7 CONCLUDING REMARKS
fully operational shortly afterwards (Figure 98).
In accordance with the ISSMGE’s guidelines, the paper focuses
on two case histories involving harbour works in a deltaic
environment. Deltas provide a fascinating but challenging
geological framework where the nature and distribution and the
materials are a direct product of recent climate, geological and
human history. The Llobregat delta is an especially well-
preserved one and it has proved possible to reconstruct its
evolution over several glaciation cycles. In the case of the
harbour works, the main foundation material was a prodelta
deposit of normally consolidated soft silty clay; so, stability
issues were dominant.
The first case involves the manifold behaviour of a
breakwater where structural, hydraulic and geotechnical aspects
interact. The random nature of the storm wave loading requires
special consideration as well as the incorporation, as far as
possible, of a more rational approach to the evaluation of safety.
It was planned to use the observational method to guide the
construction of the breakwater assisted by suitably calibrated
numerical analyses. Unfortunately, due to circumstances beyond
Figure 96. Refloating one of the caissons involved in the failure. control, the monitoring system largely became non-operational
at an early stage of the works, before the installation of the
caissons. The decision was made to proceed with the works, with
numerical analyses taking up the primary governing role. The
decision was backed by the fact that the behaviour of the silty
clay was relatively well-known, the numerical analyses had
shown a good correspondence with observations in the early part
of construction and the site investigation had indicated a large
degree of ground profile uniformity along the breakwater. In
spite of these favourable circumstances, it was ultimately a
judgment call, albeit supported by a good level of information.
There is still insufficient understanding of the effects of storm
wave loading on caissons and, especially, on the foundation.
Instrumentation installed in the vertical breakwater is providing
valuable information in this regard showing, for instance, that
pore pressures develop in the foundation clay only when a storm
exceeds a certain threshold. In any case, significant advances are
Figure 97. Prat quay at the end of the reconstruction. still needed for a good understanding and appropriate modelling
of storm loading on vertical breakwaters.
A major failure occurred during the construction of a quay due
to the flow liquefaction of the hydraulically-deposited backfill.
This required the evaluation and enhancement of the stability of
the remaining non-failed quay. Investigations after the failure
confirmed that much of the remaining backfill was also
potentially liquefiable. Analyses showed that, to complete
construction, it was necessary to mitigate the liquefaction
2 This is reminiscent of a major consulting job carried out by Terzaghi dangerous soft material had flowed away during the failure. Indeed, he
in the 1940’s concerning Necaxa dam in Mexico that had failed some concluded that the dam could be raised safely three additional meters;
years before due the liquefaction of the hydraulic fill (Goodman 1998). something that was promptly done by the owners.
Terzaghi pronounced the reconstructed dam safe because the
32
potential of the hydraulic fill and to improve the strength of the Geotechnical Aspects of Underground Construction in Soft
foundation clay. Ground, Viggiani G. (ed.), Taylor & Francis, 181-188
Preloading was the primary ground improvement method Baez, J.I. 1995. A design model for the reduction of soil liquefaction by
selected to achieve those goals mainly because it is a procedure vibro-stone columns. Ph.D. Thesis, University of Souhern
California, Los Angeles.
firmly grounded on well-established soil mechanics principles. It Beaty, M. H., and Byrne P. M. 2011. UBCSAND constitutive model
was again a judgement decision that was further endorsed by a version 904aR. Documentation report.
preload test confirming the effectiveness of the method to https://www.itascacg.com/software/udm-library
mitigate liquefaction. The implementation of the preloading Been, K., Jefferies, M. G. 1985. A state parameter for sands.
required a careful sequence of load application in order not to Géotechnique 35, 99–112.
jeopardise the stability of the quay that could be compromised, Boyer, J., Duvail, C., Le Strat, P., Gensous, B. and Tesson, M. 2005.
in the short term, by loading on and in the vicinity of the caisson. High resolution stratigraphy and evolution of the Rhône delta plain
The observational method (with associated numerical analyses) during Postglacial time, from subsurface drilling data bank.
was the primary tool to steer the various stages of construction, Marine Geology
Brinkgreve, R. B. J., Kumarswamy, S. and Swolfs, W. M. (2017).
this time adequately supported by a field monitoring system that Plaxis 2D 2017 Manual, Plaxis bv, Delft, Netherlands
operated throughout. Burland. J.B. 2017. Terzaghi: back to the future. Opening address to the
Looking back on the geotechnical engineering activities 2006 IAEG conference. Bull Eng Geol Environ, 66:29–33
related to these two case histories, it can be recognized that the DOI 10.1007/s10064-006-0083-9
essential core of the work involves geology, soil mechanics, the Cattaneo, A., Correggiari, A., Langone, L. and Trincardi, F. 2003. The
observational method and judgement. Thus, it may perhaps be
considered an appropriate tribute to Terzaghi’s legacy. pathways and supply fluctuations. Marine Geol.
Catunean u, O. 2002 Sequence Stratigraphy of Clastic Systems:
Concepts, Merits, and Pitfalls. Journal of African Earth Sciences,
8 ACKNOWLEDGEMENTS 35, 1-43. http://dx.doi.org/10.1016/S0899-5362(02)00004-0
Celigueta, M.A., Oñate, E., Suárez, B. 2007. Trabajos realizados por
The author thanks President Charles Ng for inviting him to CIMNE en relación con la rotura del muelle del Prat del Puerto de
become the 10th Terzaghi Orator. Much of the work presented is Barcelona. Informe Técnico CIMNE IT-514. CIMNE, Barcelona.
Checa, A., Díaz, J.I., Farrán, M., Maldonado, A., 1988. Sistemas
the result of joint efforts with numerous colleagues of the author, deltaicos holocenos de los ríos Llobregat, Besós y Foix: modelos
at his Department and elsewhere. The contributions of M. evolutivos transgresivos. Acta Geologica Hispánica, 23, 241–255.
Arroyo, A. Deu, J. González, N.A. González, R. Griell, D.W. Christian J.T. and Baecher G.B. 2011. Unresolved Problems in
Hight, A. Lloret, M. A. Mánica, L. Monforte, M.A. Pindado, J.M. Geotechnical Risk and Reliability. State-of-the-Art. ASCE Georisk
Strout, J. Vaunat and M.T. Yubero are gratefully acknowledged. 2011, Atlanta. 50-63. https://doi.org/10.1061/41183(418)3
Particular recognition is due to E.E. Alonso, J.L. Estrada, A. Duncan, J. M. 2000. Factors of Safety and Reliability in Geotechnical
Soriano† and J. Uzcanga, members of the design and consultant Engineering. Journal of Geotechnical and Geoenvironmental
teams of the harbour works. Very special thanks are given to my Engineering, 126 (4), 307–316.
former PhD students, Ricardo Madrid and Dani Tarragó, who doi:10.1061/(ASCE)1090-0241(2000)126:4(307)
Einstein, H.H. 1991. Observation, quantification, and judgement:
were closely involved in the two case histories reported. The key Terzaghi and Engineering Geology. Journal of Geotechnical
support of the Barcelona Port Authority is greatly appreciated. Engineering, ASCE, 117(11), 1772-1778.
Frigola, J., Moreno, A., Cacho, I., Canals, M., Sierro, F.J., Flores, J.A.,
Grimalt, J.O., Hodell, D.A. and Curtis, J.H., 2007. Holocene climate
9 REFERENCES variability in the western Mediterranean region from a deep water
sediment record. Paleoceanography, 22 (2), PA2209.
Allsop, N.W.H., Vicinanza, D and McKenna, J.E. 1996. Forces on doi:10.1029/2006PA001307
vertical breakwaters and related structures. Strategic Research Gámez, D., 2007. Sequence Stratigraphy as a tool for water resources
Report SR 443, HR Wallingford, Wallingford, UK management in alluvial coastal aquifers: application to the
Alonso, E.E. and Gens, A. 1999. Geotechnical design and construction Llobregat delta (Barcelona, Spain), PhD Thesis, Universitat
of breakwaters, Bilbao Harbour. Proc. 12th European Conf. Soil Politecnica de Catalunya, Barcelona (http://www.tdx.cesca.es/).
Mech, Geotech. Engng (Barends et al., eds), Balkema, Rotterdam, Gámez, D., Simó, J.A., Lobo, F.J., Barnolas, A., Carrera, J. and Vázquez-
489-508. Suñé, E., 2009. Onshore-offshore correlation of the Llobregat deltaic
Alonso, E.E., Gens, A. and Lloret, A. 2000. Precompression design for system, Spain: Development of deltaic geometries under different
secondary settlement reduction. Géotechnique, 50, 645-656 relative sea-level and growth fault influences. Sedimentary Geology,
Amorosi, A. and Milli, S. 2001. Late Quaternary depositional 217, 65-84.
architecture of Po and Tevere river deltas (Italy) and worldwide Gens, A. 2019. Hydraulic fills with special focus on liquefaction.
comparison with coeval deltaic successions. Sedimentary Geology, Geotechnical Engineering foundation of the future. Proceedings of
the XVII Europ. Conf. Soil Mech, Geotech. Engng. IGS, Reykjavik.
Andersen, K.H. 2009. Bearing capacity under cyclic loading — 52-82. doi: 10.32075/17ECSMGE-2019-1108
offshore, along the coast, and on land. The 21st Bjerrum Lecture. Gens, A., Di Mariano, A., Gesto J. M., Schwarz, H. 2006. Ground
Canadian Geotechnical Journal, 46, 513–535. movement control in the construction of a new metro line in
Andersen, K.H., and Lauritzsen, R. 1988. Bearing capacity for Barcelona. International symposium on Geotechnical aspects of
foundation with cyclic loads. Journal of Geotechnical Engineering, underground construction in soft ground, Taylor & Francis, London,
ASCE, 114 (5), 540–555. 389–395.
doi:10.1061/(ASCE)0733-9410(1988)114:5(540). Gens, A., Di Mariano, A. and Yubero, M. T. 2012. EPB tunneling in
Arroyo, M., DiMariano, A., Gens. A, Alonso, E., García Fontanet, A. deltaic deposits : observations of ground movements. 7th Int. Symp.
and García Germán, J. 2007. Management of third-party risk in an on Geotechnical Aspects of Underground Construction in Soft
urban deep excavation project, XIV European Conf. of Soil Ground (Viggiani G., ed.), Taylor & Francis, London, 987-993
Mechanics and Foundation Eng., Madrid, Millpress, 2, 527-532 Gens, A. and Lloret, A. 2003. Monitoring a preload tests in soft ground.
Arroyo, M and Gens, A. 2009. Engineering assessment of jet-grouted Field Measurements in Geomechanics (Myrvoll, ed.), Swets &
structures. Proceedings of the 17th International Conference on Zeitlinger, Lisse, 53-59.
Soil Mechanics and Geotechnical Engineering, (Hamza et al. eds.) Goda, Y. 1974. A new method of wave pressure calculation for the design
IOS Presss, 3, 2338-2341. of composite breakwater. Proceedings of the 14th Conference on
Arroyo, M. & Gens, A. 2021. Computational analyses of Dam I failure Coastal Engineering, Copenhagen, ASCE, Reston, VA, 1702–1720.
at the Corrego de Feijao mine in Brumadinho, Final Report for Goda, Y. 2009. Random Seas and Design of Maritime Structures, 3rd
VALE/MPF; available in http://www.mpf.mp.br/mg/sala-de- ed.;World Scientific Publishing Co. Pte. Ltd.,Singapore.
imprensa/docs/2021/relatorio-final-cinme-upc-1 González, N.A. 2011. Development of a family of constitutive models
Arroyo, M., Gens, A., Croce, P. and Modoni, G, 2012. Design of jet- for geotechnical applications. Ph.D. Thesis. Universitat Politecnica
grouting for tunnel waterproofing, 7th International Symposium on de Catalunya, Barcelona.
33
Goodman, R.E. 1998. The Engineer as Artist. ASCE Press, Reston, Lunne, T., Robertson, P.K., Powell, J.J.M., 1997. Cone penetration
Virginia. testing in geotechnical practice. Blackie Academic, EF Spon
Goulois, A., Whitman, R. and Hoeg, K. 1985. Effects of Sustained /Routledge Publ., New York.
Shear Stresses on the Cyclic Degradation of Clay. Strength Testing Madrid, R.M. 2021. Construction and post-construction performance of
of Marine Sediments: Laboratory and In-Situ Measurements (ed. R. vertical breakwaters on soft soils. The Port of Barcelona case. PhD
Chaney and K. Demars), West Conshohocken, PA, ASTM Thesis, Universitat Politecnica de Catalunya, Barcelona.
International, 336-351. https://doi.org/10.1520/STP36344S Mahoney D.P. and Kupec J. 2014. Stone column ground improvement
field trial: A Christchurch case study. New Zealand Society for
Earthquake Engineering (NZSEE) Conference, Auckland. Paper
during the Holocene (Gulf of Almeria, Western Mediterranean). number O79.
Mánica, M.A., Arroyo, M., Gens, A. and Monforte, L. 2022. Application
Goy, J.L., Zazo, C., Dabrio, C.J., Lario, J., Borja, F., Sierro, F.J. and of a critical state model to the Merriespruit tailings dam failure.
Flores, J.A., 1996. Global and regional factors controlling changes Proceedings of the Institution of Civil Engineers – Geotechnical
of coastlines in southern Iberia (Spain) during the Holocene. Engineering, doi.org/10.1680/jgeen.21.00001
Quaternary Science Reviews, 15, 773–780 Mánica M.A., Gens A., Vaunat J. and Ruiz D.F. (2018) Nonlocal
Guillocheau F., Lejay A., Homewood P. and Hoffert M. 1989. Effet de plasticity modelling of strain localisation in stiff clays. Computers
variations du niveau relatif de la mer à très haute fréquence sur and Geotechnics, 103, 138–150.
l’agencement de corps sédimentaires côtiers tidaux : importance doi.org/10.1016/j.compgeo.2018.07.008
relative des contrôles auto- et allo-cyclique. Comptes Rendues Manzano, M., Pelaez, M.D. and Serra, J. 1986. Sedimentos prodeltaicos
Acad. Sci., Paris, sérieII, 309, 1047-1053. en el Delta emergido del Llobregat. Acta Geologica Hispánica, 21–
Hight, D.W. 2002. Personal communication. 22, 205–211.
Homewood P., Guillocheau F., Eschard R. and Cross T. A. 1992. Marchetti, S. 2015. Some 2015 Updates to the TC16 DMT Report 2001.
Corrélations haute résolution et stratigraphie génétique: une Proceedings of the 3rd International Conference on the Flat
démarche intégrée. Bull. Centres Rech. Explor. Prod. Elf- Dilatometer (DMT’15), Roma, 43–65.
Aquitaine, 16 (2), 357-381. Marchetti, S., Monaco, P., Totani, G. and Calabrese, M. 2001. The flat
Hunt, D., and Tucker, M. E. 1992. Stranded parasequences and the dilatometer test (DMT) in soil investigations—A report by the
forced regressive wedge systems tract: deposition during base-level ISSMGE Technical Committee 16. Proceedings of the International
fall, Sedimentary Geology, 81, 1-9. Conference on Insitu Measurement of Soil Properties and Case
Hunt, D., and Tucker, M. E. 1995. Stranded parasequences and the Histories (Rahardjo, P.P. & Lunne, T., eds.), Bali, 1-41.
forced regressive wedge systems tract: deposition during base-level Marques, M.A. 1984. Las formaciones cuaternarias del delta del
fall - reply, Sedimentary Geology, 95, 147-160. Llobregat LXXI. Institut d'Estudis Catalans, Barcelona, 1–280.
Indraratna, B. 2017. Recent Advances in Vertical Drains and Vacuum Mayne, P.W. and Sharp, J. 2019. CPT Approach to Evaluating Flow
Preloading for Soft Ground Stabilisation. Louis Ménard Lecture. Liquefaction Using Yield Stress Ratio. Proc. of Tailings and Mine
Proceedings of the 19th International Conference on Soil Mechanics Waste 2019. University of British Columbia, Vancouver, 655-670.
and Geotechnical Engineering (Lee, W., ed.), Korea Geotechnical Meinen, N.E., Steenbergen R.D.J.M.., Hofland B. and Jonkman, S.N.
Society, 1, 141-166. 2020. Applicability of the Goda–Takahashi Wave Load Formula for
Jalut, G., Esteban Amat, A., Bonnet, L., Gauquelin, T. and Fontugne, Vertical Slender Hydraulic Structures. Journal of Marine Sciences
M., 2000. Holocene climatic changes in the Western and Eng. 8, 868; doi:10.3390/jmse8110868
M Monforte, L., Arroyo, M., Carbonell, J.M. and Gens, A. 2018. Coupled
Palaeogeography, Palaeoclimatology, Palaeoecology effective stress analysis of insertion problems in geotechnics with
the Particle Finite Element Method. Computers and. Geotechnics,
Jervey M. T. 1988. Quantitative geological modelling of siliclastic rock 101, 114–129. doi.org/10.1016/j.compgeo.2018.04.002
sequences and their seismic expressions. Sea-level changes: an Monforte, L., Arroyo, M. and Gens, A. 2022. Undrained strength from
integrated approach (Wilgus C. K. et al. eds.), SEPM Special CPTu in brittle soils: a numerical perspective. 5th International
Publication, Tulsa, 42, 47-69. Symposium on Cone Penetration Testing, CPT’22, Bologna (in
Kortenhaus, A. and Oumeraci, H. 1998. Classification of wave loading press)
on monolithic coastal structures. Proc. Coast. Eng. Conf., 1, 867– Monforte, L., Gens, A., Arroyo, M., Mánica, M. and Carbonell, J.M.
880. 2021. Analysis of cone penetration in brittle liquefiable soils.
Lacasse, S. 2013. 8th Terzaghi Oration. Protecting society from Computers and Geotechnics, 134: 104123.
landslides – the role of the geotechnical engineer. Proceedings of the doi.org/10.1016/j.compgeo.2021.104123
18th International Conference on Soil Mechanics and Geotechnical Nayak S. and Panchangb, V. 2015. Note on Short-term Wave Height
Eng, (P. Delage et al., eds.), Presses des Ponts, Paris, 1, 15-34. Statistics. Aquatic Procedia, 4, 274 – 280.
Lacasse, S., Nadim, F., Liu, Z.Q., Eidsvig, U.K., Le, T.M.H. and Lin, doi: 10.1016/j.aqpro.2015.02.037.
C.G. 2019. Risk assessment and dams – Recent developments and Nolte, K. G., and Hsu., F. H. 1979. Statistics of Larger Waves in a Sea
applications. Proceedings of the XVII Europ. Conf. Soil Mech, State, J. Waterway, Port, Coastal and Ocean Eng., 105(WW4), 389-
Geotech. Engng. IGS, Reykjavik. 1-26. 404.
doi: 10.32075/17ECSMGE-2019-1110 Olson, S.M. and Stark, T.D., 2002. Liquefied strength ratio from
Ladd, C.C. 1991. Stability evaluation during staged construction. Journal liquefaction flow failure case histories. Canadian Geotechnical
of Geotechnical Engineering, ASCE, 117 (4), 540-615. Journal, 39, 629–647.
Lafuerza, S., Canals, M., Casamor, J.L. and Devincenzi, J.M. 2005. Oumeraci, H., Allsop, N.W.H., de Groot M.B., Crouch, R.S., and
Characterization of deltaic sediment bodies based on in situ Vrijling, J.K. 1999. MAST III / PROVERBS. Probabilistic Design
CPT/CPTU profiles: a case study on the Llobregat delta plain, Tools for Vertical Breakwaters, MAS3 - CT95 – 0041. Final Report
Barcelona, Spain. Marine Geology, 222–223 (1–4), 497–510. 1. Technical University of Braunschweig, Germany.
Liquete, C., Canals, M., Arnau, P., Urgeles, R. and Durrieu de Madron, Oumeraci, H. and Kortenhaus, A. 1994. Analysis of the dynamic
X., 2004. The impact of humans on strata formation along response of caisson breakwaters. Coastal Engng., 22 (1-2), 159-183.
Mediterranean margins. Oceanography Patiño, H., Soriano, A., González, J. 2013. Failure of a soft cohesive soil
Liquete, C., Canals, M., De Mol, B., De Batist, M. and Trincardi, F., subjected to combined static and cyclic loading. Soils &
2008. Quaternary stratal architecture of the Barcelona prodeltaic Foundations, 53, 910–922.
continental shelf (NW Mediterranean). Marine Geology 250 (3–4), Peck, R.B. 1980. Where has all the judgement gone? The fifth Laurits
234–250. Bjerrum memorial lecture. Canadian Geotechnical J., 17, 584-590.
Llopis, N., 1942. Tectomorfología del Macizo del Tibidabo y valle Plewes, H.D., Davies, M.P. and Jefferies, M.G. 1992. CPT based
inferior del Llobregat. Estudios Geográficos 3 (7), 321–383. screening procedure for evaluating liquefaction susceptibility.
Longuet-Higgins, M. S., 1952. On the Statistical Distribution of the Proceedings of the 45th Canadian Geotechnical Conference, 41-49.
Heights of Sea Waves. J. Marine Research, 11(3), 245-265. Posamentier H. W., Jervey M. T. and Vail P. R. 1988. Eustatic controls
Longuet-Higgins, M. S., 1980. On the Distribution of the Heights of Sea on clastic deposition I. Conceptual framework. Sea-level changes:
Waves: Some Effects of Non-linearity and Finite Band-width, J. an integrated approach (Wilgus C. K. et al. eds.), SEPM Special
Geophys. Res., 85 (C3), 1519-1523. Publication, Tulsa, 42, 102-124.
Posamentier, H.W. and Vail, P.R. 1988. Eustatic controls on clastic
deposition II. Sequences and system tract models. Sea-level
34
changes: an integrated approach (Wilgus C. K. et al. eds.), SEPM Takahashi, S.; Tanimoto, K.; Shimosako, K. A 1994. Proposal of
Special Publication, Tulsa, 42, 125–154. Impulsive Pressure Coefficient for the Design of Composite
Puebla, H., Byrne, P.M. and Phillips, R. 1997. Analysis of CANLEX Breakwaters. In Hydro-port’94, Proceedings of the International
liquefaction embankments: prototype and centrifuge models. Conference on Hydro-Technical Engineering for Port and Harbor
Canadian Geotechnical Journal, 34 (5), 641-657. Construction, Yokosuka, Japan, 19–21 October 1994; Port and
Puertos del Estado 1990. Actions in the Design of Maritime and Harbour Research Institute: Yokosuka, Japan, 489–504.
Harbour Works. ROM 0.2-90. Ministerio de Obras Públicas y Tarragó, D. 2021. Hydraulic fills liquefaction. Effect on quay stability.
Urbanismo, Madrid, Spain. PhD Thesis, Universitat Politecnica de Catalunya, Barcelona.
Puertos del Estado. 2002. General Procedure and Requirements in the Tavenas, F., Blanchet, R., Garneau, R. and Leroueil, S. 1978. The
design of harbor and maritime structures. ROM 0.0. Ministerio de stability of staged-construction embankments on soft clays.
Fomento, Madrid, Spain. Canadian Geotechnical Journal, 15(2), 283-305.
Puertos del Estado. 2008. Geotechnical recommendations for the design Tayfun, M. A., 2004. Statistics of Wave Crests in storms, J. Waterway,
of maritime and harbour works. ROM 0.5-05. Ministerio de Port, Coastal and Ocean Eng., 130(4), 1589-1622.
Fomento, Madrid, Spain. Terzaghi, K. 1948. Address of the opening and closing sessions of the
Rabineau, M., 2001. Un modèle géométrique et stratigraphique des second international conference on soil mechanics and foundation
séquences de dépôts quaternaires de la plate-forme du Golfe du engineering. Proc. 2nd Int. Conf. on Soil Mech and Foundation
Lion: enregistrement des cycles glacioeustatiques de 100 000 ans. Engineering, Rotterdam, 6, 11-14, 17-19,
Thèse de Doctorat, Universitè de Rennes 1. Terzaghi, K. 1953. Opening Session in the Kongresshaus, Zurich. Thirs
http://www.ifremer.fr/docelec. international conference on soil mechanics and foundation
engineering. Proc. 3rd Int. Conf. on Soil Mech and Foundation
Engineering, Zurich, 3, 76.
2005. Sedimentary sequences in the Gulf of Lion: A record of Terzaghi, K. 1961. Past and future of applied soil mechanics. Journal of
100,000 years climatic cycles. Marine and Petroleum Geolo, 22, the Boston Society of Civil Engineers, 48, 110-139.
775-804. Trincardi, F., Cattaneo, A. and Correggiari, A. 2004. Mediterranean
Raptakis, D. G. 2013. Pre-loading effect on site response: Site Prodelta Systems: Natural Evolution and Human Impact
amplification and soil properties mismatch. Soil Dynamics Investigated by EURODELTA. Oceanography
Earthquake Eng., 53, 1–10. USACE 2008. Coastal Engineering Manual. EM 1110-2-1100. US
Robertson, PK. 2016. Cone penetration test (CPT)-based soil behaviour Army Corps of Engineers, Washington.
type (SBT) classification system — an update. Canadian USACE 2014. Engineering and Design. Safety of dams – policy and
Geotechnical Journal, 53, 1910–1927. dx.doi.org/10.1139/cgj- procedures, ER 1110-2-1156, US Army Corps of Engineers,
2016-0044 Washington.
Robertson, P.K. 2021. Evaluation of flow liquefaction and liquefied Voortman, H.G., Van Gelder, P.H.A.J.M., Vrijling, J.K. 1999. Chapter
strength using cone penetration test: an update. Canadian 4.1: the Goda model for pulsating wave forces. MAST III /
Geotechnical Journal (in press) PROVERBS. Probabilistic Design Tools for Vertical Breakwaters,
Robertson, PK. and Cabal, K.L. 2015. Guide to Cone Penetration MAS3 - CT95 – 0041 (Allsop N.W.H., ed.). Final Report 2a,
Testing for geotechnical Engineering. 6th Edition, Gregg Drilling Technical University of Braunschweig, Germany.
and Testing, Signal Hill, CA. Wang, Y., Leia, J., Wang, Y. and Li, S. 2019. Post-cyclic shear behavior
Sau, N., Arroyo, M and Gens, A. 2012. Site characterization of reconstituted marine silty clay with different degrees of
alternatives for numerical models of a deep excavation. Fourth reconsolidation. Soil Dynamics and Earthquake Engineering, 116,
International Conference on Site Characterization, ISC’4. Recife, 530–540.
Brasil, 1169-1178 Wheeler, S.J. 1997. A rotational hardening elasto-plastic model for clays.
Schnaid, F., de Mello, L.G.F.S. and Dzialoszynski, B.S. 2020. Proceedings of the 14th International Conference on Soil Mechanics
Guidelines and recommendations on minimum factors of safety for and Foundation Engineering, Hamburg, A.A. Balkema, Rotterdam,
slope stability of tailings dams. Soils and Rocks, 43(3), 369-395, 1, 431–434.
doi.org/10.28927/SR.433369 Wheeler, S., Näätänen, A., Karstunen, M. and Lojander, M. (2003). An
Schwarz, H., Boté, R. and Gens, A. 2006. Construction of a new Metro anisotropic elastoplastic model for soft clays. Canadian
line in Barcelona: design criteria, excavation and monitoring Geotechnical Journal, 40, 403-418.
system. International Symposium on Geotechnical aspects of Yu, H.S., 1998. CASM: A unified state parameter model for clay and
underground construction in soft ground, Taylor & Francis, sand. International Journal for Numerical and Analytical Methods
London, 757–762. in Geomechanics 22, 621–653.
Shuttle, D.A. and Cunning, J. 2007. Liquefaction Potential of Silts from Yu, H.S., 2006. Plasticity and Geotechnics. Springer Science, New York.
CPTu. Canadian Geotechnical Journal, 44, 1–19. Yu, H.-S., Zhuang, P.-Z. and Mo, P-Q. 2019. A unified critical state
Simó, J.A., Gàmez, D., Salvany, J.M., Vázquez-Suñé, E., Carrera, J., model for geomaterials with an application to tunnelling. Journal of
Barnolas, A. and Alcalá, F.J., 2005. Arquitectura de facies de los Rock Mechanics and Geotechnical Engineering, 11, 464-480
deltas cuaternarios del río Llobregat, Barcelona, España. Yubero M.T. 2021. Ejecución de túneles en terreno deltaico mediante
Geogaceta, 38, 171-174 tuneladora EPB. El caso de la Línea 9. PhD Thesis, Universitat
Solé-Sabarís, L., 1963. Ensayo de interpretación del Cuaternario Politecnica de Catalunya, Barcelona.
Barcelonés. Miscel. Barcilonensia 2, 7–54.
Soriano, A. 2005. Los coeficientes de seguridad y el Eurocódigo.
¿Cómo determinamos los parámetros resitentes del terreno?
Intevía, Madrid.
Stamatopoulos C., Petridis, P., Bassanou, M., Allkja, S., Loukatos, N.
and Small, A, 2013. Improvement of dynamic soil properties
induced by preloading verified by a field test. Engineering
Geology, 163 (1), 101 –112.
Stamatopoulos, C., Petridis, P., Bassanou, M. and Stamatopoulos, A.
2005. Increase in horizontal stress induced by preloading.
Ground Improvement, ICE, 9(2), 47–51.
Stanley, D.J. 1997. Mediterranean deltas: subsidence as a major control
of relative sea-level rise. Bulletin Institut Océanographique,
Monaco, N° special 18 (Briand, F. and Maldonado, A. eds.), 35-62.
Stanley, D.J. and Warne A.G. 1994. Worldwide initiation of holocene
marine deltas by deceleration of sea-level rise. Science, 265(5169),
228-231. doi: 10.1126/science.265.5169.228.
Takahashi, S., Tanimoto, K. and Shimosako, K. 1993. Experimental
Study of Impulsive Pressures on Composite Breakwaters-
Fundamental Feature of Impulsive Pressure and the Impulsive
Pressure Coefficient; Technical Report PARI Report 031-05-02;
Port and Harbour Institute, Ministry of Transport: Yokosuka, Japan.
35
State of the Art Papers
State of the Art 1: Transportation Geotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
State of the Art in Transportation Geotechnics
État de l'art en géotechnique des transports
Richard Kelly
Chief Technical Principal SMEC & Conjoint Professor University of Newcastle, Australia, [email protected]
Buddhima Indraratna
Distinguished Professor & Director of Transport Research Centre, University of Technology Sydney, Australia
William Powrie
Professor, University of Southampton, United Kingdom
Claudia Zapata
Assoc. Prof., Arizona State University, USA
Yoshiaki Kikuchi
Professor, Tokyo University of Science, Japan
Erol Tutumluer
Professor, University of Illinois, USA
Antonio Gomes Correia
Professor, University of Minho, Portugal
ABSTRACT: Transportation systems are undergoing a period of substantial change. Vehicles are becoming larger, heavier, faster
and longer. There is a desire to reduce environmental impact in construction and operations, and opportunities associated with
digital advancements are being created. In this paper we report the current state of the art within this context, identify elements
practitioners could adopt and opportunities for further research.
RÉSUMÉ : Les systèmes de transport traversent une période de changements substantiels. Les véhicules deviennent plus gros, plus
lourds, plus rapides et plus longs. Il y a un désir de réduire l'impact environnemental dans la construction et les opérations, et des
opportunités associées aux avancées numériques sont créées. Cet article passe en revue l'état actuel de l'art dans ce domaine et identifie
les éléments que les praticiens pourraient adopter et les opportunités de recherches futures.
KEYWORDS: Port, Airport, Road, Rail
1 INTRODUCTION. increased capacity from transportation systems and this is likely
to translate into more infrastructure as well as larger ships,
Transportation systems facilitate the operation of economies aeroplanes, trains and road vehicles. Reducing the carbon
which supports the way of life for people and communities footprint of transportation infrastructure is contributing to
(National Academies of Sciences, Engineering, and Medicine increased load carrying capacity instead of high volume.
2019). Transportation systems are the arteries of global and Increased capacity has a major effect on geotechnical
local communities. Transportation systems are required to engineering behaviour of transport systems.
move goods and people from their source to where they are HSBC estimates the financial investment over the next 20
needed or wanted to be. Ports and airports tend to act as years (in US Trillions of dollars) in roads is $30.6, rail is $10.3,
gateway and corridor infrastructures between countries while ports and airports are each $3.2. (HSBC Global Infrastructure
road and rail distribute people and goods more locally. Trends, 2019)
Roads and ports have existed for thousands of years. Rail Delivery of transport infrastructure will require balancing
has existed for about 200 years and airports for about 100 years. competing socio-economic, political, cultural, demographic and
However, the world is in a period of significant change as a result increasingly, ecological factors. This balance is reflected in
of population growth, globalisation, technology development transportation strategies published by various countries. Each
and environmental stresses. That is driving major changes in country has a different balance to achieve but there are many
transportation systems and hence will require major advances in common themes. Some key themes are:
the geotechnical engineering that supports them.
Global growth is projected to increase to 2030 with the fastest 1. Financing and building cost effective infrastructure
growth in developing countries and more moderate growth in the (National Academies of Sciences 2019, OECD 2011,
developed countries (OECD, 2011). This growth will demand ICE 2019, Australian DIRD 2016). Earthworks and
41
pavements often comprise a large percentage of speed, high volume transport systems to accommodate relatively
construction and maintenance costs. low weight vehicles. Ports, airports, major roads and rail are
2. Building integrated transportation systems (OECD required to be operational most of the time which means that
2011, CPCCC 2019, BBC 2013, ICE 2019, Australian disruption due to planned and unplanned maintenance should be
DIRD 2016). For example, greater marine traffic consistent with the operational requirements of the infrastructure.
will require greater land transport capacity to move This is driving utilisation of big monitoring and materials data
goods from the port before port storage capacity is sets to understand when and where future maintenance is
reached and to turn boats around within 24 hours. required. Some systems are designed for periodic maintenance
3. Connecting people within cities and between cities such as ballasted rail track and flexible pavements and some are
with reasonable commuting times and reduced traffic designed to require minimal maintenance such as slab track or
congestion (National Academy of Sciences 2019, rigid concrete pavements. Maintenance in constrained
CPCCC 2019, ICE 2019, Australian DIRD 2016) corridors can require people to be in close proximity with traffic
driving the need for fast 400km/hr to 600km/hr trains which increases safety risks. Overcoming these safety risks is
(CPCCC 2019) a key driver for the development of remote sensing systems.
4. Step change in capacity. For example freight trains In this paper we address three themes:
with 30,000 tonne capacity travelling at 250km per
hour (CPCCC2019) or container ships that are too 1. The future of transportation: bigger, heavier, faster,
large to navigate through existing transport routes or longer;
dock at some nation’s ports (BBC 2013) 2. Management of our environment; and
5. Improved environmental outcomes (National 3. The digital age.
Academies of Sciences 2019, CPCCC 2019, BBC
2013, Australian DIRD 2016) such as more efficient Based on this review, we discuss opportunities for practitioners
use of materials, use of recycled materials or selecting and areas of research for academics.
routes away from high value ecological land which
could favour construction through poor quality ground 2 THE FUTURE OF TRANSPORTATION: BIGGER,
conditions. HEAVIER, FASTER, LONGER
6. Resilience in response to changing environmental
conditions (National Academy of Sciences 2019, 2.1 Reclamation for ports and airports
OECD 2011, ICE 2019, Australian DIRD 2016).
Resilience is the ability to rapidly re-establish key Port facilities including seaports and airports have become
transport infrastructure after some form of major increasingly important to drive economic growth worldwide. The
disaster event. expansion of these facilities is strictly controlled by the
7. Development of intelligent systems to assess the availability of suitable nearby land, due to their locations close
performance of infrastructure assets and inform to urban areas. As a result, new port or port expansion projects
maintenance requirements (National Academy of often rely on land reclamation or swampy land where ground
Sciences 2019, CPCCC 2019, ICE 2019, Australian improvement methods are required before the construction of
DIRD 2016). This includes big data analytics and facilities. The reclamation techniques mainly employ dredged or
machine learning concepts (e.g. massive storage of borrowed materials together with the additional improvement to
wayside data that has been accumulated over many achieve the required performance criteria. Additional challenges
years to cover all aspects of construction quality and need to be addressed when dealing with the unacceptable
immediate performance, design validation and geotechnical properties such as, low shear strength and excessive
performance monitoring (short to medium term), and long term deformation which can affect the stability of built
maintenance and repairs (medium to long term). infrastructure.
8. Further development of automated driverless systems
(National Academy of Sciences 2019, CPCCC 2019, 2.1.1 Kansai International Airport
Australian DIRD 2016) and their implications on The Kansai International Airport is located 5km offshore from
traffic volumes and pavement construction. Short the mainland, as shown in Figure 1. Reclaimed zone is separated
distance commuter trains or light rail are relatively into two phases. The length of each zone is about 5km and the
common in some countries and some iron ore trains width of each zone is 500m. The ground conditions at the site are
and mining vehicles in Australia are driverless. The shown in Figure 2. The thickness of Holocene layer at the site is
concept here is that automation of a large proportion or about 20 m, and there is more than 400 m of older compressible
all traffic in large cities could occur in order to layers beneath the Holocene layer. The water depth is 18m to
maximise traffic volume while minimising travelling 20m. Reclamation of the 1st phase comprised an area of 510ha,
times. a volume of 180Mm3 and an 11km long sea wall, was started in
1988 and finished in 1991. The second phase had an area of
While there are many similarities in the engineering 545Ha, a volume of 250Mm3 and a 14km seawall (Furudoi,
requirements for ports, airports, road and rail there are some key 2010).
differences. The container trade is a high volume relatively low For the first phase reclamation, the Holocene layer was
margin business (BBC 2013) where it is important to transport in improved using sand drains and the settlement of Holocene layer
bulk and to distribute goods to and from port efficiently. New was about 6m, finished just after the reclamation. The
airports require large areas and are often located in areas with Pleistocene layer beneath Holocene layer was not treated and the
poor ground conditions. In large countries such as Canada, estimated settlement is about 9m, which is expected to continue
Australia, Russia, China, India and USA, the distance between for more than 50 years. Figure 3 show the time - settlement
the source (or destination) and port (or airport) can be several relationship of the 1st phase island. As the 1st phase of the
hundreds of kilometers which favours large capacity heavy haul airport was inaugurated in September 1994, an estimated 12m of
road or rail infrastructure. In contrast, for commuter transport, settlement was expected in the period from the beginning of
people will want to ensure the trains and buses stop at their construction to the opening of the runway. An extra 6m of
doorstep when the timetable states that they will arrive and that residual settlement was predicted over a 50-year period after the
roads are not too congested. This leads to construction of high opening of the runway.
42
The long-term settlement predictions were incorporated into
the design of the airport island, which had a design life of 50
years. The reclamation was carefully performed to minimize
uneven settlement and to ensure the thickness of reclaimed soils
by properly managing the reclamation work.
In order to predict the amount of settlement, the geotechnical
properties of the Pleistocene clay properties and layer
configurations were investigated prior to the beginning of the
first phase project. Among the boring explorations, two of which
were at the minus 400m level, because the estimated
consolidation layers were continued to that level.
The prediction of the long-term settlement was based on a
one-dimensional consolidation analysis with the boundary
conditions of pore water pressures in the Pleistocene sand layers,
which were calculated separately by a two-dimensional seepage
flow analysis. The constitutive model used for estimating the
settlement of the second phase airport island was developed from
Figure 1 Location of Kansai International Airport (Furudoi, 2010)
the results of laboratory tests and observations of ground
behavior at the first phase airport island and takes such factors as
settlement and pore water pressure into account. In this model,
the e-log p curve was simulated by two straight lines such that
one for the overconsolidated region, and the other for the normal
consolidated region. All the consolidation layers were slightly
overconsolidated. As settlement data obtained during the
construction of the first airport island showed the existence of a
secondary consolidation, secondary consolidation was also
considered.
A comparison of the measured and predicted settlements of
the 2nd phase airport island is shown in Figure 4. The estimated
data from the above model is named DB in the figure. Its validity
was confirmed by the second island settlement data.
Figure 2. General overview of seabed under airport islands (Furudoi, 20
10)
Figure 4 Comparison of measured and calculated settlement (total
settlement DB: DB model, NL: Nonlinear model; Furudoi, 2010)
When total settlement is large, differential settlement becomes
large. Minimizing the differential settlement was one of the
important issues in constructing the second phase airport island.
Figure 5 shows the change in settlement of the first and second
phase islands with time. As shown in Figure 5, the settlement in
the first island varied with the location. In contrast, the
settlement in the second phase island showed almost the same
tendency regardless of location. To avoid large differential
settlements, it was important to continuously supply reclamation
soil to the reclamation area from the initial stage of construction.
Figure 3. Settlement observed at Holocene and Pleistocene clay layers in
the first phase reclaimed island (Furudoi, 2010)
As shown black line in Figure 5, there was especially small
differential settlement, because of avoiding differences in the
construction history.
43
Figure 5. Settlement of 1st and 2nd islands (Furudoi, 2010)
2.1.2 Tokyo International Airport
An offshore extension project for Tokyo International Airport 法線
was conducted from 1984 to 2004. The offshore construction site 護岸
was used for waste reclamation facilities for dredged clay. The Figure 7 View of the TIA new runway (courtesy of Kanto Regional パラペット
surface ground condition of the site was extremely weak, Development Bureau, MLIT, Japan).
essentially liquid. To enable construction, firstly shallow H.W.L A.P.+2.10
L.W.L A.P.±0.00
solidification was used to provide access to heavy ground ケーソン
Ce
improvement machinery (Figure 6). After the surface treatment, ( 堤外側)
法線 施設
用地
境界
( 堤内側) A.P.-8.00
Rubble mound
走路
基礎捨石
CL
1: 1:
1. 2 Dr
フ
護岸 空港
1:11 1:2 5
the ultra-soft dredged clay was improved by consolidation with
滑
57.25
Mound
築堤材2
185.00
Sand mat t=1.50
先行サンドマット (A.P.-16.50)
1:3
44.00 85.00 65.00
(A.P. -19.00) 28.00 115.00 5.00
Sand
A.P.+18.00
3.70 25.00 保護砂
vertical drains and preloading. This technology requires a long
A.P.+13.81
A.P.+8.36 Parapet
パラペット
(1)-C-1
①-C-1 Block
上部ブロック
construction period.
A.P.+2.21
1:2 Sand (reclaimer 2)
揚土2
Wave-dissipating blocks
A.P.-30.0m 消波ブロック
H. W. L A. P. +2. 10
L. W. L A. P. ± 0. 00
(1)-C-2
Cover被覆石
stones
Sand
揚土11t =1. 00 (A.P.+2.50) CDM (block)
CDM改良（ブロック式）
A.P.-34.0m ①-C-2 A.P.-9.75 A.P.-1.00
2.45 5.00 1: 1:2
(11.40) (10.40)
A.P.-9.77 4/
3 4/
Cement管中混合固化処理土
treated soil (pneumatic mixing) Sand揚土1
(reclaimer 1)
1:2
12.60
1: 3 (A.P.-3.00)
1:2 Rubble
捨石1 中仕切堤22
Rubble
捨石2 Partition Sand直投2
(split barge 2)
2.00
(A.P.-5.50)
1:2 1:8
1:5 2.5 32.00 24.80
(2)-C
②-C 1: Sand (counterweight)
保護砂（カウンター）
1:8 Sand直投1
(split barge 1)
11.20
1.00
築堤材1 中仕切堤11
Partition
Original seabed
原地盤 （A. P. - 19. 50） Mound
築堤材2 (A.P.-16.50) Mound 1:8
(A.P.-16.50)
1:3 床掘置換 SCP改良60%
SCP 60% Sand (cover)
1 -C-1
(1)-C-1
A.P.-45.0m 1:2
Replaced sand
(A.P.-27.50)
(A.P.-24.50) (A.P.-24.50)
SCP改良60%
SCP 60% 保護砂 t =2. 50
Sand t =1. 50
サンドマット (A.P.-
A.P.-30.0 18.50 18.50 1 -C-1
(1)-C-1
A.P.-35.0
1 -C-2
(1)-C-2
A.P.-50.0m (2)-S
②-S SCP 30%
SCP改良30%
A.P.-30.0
A.P.-34.0
(A.P.-35.50) 1 -C-2
(1)-C-2
SD改良
SD
(A.P.-36.50)
2 -C
(2)-C
A.P.-45.0
2 -C
(2)-C (A.P.-42.00)
A.P.-45.0
36.50 23.50
A.P.-47.0 2 -S
(2)-S
2 -C
22.75 (2)-C
②-C17.50
82.25
42.00 28.00
50.75
17.50 5.25 A.P.-49.0 2 -S
(2)-S 60.00
A.P.-52.0 (2)-C 2 -C
25.00 133.00 A.P.-54.0 (2)-C
A.P.-61.0m 3 -S
3 -S
(3)-S (3)-S
(3)-S
③-S SCP(30%)
護岸
法線
SD
空港
施設
用地
境界
CDM
パラペット 説明事項：図中表記の標高は工事完了時天端高（計算値）
ただし、空港施設用地境界および滑走路CL天端高は計算測線における値
：( ) 内標高は施工時天端高または深度（計算値）
：( ) 内層厚は計算測線における値（計算値）
揚土2 Sand (reclaimer 2) ：法面勾配は施工時の値
H.W.L A.P.+2.10 揚土1(reclaimer 1)
L.W.L A.P.±0.00 Sand
ケーソン
Cement treated soil (pneumatic mixing)
管中混合固化処理土 1:2 Partition 2
中仕切堤2
A.P.-8.00
1: 1: Drainage (filter) 1:8
Rubble mound
基礎捨石 1. 2 フィルター材
1:11 1:2 5
Mound
築堤材2 Sand mat t=1.50
先行サンドマット (A.P.-16.50) Partition
中仕切堤11
(A.P. -19.00) 1:3
Sand
保護砂 Sand
保護砂(cover)
(1)-C-1
①-C-1 Sand mat
サンドマット
A.P.-30.0m SD
(1)-C-2
CDM (block)
CDM改良（ブロック式） SD改良
A.P.-34.0m ①-C-2
(A.P.-35.50)
(2)-C
②-C
A.P.-45.0m (A.P.-45.00)
A.P.-50.0m (2)-S
②-S
36.50 23.50
(2)-C
②-C 60.00
A.P.-61.0m
(3)-S
③-S
CDM SD
Figure 8 Ground improvement for TIA (courtesy Prof Kikuchi)
Figure 6 To enable trafficability, a shallow solidification method was 2.1.3 Central Japan International Airport
conducted with heavy machine on a barge (courtesy of Kanto Regional
Development Bureau, MLIT, Japan). Central Japan International Airport was opened in 2005. The
north region of the airport island was reclaimed by cement
At Kansai International Airport, prediction technology for treated dredged clay (Figure 9, Figure 10). This cement treated
consolidation settlement in a deep Pleistocene clay was method called Pneumatic flow mixing method was new
significantly advanced. This experience aided the new runway technique at the time. In this method, dredged clay was placed as
project at Tokyo International Airport (TIA), shown in Figure 7. a cement mixture. This technology does not require a long
Because, incremental consolidation pressure at TIA new runway curing time and cement treated soil itself has no settlement.
was 550 kPa or larger, which is almost the same as that in the
second phase island of Kansai International Airport, prediction 2.1.4 Application of the deep mixing method to Tokyo/Haneda
of long-term consolidation settlement derived from deep soil international airport expansion project
layers (without ground improvement) was very important.
Technologies used for soft clay/dredged clay in the previous Tokyo/Haneda International Airport was founded in 1931 as the
airport projects were 1) vertical drains to accelerate first primary airport in Japan. The airport had been expanded
consolidation, 2) replaced by good material: clay and sand, and several times to cope with the rapid increase in air transportation.
3) chemically strengthened (cement treatment) In order to cope with the recent and expected future increase in
In the TIA new runway project, 1) soft clay deposit inside the air transportation, the construction of fourth runway was
reclamation section was improved by Sand Drains (SD), 2) under commenced in 2006 and completed in 2010 (Figure 11). As soft
the rubble seawall was improved by Sand Compaction Pile grounds were stratified at the construction site, various ground
method (SCP) with low replacement ratio and 3) under the improvement techniques including the sand drain method, the
important seawalls the ground was improved by Cement Deep sand compaction pile method and the deep mixing method were
Mixing as shown in Figure 8. employed depending on the location and the requirements.
44
pressure increased linearly with the depth, which indicated the
clay was lightly over-consolidated condition of OCR of 1.3. The
second upper layer from -35 to -60 m was a clay layer underlain
locally by a sand layer. The upper two layers were improved to
increase the stability of superstructures and to reduce the residual
settlement of the man-made island for the fourth runway.
Figure 9 Satellite view of Central Japan International Airport.
Figure 12: Soil properties at construction site (Kitazume and Terashi,
2013)
Almost all of the sea revetment embankment was constructed
on ground improved by the sand compaction pile method.
However, caisson type quays were constructed at two corners of
the island (CW and CN revetments). The block type
improvement of 60m in width was constructed to a depth of
minus 45m, as shown in Figure 13 for the CW revetment. Table
1 summarizes the properties of the soil layers. The table shows
that the properties of the four soil layers were much different to
each other so that the mixing condition should be adequately
determined for each layer to assure the design strength of
stabilized soil. The mixing conditions were designed as tabulated
in Table 1 based on the laboratory mix tests, where blast furnace
slag cement type B of 110 to 165 kg/m3 was mixed with the soil
to obtain the average field unconfined compressive strength, quf,
of 3,375 kN/m2 at 28 days curing. The total of about 620,000m3
soft soils was stabilized by four DM vessels within five months.
Figure 14 shows the DMM vessels in operation.
Figure 10 Cement treated dredged clay method was used for the
reclamation of the site (Courtesy Prof Yoichi Watabe).
Figure 13: Cross section of DMM improvement at CW revetment
(Kitazume and Terashi, 2013).
Table 1: Soil properties and cement content.
binder
soil property
content
wn wl wp
Ip (kg/m3)
(%) (%) (%)
surface layer 168-177 - - - 165
clay 1-C1 132-145 132-137 51-54 78-85 140-145
Figure 11: Aerial view of construction site and existing airfield on March clay 1-C2 42-117 41-118 22-47 19-70 130-135
15th, 2009 (Courtesy Prof Kitazume). clay 2-C 35-52 32-55 18-24 14-31 110-120
sand 2-S 37 - - -
The ground condition and the major soil properties at the site
were extensively studied and are summarized in Figure 12. The
ground can be roughly divided into five layers. The most upper
layer between -20 m and around -35 m has high plasticity index
ranging from 60 to 100 and high water content ranging from 100
to 150 %. The undrained shear strength and the pre-consolidation
45
bulk cargo and container handling. Part of the reclamation lay
adjacent to a marine nature park have a constraint of zero impact
from the works.
Sea wall and
development area
Figure 14: DMM vessels in operation (Courtesy of the Tokyo/Haneda
International Airport Construction Office).
At 24 to 26 days after the construction, soil sampling was Figure 15: Expansion area of the Port of Brisbane (modified after
carried out at several points for quality assurance. The stabilized Indraratna et al., 2011a)
soils sampled were subjected to unconfined compression test to
investigate the strength of in-situ stabilized soil. A summary of The geological profile consisted of highly compressible clay
the test results is shown in Table 2. The data shows that the with undrained shear strength less than 15 kPa. The same clay
average of quf was 4,094 kN/m2, 20% higher than the target was dredged from the adjacent river and used as reclamation fill
value, and the coefficient of variation (COV) was 28.3%, lower at moisture contents in excess of the liquid limit. It was predicted
than the design value, 35%. According to that, the mix design that the consolidation could take 50 years with vertical
was modified for the subsequent construction in order to reduce settlements up to 4.0m in the absence of ground improvement.
the amount of ground heaving and the cost, where the amount of Ground improvement via surcharging alone had a high risk of
cement was decreased by 3 to 27%. Table 3 shows the quf values instability and consequent effects on the marine park.
of the stabilized soils after the modification. The table clearly Use of prefabricated vertical drains (PVDs) with vacuum
shows that the average strength of the stabilized soils was 4,066 consolidation was adopted adjacent to the park to enhance
kN/m2 and the stabilized soil constructed in-situ satisfied the stability, accelerate the consolidation process, reduce the volume
acceptance criteria. of imported surcharge fill and reduce lateral deformation
(Indraratna et al. 2011a).
Table 2: Original cement factor and field strength. The stratigraphy consisted of an upper Holocene sand layer of
approximately 2-3m thickness underlying the reclaimed soil
Field strength, quf28 Binder (dredged mud), followed by the Holocene clay layer varying in
no. of content thickness between 6m and 25m. An in situ and laboratory testing
tests average COV program was conducted including cone penetration/piezocone
(kg/m3)
(kN/m2) (%) tests, dissipation tests, boreholes, field vane shear tests and
surface 20 3,409 27.1 165 oedometer tests to determine consolidation and stability design
clay 1-C1 36 4,009 28.9 140-145 parameters. Conventional fill preloading system (WD1-WD5,
clay 2-C2 16 3,929 21.3 130-135 Figure 16) and the membrane-type vacuum consolidation system
sand 2-C 44 4,534 26.4 110-120 (VC1 and VC2, Figure 16) with circular PVDs were adopted.
Total 116 4,094 28.3 The design criterion was that the maximum residual settlement
should be no more than 0.25 m over 20 years under a service load
Table 3: Modified cement factor and field strength. of 15-25 kPa. In surcharge-only areas, both circular and band
shape drains were installed in a square pattern at 1.1 to 1.3m
Field strength, quf91
Binder spacing whereas, in the vacuum area, only circular drains were
no. of content installed under a membrane system at a spacing of 1.2m in a
tests average COV square pattern.
(kg/m3)
(kN/m2) (%) Arrays of different instruments were used, including
30 3,568 35.8 160 settlement plates, vibrating wire piezometers, magnetic
Surface 16 4,010 31.7 160 extensometers, and inclinometers. Fill placement and associated
clay 1-C1 72 4,410 29.8 120-125 settlements and excess pore pressures responses are shown in
Figure 17.
clay 1-C2 32 4,561 33.9 110-120
Very similar settlement trends can be observed among the
clay 2-C 88 3,871 26.2 80-85 different sub-areas where the settlement rate is higher at the start
Total 238 4,066 31.4 of consolidation and the magnitude of settlement varies along
with the clay thickness and embankment height. Excess pore
pressures show the effect of surcharge loading and the pore
2.1.5 Port of Brisbane pressure dissipation with time. For the same period, the reduction
of excess pore pressure in vacuum areas (VC1 and VC2) was
The Port of Brisbane is the third largest container port in greater than that in the surcharge only areas. At this site, the
Australia. The Port is located at the mouth of the Brisbane River circular drains propagated the vacuum pressure to depth without
at Fisherman’s Island (Figure 15). A new outer area (235ha) was losing suction head.
reclaimed to provide the maximum number of berths suitable for
46
vacuum pressure as well as fill height (H) to consolidate the
given clay thickness (hc), represented by the ratio (H/hc).
Although the value of b has no specific relationship to the
target degree of consolidation (U), it can act as a ‘filter’ in
distinguishing relative performance. Figure 18 shows the
variation of U with time for the membrane and membraneless
systems. Figure 19 shows U divided by b-factor with time.
Figure 16: Detail the proposed extension area at the Port of Brisbane
(Indraratna et al., 2011a)
Figure 18: Computed degree of consolidation with time (Indraratna et
al., 2013b)
Figure 19: Computed U / b with time (Indraratna et al., 2013b)
Extended running of vacuum pumps can become
uneconomical. The optimal time for turning off the vacuum
pumps depends on various factors such as degree of
consolidation, vacuum-fill surcharge ratio and required time-cost
to complete the improvement following switching off the
vacuum pumps. Based on laboratory observation supported by
field observation at the Port of Brisbane, Kianfar et al. (2015)
showed that removal of the vacuum pressure can be decided
based on flattening of the pore pressure response with time
(Figure 20).
Figure 17: Embankment responses (a) staged construction, (b)
settlements and (c) excess pore pressures (Indraratna et al., 2011a)
Normalised pore pressure dissipation rate
1 1
In order to separate the ‘clustering’ especially towards one
year, the degree of consolidation was divided by a dimensionless Axial strain
Normalised axial strain rate
0.8 0.8
Pore pressure
beta factor (Indraratna et al., 2013b), which is defined in
Equation 1 by: 0.6 Gradual pore pressure 0.6
dissipation rate
0.4 0.4
!! $
!=# $# $ (1)
"#! %" 0.2 0.2
0 0
The β-factor is independent of soil properties and is designed 0 4 8 12
Time (days)
to capture the drain and site loading conditions. The β-factor
comprises (i) the drain length (ld), (ii) the drain spacing (sd) and
its pattern (a = 1.05 for triangular and 1.13 for square spacing), Figure 20: Determination of removal time of vacuum pressure (Kianfar
et al. 2015)
and (iii) the surcharge load height including equivalent height of
47
2.1.6 New Mexico International Airport consolidation. Figure 23 shows a schematic diagram of the air
booster vacuum consolidation where the booster PVDs are
López-Acosta et al. (2019) analysed the performance of an installed in between conventional PVDs. The performance of the
embankment stabilized by vacuum pressure with a membrane- system in comparison to conventional vacuum consolidation
less system for the New Mexico International Airport (Figure system was examined at the Oufei sluice project in Wenzhou,
21). PVDs were installed at 1.2m spacing in a triangular pattern China which was one of the largest individual tideland
before the horizontal drains were installed. A suite of reclamation projects in China. At this site, the marine soils
instrumentation was installed including piezometers, settlement included silt, silt clay, silty clay, and muddy–silty clay. A
plates, magnetic extensometers, vertical and horizontal vacuum pressure of 85 kPa was applied with PVDs installed at
inclinometers. The monitored time-dependent performance 0.8m spacing in a triangular pattern. Figure 24 presents the
comprising of vacuum pressure, expelled volume of water, and variations of pore water pressure and settlement where the
pore pressure is shown in Figure 22. The nonuniform vacuum booster was activated after 72 days. It shows that the booster
pressure shows a possible leak at the connections (Figure 22b). system can further increase the dissipation rate of the pore water
Based on the measurement, the maximum applied vacuum pressure and settlement at the end of consolidation especially at
pressure was approximately 72 kPa. The rate of water extraction depths greater than 20m.
at the site decelerates with time (Figure 22c). The measured pore
pressures show that the vacuum can propagate to the depth where
the PVDs were installed (Figure 22d). At this site, longitudinal
cracks were observed around the embankment’s perimeter. They
suggested that the efficiency of vacuum consolidation of the site
depended on the prevention of the vacuum loss at the connections
and the elevation of vacuum pumps in relation to that of the
horizontal drains.
Figure 23: Cross-section of air booster vacuum system (Cai et al. 2015)
Figure 21: An embankment with vacuum preloading system (a) Vertical
cross-section (b) circular drain and the connection between horizontal
drain and circular PVDs (López-Acosta et al. 2019)
Figure 24: (a) pore water pressures and (b) settlements at different depths
(Cai et al. 2015)
2.1.8 Ningo port, China
The High Vacuum Densification Method (HVDM) was applied
at the land reclamation project at Ningbo Port, China (Xu et al.
2015). This technique combines the application of vacuum
dewatering and dynamic compaction. Vacuum pressure
facilitated by PVDs is applied to shallow depth of the dredged
soil to increase the undrained shear strength. Subsequently,
dynamic compaction can be applied to create positive excess
pore pressures which can be dissipated after a second round of
vacuum application. The process can be repeated until the
desirable results in terms of shear strength and compressibility
are obtained. At this site, the shear strength of soil was increased
Figure 22: (a) staged construction (b) measured vacuum pressure (c)
by 30-40kPa to a depth of 5m after the first cycle of the treatment.
volume of discharged water (d) pore pressure (López-Acosta et al. 2019)
Figure 25 shows the CPT profiles before and after treatment. It
can be seen that the cone resistance increased as a result of the
2.1.7 Oufei sluice project, Wenzhou, China
treatment.
Cai et al. (2015) introduced a new approach to improve deep
marine clay where a booster prefabricated vertical drain is
employed to channel compressed air into the deeper soil layer to
assist pore pressure dissipation close to the end of vacuum
48
10-wheel
Figure 27: possible landing gear configurations (Sarker and Tutumluer,
2020)
Construction Cycle 5 (CC5) tests are summarized by Sarker
and Tutumluer (2020). In these tests the complex deformation
Figure 25: Cone resistance before and after treatment (Xu et al 2015) behavior of unbound layers due to heavy aircraft loading applied
with wander was characterised. CC5 sections were built with 5in
2.2 Airport pavements hot mix asphalt overlying 8in P209 granular base material
Airfield pavement represents almost 30% of initial airport overlying different subbase materials (crushed quarry screenings
construction cost and between 30% and 40% of annual airport and dense graded aggregates) with 34in and 38in thicknesses
infrastructure maintenance costs. Pavement is the biggest Airport over a low-strength subgrade. These pavement sections were
airside asset (Fabre, 2021). Therefore, constructing durable instrumented with various sensors such as multi-depth
pavement and extending the life of existing pavements are deflectometer (MDD), pressure sensor, and asphalt strain gauge.
important considerations. Because both wide body passenger and The MDD data indicated that load wander affected the
cargo aircraft gear/wheel loads are quite heavy, protecting the residual (non-recoverable) deformation accumulations in the
pavement foundation, i.e. subgrade and granular subbase and unbound aggregate layers. The first pass on each wander position
base from excessive permanent deformation has been the key in the west to east direction typically caused the highest
consideration in airport pavement design. deformation response and the return pass along the same wander
Recent airport pavement research has been linked to the position showed significantly less residual deformation. This
concept of building airport pavements to protect the subgrade shakedown behavior was more evident when the wander width
layer. Key elements in the work include better prediction of was kept narrow (Figure 28). In contrast, accumulations of
unbound aggregate base/subbase permanent layer deformations permanent deformation in the subbase layers appeared to
and their accumulation rates. This requires a better increase as traffic progressed, however, this may have been a
understanding of the magnitudes and variations of stress states in function of the relative positions of the sensors with respect to
the granular layers, number of load applications, gear load the load. Figure 29 shows that the subgrade layers in sections
wander patterns, previous loading stress history effects, with 38in. subbase layers did not show any significant rutting but
trafficking speed or loading rate effects, and effects of principal did occur to some extent in subgrade layers of sections that were
stress rotation due to moving wheel loads. built with 34in. subbase layers once the load exceeded 58kips.
State-of-art resilient and permanent deformation models have The HMA layer did not rut.
been developed based on the full-scale test data from
Construction Cycles 1, 3, 5 and more recently 9 (CC1, CC3, CC5 0.5
and CC9) at the USA Federal Aviation Administration (FAA)
National Airport Pavement Test Facility (NAPTF). NAPTF is 0.0
Subgrade
an indoor facility for construction and testing of full-scale airport
MDD Deformations (in)
pavement test sections under actual full size multiple gear/wheel -0.5
loading scenarios (Figure 26 and Figure 27).
HMA and Base
-1.0
-1.5
P-154 Subbase
-2.0
-2.5 Total
-3.0
0 1000 2000 3000 4000 5000
Pass
Figure 28: MDD permanent deformation accumulations, LFC1-SE
section, 6-wheel (Sarker and Tutumluer, 2020)
Figure 26: NAPTF test vehicle (Sarker and Tutumluer, 2020)
49
-8
-6
Highest Rut (in)
-4
-2
Predicted (Sigmoid)
Predicted (Power)
0 Observed in Field
0 5000 10000 15000 20000 25000 30000
Figure 29: Maximum rut depth, section LFC1-NW, 34in subbase, and
section LFC2-NE, 38in subbase (Sarker and Tutumluer, 2020) Pass
A stress history approach (Donovan et al 2016) was Figure 30: Comparison of observed and predicted highest rut amounts
developed to predict rutting. This approach considers that if there with traffic passes, section LFC3-N (Sarker and Tutumluer, 2020)
was rut due to a previous pass and if there was rut in same
location due to current pass, then the incremental increase in rut One of the objectives for the CC9 tests was to assess the
depth would reduce. Similarly, incremental heave also decreases performance of geogrids in airport pavements. Geogrids act by
with subsequent passes. In contrast if the element was constraining aggregate particles from moving and hence increase
compressed in the previous pass, it would heave more in the next their strength and their stiffness. The effect that geogrids have
offset pass. Likewise, if the element heaved in the last pass, it on small strain stiffness in the CC9 tests has been reported by
would rut more in the next pass when the load was directly over Kang et al (2022). The pavement profile consisted of a 5in HMA
the element. Multipliers for heave and rutting as a function of layer underlain by an 8in granular base material which in turn
load offset were generated from the CC1 and CC3 tests and used was underlain by a 29in granular sub-base material. A biaxial
to predict pavement response. However, the developed model geogrid was installed at the interface of the base and the sub-base
showed somewhat limited success in determining the transverse layers. Bender elements (BE) were installed just above the
surface profiles of the CC5 full-scale accelerated flexible geogrid layer and at the same depth in a control section
pavement test sections. It was observed that while this approach constructed without the geogrid. A pair of bender element shear
could predict the deformations in the CC5 test sections at the wave transducers attached at each end of the BE field sensors
most trafficked location on the pavement, it could not predict the generates the shear wave and collects the shear wave velocity
complete transverse surface profiles of the pavement sections propagating through the constructed aggregate material.
due to the complexities arising from factors such as random Tests performed after pavement construction prior to loading
loading sequences, complex landing gear configurations, and showed that the shear wave velocity and small-strain modulus at
load wanders. the geogrid-stabilized section were 9.5% and 19.7% higher than
This model was improved by Sarker and Tutumluer (2020). that measured in the unstabilized section, respectively. After the
First, critical wander locations were established and their moving wheel loading was applied the shear wave velocity
contributions to transverse rut profiles were studied for multiple reduced more in the unstabilized section than in the geogrid-
passes. Then, by utilizing the measured MDD data, individual stabilized section. The geogrid appeared to reduce the dilation
pavement layer deformations, and the periodic transverse field caused by aircraft wander. During static load tests the stiffnesses
surface profiles, a rut prediction model was developed using with and without geogrid were similar. Kang et al (2022)
general linear models in the forms of power and sigmoidal suggested that this may indicate that the geogrid stabilization is
function distributions to determine realistic surface profiles of less effective when the confining pressure is high.
the CC5 test sections. A quadratic model was established based
on the MDD calculated theoretical transverse rut basins and 2.3 Railways
using general linear models, the relationships between the MDD
and field transverse rut profiles were developed. Figure 30 shows 2.3.1 Loads and stresses
the power model predictions were reliable up to 15,000 passes. Considerable advances have been made over the past 10-15 years
However, at higher traffic, the rut prediction model using the in our understanding of the loads and stresses applied to the
sigmoidal function could result in more accurate predictions of subgrade by trains as they pass. Six key areas are (1) the
the field rut depths. understanding that stress paths in the field involve a rotation in
These results confirm results of previous studies, that the direction of principal stresses during load cycling, which may
although wander reduces the number of repetitions of maximum affect subgrade; (2) the role of trackbed stiffness and propensity
load applied to the most heavily trafficked pavement location, for plastic settlement on the performance of railway track; (3)
wander does not necessarily increase the pavement life. Rather it appreciation of the role of grain shape and mineralogy on ballast
causes constant particle shuffling and rearrangements in performance, and the implications for re-using ballast; (4) the
unbound layers, ultimately leading to a weaker and more ability to characterize and reproduce in experiments the complex
unstable granular layer. loading associated with passage of a complete train; (5)
One limitation of this approach is that it requires at least 5,000 recognition of the role of train speed in enhancing track
passes and/or a total of eight MDD residual basins calculated at deflections as a result of resonance or so-called “critical velocity”
different stages of traffic for accurate predictions, hence is not effects; and (6) environmental loads, the need for resilience to
yet suitable for use in practice. climate change, and to be aware of the greenhouse gas emissions
50
(“carbon footprint”) associated with new geotechnical evolution of increased differential settlement in a vehicle-track
transportation infrastructure and maintenance strategies. interaction (VTI) model through the use of an appropriate plastic
settlement law for the trackbed is feasible, but the quantitative
2.3.1.1 Principal stress rotation outcome depends crucially on the model and parameters adopted.
While we probably have, in principle, the computing power
Brown (1996) recognized that elements of soil below a pavement needed to predict the evolution of track geometry over tens of
or a railway are subjected to complex loading involving shear as thousands train passes, experimental (field) validation and model
well as normal stresses, resulting in a rotation of the direction of parameterisation will be the key challenges for the next few
principal stress as a vehicle or train approached and receded. years.
Only when the axle load is directly above the soil element is the
major principal stress vertical, as in a conventional triaxial
apparatus. This led to a concern that soils exhibiting stability or
shakedown under conditions of uniaxial cyclic loading might be
susceptible to fatigue failure when loading cycles involving the
application of a shear stress and rotation of the major principal
stress direction are applied. Investigation of the stress paths to
which soil elements below railway tracks are subjected as trains
pass using finite element analysis by Yang et al (2009) showed
that principal stress rotation, which arises from the interaction of
individual axle loads at moderate depths, was potentially an issue
mainly near the top of the subgrade. The ballast is generally
shallow enough for interaction of loads and hence principal stress
rotation not to be significant, while at greater depths the stresses
associated with train passage are small. Yang et al (2009) also
showed how stress paths comprising cycles of deviatoric and
torsional stress 90° out of phase with each other could be
imposed in a laboratory hollow cylinder test to simulate those
experienced by soil elements in the field.
The effect of principal stress rotation during load cycles has
been investigated by a number of researchers, including Yang et
al., (2009); Gräbe and Clayton (2009, 2014), Xiao et al., (2014),
Qian (2016), Wu et al., (2017), Mamou et al (2017, 2018,
2019a,b, 2021), Guo et al. (2016, 2018), Blackmore et al. (2021),
Fedekar et al. (2021), and Zhang et al. (2021). Mamou et al.
(2017, 2018, 2019a,b, 2021) showed how the susceptibility of a
saturated mixed material (comprising sand, clay and silt sized
grains) to failure during cyclic loading with principal stress
rotation depended on the magnitude of the cyclic shear stress
(Figure 31), the clay content of the material and the drainage
conditions.
When shear stresses were below a threshold stress, the
material was found to reach a stable state during load cycling.
Above the threshold shear stress, strains would increase
gradually (ratchet) with increasing number of load cycles, until Figure 31: Axial strain vs number of cycles for a typical engineered
eventually failure occurs. The cyclic threshold stress was found trackbed material in cyclic loading with principal stress rotation; cyclic
to increase with an increase in clay content up to 14% and was axial stress was constant Δσz=30kPa and the cyclic shear stress, applied
higher in free to drain than in undrained conditions. Tests by Guo 90° out of phase to the axial stress, was increased in steps (Mamou et al.,
et al. (2016, 2018) confirmed that, above certain critical cyclic 2017)
stress ratios, principal stress rotation accelerates excess pore
water pressure and permanent strain development in saturated,
intact, K0-consolidated soft clay. Blackmore et al. (2021) showed
that unsaturated materials were much less susceptible to adverse
effects of principal stress rotation.
2.3.2 The role of trackbed stiffness and propensity for plastic
settlement on the performance of railway track
Field measurements at a number of sites have demonstrated that
even on well-performing railways, the track support stiffness is
likely to vary from sleeper to sleeper (e.g. Figure 32, from Le Pen
Figure 32: Average of 12 trains (a) sleeper deflections and (b) system
et al. 2020 and/or Figure 33, from Track Stiffness Working
modulus and trackbed modulus (assuming pad stiffness of 84 MN/m and
Group 2016). Numerical modelling has suggested that significant 0.6m sleeper spacing). (after Le Pen et al., 2020)
variations in dynamic wheel load are associated with variations
in the initial track level rather than minor variations in sleeper Transition zones are areas where the track stiffness changes
support stiffness (Milne et al. 2019). This also applies at under- rapidly such as at the locations of structures or level crossings.
track crossings, where the difference in track support stiffness Typically, problems related to railway transitions are enhanced
may be more pronounced; however, if the change in track support dynamic load, differential settlement and accelerated track
stiffness over a relatively short distance (a few sleeper spacings) degradation (Indraratna et al., 2019). These problems are highly
is too high, a perceptible increase in vehicle dynamic load inter-related, so any increase in one parameter causes a
becomes apparent (Powrie et al, 2019a). Translating this into the corresponding increase in the others, especially when faster
51
speeds and increased axle loads are involved. The behaviour of a Rubber mats have recently been trialled for track substructure
track transition subject to dynamic load, rail deflections and track having stiff foundations to reduce deformation and degradation
acceleration is summarised in Figure 34. of ballast, while enhancing the overall track stability (Costa et al.
2012a, Lakuši et al. 2010). These studies found that the rubber
mats could provide better load distribution by increasing contact
areas, reducing the contact forces, thus reducing track damage.
A high-capacity drop weight impact testing equipment was used
to evaluate the ability of rubber mat to attenuate dynamic impact
loads and mitigate ballast degradation (Figure 35) where a
schematic diagram of a ballast assembly is shown. Tests were
conducted on ballast with and without the inclusion of rubber
energy absorbing mats (READS) placed on a soft subgrade and
a stiff base (concrete) subjected to varying impact loads.
Figure 33: Data from well performing track: (a) measured sleeper
movements (b) inferred support system modulus seen by the rail (after
Track Stiffness Working Group, 2016)
10
Track acceleration (m/s 2)
4
90
2 80
1.4
Railpad force (kN)
70
Rail deflection (mm)
1.2
60
1
50
0.8
0.6 40 Figure 35: High capacity drop weight impact apparatus: schematic
Soft Stiff
track track diagram of ballast sample (Ngo et al., 2019)
0.4 30
0 8 16 24 32 40 48 56 Measured test data shows that the inclusion of rubber mats
Distance along the track (m) helps to reduce the dynamic impact load transferred to the ballast
layer resulting in significantly less permanent deformation and
Figure 34: Variation in track acceleration, rail deflection and railpad force
at track transition under moving loads (Indraratna et al. 2019a)
degradation of ballast, as well as significant attenuation of load
magnitude and vibration to the underlying subgrade layers (Ngo
The theory of beams on elastic foundations (BOEF) has been et al. 2019). The beneficial effects of the mats are more
extensively used to model railway tracks and transitions (e.g. pronounced on the stiff subgrade, and this corroborates the
Kang et al. 2008). Some studies used BOEF theory considering energy absorbing nature of rubber mats whereby less energy is
Euler-Bernoulli beam on a Winkler foundation for transition transferred to the ballast and other substructure components, thus
zone modelling. Others adopted a Timoshenko beam for this reduced deformation and degradation.
purpose. The BOEF model, in spite of some limitations and the Typical acceleration of ballast under soft and stiff subgrade
lack of a comprehensive model for predicting track response for (with and without READS) measured at the 10th drop (N=10)
complex track substructures, is widely used in practice and offers subjected to a drop height of hd = 100mm is shown in Figure 36.
considerable insights. Its main shortcoming is the limited ability It is seen that the inclusion of recycled rubber mats reduces the
to represent sudden, large variations in track support stiffness, peak acceleration. Measured accelerations for stiff subgrade are
such as a hanging sleeper or as may occur at transitions onto or always larger than those for the soft subgrade indicate higher
off stiff substructures. levels of vibration at the sleeper-ballast interface, as expected.
Dynamic impact loads are commonly caused by wheel or rail With soft subgrade, maximum accelerations are around 66 g and
abnormalities such as flat wheels, imperfect rail welds, rail 105 g for the ballast assembly with and without rubber mat,
corrugations, expansion gaps between two rail segments, and respectively. The accelerations measured approximately 110 g
transition zones (e.g. Priest et al. 2010). When a train passes (with shock mat) and 169 g (without shock) for stiff subgrade. In
through a rail crossing and through turnouts or transition zones, addition, the inclusion of rubber mat helps to attenuate vibration
the rapid change in wheel rail contacts coupled with sudden faster, as shown in Figure 36b. Based on these results, rubber
variations in track stiffness causes the wheels to displace up and mats could be used at track transition zones to reduce vibration
down, giving rise to impact loads (e.g. Paixão et al. 2014). At and prevent excessive ballast deformation and breakage.
transition points such as bridge approaches, road crossings and
slab track to ballasted track, a considerable change in track
stiffness causes high impact forces that accelerate track
deformation.
52
Walker and Indraratna, 2018). These models have been used to
study the instantaneous response of railway track under moving
load in the frequency and time domains. However, these models
do not capture the wave propagation and mechanical response of
the track-ground system.
In order to better understand the effect of wave propagation,
several studies have considered the track as a beam resting on an
elastic half-space or layered half-space. Though various
solutions have been implemented, they are mostly based on a
Fourier transform in the frequency-wave number domain. These
methods are widely applied ground vibration predictions due to
their computational efficiency, they are rarely applied to study
the complex dynamic response of in the track due to geometric
and material complexities (Dong et al., 2018).
In order overcome the limitations of semi-analytical
techniques, advanced numerical modelling for moving loads
Figure 36: Measured acceleration of ballast under a drop height of hd = using Finite Element Method (FEM) or Discrete Element
100mm, at the 10th drop (N=10) placed on: (a) soft subgrade; and (b) Method (DEM) can be used. The FEM has been employed to
stiff subgrade (Ngo et al. 2019) study the effect of increasing train speed on ground vibration and
instantaneous track deflection (Connolly et al., 2013; Costa et
2.3.3 Characterising train loading al., 2012b; El Kacimi et al., 2013; Kouroussis et al., 2014;
Sayeed and Shahin, 2016; Shih et al., 2016), and dynamic stress
Advances in the measurement and analysis of railway track amplification (Li et al., 2018; Tang et al., 2019a; Varandas et al.,
deflections as trains pass have led to the ability to quantify train 2016), among others. In addition, the FEM allows complex
loading as the superposition of a series of different magnitudes geometries and implementation of advanced constitutive
and frequencies, as illustrated in Figure 37. When normalised relationships in order to incorporate non-linear response and
with respect to the vehicle or car passing frequency, the load long-term settlement and track degradation (Correia and Cunha,
spectrum depends on the characteristics of the train including 2014; Li et al., 2016; Nguyen et al., 2016; Varandas et al., 2020).
axle and bogie spacing, axle loading, and car length and number. The FEM considers ballast as an elastic or elasto-plastic
Characterisation of train loading in this way has implications for continuum, while semi-analytical spring-mass-damper models
soil element, ballast and component testing. For example, tests in considers all track layers as multi-rigid body dynamics.
which a specimen is subjected to the full axle load at a frequency Mesomechanical behaviours, such as discontinuity, angularity,
based on the passage of adjacent axles will likely significantly breakage, randomness, of the ballast layers are highly simplified
overestimate the actual response. Furthermore, the higher (Bian et al., 2020; Zhang et al., 2017). The DEM has also been
frequencies of loading attenuate rapidly with depth (Powrie et al. successfully implemented to study particle breakage/degradation
2019b). and subsequent settlement of ballast layer under moving load
The dominant frequency for pore pressure generation, and the (Zhang et al. 2017, Indraratna et al. 2014a). In summary, the
ability to calculate realistic residual pore pressures during and choice of analytical and numerical prediction approach should be
after train passage, remain an ongoing debate and challenge. aligned with the expected outcomes that to be investigated.
These various forms of model have been developed in recent
years in an attempt to capture aspects of the critical velocity
problem. Models that account for the development of strains
during train passage are perhaps the most comprehensive but are
very computationally demanding and time consuming. A method
gaining in popularity for large-scale infrastructure is based on the
dispersion curve, which inverts the process followed by
geophysical investigations. A phase velocity versus wavelength
relationship is estimated from an expected profile of shear wave
velocity versus depth for the ground and earthworks, together
with other basic soil characteristics. The minimum estimated
phase velocity is then taken as representative of the critical
Figure 37: Disaggregation of train loading at the sleeper into a series of
velocity of the ground and earthworks system
loads of different magnitude and frequency, for a 20m long vehicle with An approach developed by Sheng et al. (2004) involves
14.2m bogie centres and 2.6m axle spacing, 75 kN wheel load, 25 MPa plotting the dispersion curves (i.e., the frequency-dependent
rail support system modulus and UIC 60 rail (Milne et al. 2017) wavenumbers or wave speeds of free waves that can occur at the
ground surface in the absence of damping) for the ground and for
2.3.4 Resonance (critical velocity) effects the track with a free base. The point at which the wave speeds for
the track and the ground coincide gives the wavenumber /
The amplification of dynamic track deflections below trains as frequency at which waves can propagate in the coupled structure;
the train speed approaches that of (surface) Rayleigh waves in hence the critical velocity of the system and the frequency of
the underlying ground, owing to a resonance effect, is well oscillations (“resonance”) seen behind the load. This approach
established. A train speed at which dynamic response is has recently been taken up by others (e.g. Alves Costa et al. 2015,
significantly amplified due to resonance is called the ‘critical 2020; Connolly et al. 2015), while Kausel et al. (2020) showed
speed’. Several semi-analytical and numerical approaches have how a dispersion curve could be developed for layered ground.
been implemented to predict the dynamic response of high-speed However, it is difficult to apply in practice because of the
railway track at the critical speed. In the semi-analytical methods, information required and the stage in the design process at which
the track and foundation are modelled as variants of beam, it is needed (for example, ground improvement, should it be
masses and spring-dashpot assemblies to represent rail, rail pads, necessary, can only be applied prior to the placement of the track
sleeper, ballast and foundation (Kouroussis and Verlinden, 2015;
53
form). This and other practical design issues are discussed by
Duley (2021). 3.5
Amsterdam - NL: sand,soft clay, peat
Dynamic Displacement / Static Displacement
2.3.5 Dynamic amplification 3 Ledsgard - Sweden: Soft organic clay
Stilton Fen - UK: Very soft silty clay / Peat
As the train speed increases, the dynamic loading due to 2.5 Gravel Hole - UK: Very soft organic silt
acceleration from the components of the vehicles becomes more
significant. Irregularities at the wheel-rail interface amplifies the 2
acceleration the unsprung masses (the wheels and axles).
1.5
Acceleration of both sprung and unsprung masses also occur due
to longer wavelengths irregularities such as track support 1
stiffness and unsupported sleepers (Esveld, 2001; Powrie et al.,
2019b). These dynamic loads are taken into account in track 0.5
design by the Dynamic Amplification Factor (DAF), which is the
ratio of dynamic to the static response (deflection and stress). In 0
0 0.2 0.4 0.6 0.8 1
DAF calculations the track substructure is modelled with a single Normalised Speed ( vtrain/vcrit)
parameter called track modulus. Measurements for field and full-
scaled laboratory tests have shown that the dynamic Figure 39: Normalised plot of track vertical displacement (dynamic
amplification is primarily affected by the Rayleigh wave displacement ÷ static displacement) against normalised train speed (train
propagation speed, type of track (ballasted and slab track) and speed ÷ Rayleigh wave speed) for various case study sites. Data from
stiffness variation (Bian et al., 2014; Indraratna et al., 2010). Woldringh & New (1999) and Duley (2018)
Figure 38(a) shows that the DAF of the vertical displacement
of the rail as a function of the ratio of the train speed, c, to the Other studies conducted on the field and prototype models
shear wave speed of homogeneous ground for an elastic modulus have shown an increase in vertical, lateral stresses and
of 200MPa. Figure 38 (b) shows the DAF of the same track subsequent deformation with increase in the speed of the train
structure modelled with a softer subgrade layer of 3m thickness (Bian et al., 2014; Hendry et al., 2010; Indraratna et al., 2010;
is introduced (Etop = 50MPa, Ebottom = 200MPa). The Lamas-Lopez et al., 2016; Tang et al., 2019b). As a result, the
amplification increases with flexibility of the track (bending track layers not only experience amplified vertical stresses, but
stiffness of slab track V3 is twice of slab track V2). also longitudinal and shear stresses (Malisetty et al., 2020;
Powrie et al., 2007; Yang et al., 2009). The magnitude of these
stresses is found to increase with the speed and axle load of the
train. This complex stress conditions in the track layers leads to
excess deformation and settlements in track substructure layers
including ballast and subgrade soils (Grossoni et al., 2020;
Varandas et al., 2020). Therefore, an accurate prediction of the
complex behavior of railway tracks subjected to an increasing
speed of trains must be achieved through rigorous field
measurements, analytical and numerical analysis.
2.3.6 Novel multi-laminate model for estimating permanent
Figure 38: Dynamic amplification factor of vertical rail displacement (a) deformation at high speeds
homogeneous subgrade (b) layered subgrade (M = c/CS, upper layer) A novel constitutive model in multi-laminate framework has
(Costa et al., 2015)
been developed by Malisetty et al. 2020a and Malisetty et al.
2020b. By considering the evolution of plastic strains at the
A comparison of Figure 38 (a) and (b) shows track stiffness contact planes, the multi-laminate model incorporates stress
and type of track have a significant effect on the DAF and critical
induced anisotropy caused by principal stress rotation inherently
speed, in addition to the influence of subgrade stiffness. The in its framework (Lee and Pande, 2004). The model assumes
relationship between CR (Rayleigh wave speed) and CS (Shear that deformations are induced at the inter-particle contact planes
wave speed) is approximately CR = 0.934CS. According to
irrespective of the loading condition. The overall behaviour of
Sayeed and Shahin (2016) and Dong et al., 2018, the critical the material is captured by numerical integration.
speed increases with the top subgrade stiffness while the DAF Train induced stress paths can be characterized by Cyclic
exhibits a decreasing trend. Therefore the performance of the
Stress Ratio (CSR), shear tress ratio ht and modified CSR*,
track can be improved by stiffening the subgrade using ground
where, ∆&&' is the amplitude of cyclic vertical stress, '(' is the
improvement in soft soil deposits, and installation of piles.
initial mean effective stress and ∆()& is the amplitude of cyclic
New data from Gravel Hole on the UK West Coast Main Line
shear stress.
confirm the overall trend of normalised displacement against
train speed shown in Figure 39.
∆+#$
Presentation of the data in normalised form masks the fact #)*+ = $ (2)
,-%$
that, at low train speeds on a soft subgrade a dynamic
amplification factor of 3 applied to a static displacement of 5 mm
would be a problem; whereas the same factor applied to a static ∆.&#
#,. = $ (3)
displacement of 0.5 mm on a much stiffer subgrade at a much ∆+#$
higher train speed would probably not.
Further challenges remain relating to the accurate
#)*+∗ = )*+-1 + ,., $ (4)
determination of effective soil stiffnesses, Rayleigh wave speeds
and system critical velocities, which depend on a combination of
the characteristics of potentially multiple and / or discontinuous The shear stress ratio ,. is 0 for triaxial compression with no
soil strata, the track form and any earth structures (embankments principal stress rotation and )*+∗ reduces to CS+. Based on
or cuttings) that may be present. the model predictions from Malisetty et al. (2020a), the
54
permanent axial strains (30,#% ) of ballast at shakedown under exceeded 1 for all axle loads. The train speed for which ψ≥1
these stress paths are replotted along with CSR∗ in Figure 40. decreased with increasing axle loads. This implies that freight
For a typical passenger train with sleeper-ballast interface trains with higher axle loads must be limited to lower speeds than
vertical stress amplitude of ∆&′& = 230 ;<= , increasing the that of passenger trains to avoid a rapid increase of shear stress
)*+∗ reduces the stability of ballast leading to a steady within the ballast.
increment of permanent axial strains at shakedown. The
permanent axial strains which reached a stable shakedown state
at low )*+∗ do not tend to reach shakedown, eventually
reaching ratcheting failure at )*+∗ = 5.4. It is evident that the
rotation of principal stress axes at higher train speeds can induce
instability without changing the amplitude of vertical stress.
Further, increasing the vertical stress amplitude ( ∆&′& >
230 ;<=) led to quicker failure (steep slope) when compared to
low ∆&′& which signifies that heavy axle loads imparted by
freight trains increases instability when compared to passenger
trains travelling at the same speed. The modelling outcomes
show that if the influence of Raleigh waves and principal stress
rotations are not incorporated in the design of tracks for high
speed trains, the deformation of ballast can be underestimated
and can affect track stability causing unevenness of the track
profile, ultimately increasing the frequency and cost of Figure 41: Dynamic stress amplification in the ballast layer at different
maintenance operations. train speeds (Malisetty et al., 2020a).
The corresponding CSR* was computed for all the stress path
at different train speeds and axle loads the results are plotted in
Figure 42. Using the multi-laminate model predictions of ballast
(See Figure 40), the CSR* corresponding to greater than 20%
permanent axial strains was considered as failure and is shown in
Figure 42. The region below the failure plane can be considered
as a stable loading zone and the region above the failure plane as
an unstable loading zone. The intersection line represents the
critical train speed for different axle loads. The critical train
speeds corresponding to 20% axial strain for different axle loads
are very close to that of those estimated for ψ≥1, which indicates
that failure or excess deformations occur when the peak stress
applied on the material is exceeded.
Figure 40: Influence of cyclic principal stress rotation on shakedown
axial strain response of ballast (data sourced from Malisetty et al. 2020a)
2.3.7 Estimation of critical train speed for different axle loads
Using the multi-laminate model predictions from Malisetty et al.
(2020a), the critical train speed for a ballasted railway track can
be analyzed for different axle loads. The axle loads considered
for analysis represent passenger (P=15 tonnes), coal freight (P =
25 and 30 tonnes) and iron ore freight wagons (P = 40 tonnes).
The dynamic stresses in the ballast layer were analysed using
beam on layered elastic half-space model, considering Rayleigh
wave propagation and dispersion in the ballast and underlying
subgrade layer. The elastic modulus for ballast (Eb=100 MPa)
and alluvial subgrade (Eb=30 MPa) are adopted from field data
(Nimbalkar and Indraratna, 2016). Two approaches for
estimating the critical speed are discussed here using: peak static
shear strength of ballast and maximum allowable permanent Figure 42: Critical speed estimation in ballasted railway tracks for
axial strains under repeated loading. different axle loads based on multi-laminate model predictions (Malisetty
For each axle load, the stress response at the mid-depth of the et al. 2020a)
ballast layer was calculated and the ratio of dynamic shear stress,
qdyn, to peak static shear stress, qsta_peak, (y) developed for a While the amplification factor is same for all axle loads, the
confining stress of 30 kPa. This ratio computed for different train high magnitude of dynamic stresses at high axle loads plays an
speeds is shown in Figure 41. important role in determining the critical speed. In essence, it
The dynamic shear stress amplifies with increasing the train can be interpreted that the critical train speed of the track not only
speed for all axle loads. At a train speed of V=270 kmph, the depends on the track material properties and geometry, but also
Raleigh waves generated by the train reached a resonance state on the axle load; higher the axle load, lower will be the allowable
causing maximum stress amplification and this speed was found maximum train speed.
to be unique for all axle loads. However, when the train speeds However, these critical speeds were only determined for the
lie in the sub-critical zone (V<270 kmph), the value of ψ shakedown performance of ballast. The influence of wave
55
propagation on the number of loading cycles after which this
shakedown state is reached needs further research. The influence
of Raleigh waves on the instability of undrained subgrade soils
needs to be addressed in more detail using the multi-laminate
constitutive model and needs to be integrated with Finite Element
Modelling. This is required to investigate the influence of
principal stress rotation due to moving trains on mud pumping
behavior and stability of subgrade soils, where excess pore water
pressures play an important role.
2.3.8 Dynamic stresses in the track under high-speed train
loading
The state of stress in the track subjected to moving train load
exhibits a complex stress path as the load approach and passes a
given location. As the train moves towards an observation point,
the shear stress gradually increases, and then decreases to pass
through zero when the load is exactly above the observation
point, while the vertical stress peaks. The shear stress then
changes its sign and increases as the load moves away from the
observation point before gradually returning back to almost zero.
This makes the shear stress to be 90 degrees out of phase with
the normal stress, which in turn results continuous rotation of
principal stress axis. Several studies have shown that principal
stress rotation accelerates permanent deformations compared to
a single point cyclic loading (Bian et al., 2020; Cai et al., 2018).
Yang et al. (2009) showed, using a 2D FE model, that the Figure 43: Finite Element Modelling (a) Section (b) Deformation
vertical and longitudinal shear stresses increase with the train contours at 60km/h (c) Mach effect (vertical deflection) at 300km/h
speed . Numerical results indicated that the vertical and shear (unpublished part of a current program at the University of Technology
stresses at the critical speed are about 20% and 80% higher than Sydney through funding from the Australian Research Council, courtesy
those calculated in a static analysis, respectively. The shear stress A. Tucho, PhD student)
at 50% of the critical speed was underestimated by 30% while
the error in the vertical stress was about 15% in static analyses.
Three dimensional FEM has been used to compare the stress
state in the ballast layer at low and high speed (Indraratna et al.
2010). The axle load configuration mimics the RAS 210
Australian Standard Wagon equivalent to 20.5t axle load. The
ballast and capping material parameters were adopted from
triaxial test results reported by Indraratna et al. (2010) while the
subgrade was assumed to have a resilient modulus of 50MPa.
The track was modelled with solid elements and infinite element
boundaries were applied at the all sides except the centerline
(Figure 43a). Figure 43b and 43c show the difference in
deformation contours when the speed of moving load changes Figure 44: Stress Distribution at a speed of 60km/h and 300km/h
from slow speed to high speed, where the critical speed of the (unpublished part of a current program at the University of Technology
track ground system which is about 85m/s. At low speed the Sydney through funding from the Australian Research Council, courtesy
vertical transient vibration is symmetrical while asymmetrical A. Tucho, PhD student)
transient displacement is observed at high speed (e.g. Connolly
et al., 2013). 2.3.9 Ballast structure, grain shape and mineralogy
Figure 44 shows the corresponding stress changes at low
speed (0.2Vc) and near the critical speed of the track-ground Ballast is gravel to cobble-sized hard granular material that is
model computed at the middle of the ballast layer (Point A of placed below and surrounding sleepers. The ballast provides
Figure 43a). Simulated results show that at the critical speed the the following functions (Li et al, 2015):
vertical stress increased about 50% while the shear stress along
the longitudinal direction increased about 200% in the opposite • Supports the rail and sleepers by providing vertical,
direction. lateral and longitudinal resistance
For low speed, almost all the stress curves are symmetrical as • Transmits and reduces rail wheel forces
shown in Figure 45. For high speed, the peak value of each stress • Provides drainage
is higher indicating the existence of vibration amplification. The • Provides resilience and damping of wheel/rail forces
symmetry of stress response is lost except for the vertical normal
stress. Yang et al. (2019) and Dong et al. (2019) reported similar Heavy cyclic loads cause progressive deterioration and
asymmetry of stress responses which is attributed to the load breakage of ballast, and consequently the loss of track geometry
reaching the reference point before the stress wave has fully leading to frequent track maintenance. During train operations,
propagated to the same point. Comparison of dynamic stress path ballast deteriorates due to the degradation of angular corners and
from the double-bogie load in Figure 44 shows that the deviatoric sharp edges, the intrusion of fines, and mud-pumping from the
stress and shear stresses increase significantly with the train underlying soft soil layers. In addition, impact forces generated
speed. Similarly, high speed results in significant upward by wheel and/or rail irregularities or imperfections or at
displacement and negative mean stress, which could lead to loss transitions zones (e.g., bridge approaches and exits, road
of confinement and instability (Dong et al., 2019). crossings) may exacerbate track deterioration (Powrie et al.
2007; Indraratna et al. 2019a). As the sharp corners of aggregates
56
break and weaker particles are crushed under heavy cyclic changes in both angularity and form. Abadi et al. (2021)
loading, differential track settlement can occur (Indraratna et al. quantified angularity from the dimensions and volume of solid
2014a). As the ballast assembly becomes degraded its grain ellipsoids fitted to 3D images of ballast grains captured using a
angularity, overall strength and drainage capability may be laser scanner. Figure 46 shows data of form for fresh and used
reduced (Indraratna et al. 2011b, Tutumluer et al. 2012, ballast grains; the used ballast is generally more spherical than
Tennakoon et al. 2012). the material.
0.6
0.5
0, 0.4
ss
e0.3
yn
ta
50-63mm
l 40-50mm
P0.2
31.5-40mm
0.1
22.4-31.5mm
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Elongation, ?
Figure 46: Plot of Platyness vs Elongation (Potticary et al., 2015) for
fresh and used ballast. Orange data points represent fresh ballast and blue
data points represent used ballast of different sizes (data extracted from
Abadi et al. 2021)
Figure 45: Stress paths at a speed of 60km/h and 300km/h based on the On loading, the local surface roughness seems to be affected
stresses shown in Figure 44 (courtesy A. Tucho) immediately at the inter-grain contacts (Gupta et al. 2021);
Figure 47.
The past decade has seen a significant number of studies
investigating the performance of railway ballast at the grain
scale. These have included experimental investigations of 2.85mm
2.85m
structure using X-ray computed tomography (CT scanning) of
samples whose structure was fixed by impregnation with resin, m
2.16m
2.16mm
detailed assessment of grain mineralogy and shape, and
numerical analysis using the discrete element method. m
Studies such as those reported by Le Pen et al (2014) and
Aingaran et al (2018) have shown that during trafficking, ballast
develops a dense structure through a reduction in void ratio and
an increase in average coordination number. Maintenance by
tamping disrupts this structure by disturbing the grains;
subsequent trafficking does not then ever quite recover the
maximum bulk density originally attained.
Ballast grain shape may be characterised by three measures 1.19mm (a)
(in decreasing order of scale) of form, angularity and roughness.
Grain shape influences the density, void ratio and hence the 2.85mm (a)
stiffness and (peak) strength of the assembly. 2.16mm
Le Pen et al (2013) showed that form and angularity varied 0.6mm
slightly but not significantly across particle scale and introduced
the concept of “ellipseness” (rather than the traditional
“roundness”) to quantify angularity. Potticary (2018), Potticary
et al. (2014, 2015, 2016) and Harkness & Zervos (2019) (b)
quantified ballast form by combining the Longest (L),
Intermediate (I) and Shortest (S) orthogonal dimensions of
ballast grains into two parameters they termed platyness (α) and Figure 47: Surface characterization of ballast particles using varying
elongation (ζ) (defined in Figure 46). This results in individual focus microscopy. a) fresh ballast surface (average surface roughness =
grains plotting onto a triangular diagram (Figure 46), which is 50µm); b) used ballast surface (average surface roughness = 20µm) (after
considered to be an advance on the conventional Zingg plot of Gupta et al. 2021)
I/L vs S/I because (i) each grain is characterised as one of three
distinct shape types (spherical, columnar or platey: the Zingg plot Monotonic triaxial tests, and cyclic loading in sleeper rig
distinction between platey and blade-like is rather subtle), and tests, show no appreciable difference between fresh and
(ii) each of the shape parameters include all three of the principal recovered granite ballast, suggesting considerable potential for
dimensions. re-use (Abadi et al., 2021).
Recent work on ballast degradation has focused on how the Ballast fouling is a potential concern and has been
three measures of grain shape change with loading. Mineralogy investigated through triaxial and laboratory rig tests and discrete
is important: much empirical evidence of the nature and element analysis (e.g. Indraratna et al. 2014a, Ngo et al. 2014).
timescale of ballast degradation by grain attrition and breakage Experience in the field and physical test data suggest that dry
is based on relatively low quality ballast (particularly limestone), fouling, for example by sand grains small enough to fall through
and is unlikely to be applicable to modern high quality ballast the voids to the bottom of the ballast bed, makes no difference to
such as granite. For granite ballast, trafficking seems to result in the resilient stiffness until the voids in the ballast are filled to
57
sleeper soffit level. At this point, the resilient stiffness increases Measures that have been shown to reduce the rate of
significantly and ongoing plastic settlement substantially stops development of permanent settlement of the ballast bed, at least
(University of Southampton, 2013). However, fouling by wet, in full-scale laboratory rig tests (Abadi et al. 2016, 2019;
fine (clay-like) particles is generally held to reduce the resilient Navaratnarajah et al. 2018; Jayasuriya et al. 2019) include the
stiffness and accelerate plastic settlement. This could be a result following:
of associated subgrade deterioration, and / or the fine, wet fouling • under sleeper pads (which increase the ballast to
particles reducing the stiffness and strength of the ballast grain sleeper contact area when a standard concrete sleeper
contacts. While there have been several individual or case is used, providing a more stable and less aggressive
specific studies, a unified theory of the effects of ballast fouling interface)
accounting for the presence or absence of water has not as yet • the use of duoblock rather than monoblock sleepers
been postulated. (which eliminates the possibility of “centre binding”,
but for other reasons seems to be falling out of favour
2.3.10 Ballast confining pressure in railway administrations around the world)
• changing the grading to include a greater proportion of
The confining pressure is a significant factor in ballast slightly finer grains, improving mechanical
performance. Large-scale cyclic triaxial tests have been interlocking.
performed to investigate how confining pressures affect ballast
breakage under high cyclic loading (Lackenby et al. 2007, Sun
et al. 2019). Ballast was tested under different effective confining
pressures (s3¢) ranging from 10kPa to 240kPa with deviator
stresses up to 500kPa. Figure 48 shows the effect of confining
pressure on the axial and volumetric strains in the ballast; here
the axial strains decrease with an increasing confining pressure
and the ballast assemblies dilate at a low confining pressure (s3¢
< 30 kPa), but become progressively more compressive as the
confining pressure increases from 30 kPa to 240 kPa. The effect
of s3¢ on particle breakage is shown in Figure 49, where breakage
is divided into three regions: (I) a dilatant unstable degradation
zone (DUDZ); (II) an optimum degradation zone (ODZ); and
(III) a compressive stable degradation zone (CSDZ).
Figure 49: Effect of confining pressures on particle degradation
(Lackenby et al. 2007)
Methods of reducing ballast settlement by preventing or
restricting lateral spread include
• reinforcement with random fibres (Ajayi et al. 2017);
the development of tensions in the fibres increases the
effective confining stress on the ballast
• reinforcement with geogrids (Indraratna et al. 2011b)
• physical confinement
• reducing sleeper spacing, widening the sleeper at the
ends or using intermittent lateral restraints (Indraratna
et al. 2011b) to increase resistance to ballast lateral
movement and possibly the confining stress
• Increasing the height of a ballast shoulder raised above
the level of the sleeper surface (Lackenby et al. 2007), or
• reducing the slope of a ballast shoulder that falls away
Figure 48: Effect of confining pressures on axial and volumetric strains below sleeper surface level, where (as is almost
(Lackenby et al. 2007)
invariably the case) the ballast bed is raised (Abadi et
al. 2016).
2.3.11 Ballast permanent settlement
It is well known that railway track gradually settles over time With the possible exception of under sleeper pads, few if any
with trafficking. Both the subgrade and the ballast may of these techniques have been adopted in practice. This may be a
contribute to this settlement; but it should be possible to result of concerns regarding cost, interference with maintenance
minimise at least the component attributable to the ballast by by tamping (e.g. the use of geogrids), the lack of available land
appropriate methods of installation and maintenance. After initial along the trace width (reducing the ballast shoulder), or a lack of
densification (which may occur because the track is lifted to field demonstration of the efficacy of the technique.
above its final geometry by tamping), ongoing settlement is
usually attributed to ballast grain breakage (often due to
maintenance tamping) and lateral spreading of the ballast bed.
58
2.3.12 Characterising fouled ballast In Equation 7, qpeak,b and qpeak,f are peak deviator stresses for
fresh and fouled ballast respectively, and b is an empirical
Methods to quantify the degree of fouled ballast include Selig parameter for fouled ballast. The values of b for different fouling
and Waters (1994) and Feldman and Nissen (2002). These agents can be determined from permeability tests.
methods do not account for the differences of specific gravities
of various contaminants (e.g., coal, clay, mud, pulverised rock). 10
0
A void contaminant index (VCI) (Tennakoon et al. 2012) was Coal-fouled ballast: Experimental
developed to include contributions from various materials. -1 Coal-fouled ballast: Theoretical
10 Sand-fouled ballast: Experimental
Hydraulic Conductivity, k (m/s)
Sand-fouled ballast: Theoretical
234' 5 6' -2
10
B)C = 4(
× 5)( × 6 × 100 (5) Bellambi Site
)' (
-3 VCI=33% Rockhampton Site
10 VCI=72%
where, ef, eb are void ratio of fouling and ballast; Gsb, Gsf are
-4
specific gravity of ballast and fouling; Mf, Mb are weight of 10 hydraulic conductivity
of coal fines
fouling and ballast. -5
Sydenham Site
VCI=22%
Tennakoon et al. (2012) studied the effect of the degree of 10
hydraulic conductivity
fouling on the overall hydraulic conductivity of fouled ballast -6 of clayey fine sand
using a large scale permeability apparatus. Based on test results, 10
a relative hydraulic conductivity ratio (kb/k) was introduced to -7
10
understand how the overall hydraulic conductivity (k) varies in 0 20 40 60 80 100
comparison to clean ballast (kb), as shown in Equation 6.
Void Contaminant Index, VCI (%)
7( 89: 7 Figure 51: Variation for hydraulic conductivity vs. void contaminant
7
= 1 + 2(( E7( − 1G (6) index for coal/clayed fine sand (Tennakoon et al., 2012)
'
The study showed that a 5% increase in VCI decreases the The allowable train speed Vf for fouled ballast can be
hydraulic conductivity by a factor of at least 200 and 1500 for estimated as:
ballast contaminated by coal and fine clayey sand, respectively.
/ 1 /
Variations of relative hydraulic conductivity of ballast with ;*+,-,'?@3 A? 3 A B
VCI are shown in Figure 50. The drainage ranges are subjective B> = H 3)
0 2 0
− 1I (.((D,
4
(8)
0
and will vary depending on the local climate, track use, etc. For
example, a hydraulic conductivity of 10-5 m/s for a section of Where Pd is the design wheel load, qr is the rail seat load, B is
track may be acceptable (due to the low rainfall etc.), but sleeper width (m), l is sleeper length (m), h is ballast thickness,
unacceptable for a track section subjected to very heavy rainfall. Dw is the diameter of the wheel (m) and V is the velocity of the
Figure 51 shows the variation of hydraulic conductivity for train (km/h). Figure 52 shows the predicted maximum train speed
fouled ballast with increased VCI. The overall hydraulic for coal-fouled ballast at varying levels of VCI. Both the
conductivity always decreases with an increase in VCI. When the confining pressure and level of fouling affect the operational
VCI > 75%, further reduction in hydraulic conductivity becomes train speed (Vf). Train speed can be slightly increased with
marginal as it approaches the hydraulic conductivity of the increased track confinement but is significantly reduced with an
fouling material itself. These observations agree with the increase in fouling, in particular when VCI is greater than 30%
laboratory measurements of sand-gravel mixtures reported by at which stage track maintenance could be initiated to avoid
Jones (1954), whereby a high percentage of sand (greater than potential track problems and to increase passenger comfort
35%) in gravel provides a hydraulic conductivity close to that of (Indraratna and Ngo 2018).
the sand alone.
Figure 50: Variation of relative hydraulic conductivity with Void
Contaminant Index for coal fouled ballast (Tennakoon et al., 2012)
Figure 52: Predicted allowable train speed with VCI for a heavy haul
of 30 tonne axle load (Tennakoon et al., 2012)
VCI has also been related to the shear strength of fouled
ballast (Indraratna et al. 2013a). 2.3.13 Measurement and modelling of ballast particles
;*+,-,' 2 Settlement of ballasted tracks occurs due to rearrangement of
= (7)
;*+,-,( 23<√89: ballast particles as well as their lateral spread under cyclic train
loads. The mechanisms controlling ballast settlement have been
investigated by experimentally and using the numerical discrete
59
element method (DEM). Zhai et al. (2004) placed multiple of different sizes and positions via a clump approach. Indraratna
accelerometers in ballast layer to measure the particle et al. (2014a) adopted the clump approach and simulated the
accelerations for modelling ballast vibrations under moving angular ballast aggregates in DEM. The simulated particles were
trains. Kumara and Hayano (2016) used a method of particle generated in a simulated shear box at random orientations to
image velocity (PIV) to track local deformations induced by resemble experimental conditions. Huang and Tutumluer (2011)
tamping tools. A series of studies have been recently conducted presented a method to simulate fouled ballast in DEM by
to monitor the movement of ballast particles using a wireless reducing the surface friction angle of ballast. Indraratna et al
device called “Smart Rock”. Liu et al. (2016a) developed this (2014a) introduced an approach to model fouled ballast by
device by encasing a motion-sensing unit in a ballast shaped PLA injecting various amounts of fine particles into the voids to
plastic object and used it to study the motion of ballast particles represent different values of void contaminant index (VCI).
reinforced with different geogrids. Zeng et al. (2019) used small Liu et al (2016a) compared the particle displacement and
cubes with embedded motion sensors to monitor ballast orientation data with their distinct element model (DEM) and
movement for the identification of its condition. In these studies found both to be in good agreement as shown in Figure 55. In
the SmartRock did not either have the density of an actual rock another study, they have also shown that their DEM model was
or the shape of a realistic ballast particle as shown in Figure 53. able to capture the ballast particle accelerations that could
Ballast particles with realistic geometry and density have compare well with the data obtained through “Smart Rock” as
been created by 3D printing using a special kind of filament shown in Figure 56.
infused with metal powder to provide a realistic weight. The final Discrete element analysis (DEM) has become increasingly
product as compared to actual ballast particle is shown in Figure popular for modelling the behaviour of railway ballast at the
54. grain scale, not least because the relatively large grain size
compared with soils means that a realistic number of individual
grains can be included in simulations. Modelling approaches
have moved from 2D to 3D, and away from simple single or
bonded pairs of spheres to include agglomerates of large numbers
of spheres clustered together to represent “catalogues” of ballast
grain shapes obtained by 3D scanning (e.g. Ngo et al. 2016, Li
and McDowell 2019, Shi et al. 2020); polyhedra (e.g. Tutumluer
et al. 2013); or level-set defined “potential particles” (Harkness
2009), which take the form of slightly rounded polyhedra, and
super-ellipsoids (e.g. Duriez & Galusinski 2021).
Ways of modelling the effects of ballast attrition in DEM have
been developed, primarily involving the breakage of asperities
(McDowell and Li, 2016) and the “conical damage” approach
(Harkness et al. 2016), also adapted by Suhr et al. 2018.
Figure 53: Smart Rock Liu et al. (2016a) Zeng et al. (2019) Modelling attrition in some way seems necessary in DEM to
replicate the transition from dense to loose behaviour of a
granular medium as the confining effective stress is increased
(Ahmed et al. 2016).
DEM has been used to investigate various aspects of ballast
behaviour: geogrid reinforcement (Ngo et al. 2014, 2016, Qian
et al. 2015), effects of layers of differently-sized particles (Li and
McDowell, 2020), lateral resistance (e.g. Chen et al. 2015),
differences between fresh and recycled ballast (e.g. Jia et al.
2019), ballast settlement (e.g. Tutumluer et al. 2013), under-
sleeper pads (Ngo and Indraratna, 2020) and tamping (e.g. Shi et
al 2020, Guo et al. 2021).
With DEM simulations and other advanced geotechnical
modelling studies, the computed behaviour is increasingly
dependent on the underlying constitutive models and input
parameters used. In the case of DEM, there is an evidence gap
between a phenomenon and how it is represented in the
simulation (for example, representing ballast grain attrition by a
change in surface roughness, although both the conical damage
model and the breakage of asperities does do offer a more direct
link); with the result that DEM may often indicate a plausible
rather than a definitive explanation. As is often the case,
empirical validation through high quality field measurements
Figure 54: Smart Ballast particle (unpublished part of a current program made over a long period of time remains both essential and a
at the University of Technology, Sydney through funding from the major challenge.
Australian Research Council), courtesy A. Siddiqui, PhD student) DEM simulations can assess the evolution of contact force
distributions within the fresh and fouled ballast assembly.
Liu et al. (2016a) used the “Smart Rock” particles in a ballast Distributions of contact force chains of fresh and fouled ballast
box representing half section of a rail track structure and (VCI=25%, and 50%) at an axial strain of εa=10% under a given
concluded that the particle translation and rotation are important confining pressure of σ3=30 kPa are shown in Figure 57. Contact
modes of ballast particle movement that affect ballast forces between grains are plotted as lines whose thickness is
performance. proportional to the magnitude of the force. Contact forces were
The discrete element method (DEM) has been widely used to distributed non-uniformly throughout the assembly and
study the micro-mechanical behaviour of ballast (e.g. Zhang et transmitted vertically from the top plate to the bottom of the
al. 2017). In DEM, irregularly-shaped ballast aggregates can be simulated triaxial apparatus which was aligned in the loading
modelled by connecting and overlapping several spherical balls
60
direction. The number of particle contacts increased significantly which provide a subsequent input to the DEM (Ngo et al 2017).
with an increase of VCI and consequently the magnitude of the The coupled DEM-FEM model was used to analyse the load-
contact forces decreased hence decreasing ballast strength. deformation responses of an instrumented track and the
numerical predictions are compared with the field data in Figure
58. The field study was carried out along a section of
instrumented track constructed over general fill and rock
subgrades. The track substructure consisted of a 300 mm thick
layer of ballast placed over 150 mm thick capping layer,
overlying structural fill (thickness varying from 500 to 900 mm).
Figure 55: Comparison of Ballast Particle Displacement and Orientation
Data with DEM (Liu et al. 2017a)
Figure 57: Contact force distributions of fresh and fouled ballast in
simulated triaxial tests (Ngo et al. 2017)
2 Field data (hard subgrade)
4 Model prediction
Settlement (mm)
6 Field data
Soft subgrade
8
10
Singleton track, soft subgrade Model prediction
12 Singleton track, hard subgrade
Model prediction, soft subgrade
Model prediction, hard subgrade
14
0 2000 4000 6000 8000 10000
Load cycles, N
Figure 56: Comparison of Ballast Particle Accelerations with DEM (Liu Figure 58: Comparisons of settlement between the combined model and
et al. 2016b) field data (Ngo et al. 2017)
DEM simulations of complete track systems are too Comparisons of predicted settlement (Sv) of ballast for two
computationally intensive for practical (or academic) use. The types of subgrades obtained from the model and field
method can be extended by coupling it with the finite element observations are presented in Figure 58. It is seen that the model
method (FEM). The coupled DEM-FEM method has been used predicts the vertical deformation of ballast well in relation to the
to study the load-deformation response of the ballast assembly measured field data. The larger settlement predicted for soft
while considering interaction between the ballast aggregates and subgrade (Sv =10.2mm) compared with the field observation of
the subgrade layer. In this coupled model, the discrete ballast Sv = 8.7mm can be attributed to inevitable discrepancies between
grains were modelled by DEM and the capping/subgrade layers the field conditions and simplified boundary conditions applied
are modelled as a continuum by FEM. Interface elements are for the coupled DEM-FEM model.
introduced to transmit the interacting forces and displacements The simulation also showed that ballast directly underneath
between adjoining material domains. The DEM transfers contact the sleeper experienced higher contact forces than ballast at the
forces to the FEM, and then the FEM updates the displacements shoulder of the embankment. This caused vertical stress at the
61
interface between the ballast and capping to be distributed non- The resilient modulus (MR) is also frequency dependent.
uniformly along the interface and was greatest towards the MR is the ratio of the applied deviator stress (Dqcyc= qcyc,max -
centreline of track. These simulations were consistent with the qcyc,min) to the recoverable (resilient) axial strain (εa,rec) during
data measured in the field by Indraratna et al. (2014b). unloading.
2.3.14 Effect of load frequency on ballast ∆;"5"
J+ = (9)
E,,6+"
A series of large-scale triaxial cyclic tests on ballast was
conducted to study the load-deformation behaviour of railway
ballast under varied loading frequencies. Tests were carried out
for ballast under different frequencies of f=10-40Hz subjected to
a confining pressure of 20kPa, mimicking low confinement in the
field. Cyclic tests were conducted within the bounds of qcyc,max
=230kPa and qcyc,min =30kPa. Fresh latite basalt was used in this
study. Stress-controlled cyclic loading was applied and the tests
were ceased after N=500,000 cycles. During these tests, the axial
strain, ea and volumetric change, ev were recorded by automated
data loggers. Ballast aggregates were recovered at the end of each
test, and the changes in particle size distributions were measured
to analyse the extent of breakage.
The axial strain increases with an increase in loading cycles
and frequency. The rate of increase in axial strain reduces with
the number of cycles. When the ballast aggregates were
compressed to a threshold packing arrangement, subsequent
loading would initiate volumetric dilation and particle breakage.
Figure 59 presents volumetric strain versus number of cycles
and shows that all tests exhibited volumetric compression at
initial loading cycles (N=1000) progressing to a threshold
compression, followed by a decreasing rate of compression at Figure 60: Measured ballast breakage subjected to varying loading
subsequent loading cycles. Under low frequencies of f=10, 20Hz, frequencies f, with and without the inclusion of recycled rubber mats
ballast exhibits volumetric compression throughout the tests. On (Indraratna et al. 2019b)
the other hand, at high frequencies of f=30 and 40Hz, ballast
shows dilation behaviour after N=10,000 cycles. The resilient modulus of granular materials can be expressed
by a simple hyperbolic model as a function of the bulk stress, θ
which is the sum of the principal effective stresses σ1′+ σ2′+ σ3′.
In Equation 10, k1 and k2 are empirical coefficients.
J+ = ;2 K70 (10)
Figure 61 presents the variation of MR at different load steps
with different frequencies for comparison with data conducted
earlier by Lackenby et al (2007) at f=20 Hz. It is seen that
Equation 10 can be used to estimate the strain hardening and
softening of ballast at various frequencies.
For example, a standard gauge track with 25t axle load
(qmax,cyc=230kPa) operating at 220km/h (f=30 Hz) often has a
designed track confinement of less than 30kPa. With θ=420kPa
and f=30Hz, the ballast layer will experience strain softening; for
example, k1=4.10 and k2=0.55 is selected to calculate a value of
MR =113.65MPa. The study also indicated that at a higher stress
level, the increased contact area between particles decrease the
freedom of ballast aggregates to deform elastically, which is
strain hardening. As frictional failure occurs and significant
particle sliding and rotating results in a more open structure
Figure 59: Measured volumetric strain of ballast and ballast breakage causing a larger elastic strain indicating strain softening.
index with frequency (Indraratna et al. 2019b)
2.3.15 Substructure Capping (sub-ballast) Layer: Role of
Ballast breakage can be reduced by installing rubber mats
Permeability, Drainage and Filtration
below the ballast. The amount of ballast breakage with and
without the inclusion of a rubber mat is presented in Figure 60 A capping layer is required for protecting erodible subgrade fines
for various frequencies. The ballast breakage increases with an while performing its normal function of sustainable load transfer
increase in frequency and the ballast assemblies having the to the underlying subgrade. Most compacted capping layers tend
rubber mat show a decreased breakage compared to those to be very stiff after construction and become even stiffer over
without rubber mats. The reduction of ballast breakage is time due to cyclic (repeated) loading. Increased compaction of
primarily caused by the energy absorbing characteristics of the sub-ballast (capping) materials can lead to a higher propensity
mat which results in less load being transferred to ballast grains for tensile cracking as well as reducing its permeability thus
(Navaratnarajah et al. 2018). The inclusion of a rubber mat preventing the dissipation of any dynamic excess pore water
decreases the BBI up to 30.8% and 35.3% for ballast subjected pressures that may inevitably build up during the passage of very
to a frequency of f=10 and 20Hz, respectively. heavy (e.g. 40t axle load) and very long freight trains (up to
62
5km). This situation may be exacerbated if the cracks retain ballast capping material (designed to act as an effective filter) to
infiltrated water. be internally stable.
GI >
MGHD /OJD,#0 ≤ 1 (12)
Empirical equation
800 f=5 Hz Current study
f=30 Hz Current study 2.3.16 Subgrade response to high-speed train loading
f=20 Hz Lackenby et al. (2007)
Englund (2011)
Subgrade plays an important role in maintaining satisfactory
600 performance of railway track. Experimental investigations on its
Resilient modulus, MR (MPa)
cyclic behavior have been performed by many researchers under
undrained conditions in the past several decades (Liu and Xiao
2010; Cai et al. 2018; Indraratna et al. 2020a), among others.
400 Test results show that soft subgrade mainly exhibits stiffness
degradation and plastic strain accumulation under cyclic loading
MR=k1q
k2
induced by passage of moving trains, which become more critical
k1 k2 when freight trains with higher speed and heavier axle loads are
200 f=5Hz 8.61 0.59
f=30Hz 4.10 0.55 introduced.
Englund (2011) 35.44 0.41 The load induced by a moving train on a saturated fine grained
soil results in principal stress rotation and accumulation of pore
0 water pressure, which in turn results stiffness degradation and
0 200 400 600 800 1000 1200 1400 1600 1800 2000 plastic strain accumulation (Dareeju et al., 2017; Gu et al., 2019;
Bulk stress, q (kPa)
Jefferies et al., 2015). Some drainage and dissipation of excess
pore water pressures may occur during and after the train load
passage. The time rate of dynamic pore pressure accumulation is
Figure 61: Measured resilient modulus of ballast (Sun et al. 2019)
influenced by train speed/frequency, soil permeability and
stiffness of the subgrade (Bian et al., 2019; Powrie et al., 2019b;
An alternate approach is to provide capping with a more open Wang et al., 2019). Therefore, the speed of the train relative to
grading so that it drains freely. The capping needs to act as a the time scale of pore water pressure dissipation/ permeability is
filter to prevent subgrade pumping and the subgrade needs to be important to understand the mechanism. In Figure 62, as the ratio
relatively insensitive to water. of load speed, c, to subgrade Darcy permeability, kD increases,
For a compacted sub-ballast material to act as an effective the maximum excess porewater pressure in the subgrade
filter, it must satisfy three conflicting requirements: (a) increases then decreases after the critical speed (Bian et al.,
sufficiently fine to arrest the pumped subgrade fines without 2019). Little excess pore pressure develops when the non-
clogging, (b) sufficiently coarse to avoid development of excess dimensional ratio (c/kD) is less than about 100. Above this
pore water pressure, and (c) to ensure that the finer fraction from value train speed starts to generate additional excess pore
the compacted material itself is not washed out under elevated pressure.
hydraulic gradients induced by dynamic loads (Indraratna et al.
2018). The original filter selection criteria were established for
static conditions based on soil gradations alone and they cannot
always ensure effective filter performance for severe dynamic
conditions imposed by heavy haul trains (Indraratna et al. 2018).
Sinusoidal loading up to 25Hz frequency (175km/hr) was
simulated in the laboratory. The time-dependent reduction in
permeability of subballast acting as a filter due to captured
subgrade fines and compaction due to cyclic loading was
quantified. Based on these results, the following retention
criterion was proposed to select effective subballast filter (Trani
and Indraratna, 2012).
B"27
F87
≤3−4 (11)
where, Dc35 represents the particle size in the filter corresponding
to the 35th percentage passing by mass and d85 the particle size
of the underlying material corresponding to the 85th percent Figure 62: The effect of speed and c/kD on maximum pore water pressure
passing. (Bian et al., 2019)
Having examined the effects of compaction and loading
conditions on various specimens of sands, sand-gravel mixtures, The dynamic response of saturated subgrade has been studied
and silt-sand-gravel mixtures under internal erosion tests (Israr using poro-elascticiy (Bian et al., 2019; Li et al., 2019).
and Indraratna, 2019), two major phenomena have been However, poro-elastic dynamics cannot capture the stiffness
identified as governing factors of internal erosion under cyclic degradation and plastic strain accumulation caused by a moving
loading; (1) excessive perturbation due to dynamic loading, and load. Zhao et al (2021) presented a numerical study using finite
(2) accumulation of excess pore pressure. The former causes element method that investigated the dynamic performance of
constrictions to vary between the loosest and densest states, saturated subgrade under moving train loadings, in which a two-
while the latter causes the local hydraulic gradients to phase porous medium consists of soil skeleton and pore fluid was
exponentially increase and reach the critical hydraulic gradient considered in the analysis. Figure 63 shows the model
causing internal instability. The ratio of controlling constriction predictions in terms of normalized pore water pressure versus
size of the stable coarse fraction in the loosest state of vertical strains. The simulation results show that the model can
compaction and representative particle size by surface area of the capture the build-up of pore water pressure with increased
potentially erodible fraction must satisfy Equation 12 for a sub- loading cycles that have been widely observed in the laboratory
63
tests. Figure 64 presents the stress path that showing a continuous fluidisation of fine particles, some studies indicate the role of
rotation of principal stress orientations. This kind of heart-shape varying accelerating factors such as the degraded contact
envelope has also been reported by other studies via numerical between ballast and the subgrade, freeze-thaw cycles causing the
investigations or analytical approximations (Li et al., 2019, Bian ponding of water under the tracks in cold regions and stress-
et al., 2019), among others. With continuous passage of moving strain concentrations under rail joints, switches and transition
trains, the stress path gradually moves to a lower stress state due zones in exacerbating mud pumping (Powrie, 2014; Nguyen et
to the build-up of pore water pressure. al., 2019).
Subgrade with low-to-medium plasticity is more vulnerable
0.4 to mud pumping. Figure 65 shows that the plasticity chart for
vulnerable subgrade lie in close vicinity of the A-Line, and high
plasticity soils have not exhibited signs of mud pumping. It is
Pore water pressure ratio, (p/à 'z0)
believed that highly plastic clays undergo undrained shear failure
0.3 as opposed to subgrade fluidisation. With the addition of 10%
plastic fines (artificial kaolin) by dry weight, Indraratna et al.
(2020b) observed the change in the mode of failure of subgrade
from fluidisation to undrained shear failure.
0.2
0.1
1st cycle
0.0
-0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10
àz (%)
Figure 63: Pore pressure ratio with vertical strain (Zhao et al. 2021)
Figure 65: Plasticity chart of subgrade that experienced mud pumping
(Indraratna et al. (2020a))
1500
1st cycle
Indraratna et al. (2020a) investigated the undrained cyclic
1000 response of saturated subgrade compacted at different initial dry
densities and subjected to varying cyclic stress ratio (CSR) and
loading frequency (f). CSR is defined as the ratio of the applied
500 cyclic stress to twice the effective confining pressure during the
consolidation stage. When the specimen was subjected to higher
) xz (Pa)
CSR values there was a rapid accumulation of cyclic axial strains
0 and mean excess pore pressure. For example, when the specimen
compacted at an initial dry density of rd = 1790kg/m3 was
-500
subjected to a CSR = 1.0 and f = 1.0Hz, the residual strains
increased to 5 % in about 10 cycles and the mean excess pore
pressure ratio increased to 0.6 in the same time Figure 66. In
-1000 addition, there was an increase in the fines content towards the
top of the sample owing to the formation of slurry at the top of
the sample Figure 67.
-1500 Figure 68 indicate that presence of plastic kaolin fines can
5000 6000 7000 8000 9000 10000
significantly reduce the build-up of excess pore pressure and
' '
x-x z)/2 (Pa) associated subgrade fluidisation. After adding 10% kaolin, the
subgrade soil does not experience an early softening and
Figure 64: Deviatoric stress path (Zhao et al. 2021) fluidization with a CSR of 0.2. When the specimen subjected to
a cyclic stress ratio of 0.4 it fluidised. Figure 69 shows that the
2.3.17 Mud pumping change in liquidity index at the top of the specimen subjected to
a CSR of 0.5 is about 75%. On the other hand, when the
With the continuous passage of trains on ballasted tracks, the specimens were stable (at CSR ≤ 0.3), the specimens had a small
upper layers of a saturated subgrade can soften and form a slurry uniform liquidity index throughout the sample.
near the ballast interface. Repeated application of railway loads Rapidly increasing excess pore water pressure (EPP) in the
pushes the softened subgrade slurry upwards and can reach the subgrade under cyclic loads is the main reason causing mud
top of the ballast layer. The upward migration of soft subgrade pumping beneath rail tracks. In order to gain some
fines into the coarser ballast layers under the action of moving micromechanical insight into this phenomenon, a numerical
loads is termed as mud pumping or subgrade fluidisation. The simulation based on the discrete element method (DEM) coupled
increased hydraulic gradient due to the high excess pore with computational fluid dynamics (CFD) where the soil is
pressures results in subgrade fluidisation which then triggers an discretised in the form of a granular fabric subjected to fluid
upward migration of fines under cyclic railway loading (Duong pressure can be used. Soil particles are simulated in DEM
et al., 2014; Indraratna et al., 2020a). obeying Newton’s law while the fluid flow behaviour is governed
In most cases, poor drainage of railway ballast is found to be by Navier-Stokes theories following computational fluid
the leading source of accumulated excess pore pressure and is dynamics. The mutual interaction between soil particles (i.e.,
associated with softening of the subgrade (Tennakoon et al., solid phase) and pore water (i.e., fluid phase) is best carried out
2012). While this external loading is vital in initiating the
64
in a manner whereby the variables of the solid and fluid phases
are updated constantly over time.
Figure 66: Development of (a) Residual axial strain and (b) mean excess
pore pressure ratio at different cyclic stress ratio (rd = 1790kg/m3, f =
1.0Hz) (Indraratna et al. (2020a))
Figure 68: q-p’ plots of the test specimen subjected to undrained cyclic
triaxial testing (Indraratna et al., 2020b)
Figure 67: Subgrade soil fluidisation with the migration of fines under
undrained cyclic triaxial loading (Indraratna et al., 2020a)
Nguyen and Indraratna. (2020) modelled a 6 mm wide and
17.5 mm high specimen of soil was modelled in DEM Figure 70. Figure 69: Change in the liquidity index throughout the height of the
For computational simplicity, the particle size distribution (PSD) specimen (Indraratna et al., 2020b)
of a typical Ottawa sand (cohesionless) was used. In this
approach, a periodic boundary is applied on the soil specimen in Figure 71 shows the axial strain developed with increasing
DEM (i.e., infinite soil element) while the adopted slip fluid-wall hydraulic gradient. In this analysis, the vertical displacement is
interaction enables the boundary effects to be eliminated. The computed over the simulated time while increasing the hydraulic
hydraulic gradient (i) is assumed to increase gradually with time, gradient, i, thus enabling the axial strain to be obtained. The axial
which mimics the accumulated excess pore water pressure under strain begins to rise gradually with increasing i, and there is a
a constant cyclic load. The increasing excess pore water pressure swift change when i reaches approximately 1.4 where soil
along the depth of soil results in increasing hydraulic forces particles move rapidly. Increasing the hydraulic gradient results
acting on the particles. The densities of solid particles and fluid in correspondingly elevated hydraulic forces acting on the soil
are taken as 2650kg/m3 and 1000kg/m3, respectively, while a particles, causing their instability and internal migration.
Young’s modulus of 6.8x10-9Pa is assumed for the granular
assembly. The coefficients of sliding and rolling friction are
taken as 0.5 and 0.1, respectively.
65
samples) was satisfactory or slightly conservative for the plastic
materials, but conservative for non-plastic or slightly plastic
materials. Additional studies of the predictions of subgrade
moisture conditions for purpose of pavement design was reported
by Gomes Correia (1996a).
Figure 70: 3D soil specimen established in DEM coupled with CFD
(Nguyen and Indraratna, 2020)
Figure 72: Degraded contact network of soil particles (i.e., soil fabric) at
the inception of fluidization (Nguyen and Indraratna, 2020)
A first step for a rational design of flexible pavements was
developed by Burminster (1943) who introduced the study of
stress distribution that subsequently was extended to multi-layer
systems and numerical modelling by FEM (e.g., Balay et al.,
1997). From this point, the state of the art moved from empirical
rules used in routine design of pavements to a mechanistic
approach making use of the knowledge of the mechanical
behaviour of soil mechanics framework, non-saturated soil
mechanics and soil dynamics. Several contributions were
promoted by ISSMGE related activities and particularly in the
framework of ECTC11, TC3/ TC202. Some pioneer
contributions include Brown (1996), Gomes Correia (1996b,
1999, 2001), Gomes Correia and Lacasse (2006). Zapata (2018)
notes that most of all pavement layers will not exist in a fully
saturated state and that the initial Mechanistic Empirical
Pavement Design Guide (MEPDG) by the American Association
of State Highway and Transportation Officials (AASHTO) has
incorporated a design methodology for pavement systems that
Figure 71: Axial strain development of specimen with increasing considers unsaturated material response and prediction
hydraulic gradient (Nguyen and Indraratna, 2020) techniques based upon the site-specific environmental conditions.
This effort represented a huge step forward from the empirical
Figure 72 shows how the contact network of soil particles Structural Number used in the AASHTO 1993 pavement design
degrades under an increasing hydraulic gradient. For example, guide. However, and despite many efforts and advances through
when the hydraulic gradient reaches 1.4, the maximum contact the world, there exists and important gap between the state of the
force drops severely from 14N to around 5.6N. When the contact art and the state of practice.
forces become equal to the hydraulic forces soil fluidization Environmental conditions affect fatigue and thermal cracking,
occurs. Despite sustaining a certain degree of particle contact permanent deformations, and the smoothness of flexible
forces when i reaches the critical level, these contacts are shown pavements. In rigid pavements, faulting and fatigue cracking,
to be in a chaotic state, hence do not necessarily represent the curling, warping, shrinkage, punchouts, and roughness are
directional shear strength in a soil which we often assume in affected by moisture and temperature gradients. Mechanisms of
continuum mechanics. failure such as the initial crack width are also directly related to
these factors (Zapata, 2018).
2.4 Roads Although the bound pavement layers act as a cover of the
unbound materials, external factors such as precipitation,
2.4.1 Overview temperature, wind speed, solar radiation, relative humidity, and
Empirical pavement design by CBR procedures (soaked sample depth to groundwater table are environment parameters that
for 4 days to take care for the worst moisture conditions after affect pavement performance, as the internal stress state is
construction) were originally evaluated by a field study (Corps continuously seeking an equilibrium condition. In a pavement
of Eng. 1948, 1955) showing that this procedure (4 day soaked structure, moisture and temperature gradients and also
66
freeze/thaw cycles are environmentally driven variables that can
significantly affect the pavement layer and subgrade properties
and, hence, its stiffness (Zapata, 2018).
Ward Wilson (1990) proposed a one-dimensional coupled
thermo-hydro-mechanical model that incorporates the
conservation of mass, allows the use of flow equations such as
those proposed by Sheng et al. (2008) and imposed boundary
conditions (Caicedo 2019). However, the model is complex,
requires high computational effort and it is difficult to implement.
A simplified set of mechanistic-empirical models to estimate the
equilibrium moisture (suction) conditions beneath a pavement as
a function of the Thornthwaite Moisture Index (TMI) were
proposed by Zapata et. al. (2009). The model for granular
materials is presented in Figure 73 as a function of percent
passing the #200 sieve and the model for subgrade materials is
presented in Figure 74 as function of the percent passing the #200
sieve and wPI, which is the product of the passing #200 (in
decimal form) and the Plasticity Index. Mathematical models are Figure 75: Combined family of SWCCs when using the Torres-
available for implementation (Perera et al. 2004). Hernandez (2011) D10 model for nonplastic soils and Zapata (1999)
model for plastic soils (Fredlund et al. 2012)
2.4.2 Resilient modulus of subgrades
A recent summary of the empirical relationships established for
the resilient modulus has been presented by Gomes Correia and
Ramos (2021), including the effects of stress variables and state
conditions. Cary and Zapata (2011) extended the universal model
to include effects of suction on resilient modulus, as shown in
Equation 13.
( ++( +.(
%&'(! &'∆("#$%!& +' ,)*& (.,- &/., )
!! = #"# $$ % & % + 1& % + 1& (13)
*! *! *!
Where pa is the atmospheric pressure; k1’≥0, k2’≥0, k3’≤0 and
k4’≥0 are regression constants; θ is the bulk stress, ua is the pore
Figure 73: Suction at equilibrium for granular base materials (Zapata et air pressure, Δuw-sat is the build up of pore water pressure under
al., 2009) saturated conditions; τoct is the octahedral shear stress; ψm0 is the
initial matric soil suction; Dψm is the relative change of matric
soil suction with respect to initial suction due to build up of pore
water pressure under unsaturated conditions. This model has
been validated and compared to back-calculated stiffness
obtained from Falling Weight Deflectometer (FWD) tests
conducted at different moisture conditions (Salour et al. 2014,
Salour et al. 2015).
Alternatively, changes of resilient modulus with moisture
content can be estimated using Equations 14 and 15.
!! = )( !!1*2 (14)
*+, )3 = (. + /0 &456 )&" + 234 + 5√789: − (. + /0 &456 )&" </ >1 +
$/012√#456 ;$;)<&
7897 $: :;<(=;>? #45 )@ :-- A:
/7189$#456 ?
Figure 74: Suction at equilibrium for subgrade materials (Zapata et al., 0 (15)
2009)
The suction obtained with these models can now be related to where, FU is the ratio of MR at a given time to a resilient
the soil degree of saturation (or any soil water content modulus at optimum conditions MRopt, α = -0.600, β = -1.87194,
measurement) via the Soil-Water Characteristic or Retention δ = 0.800, γ = 0.080, ρ = 11.96518, ω = -10.19111, S is the degree
Curve (SWCC). An empirically predicted model available is that of saturation, Sopt is the saturation at the optimum moisture
proposed by Torres-Hernandez (2011). This model is based on content, and wPI is the product of the passing #200 (in decimal
information for more than 1,227,000 soils collected and grouped form) and the Plasticity Index.
in a catalogue of properties for 31,876 plastic soil units and 4,518 A simplified model was developed for non-plastic materials
non-plastic soil units collected as part of the NCHRP 9-23A in Equation 16 where FU-STD refers to materials prepared at
project for the implementation of the MEPDG. standard compaction conditions.
2.,(OPH
PQR SKLMNB = −0.40535 + ABAC*D
(16)
9%.;8<8=><.22<?=@ <%% EF
234
67
Results of Equations 15 and 16 are summarised in Figure 76 unsaturated conditions, because pores available for water
where it can be seen that FU increases (i.e., MR increases) as the pressure dissipation are scarce. However, Figure 79 shows that
degree of saturation reduces below the saturation at optimum exactly the opposite effect was obtained. Specimens 3 and 4
moisture content and vice versa. Although the model predicts tested under saturated conditions with a dwelling time (tD) of 4 s
that the resilient modulus under unsaturated conditions can be and effective bulk stresses (qnet) of 509 and 344 kPa, respectively,
much greater than at optimum moisture content, care should be developed lower pore water pressures in excess than Specimens
taken before adopting these values in design. Firstly, the 1 and 2, (which were tested at the same tD and qnet), respectively,
equilibrium moisture content and potential variations in moisture but under unsaturated conditions. Explanations for this include
should be estimated. Secondly, soils compacted dry of the presence of pressurized air within the pores exerting pressure
optimum tend to have a bi-modal void structure that can collapse on the pore water (suction stress), a reduction in voids as pore air
on wetting (e.g. Alonso et. al., 2012). It is preferable to compresses during the test increasing the chances for higher
compact soils near to the optimum moisture content to avoid this excess pore pressure build-up and low hydraulic conductivity in
problem. unsaturated specimens hindering excess pore pressure
dissipation compared with saturated specimens.
12 The pore water pressure build-up gradually reached an
wPI = 50
asymptotic value for all specimens tested under the conditions of
10
wPI = 45
the study. The unsaturated specimens did not develop positive
pore water pressures even after thousands of repetitions.
F U-STD - Environmental Factor
wPI = 40
8
wPI = 35
However, under extreme loading conditions (i.e., under unusual
wPI = 30
high loads), it cannot be ruled out that the rate of pore pressure
wPI = 25
6
build-up becomes high enough to transition into the saturated
wPI = 20
regime and significantly reduce the effective confining stresses
wPI = 15 leading the specimen to failure.
4
wPI = 10
wPI = 5
2 wPI = 0
0
-60 -50 -40 -30 -20 -10 0 10 20 30
S-Sopt (%)
Figure 76: Variation of the Environmental Factor (FU) as Function of
Degree of Saturation and Soil Index Properties (Cary and Zapata, 2011)
Suctions are estimated from moisture content using the soil
water characteristic curve (SWCC). Currently the MEPDG
uses the drying arm of the SWCC and does not consider
hysteresis. Rosenbalm and Zapata (2013) presented a
stochastic evaluation of the models used in the MEPDG to
estimate the water content of the unbound/granular materials
considering the hysteretic behavior of the SWCC. They found
that the prediction of the resilient modulus is overly conservative Figure 77: Pore Water Pressure Response to One Traffic Loading Cycle
for any climatic condition when hysteresis is not considered and (Cary and Zapata, 2016)
therefore, consideration of the SWCC hysteresis in pavement
design and analysis is necessary and recommended, particularly
in areas with lower TMI (arid climate). Global Cycle End Global Peak Excess
Clearly, the resilient modulus changes as moisture and suction Excess Pore Pressure Pore Pressure
Curve "Δu2" Curve "Δup"
changes. Build-up of pore pressures in unsaturated soils as a
result of traffic loading will have an effect on the resilient Δup
modulus. Cary and Zapata (2016) performed laboratory testing
Excess Pore Pressure
Δu1 up'
where a dynamic load was applied to unsaturated soil samples. Δu2-2 u1'
The form of the pore pressure response is shown in Figure 77 and u2'
Δu2-1
the accumulation of peak and residual excess pore pressures is
Δu2-2 = Δu2-1 + u2'
shown in Figure 78. Figure 79 shows a plot of the predicted
Δup = Δu2-1 + up'
global peak pore water pressure curves obtained as a result of Global Curves
Δu2-1 = u0'
fitting the data to a nonlinear function. Note that the values General Function:
plotted in the figure correspond to the actual pore water pressure
rather than the excess pore pressure. For the unsaturated
specimens, the origin of the predicted excess pore pressure was
translated to the initial negative pore water pressure imposed on
each specimen as the initial equilibrium condition. Time
Figure 79 shows that, for constant initial matric suction (cm0)
and dwelling time (tD), the greater the bulk stress (unet or ueff) Figure 78: Excess and Peak Pore Water Pressures in Response to Traffic
applied, the greater the pore water pressure developed. Also, Loading (Cary and Zapata, 2016)
under unsaturated conditions, the shorter the dwelling time, the
greater the pore water pressure build-up, which means less
dissipation between cycles.
It is a common belief that specimens under saturated
conditions develop higher excess pore pressures than those under
68
suction reduces and hence the resistance to stress reduces. All
100 this is also affected by the fine and coarse contents of the soil (Su
θ' = 509 kPa Saturated et al., 2022).
Positive Pore Water
80 tD = 4 s Conditions
θ' = 344 kPa The general state of practice is to adopt empirical laws
Pressure: +uw
60 tD = 4 s
relating permanent strain to strain after one cycle and various
40 material parameters, shear and mean stresses and a power
20
function of the number of cycles (Caicedo, 2019, Ramos et al.,
2020). The power function is generally stabilising and the ratio
Peak Pore Water Pressure (kPa)
0 of stress to strength is often used to assess a plastic failure range.
Negative Pore Water Pressure: -uw = ψmo
-20 Unsaturated Using these approaches in normalised conditions is possible to
Conditions
-40
θnet = 509 kPa rank materials against permanent strains, what can be a very
tD = 4 s
practical tool in the selection of materials (Ramos et al., 2020).
-60 Coupling this behaviour with the resilient modulus (both
θnet = 344 kPa
-80 tD = 4 s parameters controlling the pavement behaviour), it is possible to
-100
rate subgrade performances, as illustrated for pavements and
railroad by Coronado et al. (2011) and Gomes Correia and
-120 Ramos (2021).
-140 These empirical models are still commonly used in routine
θnet = 509 kPa θnet = 344 kPa pavement design addressing the problem of rutting of flexible
-160
tD = 8 s tD = 8 s pavements. However, with the advance of numerical pavement
-180 modelling, the permanent deformations occurring mainly in
0 5,000 10,000 15,000 20,000 25,000 30,000
subgrade and unbound layers are taken into account by
Number of Repetitions constitutive modelling of cyclic plasticity of these geomaterials.
Specimen 1 Specimen 2 Specimen 3 Several of these constitutive laws have been developed with both
Specimen 4 Specimen 5 Specimen 6 isotropic and kinematic hardening, as well as using shakedown
theory, to take account for the mechanical behaviour of these
Figure 79: Peak Pore Water Pressures Observed on Samples Tested at materials for large number of cycles, which is very different to
Different Bulk Stresses, Dwelling Times, and Saturation Conditions the conventional soil dynamics modelling (Jing et al., 2018,
(Cary and Zapata, 2016) Allou et al. 2007, Chazallon et al., 2006).
The results of the study were used to develop functions to 2.4.4 Volume change behaviour of expansive soils and
assess changes in pore water pressure under saturated conditions reactive fills
(∆u_(uw-sat)) and suction changes (Δψ_m) for unsaturated soil
Shrinking and swelling of fills and subgrade soils causes
conditions with the number of cycles, that can then be used in
cracking of road shoulders, potentially tension cracking of
Equation 13 to estimate the resilient modulus.
pavements and settlement of fills, which can affect pavement
The results suggest that compacting unbound materials at or
shape and create bumps at the approach to bridges and culverts.
near optimum conditions could allow generation of pore
Whether or not a material compresses or swells, its magnitude
pressures that degrade the resilient modulus. It is possible that
and time rate of volume change, is a function of its mineralogy,
compacting subgrades at conditions below optimum might be a
density, load, grading, compaction moisture content, and
better practice, albeit factoring in other effects of compacting
subsequent changes in moisture content (Muttuvel et al., 2020).
materials at low moisture contents.
A soil compacted at low density may expand when soaked at low
applied stress but will collapse when the load increases. Soils
2.4.3 Plastic deformations
compacted at intermediate densities may expand or collapse
The resilient modulus of road materials has been extensively when loaded albeit to a lesser extent. Soils compacted to high
assessed because it is related to fatigue cracking of pavements densities may expand when soaked (Caicedo, 2019). In
(Caicedo, 2019). Research into permanent plastic deformations addition, soils and rocks derived from shales may break down
is more limited. A review of mechanistic-empirical permanent over time as a result of changes in humidity or water content
deformation models, including laboratory testing, modelling and (Cardoso et al., 2012; Pineda et al., 2014) and microstructural
ranking was recently presented by Ramos et al. (2020). Caicedo changes (Alonso et al., 2012).
(2019) states that accumulation of plastic strains results from While trends in qualitative behaviour are reasonably clear,
particle displacements caused by shear and compressive stresses quantification of the magnitudes of swell or settlement and time
and controlled by friction and suction pressure. Similar to other rate of swell or settlement for engineering purposes is difficult to
materials there are three ranges of behaviour, the boundaries of assess. In recent years, unsaturated soil models have been
which are controlled by stress level and water content. These are: developed that can account for some of these behaviours such as
the Barcelona Basic Model (Alonso et al., 1990) and the Sheng,
1. Plastic shakedown range where plastic strains Fredlund, Gens model (Sheng et al., 2008).
stabilise after a number of cycles;
2. Intermediate range where plastic strain increases 2.4.4.1 Approaches to Estimate Shrink/Swell
at a decelerating rate but does not stabilise; and
Methods for estimating shrink/swell soil volume change can be
3. Plastic failure range where strains increase until
generally subdivided by empirical approaches using soil index
failure occurs.
properties such Atterberg Limits, particle size distribution, etc.,
and mechanistic approaches using engineering properties such as
Motorway roads would aim for the plastic shakedown range,
SWCCs, hydraulic conductivity, and results of 1-D oedometer
more lightly trafficked roads could aim for the intermediate range
tests.
and ideally, the plastic failure range would be avoided.
Published studies which empirically relate soil index
Quantification of these ranges can be difficult because suction is
properties to shrink/swell soil volume change potential include
an important parameter and hence drainage and infiltration are
but are not limited to: Seed et al. (1962), van der Merwe (1964),
also important. When pavement materials become undrained,
Ranganatham and Satyanarayana (1965), Nayak and Christensen
the air voids reduce, the degree of saturation increases, the
69
(1971), Schneider and Poor (1974), Chen (1975), Johnson and and evapotranspiration for a particular region. TMI can be found
Snethen (1978) Weston (1980), Picornell and Lytton (1984), from the map for the continental United States presented by
Dhowian (1990), Basma and Suleiman (1991). Lytton et al. (1990) or from the models presented by Perera et al.
Direct laboratory measurements of the soil volume change (2004) and Witczak et al. (2006). Olaiz et al. (2016) reviewed the
potential help improve the estimation of volume change used for differing published procedures which attempt to simplify the
design (Olaiz, 2017; Houston and Zhang, 2021). The 1-D TMI calculation from the original proposed by Thornthwaite in
oedometer “Response to Wetting Test” (ASTM D4546, 2014) is 1948 and concluded that the simplified equation by Witczak et al.
a common laboratory test for volume change determination. The (2006) provides values similar to those determined by the
following published studies include 1-D oedometer test-based original Thornthwaite (1948) procedure.
relationships to unsaturated soil volume change: Jennings and The TMI has been used by geotechnical engineers to estimate
Knight (1957), De Bruijn (1961, 1965), Burland and Jennings soil suction envelope parameters, required for suction-based
(1962), Sampson et al. (1965), Nobel (1966), Sullivan and shrink/swell soil volume change estimations, such as the depth to
McClelland (1969), Komornik and David (1969), NAVFAC stable (commonly referred to as equilibrium) suction, magnitude
(1971), Wong and Yong (1973), Gibbs (1973) Jennings et al. of stable suction, and total potential suction change at the surface.
(1973), Smith (1973), Teng et al. (1972, 1973), Teng and Clisby Published studies which relate TMI to suction envelope
(1975), Porter and Nelson (1980), Fredlund et al. (1980), parameters include but are not limited to: Mitchell (1980, 2008,
Sridharan et al. (1986), Erol et al. (1987), Shanker et al. (1987), 2013), McKeen and Johnson (1990), Fityus et al. (1998, 2004),
Nelson et al. (1992, 2001), Al-Shamrani and Al-Mhaidib (1999), Barnett and Kingsland (1999), Cameron (2001), Jaksa et al.
Basma et al. (2000), Subba Rao and Tripathy (2003) (2002), Briaud et al. (2003), Aubeny and Long (2007), Chan and
One key difference from the laboratory oedometer test Mostyn (2008), Vanapalli and Lu (2012), Karunarathne et al.
compared to the field conditions the unsaturated soil will (2012), Li et al. (2013), Sun et al. (2017), Lopes and
experience is the final degree of saturation. The response to Karunarathne (2017), the Post-Tensioning Institute 2nd and 3rd
wetting test inundates the sample, driving to saturation. editions (2004, 2008), Vann (2019), Vann and Houston (2021),
However, with proper drainage, the probability that the soil will and Olaiz et al. (2021). Figure 80 below presents an example
reach this moisture level over the period of the pavement design of TMI correlations to depth to stable suction using a TMI
life is very low (Houston and Houston 2017). Suction-based contour map for Australia (Mitchell 2008).
volume change approaches are required for the estimation of the
unsaturated soil volume change at moisture levels below
saturation.
Several methods which account for the moisture/suction state
of the soil include the Barcelona Basic Model (Alonso et al.
1990), the Lytton et al. (2005) approach adopted by the Texas
DOT and the Post-Tensioning Institute (PTI, 2004, 2008), the
Sheng, Fredlund, Gens model (Sheng et al. 2008), the Modified
State Surface Approach (MSSA) first introduced by Zhang and
Lytton (Zhang and Lytton, 2009), and the Surrogate Path Method
(Singhal 2010; Vann et al. 2018).
2.4.4.2 Incorporating the Effect of Shrink/Swell on Pavement
Ride Quality
The ability to estimate soil volume change as a function of time
is a valuable tool in the design of pavement structure as it allows
for the estimation of the cumulative International Roughness
Index (IRI) impacted by environmental factors (Olaiz et al. 2021).
Time-dependent estimations of shrink/swell soil volume change
must quantify the changes in soil moisture/suction due to the
varying climate. For a relatively near-surface groundwater table,
significant potential exists for capillary rise into subgrade soils.
The conventional assumption that negative pore water pressures
(i.e., soil suction) can be estimated by backward extrapolation
above the groundwater table of a line of slope equal to the depth
times the unit weight of water, is appropriate in a thin region
above the groundwater table, where soils are wetted to a degree
of saturation of 85% or more (Zapata et al. 2007). When the
groundwater table is relatively deep, the environmental factors
such as precipitation, temperature, wind speed, solar radiation,
and relative humidity affect the moisture state of the soils near
the surface, commonly referred to as the active zone (Nelson et Figure 80 Distribution of TMI through Mainland Australia and the
al. 2001). General TMI versus Equilibrium Soil Suction Relationship (Mitchell,
Given the complexity and large number of climatic 2008).
parameters affecting the climate-driven flux boundary conditions,
the Thornthwaite Moisture Index (TMI) (Thornthwaite 1948; Further complications arise when soils are subjected to cycles of
Thornthwaite and Mather 1955, Witczak et al. 2006) is wetting and drying. Rosenbalm and Zapata (2017) reported the
commonly used by geotechnical engineers as an index that results of laboratory tests showing that shrinking and swelling
quantifies climate variability at a given location (McKeen and increased with number of cycles, albeit at a decreasing rate
Johnson 1990). The TMI represents the aridity or humidity of a Figure 81 and that the swell pressure reduced with the number of
soil-climate system by summing the effects of annual cycles Figure 82. The vertical deformations observed on the soils
precipitation, evapotranspiration, storage, deficit, and runoff. To subjected to multiple wetting and drying cycles appeared to reach
a significant degree, the TMI index balances lateral infiltration an equilibrium condition. At lower net normal stresses, the
70
vertical deformation is recoverable and at higher net normal • Rigorously mix the water into the fill to maximise its
stresses the vertical deformation is irrecoverable. The distribution. The rotary hoe method adopted by
recoverability of the vertical deformation can be attributed to the Hopkins and Beckham (1998) aims to achieve a high
soil fabric and changes in the void ratio. As the cycles progressed, degree of mixing.
the soil reached an equilibrium swell strain (recoverable) or an • Adopt thin layers and use heavy compactors in order
equilibrium collapse strain (irrecoverable) due to the applied net to reduce the average particle size, maximise density
normal stress and changes in matric suction. In general, it was and fill voids with particles.
observed for the soils evaluated as part of this study that an
equilibrium condition was reached after the fourth or fifth cycle 2.4.4 Frost action
for the wetting cycle, depending if the vertical deformation was
swell or collapse, respectively. Conversely, the shrinkage strains Swelling on freezing and loss of bearing capacity on thawing
reached an equilibrium state around the second or third drying need to be considered in cold regions. Caicedo (2019) describes
cycle. the freezing process as fast desaturation which creates a
cryogenic suction. The suction causes potentials that move
water towards the freezing fringe. In sands, most of the water
freezes leaving only a small amount of unfrozen water.
Accumulation of water is small and frost susceptibility is low.
In silty soils and fine-grained sandy soils, some unfrozen water
remains allowing water migration and accumulation into the
freezing fringe. Frost susceptibility is high. In clays, a large
amount of water remains unfrozen, but it cannot accumulate due
to low permeability and frost susceptibility is low.
Cycles of freezing and thawing increase the soil’s void ratio
(Caicedo, 2019). The increase of water content during freezing
reduces the resilient modulus during thawing. The water
content before freezing and the number of cycles also reduces the
resilient modulus. Like unsaturated soils, the rate of reduction in
resilient modulus reduces with the number of freeze-thaw cycles.
Figure 81: Swelling/collapse Results Observed on Soils Subjected to The MEPDG provides methods to estimate the reduction in
Wetting and Drying Cycles (Rosenbalm and Zapata, 2017) resilient modulus as a function of material properties, provides
resilient moduli for frozen soils and suggests that the resilient
modulus returns to its original value during thawing after a
recovery period, the length of which depends on the material
properties.
Frost heave in the cold climate region may cause serious
damage to the pavement infrastructure (Schaus et al. 2011; Janoo
and Berg 1990; Dore 2002; Roy et al. 1992; Dore et al. 1997).
The resulting cracks, dips, heaving, and potholes make driving
uncomfortable, damage vehicles and increase the risk of car
accidents.
Three basic factors work together to generate frost heave: 1)
freezing temperatures; 2) frost susceptible soils, such as silty
soils or clayey soils as described above; and 3) water in the
subgrade soils (Oswell 2011). Frost heave of soils is a complex
coupled multi-physical process involving the coupling of the
thermal-hydro-mechanical (THM) field (Li et al. 2000;
Figure 82: Swelling Pressure for Soils Subjected to Wetting and Drying
Neaupane and Yamabe 2001; Liu et al. 2012; Zhou and Meschke
Cycles at Different Initial Compaction Conditions (Rosenbalm and
2013; Zhang and Michalowski 2015). The THM process is
Zapata, 2017)
generally coupled by extending the soil mechanics under
isothermal conditions theories to incorporate the effects due to
Practical solutions aimed at minimising volumetric changes
thermal expansion of solid skeleton as well as the pore fluid
of compacted fills are provided by BRE (1998) and Hopkins and
migration (Rabin and Steif 1998; Brownell et al. 1977).
Beckham (1998). Some recommendations are: Frost heave is mainly caused by the formation of ice lenses.
• Compact the lowest soil layers to modified maximum
The first widely accepted theory for explaining the ice lens
dry density. Increasing density has been shown to
formation is capillary theory, which states that water is draw by
reduce compression; the matric suction upwards to form the ice lenses (Taber, 1930).
• Progressively reduce the compactive effort as the
The drawback of capillary theory is that it does not account for
height of the fill increases and load applied to the fill
the initiation of new ice lenses. Another model that explains the
layer reduces. However, higher compactive effort in ice lens formation is the frost fringe model which proposes that
the upper layers of fill provides some risk management
frost heave can occur after ice has formed a frozen fringe by
as it would be more prone to swell hence counteracting
penetrating into the pores of the soil (Miller, 1972). The particle-
any compression that occurs deeper in the fill. engulfment model and geometrical supercooling model are
• Compact soils at optimum moisture content or wet of
recently developed models to simulate the periodic formation of
optimum. The intention is to maximise compression
ice lenses that have been observed in experimental work (Style
on loading by reducing the unsaturated stiffness of the et al. 2011).
soil and to maximise volume changes during
Various models have been developed to simulate the variation
construction rather than post construction.
of soil properties under frost effects. The distributions of
Compacting dry creates a stiff fill but long-term temperature and water content as well as the associated volume
changes in moisture content and humidity will increase
change have been the focus of investigations. Hydrodynamic
post construction internal compression.
models and rigid ice models are two of the most common types
71
of models for this purpose. These include the CRREL model mitigation of natural disasters, treatment and recycling of
(Guymon et al. 1986), the Hydrodynamic model (Harlan, 1973; industrial wastes, remediation of polluted soils, development of
Taylor and Luthin, 1978; Jame and Norum, 1980; Hromadka and brownfield sites and maintenance, and rehabilitation of existing
Yen, 1986; Noborio et al. 1996; Newman and Wilson, 1997; structures.
Hansson et al. 2004; Nishimura et al. 2009; Thomas et al. 2009), The United Nations Rio+20 website (www.uncsd2012.org)
the Rigid model (Miller, 1978; O’Neill and Miller, 1985), the defines sustainable development as development that meets the
Semi-empirical models (Konrad and Morgenstern, 1980, 1981, needs of the present without compromising the ability of future
1982a, 1982b; Kujala, 1997), Poromechanical models (Coussy, generations to meet their own needs. Sustainable development
2005; Alonso et al. 1990; Lu and Likos, 2006; Coussy and consists of three pillars: economic development, social
Monteiro, 2008), Thermomechanical models (Duquennoi et al. development and environmental protection. There is tension
1989; Fremond and Mikkola, 1991; Li et al. 2000, 2002), and between these pillars and between achieving simultaneous
Multiphysics models (Song and Yu, 2018). economic development and environmental protection in
Methods for mitigating the frost heave in pavements include particular. For example, there is a trend for incorporation of
increasing the thickness of the base layer that are less susceptible principles of sustainable development into projects, particularly
to the influence of freezing/thawing process, thermally insulating in government contracts. However, it is very difficult for
the subgrade soil or base layer, and reducing the water content in governments tasked with obtaining value for money for
subgrade soils by good drainage design (Mackay et al. 1992; taxpayers to award contracts incorporating high cost but
Dore et al. 1999). Frost heave is an important form of sustainable construction, and difficult for contractors to win
environmental load that affects ride quality on pavement. In the projects based on sustainable principles when clients award work
MEPDG, the frost heave related distress is counted by the based on lowest bid (Mitchell and Kelly, 2013).
equations that predict the International Roughness Index (IRI). Mitchell and Kelly (2013) identify that quantification of
In a recent study sponsored by NCHRP, the team examined sustainability is a key challenge along with methods to reduce,
the performance of the existing IRI prediction equations for road reuse and recycle. Basu and Puppula (2015) expanded on these
sections subjected to frost heave. The following observations themes and provided examples for methods to quantify
were found for flexible pavement: sustainability and sustainable applications.
• Overall, current IRI prediction equation overestimates 3.2 Carbon accounting in geotechnical infrastructure
the IRI values for flexible pavement.
• The higher the frost susceptibility of the road section, Carbon accounting is a relatively new but important area.
the larger the prediction error with the current IRI Modern, robust geotechnical solutions that address the perceived
prediction equation. issues of the past (especially but not uniquely associated with
• The larger the Freezing Index for the site, the higher Victorian era embankments and cuttings) come at a cost – not
the segregation potential, and the larger the prediction only in terms of money, but also in terms of carbon. While a
error by the current IRI equation. financial payback over a period of decades might be acceptable,
in carbon terms it is not. Given the imminence of the threat of
Efforts to develop and calibrate IRI equations for pavements climate change and the urgent need to mitigate the worst effects
subjected to the effects of frost heave are underway at Case in the next few years, the carbon payback time is more
western Reserve University and Arizona State University. The problematic. This will increasingly and urgently need to be
current effort incorporates heuristic models for Asphalt Concrete factored into geotechnical calculations: methodologies will need
(AC), AC overlays, and Continuously Reinforced Concrete to be developed discussed and generally accepted.
(CRCP) pavements. The new Site Factors showed improved Current research sponsored by Network Rail in the UK is
performance when compared to the historic Site Factor equation investigating, amongst other carbon related projects, the carbon
used in the current IRI models. The calibrated equation for AC impacts of cuttings and embankments. Aims include the
pavements is presented below: assessment of greenhouse gas emissions associated with
different interventions (maintenance, refurbishment and
-.:? renewal), and the formulation of strategies for carbon mitigation
9C9BC = 9C9BC,!= × EℎGH#I0JJD × [L,0 &@A*BC9%% + 0.0005 × )9$EE(F × over the whole life of the asset.
JGI(2P × L,0 + 0.04)] (17)
3.3 Reduce, reuse, recycle in Transport Infrastructure
EℎGH#I0JJD = −0.188 × *I(UℎGH#I0JJ) + 0.131 (18)
3.3.1 Reductions in materials hence carbon footprint
9C9BC,!= = 9.63 × *I(0.175 × )9 + 1) × *I(0.086 × 8X0HG$ + 1) ×
*I30.05 × YG*U*?GE?92$H? + 1: (19) Reductions in quantities can be achieved in many ways. Many
forms of admixtures for soil stabilization have been used over the
where, Thickness is the AC thickness, FI is the Freezing Index, years including, but not limited to, lime, cement, slag, bio-grout,
age is the age of the pavement, Precip is the precipitation and Silt coal byproducts, iron oxides, lignins, zeolites and rice husk ash
is the percentage of silt in the soil. Results showed a great (Mitchell and Kelly, 2013; ICE Ground Improvement Journal
improvement over the current MEPDG equation. more generally). Geosynthetic stabilization has developed
considerably in recent years. Bio-degradable geosynthetics and
3 MANAGEMENT OF THE ENVIRONMENT vertical drains constructed from jute fibres have also been
developed. Fibre reinforced soils have been explored.
3.1 Sustainability in geotechnical engineering These techniques are used to maximise use of poor quality
subgrades and site won fills, minimize import of quarried
Simpson and Tatsuoka (2008) reflected on discussions of the products and reducing the quantity of materials used.
Intergovernmental Panel on Climate Change (IPCC, 2007) and Reductions in fill materials can be achieved using vacuum
suggested that production and use of energy along with climate consolidation. The vacuum pressure provides a similar surcharge
change would be dominant themes in the near future. They also load to an equivalent weight of fill. Reductions in fill quantities
suggested that ground improvement will be critically important can also be achieved through steepening batter slopes using
in future geotechnical practice to achieve reductions in quantities geosynthetics.
of material used, reduction in carbon footprint, prevention and
72
Reductions in diesel burnt during compaction can be achieved road construction such as embankment fill, base and subbase
by optimizing processes using intelligent compaction. materials, flowable stabilized soils and concrete additives.
There is significant research being performed in this area as
3.3.2 Reuse of transport infrastructure reflected by the 31 papers published in the ASCE GeoInstitute
Geocongress 2012 (Hryciw et al., 2012) in a session titled
Reuse of brownfield sites, maintenance and rehabilitation of ‘Beneficial reuse of waste and recycled materials in sustainable
existing infrastructure and remediation of contaminated sites is geotechnical construction’. Much of the research is aimed at
becoming increasingly common. An example of remediation of characterising the mechanical properties of the waste materials,
failing earthworks slopes using discrete laterally loaded piles is an example of which is the Port Kembla Harbour reclamation.
provided here but there are many other examples of infrastructure In the past decade, the existing facilities at the Port had
reuse in the literature. reached its capacity in the Inner Harbour area. The Port Kembla
Remediation of a failing infrastructure slope using individual Port Corporation (PKPC) planned for the development of the
(discrete) piles is potentially attractive in terms of reducing cost, Outer Harbour to create additional berths and storage areas by
carbon and disruption but raises several questions relating to the reclaiming additional land.
maximum pile spacing at which the approach will still be For the fill selection, the use of locally available granular
effective, the pile depth needed, lateral loads and mechanisms of waste materials (i.e. coalwash and steel furnace slag) was
behaviour. Field monitoring, centrifuge modelling and numerical considered an economical and environmentally sustainable
studies have been used to shed light these issues. Geotechnical option to the traditionally quarried aggregates (Rujikiatkamjorn
centrifuge studies by Hayward et al. (2000) showed that the pile et al., 2013, Tasalloti et al., 2015a). However, the use of
spacing could be increased up to 4 pile diameters and the method heterogeneous compacted waste materials such as slag and
still be effective in preventing failure of the slope; a model with coalwash created some challenges related breakage of coalwash
piles at 6 pile diameters failed with quite large slope (CW) and swelling for steel furnace slag (SFS). Therefore, it was
displacements, although there was still enough interaction necessary to examine the geomechanical behaviour of these
between adjacent piles to prevent flow-through of the soil. The blends using a compaction field trial. The trial successfully
inferred resultant stress acting on the upper-most sections of the demonstrated that the desired density, shear strength and CBR
piles increased monotonically with increasing pile spacing. This could be achieved. The fill with a higher percentage of steel
is in contrast to the conceptual behavioural models proposed by slag stiffened more than fill with less slag from 30 and 170 days
Ito and Matsui (1975, which indicates a decrease in the force after compaction (Tasalloti et al., 2015b)
beyond a pile spacing of two pile diameters), and by Wang and Tyre derived aggregates (TDA) or rubber crumbs (RC)
Yen (1974), which indicates a maximum force at a pile spacing present excellent engineering properties such as low unit weight,
of about five pile diameters). high damping and high energy absorbing capacity (Senetakis et
Field studies have revealed two potential behavioural modes al., 2012; Indraratna et al., 2019c; Qi et al., 2019a), which prevail
(Powrie and Smethurst 2015): a pile embedded into deeper, their applications in dynamic loading projects (e.g. railways,
stable ground essentially resists the sliding of a top “slab” of soil, airport runways, seismic isolation projects) to help reduce
of thickness 2-3 metres; while a pile supporting a rotational slip vibration and track degradation. Esmaeili et al. (2017) indicated
in a scheme incorporating a surface stone blanket, or a pile adding 5% TDA into fully fouled ballast could reduce 34%
responding to inward shrinkage of the embankment as a result of ballast breakage. Sol-Sánchez et al. (2015) found that by adding
vegetation effects, exhibits a rather more complex bending 10% RC (by volume) into ballast layer could increase the
behaviour (Smethurst et al., 2020). In neither the centrifuge dissipated energy significantly and reduce the ballast breakage.
models of Hayward et al. (2000) nor the field studies reported by Signes et al. (2015) investigated the performance of mixed
Powrie and Smethurst (2015) did the pile bending moments coarse aggregates with TDA as subballast through laboratory and
approach anything like the values associated with classic limiting field tests and found the inclusion of 1-10% (by weight) TDA
lateral pressures and flow through between the piles; although can improve the resistance of track degradation efficiently albeit
the spacing was deliberately low enough to prevent this. with a little reduction in bearing capacity.
Numerical analyses by Summersgill et al. (2018) of clay Comprehensive laboratory tests have been carried out to
cutting slopes stabilised with a pile wall suggested that the piles investigate the geotechnical behaviour of SFS+CW+RC
will only ever delay the failure of slopes that are constructed mixtures by Indraratna et al. (2017a) and it has been found that
over-steep, hence would not provide long-term stability. While the addition of RC in the waste mixtures can further reduce the
consistent with the long-held view that first time failure of cut particle breakage of CW and the swelling potential of SFS,
slopes in clay is governed by the critical state soil strength, the meanwhile increasing the ductility, damping property and energy
computed behaviour is dependent on the soil model used in the absorbing property of the waste mixtures (e.g. Qi et al., 2019b).
analysis, particularly the post-peak rate of softening, as also An alternative mixture using CW+RC was proposed by
found by Rouainia et al. (2020). The geometric arrangement of Indraratna et al. (2019d) for use as capping for railways. In this
the piles relative to the softened bands that form in the case, the compaction energy was increased to compact the waste
unstabilised analysis may also have had an effect: longer or mixtures to a denser condition to compensate for the absence of
differently arranged piles might have been more effective. This SFS. The optimum compaction energy range for the CW+RC
is a further example of the behaviour determined in a mixture was 900-1600 kJ/m3. Particle breakage was less than that
sophisticated model being crucially dependent on input observed for 100% CW compacted to maximum standard dry
parameters that are difficult to determine or characterise with density. Moreover, adding 10% RC by weight will significantly
confidence. increase the energy absorbing property and provide comparable
strength to traditional subballast materials (Indraratna et al.,
3.3.3 Recycling of waste materials in transport infrastructure 2019b).
Mitchell and Kelly (2013) report that recycled waste materials Constitutive models that have been tested and verified for
that have been used as admixtures to stabilise soils or substitutes traditional materials, cannot be directly applied when elastic and
for natural aggregates include building rubble (crushed concrete, compressible recycled rubber is introduced into a matrix of
brick, glass, etc.), coal ash from power stations, rubber tyres, relatively rigid aggregates. Recently, models have been modified
plastics, slag and rice husk ash. Collins and Ciesielski (1994) list to describe the behaviour of aggregate-rubber mixtures. Youwai
38 types of waste materials that have been considered for use in and Bergado (2003) adopted a hypo-plasticity model within the
critical state (CS) framework by modifying the original equation
73
for dilatancy to capture the increase in initial dilatancy with inclusions (e.g. UBM and geogrids) at Singleton (near
increasing mean stress when rubber is added to sand. In general, Newcastle, Australia). Geogrids and a 10 mm thick UBM were
sand-rubber mixtures fail to each a clear CS unless a very large installed at the bottom of ballast with changing subgrade
axial strain was attained which explained the deviation of this conditions (i.e. soft alluvial deposit, concrete bridge deck and
model from the experimental data at the CS. Mashiri et al. hard rock). The test result shows that the ballast breakage index
(2015) proposed a constant stress ratio (CSR) framework instead (BBI) for the track incorporated UMB on the concrete bridge
of the CS framework for sand-rubber blends. deck (0.039) is significantly smaller than other conditions
These models do not fully capture the effect of rubber on the (0.104-0.136). This verifies the effectiveness of using UBM to
total energy absorbed by the mixture. Qi et al. (2019a) proposed mitigating the ballast degradation on a stiff subgrade condition.
a modified elasto-plastic model to describe the behaviour of a Indraratna et al. (2017b) has investigated the use of recycled
mixture of coal wash (CW), steel furnace slag (SFS) and rubber truck tyres for use in rail track. They conducted plate load tests
crumbs (RC) within the critical state framework. In this model, on unreinforced and tyre cell-reinforced capping layer using a
the critical stress ratio, MCS*, is a function of total work, Wtotal, large-scale prismoidal triaxial apparatus. With the increase in the
determined when the specimen achieved its point of failure. The thickness of the capping layer, the stress transmitted to the top of
critical state line (CSL) in the e-lnp' space also rotates clockwise the subgrade layer was substantially reduced. The same benefit
as rubber content increases. was achieved with a thinner capping layer reinforced with a tyre
cell. Strain measurements on the tyre walls were small even at
high loads showing the ability of the tyre cell to maintain its
shape while providing the confinement. The hoop stresses in the
tyre were estimated to produce an additional confining stress as
high as 500 kPa at the maximum applied load of 6500 kPa.
Indraratna et al. (2017c) further evaluated the effect of tyre
reinforcement on the ballast degradation, time-dependent
resilience, damping characteristics and track deformations
during cyclic load tests. Crushed basalt and spent ballast were
chosen as the infill materials for the tyre cell. The number of load
cycles applied in each test was 500,000 with the frequency of
15Hz and the maximum cyclic load of 370kPa as exerted by a
40-ton axle load. Settlement of the reinforced crushed basalt was
smaller than the unreinforced material. However, for the spent
ballast, the settlement for the reinforced specimen was more than
the unreinforced specimen Figure 85. The reason for that might
be that i) less particle interaction due to the internally smooth
rubber tyre cell and the rolling, sliding of the particles, and ii) the
Figure 83: Relationship between the total work, Wtotal, and the critical size ratio.
stress ratio, MCS, for CW+SFS+RC mixtures (Qi et al., 2018)
Figure 84: Effect of rubber content on the critical state of CW+SFS+RC
mixtures (Qi et al., 2018)
Recycled rubber mats/pads can be used to modify rail track
stiffness. Mats installed underneath ballast are called under
ballast mat (UBM) and those placed at the interface of ballast and
sleeper are called under sleeper pads (USP). The biggest
advantage of using the rubber mats/pads is that they can
significantly increase the contact area at the interface, and thus
attenuate the dynamic and impact force, thereby reducing the
ballast degradation and track deformation. Nimbalkar et al.
Figure 85: (a) Lateral displacements and; (b) vertical settlements of the
(2012) evaluated the performance of the track specimen testing specimens (Indraratna et al. 2017c)
subjected to the impact force by changing the position of rubber
mats and subgrade conditions. The experimental, numerical and Lateral tension and vertical compression forces were
field tests results revealed that under ballast mats are more measured on the tyre. Although, the circumferential tensile
efficient to attenuate the impact stress and mitigate the ballast strains were small, they helped develop additional confining
degradation for a stiffer subgrade condition (e.g. tunnels and stress to reduce lateral spreading. Use of rubber tyres also
bridges), while for a softer subgrade condition, placing the rubber reduced the track modulus. Although a higher track modulus is
pads on top of ballast reduced ballast breakage and track generally desired, a lower modulus can be beneficial over a rigid
settlement. substructure which must have less stiffness so that the transition
Field tests were carried out by Indraratna et al. (2014c) to
investigate instrumented track sections with different geo-
74
to the rail embankment is not accompanied by the large stiffness hence effectively under-drained, pore pressures will remain
changes. much lower – near zero in an embankment of reasonably uniform
Recycled tyres have also been used to construct tracks over permeability with depth (Briggs et al. 2013b). If the embankment
poor subgrade at the Ballina Soft Soil Field Testing Facility in is of very low permeability, it is effectively impervious and
Australia. The access track was built with 940 tyre units along suctions can be retained even in wet winter conditions (Briggs et
with 546 tonnes of the crushed rock Figure 86 over a 10m thick al. 2013b).
soft soil deposit. Similar applications have been used for The mechanical reinforcement effect of roots has been
temporary platforms for piling rigs and cranes. investigated experimentally using a geotechnical centrifuge (e.g.
Liang et al. 2017) and novel shearing experiments carried out in
an X-Ray CT scanner to observe root and soil behaviour (Bull et
al. 2020). The fundamental understanding of the root and soil and
behaviour gained from these experiments has allowed further
development of appropriate analytical and numerical models to
suitably characterise the increase in strength given by the root
system (Woodman et al. 2020, Meijer et al. 2021).
3.4.2 Failure modes likely to become more prevalent owing to
climate change
Experience over the past few years of more extreme summer and
Figure 86: A view of the reinforced access track at Ballina site: (a-b) especially winter weather has suggested we will see a rise in
before and after infilled with crushed rock (unpublished part of a current instances of earthwork failure through
program at the University of Technology Sydney through funding from
the Australian Research Council, courtesy Ms. F. Mehmood, PhD student) • Debris flow: possibly from a degraded / weathered near
surface zone, or possibly from a downslope drain
3.4 Effects of and on the environment subjected to inundation flow as at Carmont (Scotland) in
2020
3.4.1 The influence of vegetation and weather on earthworks • “Rapid drawdown” effects, where a river level has risen
Interest in the effects of vegetation on geotechnical infrastructure adjacent to an embankment, remained high for a period
for transportation (particularly embankments and cutting slopes) of several days or weeks, and then fallen relatively
arose for two reasons: (1) pore water abstraction in summer quickly such that the stabilizing surface load effect is
followed by re-wetting in winter causing seasonal cycles of removed while high pore pressures remain within the
shrinkage and swelling of clay soils that are disruptive to the embankment.
infrastructure supported, and (2) the effect of vegetation on slope These mechanisms are associated with changing patterns of
stability, through reduced pore water pressures as a result of land use, increased intensity of rainfall and storminess, and
water abstraction and root reinforcement effects. possibly the weakening of a surface zone over a period of
Smethurst et al. (2012) showed that it is possible link climate decades as a result of seasonal cyclic changes in pore pressure
data and observed pore water pressures and moisture contents in and water content. The latter mode of deterioration has been
a grass or scrub-covered slope, using a simple soil water balance investigated using numerical models calibrated against site
model of the rooting zone and surface abstraction. Grass and observations by Rouainia et al. (2020). Application of such an
small shrub vegetation was not able to create a moisture deficit approach, even at the asset scale, is computationally intensive.
in summer that could be maintained through a wetter winter – Svalova et al. (2021) show how this difficulty can be overcome
relatively shallow summer drying is removed by infiltration in by carrying out a limited ensemble of simulations for a range of
winter. slope strengths, geometries etc., from which a statistical emulator
Trees have a deeper root zone and abstract water from greater was created to characterise the behaviour of the entire parameter
depths. Modelling moisture abstraction by trees requires removal space (i.e., to predict the way in which the factor of safety of a
of the water from deep within the soil, modelling the effects of slope of any combination of strength and geometry changes over
the deeper tree roots (e.g. Briggs et al. 2016). The greater time). This can then be used to consider vulnerable earthworks at
abstraction and deeper root system of trees means that suctions the route and network scale.
in a low permeability embankment can persist through a typical
winter, as infiltration does not then reach the full depth of the 3.5 Resilience to flooding
root zone. Tree removal from the slope, especially from the toe Infrastructure resilience is the ability to withstand, adapt to
of an embankment, risks losing the deep suctions that maintain changing conditions, and recover positively from shocks and
stability in winter. Removal of trees from the crest while leaving stresses.
trees in place at the toe enables the deep suctions below the toe Flooding can affect road and rail in several ways. A flood can
to be retained even through a wet winter, and should mitigate the washout ballast, cause scour and saturate the pavement and
worst effects of seasonal shrinkage / swelling in a high plasticity subgrade leading to loss of strength and degradation in pavement
clay embankment (Smethurst et al 2015). Avoidance of high performance as well as batter failure.
water demand tree species within 1.5 times the tree height from The after-effects of flooding are managed in practice in
the track also helps mitigate damaging seasonal cycles of several ways:
shrinkage and swelling (Briggs et al, 2013a). • Shut road or rail
Pore water pressures in a clay fill embankment are generally • Reduced speed to reduce load
most detrimental to slope stability (highest) in winter. It was • Load limits
shown by Briggs et al (2013b) that the embankment foundation Key considerations are
plays a crucial role in the wet winter pore pressures within a clay • Adopting a flood return period for design that provide
structure of permeability typical of old tipped fill embankments. satisfactory immunity at reasonable cost
If the embankment is founded on low permeability clay, pore • Limit damage and safety risks in larger floods
pressures could rise towards hydrostatic below a water table • Re-open the road or rail as soon as practicable with
close to the soil surface. If the embankment is on gravel or chalk, minimal cost.
75
Guidelines for the protection of flood affected road method, considering various groundwater profiles and scour
pavements have been developed by the Queensland Department depths at the toe to quantify the deformation of the road surface.
of Transport and Main Roads (DTMR). These guidelines were A probabilistic assessment of the magnitude of deformation and
developed based on experience of flooding in western the groundwater level and scour depth was undertaken to derive
Queensland and data from moisture probes installed in unsealed fragility functions for the prediction of damage to assets exposed
pavements. Their procedure can be summarized as: to these multiple hazards.
1. Assess how many days the pavement was under
water 3.6 Bio-geotechnics
2. Assess the structural capacity of the pavement
a. Minor inundation: visual One of the areas that has advanced in the past two decades is bio-
b. Moderate inundation: inspection or geotechnics. Bio-grouting processes are sustainable solutions
deflection testing that have the potential to replace Portland cement as binder for
c. Severe inundation: Deflection testing different soil improvement applications and reduce the CO2
3. Compare with historic structural data if available emissions associated with its production (Khodadadi Tirkolaei et
4. If inspection indicates poor surface seal then adopt al., 2017).
a higher inundation condition. Khodadadi et al. (2017) present a comprehensive review of
5. Decide whether to open the road or impose load the latest technologies developed on bio-mediated and bio-
restrictions inspired improvement of granular soil via microbially induced
6. Depending on structural assessment main load calcium carbonate precipitation (MICP), enzyme induced
restriction for a period of 3 to 4 weeks. calcium carbonate precipitation (EICP) and microbially induced
The period of load restriction is based on a 3 month pavement desaturation and precipitation (MIDP). These technologies are in
dry back model where the subgrade modulus increases linearly the research phase, but commercialization looks promising based
with time. Based on this model and an 80% load restriction, on field trials and laboratory results at different scales.
loss of pavement life is reduced to 33% after 1 week, 58% after MICP has proved to reduce the hydraulic conductivity, and
2 weeks, 74% after 3 weeks and 84% after 4 weeks. The 3 increase the strength and stiffness in granular soils by
month dry back model was validated by unpublished field mechanisms that range from urea hydrolysis (Burbank et al.
moisture probe data. 2011, Gomez et al. 2014, DeJong et al. 2010; Cheng et al. 2013)
This short case history shows that asset owners have to denitrification, iron reduction and sulfate reduction (Almeida
developed some procedures to address resilience to flooding. et al. 1995, Hamdan 2013; Karatas et al. 2008, O’Donnell 2016,
In contrast, the academic world has only recently started to van Paassen et al. 2010. Challenges to this technology that need
address this issue. to be addressed before implementation at a large-scale include
Elshaer and Daniel (2018) investigated the structural response the cost optimization of the technology, the environmental
of pavements that have been inundated and the foreseen changes impact due to ammonium by-products and non-uniformity in the
in capacity using a mechanistic approach using layer elastic treatment due to the complexity of natural soil environments.
analysis and the AASHTO empirical approach to determine the EICP induces carbonate precipitation by using free urease
structural number. The relative impact of parameters such as enzyme generally obtained from agricultural sources (Yasuhara
unbound material type, layer thickness, traffic loads, and et al. 2012, Hamdan et al. 2013, Neupane et al. 2015, Hamdan
interlayer bond conditions on the reduction in expected strain 2015, Kavazanjian and Hamdan 2015, Putra et al. 2016). This
values at critical locations were evaluated. The results showed variant in the process eliminates the need for microbial
increases of 15–80% in vertical strains at the top of subgrade transportation to soils and oxygen availability soils at greater
layer for low volume and interstate sections and 6–15% increase depths. Challenges to the technology include the high cost of the
in horizontal strain at the bottom of asphalt layer for low volume enzyme and uniformity in the treatment.
sections and 3–8% for interstate sections. Accurate information The use of microorganisms to induce desaturation of granular
on the layer thicknesses, traffic type, and interlayer bond soils via denitrification (MIDP) results in nitrogen and carbon
condition were found to be most important for the evaluation of dioxide gas generation, which in turn lowers the degree of
the change in expected horizontal strain. The types of base and saturation of the soil matrix. Given sufficient substrates, the
subgrade materials are the most important factors for evaluating stimulated bacteria will produce enough gas to develop a
the change in expected vertical strain. continuous gas phase. Introducing gas into the soil to reduce the
Asadi et al (2021) assessed the post-flood water movement degree of saturation is shown to increase the soil resistance to
throughout its granular layers using unsaturated flow principles. dynamic loading (Rebata-Landa and Santamarina 2012, He h.
They developed a numerical model for simulating the partially 2013, Stallings Young et al. 2020). The process of microbially
saturated water flow in pavements caused by surface infiltration. induced desaturation and precipitation (MIDP) involves injecting
Hasnayn et al (2020) performed laboratory tests to assess the an aqueous solution containing calcium, nitrate and a source of
performance of railway subgrade after flooding. They performed dissolved organic carbon (DOC). As a result, indigenous nitrate
full-scale experimental investigation to study the behaviour of reducing bacteria are stimulated in-situ and convert the nitrate
subgrade in both saturated and unsaturated conditions, and how (NO3-) into nitrogen gas (N2), while oxidizing the organic carbon
this behaviour changes with soil suction. Further, the to dissolved inorganic carbon (DIC). In the presence of dissolved
investigation also studies the role of sand-blanketing during and calcium, production of DIC results in precipitation of calcium
after repeated flooding events. The results showed that as soil carbonate minerals, while the energy released from the oxidation
suction reduces, flooding results in a continual reduction in both of DOC is used by the micro-organisms for growth and
soil stiffness and track stiffness. It was shown that the maintenance.
introduction of a sand-blanket has limited effectiveness as a Ground improvement through MIDP occurs in two-phases
drainage material, particularly after prolonged and repeated (Karatas et al. 2008; Kavazanjian et al. 2015; O’Donnell et al.
flooding. 2017a, 2017b). The first phase is microbially induced
McKenna et al (2021) develop fragility functions to express desaturation (MID). Previous research implementing MID in
the probability that an asset exceeds some serviceability or limit homogeneous sandy soils have demonstrated that a single
state as components in the quantitative risk analysis of treatment with a low concentrated solution of calcium nitrate and
infrastructure exposed to natural hazards. A generic granular calcium acetate is sufficient to mitigate liquefaction via
highway embankment was modelled using the finite element desaturation (Rebata-Landa and Santamarina 2012; He et al.
2013; Stallings Young et al. 2020; Wang et al. 2020; Stallings
76
Young 2021; Wang et al. 2021). The second phase of MIDP induction based strain coil sensors, subgrade pressure cells, etc.
involves microbially induced carbonate precipitation (MICP), in More recently, with the recent emphasis to develop smart
which the sand is cemented with calcium carbonate (CaCO3) sensors, such as smart rocks and bender element transducers
minerals. based on shear wave propagation for determining layer stiffness
Column studies on MIDP demonstrated that the combined properties, more detailed perspectives on the layer stiffness
formation of biomass, biogas and biominerals can reduce the characteristics and rutting failure mechanisms of pavement
hydraulic conductivity by several orders of magnitude (Pham et foundation layers are gained to provide some unique
al. 2018). Stallings Young et. al. (2021) performed experiments perspectives, such as the depth of failure surface in a thick
using a relatively thin tank of soil to simulate planar flow through granular subbase by the use of strain coils installed with depth
a granular soil treated with MIDP. The treatment successfully profile or the quantification of stiffened zone in a geogrid-
desaturated the soil and once a continuous gas phase was formed, stabilized unbound aggregate base and subbase layers (e.g. Kang
it remained relatively constant throughout the remainder of the et al. 2022).
experiment. Flushing with additional water did not remove a
significant amount of the biogenic gas. Desaturation resulted in 4.2.2 Road pavements
the reduction of hydraulic conductivity and the presence of gas
affected the measured excess pore pressure and demonstrates gas Rana et al (2016) summarised some recent advances in sensing
formation through the system. The mechanical properties of the for road pavements. Fibre optic sensors can be used to monitor
material were not investigated in this testing program. various parameters such as strain, displacement, vibration,
The main challenge for MID is to predict and control the cracks, corrosion, and chloride ion concentration, etc. An
uniformity of the treatment in field conditions. example is shown in Figure 87. Piezoelectric sensors can be
incorporated in remote and inaccessible locations. They can also
4 THE DIGITAL AGE be used to harvest energy from pavements due to the movement
of the vehicles and generated pressure (Zhao et al., 2013;
4.1 Intelligent construction Gholikhani el al., 2019).
A self-sensing composite has the ability to sense its own
Earthworks construction refers to the tasks of excavation, deformation and damage. Strain and damage sensing in a
transportation, spreading and compaction of geomaterial (e.g. composite material is usually achieved through detecting change
soil, rockfill and soil-rockfill mixture). Whereas relying heavily in their electrical resistivity (Rana et al, 2016). To achieve piezo-
on machinery and repetitive processes, these tasks are highly resistivity in a composite material, it should contain a conducting
susceptible to optimization. In this context Artificial Intelligence element. Different types of conducting components have been
techniques (AI), such as Data Mining and modern optimization used in the existing self-sensing composite materials. Short and
can be applied. Parente et al (2016) developed a novel intelligent continuous carbon fibres (CFs), carbon particles as well as
earthwork optimization system, capable of integrating AI and carbon nanomaterials such as carbon nanofibers (CNFs) and
modern optimization and GIS technologies in order to optimize nanotubes (CNTs) have been utilized for this purpose. These
the earthwork processes throughout all phases of design and conducting components form a conducting electrical network
construction work. This integration system allows significant within the composites. When the composites are subjected to
savings in time, cost and gas emissions contributing for a more deformation or damage, this conducting network is disturbed
sustainable construction. leading to a change in the electrical resistivity.
Intelligent compaction is a process where a smooth drum
roller is instrumented and the stiffness of the soil inferred from
the data. Variations to vibration amplitude, frequency, roller
speed, roller weight and number of passes can be varied to
determine the optimum combination of factors to achieve
compaction requirements while minimizing effort, time and
diesel burnt. Rollers equipped with GPS can also track their path
and height helping with quality control and density assessments.
Data can be presented as CMV (compaction meter value) based
on Fast Fourier Transform peaks from accelerometer signals or
by measuring a stiffness from inferred load, displacement
response obtained from the accelerometer (Briaud and Seo, 2003;
Anderegg et al, 2006). Experience with CMV suggests some
averaging (over say a 0.5 m square area) of CMV is needed to
allow for measurement variability.
The fundamental problem with the stiffness measures is that
they do not provide a good measure of compaction because
density is not strongly correlated to stiffness. Tatsuoka et al Figure 87: (a) Schematic design of fine aggregate asphalt mixture
(2021) show that stiffness can be near constant as density encapsulated fibre optic sensor (b) real picture (c) strain sensing
increases at near constant moisture content. Developing capability (Rana et al. 2016)
methods to assess density remotely is an area requiring further
research. CNT/cement composite sensors have been developed and
If method compaction is used then the challenge of measuring applied in pavement monitoring (Yu and Kwon, 2012). These
density can be overcome and the benefits of using intelligent CNT sensors were installed in the road for testing the pavement
compaction can more readily be realized. monitoring capability (Figure 88). Figure 89shows the response
of pre-cast and cast-in-place CNT sensors while a truck passes
4.2 Smart sensors over the road and compares the response with that obtained in
case of strain gauges. It can be observed that an abrupt change in
4.2.1 Airport pavements voltage occurs when a wheel passes over the road and each wheel
represents one signal peak.
Many of the traditional sensors installed in airport pavement A recent laboratory proof of concept was established by
foundation layers have been multi-depth deflectometers, Abedi et al. (2020) using originally hybrid combination of
77
carbon nanotubes (CNTs) and graphene nanoplatelets (GNPs) in 4.3 Data analytics
a cement mortar achieving high mechanical, microstructural,
durability and self-sensing performances. For concentrations 4.3.1 Predictive maintenance
of 0.1, 0.3, 0.5, 0.7, and 1% CNT+GNP in the cement mortar
(W/C 0,5) tested and for the different hydration times (7, 28 and Use of instrumented vehicles or pavements coupled with inverse
90 days) an optimum concentration of CNT+GNP around 0.5% analysis or other stochastic numerical methods to predict future
shows the best performance in terms of durability (resistance maintenance is an area of research that is rapidly expanding.
against freeze-thaw cycles), microstructure and mechanical Drivers for the research are to eliminate safety risks to
behavior under compression and flexural cyclic loadings. Figure maintenance workers inspecting infrastructure in heavily
90 shows an example of the fractional change in resistivity trafficked roads and railways, earlier identification of damage,
(sensing sensibility) under cyclic flexural loading. This optimum increased inspection productivity and reduced whole of life
percentage (0,25% CNT+0,25% GNP) reveals to have maintenance and operational costs. Xie et al (2020) provide a
significant higher gauge factors when compared with the comprehensive summary of techniques being utilized in the rail
available results in literature using individual nanoparticles. industry.
Mechanical and data-driven models are used to evaluate the
track conditions for diagnostic and prognostic purposes.
Mechanical models are based on engineering mechanics which
allow physical understanding of a system but often rely on
simplifying assumptions. Data-driven methods correlate data
with maintenance events and include statistical models and
machine learning models. Statistical models make inferences
about the relationships between variables often described in a
mechanical model, whilst machine learning models focus on
making the most accurate correlation between data and
maintenance events without needing a mechanical model. Both
Figure 88: Installation of CNT/cement sensor in roads for testing approaches can handle high dimensional and multivariate data
monitoring capability (Rana et al. 2016) and identify relationships between the track status and
measurement data. The performance of the data-driven methods
depends on the appropriate choice of data pre-processing and
analysis models.
Collecting information is the first step in the application of
data-driven models. Some commonly used measurement
methods for railway track are summarized in Table 4.
Table 4 Methods of rail monitoring (after Xie et al. 2020)
Monitoring
Methods Technologies
Objects
Visual, ultrasonic Ballast section, ties,
Walking patrols
testing fasteners, rail head
Figure 89: Response of cast-in-place and pre-cast CNT sensors (a,b) and
strain gauges (c, d) while a truck passes over the road (Rana et al. 2016) Camera-based
measurement,
Track geometry, rail
Specific inspection ultrasonic testing,
head, ties, fasteners,
600 0.05 cars vibration acoustic,
Flexural Load CG (1%) substructure
eddy, laser,
CG (0.7%) CG (0.5%)
magnetic
CG (0.3%) CG (0.1%)
Fractional Change in Resistivity
500
0.04 Camera-based
measurement,
400 Track geometry, rail
In-service vehicles ultrasonic testing,
0.03 head, ties, fasteners
vibration,
Load (N)
300 displacement
Fiber, Rail head, wheel-rail
0.02
meteorological interactions,
200 Wayside detectors
sensor, impact load temperature,
detector weather
0.01
100
An approximate distribution of methods used for predictive
0 0.00 maintenance is provided in Figure 91. Data driven models
0 200 400 600 800 1000 1200 comprise about 74% of the total and statistical methods about
Time (s) 26%.
Figure 90: The fractional change in resistivity together with the cyclic
flexural response for reinforced cementitious composite by different
CNT+GNP concentration (Abedi et al, 2020)
78
Table 6 Data driven models in predictive maintenance of railway track
Classical machine learning model (Xie et al., 2020)
Ensemble model
Statistical model
Unsupervised learning model
Deep learning Deep learning model Classical
model 9/109 (8%)
Tree-based machine
model 21%
Advantage Disadvantage Application
Unsupervised learning
Classical machine learning
KNN learning
4%
model 9/109 (8%)
model 57/109 (52%)
SVM
model
33%
Other
16%
Robust; no Time- Prediction of
expert consuming; average track
ANN ANN
Statistical model
26%
knowledge is poor degradation
28/109 (26%) Ensemble model
Distribution in classical macahine learning model needed interpretability rates
6/109 (6%)
Figure 91: distribution of data-driven methods in predictive maintenance Fast
of railway track (Xie et al., 2020) Efficiency in
Poor classification
small data
interpretability and evaluation
Some examples of statistical models are provided in Table 5 size; ability to
SVM , sensitive to of rolling
and of data driven models in deal with
kernel contact fatigue
nonlinear
function (RCF) defects
characteristics
Table 6. Examples of rail defects addressed by the models in tracks
are provided in Table 7.
Table 5 Statistical models in predictive maintenance of railway track (Xie Prediction of
et al., 2020) Tree-based Overfitting on tram track
Interpretation
model noisy data degradation
Statistical index
Advantage Disadvantage Application
models Not easy to
Regression determine
Prior Estimating K-nearest hyper- Classification
modeling Simple;
knowledge of the neigh-bours parameter k; of tamping
(e.g., interpretation
Simple; the data is remaining (KNN) sensitive to effectiveness
multivariable
interpretability needed to useful life of data
regression,
select the best railway distribution
multi-stage
fitting model track
regression)
Table 7 Some recent applications (Xie et al, 2020)
Probability
distribution
model (e.g., Data-Driven
Based on Prediction of Track Defects Equipment
Weibull, Simple; Method
specific track failure Track geometry
normal, interpretability
hypotheses time recording car,
lognormal and Bayesian method
extreme value operation records,
distributions) maintenance records
Ensemble of gamma
Track geometry process, logistic
Time series recording car regression, and
model (e.g., SVM
Geometry
autoregressive Prediction of
irregularity Track geometry
(AR), Time series is the short- ANN
autoregressive Interpretability required to be term trend of recording car
integrated stationary track
moving irregularity Track geometry
Random forest
average recording car
(ARIMA)) In-service vehicles
Bayesian Predictors are Investigation on-board sensing SVM
methods (e.g., required to be of the rail device
Stable; better
Bayesian independent; squat failure
performance Camera CNN
inference, prior probability
for small
Markov Chain distribution using Laser-ultrasonic
dataset size SVM
Monte Carlo assumption Bayesian technology
(MCMC)) needed inference
Stochastic Based on Evaluating Rail head defects Ultrasonic and rail Bayesian inference
process (e.g., specific the surface photos method
Better
Markov hypotheses; deterioration
performance Sperry’s eddy
process, not suitable of track
in process current walking Clustering
Gaussian for mid to geometry
status stick
process, long-term based on
prediction
Gamma system Markov SVM
process) prediction process Missing rail
Camera Multiple signal
component
classification
79
Continuous wavelet
used. Various algorithms are being explored to speed up the
In-service vehicles posterior computation along with parallel computing methods.
Rail break transform
Examples where Bayesian methods have been used are tunnel
Eddy current sensor Bayesian network convergence (Bjureland et al., 2017), braced excavations (Hsein
Variational Juang et al., 2013) and embankment settlement (Zheng et al.,
Fiber bragg grating heteroscedastic 2018).
Substructure failure:
sleeper and ballast Gaussian process Zheng et al. (2018) assessed predictions of future settlement
Camera Bayesian method performance for various amounts of monitoring data for a trial
embankment constructed at Australia’s soft soil field testing
facility. They found that accurate predictions could be obtained
Predictive maintenance is an emerging field with many after about 116 days of data despite consolidation settlement not
unresolved issues. Some of these are: being completed nearly 1000 days after construction. Such
• Identification of the most accurate or most practical early accuracy allows decisions to be made regarding the
methods to assess maintenance requirements performance of the embankment and whether risks need to be
• Reducing the rate of false positives and negatives mitigated or opportunities can be taken advantage of, during the
• Interpretation of black box methods construction period with confidence.
• Pre-processing costs of noisy input data and incorrect
records
• Computational power and the time required to analyse
big data sets
4.3.2 Little data: Bayesian statistics and the observational
method
Predictive maintenance is an example of big data techniques
being used in practice. However, much of the time
geotechnical engineers need to make decisions based on limited
information. The statistical and data driven techniques can also
be used to derive best estimates and co-efficients of variance
which can help understand levels of certainty and assist decision
making. Bayesian statistics is one method that has particular
application in the interpretation of monitoring data during
construction. The Bayesian approach is essentially a
mathematical version of the Observational Method.
Kelly and Huang (2015) explain that Bayes’ formula can be Figure 92: Long term settlement predictions at various times (Zheng et
written as follows: al., 2018)
<(T|V) ∝ X(V|T)<(T) (20) Zheng et al. (2020) also used the method to back analyse,
magnetic extensometer and vibrating wire piezometer data as
where P(q) is the prior probability distribution of the material shown in Figure 93 and Figure 94. The Bayesian method
parameters, L(y|q) is the probability of measurements (y) created a close match between calculation and measurement,
conditional on the material parameters (q) and P(y|q) is the albeit that some of the posterior material parameters lay outside
posterior distribution of the material parameters updated by ranges obtained by laboratory testing.
measurements.
The prior distribution can be obtained by assumption,
measurements or a combination of both. The prior distribution
can be developed at the design stage of a project and provides a
measure of uncertainty in the design predictions. The
measurements can be written as:
YQ = R(T) + Z (21)
where e is the “error” or difference between measurement and
calculation g(q). The calculation can be of any type ranging
from analytical to three dimensional FEM.
If the measured error is assumed to be normally distributed,
the likelihood function of measurement yi can be written as
RG LS(U)LWH
X(YQ |T) = [ # +H
$ (22)
Figure 93: Back analysis of magnetic extensometer data (Zheng et al.,
where µe and se are the mean and standard deviation of 2018)
measurement errors, f is the probability density function of the
standard normal distribution.
The posterior distribution is typically solved using a form of
Markov Chain Monte Carlo and can be extremely
computationally expensive. Often many thousands to millions
of simulations need to be run to obtain accurate posterior
parameters. The computational time is shorter if analytical
calculations are adopted and (much) longer if complex FEM is
80
solidification methods such as cement treated clay was one of the
answers to the introspection.
To avoid large differential settlements, it was important to
continuously supply reclamation soil to the reclamation area
from the initial stage of construction.
5.3 Airport pavements
Increase airport pavement life by incorporating low utilization
due to COVID into whole of life cycle plans.
Increase aircraft pavement life by instrumenting pavements
and directing aircraft along low trafficked paths, taking into
account dilation caused by aircraft wander.
Geogrids help reduce the thickness of granular layers and
reduce settlement and dilation of granular layers. Construction of
runways using more ductile materials with damping will offer
less noise and vibrations.
Figure 94: Back analysis of piezometer data (Zheng et al., 2018) Field scale comparisons of channelized and wandering traffic
could provide opportunity to further develop prediction models
for airport pavement prediction methods.
5 DISCUSSION
5.4 Low speed rail
5.1 Outcomes for practitioners and academics
Inclusion of rubber elements in the substructure can help reduce
The USA Transportation Research Board is in the process of
ballast degradation and transition performance and increase track
finalizing its critical issues report (https://www
longevity. They should be placed in track in such a way as to
ttnews.com/articles/trb-finalize-critical-issues-report). The
prevent interfering with maintenance equipment that needs to
critical issues include:
replenish ballast during maintenance cycles and occasionally
damaged sleepers.
• Transformational technologies and services
Ballast performance can be improved by reinforcement with
• Population trends
random fibres or geogrids, physical confinement, increasing
• Energy and sustainability
resistance to lateral movement by reducing sleeper spacing,
• Resilience and security
widening the sleeper at the ends or using intermittent lateral
• Safety and public health
restraints; or by reducing the shoulder slope below sleeper
• Equity
surface level, in cases where the ballast bed is raised and also by
• Governance
adjusting the ballast to a more well-graded distribution with
• System performance and asset management
greater internal friction.
• Funding and finance
In Australia and some other nations, especially along the
• Goods movement
coastal areas with high population, the possibility of constructing
• Institutional and workforce capacity
very long straight stretch of track will be limited. This makes it
• Research and innovation.
unfeasible to have very long heavy haul trains (2-3 km) operating
at very high speeds on concrete slab tracks. Ballasted tracks
The state of practice (SOP) requires practitioners to help asset
will remain the most economical choice for heavy haul despite
owners optimize a balance between cost, time, maintenance
annual maintenance requirements. Optimum speeds should still
considerations, depleting natural construction materials,
be less than 100km/h for trains that are 2-3 km long.
community requirements and increasingly environmental
For effects of track stiffness, experimental (field) validation
constraints for bigger, heavier, faster, longer and more automated
and model parameterisation will be the key challenges for the
vehicles.
next few years.
Implementation of next-generation transportation
For effects of track stiffness, 1:1 scale physical modelling or
technologies, such as smart sensors and energy harvesting, will
field validation and model parameterisation will be the key
mandate principles and guidelines for the design, maintenance,
challenges for the next few years. Research on new materials for
and rehabilitation of future physical assets such as sustainable
construction including waste granular assemblies require more
infrastructure materials capable of directly communicating and
intense fundamental studies through appropriate numerical
interacting with vehicles and other types of urban infrastructure
(DEM) micro-mechanical studies and greater efforts on field
facilities. A well designed, constructed and connected
performance monitoring of new designs adopting recycled
transportation infrastructure is essential for smart systems,
materials and mineral byproducts. Applied research on transition
resiliency, economic competitiveness and environment-friendly
zones will require greater focus as more heavy haul networks in
operation.
the future will be built over soft estuarine terrains, bridges and
Elements that practitioners and academics could adopt from
tunnels.
this paper are summarized in the following sections.
5.5 High speed rail
5.2 Reclamations for ports and airports
As the train speed increases the track support stiffness needed to
When a port or offshore airport is constructed far-offshore, the
avoid significant load and displacement magnification may also
depth of water will be deep. In such a case, a large amount of
increase. In this regard, however, the use of normalised datasets
reclamation material will be required. For Kansai International
can be misleading: at high speeds on a stiff trackbed, the basic
Airport, the total geo-material volume was almost the same as
static deflection being amplified is very much smaller than at a
the volume of two small mountains. It was pity to use such large
low speed on a soft trackbed, even if the ratio of train speed to
amount of natural resources. From this introspection, the
critical velocity is the same.
engineer involved in the project considered utilizing recycled
materials such as dredged soil or broken concrete. Geomaterial
81
With intensive trafficking at high speeds, ballasted track may pavement systems that considers many or most elements of the
reach its performance limit owing to issues of ballast flight and state of the art. Complex coupled thermo-mechanical-hydro
increased maintenance requirements. These effects are further models as well as simplified methods will continue to be
magnified as the axle load increases. Thus slab track may be developed.
required more frequently in the future, particularly for higher Resilient and plastic deformations are related to changes in
speed, heavier trains. This will increase construction costs, hence environmental effects and water. These are complex areas that
the benefits of higher speeds will need to be clear. Most slab track will continue to be explored over the next few years.
forms do not have continuous bending stiffness across panels: Quantification of the magnitudes of swell or settlement
this seems to be a potential advantage that is not currently being (including frost heave) and time rate of swell or settlement for
exploited, perhaps owing to a combination of a lack of engineering purposes is difficult. Further research and
understanding of how stiffer track systems really interact with development of unsaturated soil models are likely to contribute
the ground, and concerns over construction methods and quality. to this area. Practical methods have been developed to partially
Challenges remain relating to the accurate determination of manage volumetric changes in soil but new solutions are required
effective soil stiffnesses, Rayleigh wave speeds and system as the volume changes are only partly predictable and
critical velocities, which depend on a combination of the controllable using low cost methods.
characteristics of potentially multiple and / or discontinuous soil
strata, the track form and any earth structures (embankments or 5.7 Environmental management
cuttings) that may be present.
The dominant frequency for pore pressure generation, and the Two main principles of green construction involve using
ability to calculate realistic residual pore pressures during and recycled materials and utilizing renewable resources to supply
after train passage, remain an ongoing debate and challenge. the energy needs. Current demand for recycled aggregate
The influence of wave propagation on the number of loading products like crushed concrete, glass and reclaimed asphalt
cycles after which the shakedown state is reached needs further aggregate, and secondary aggregates such as fly ash and slag, and
research. The influence of Raleigh waves on the instability of mined materials like clay and shale is expected to continually
undrained subgrade soils needs to be addressed in more detail grow into the future. Advances will be spurred by environmental
using the multi-laminate constitutive model and needs to be and land use concerns favoring recycled and secondary aggregate
integrated with Finite Element Modelling. This is required to materials, and by rising demand for specialty products such as
investigate the influence of principal stress rotation due to expanded clay and shale, which are used in applications where
moving trains on mud pumping behavior and stability of light weight is of extreme importance. The use of lightweight
subgrade soils, where excess pore water pressures play an aggregate will contribute significantly to sustainable
important role. development by conserving energy, lowering transportation
requirements, maximizing structural efficiency and increasing
5.6 Roads service life, and, as a result, is expected promote the ability of
future generations to meet their needs.
Advances in computing, robotics, artificial intelligence, and Sustainability and resilience will become increasingly
Internet of Things with 5G communication have enabled important. Engineers need to optimize many competing criteria.
revolutionary technologies such as autonomous and connected Practice is leading academia with respect to creating
vehicles to emerge, and these next-generation systems hold the infrastructure resilient to floods and drought. Potentially,
promise to completely change the paradigm of transportation critical infrastructure will attract resilience requirements in a
efficiency, sustainability, resilience, and safety. The principal similar fashion to areas subject to earthquakes to ensure they
opportunity from autonomous and connected systems, for remain operational in the face of natural disasters. The
example, arises from the ability to monitor both infrastructure insurance industry may also drive a higher degree of resilience in
and vehicle components of the system, and operate them in ways certain locations to enable risk cover to be provided at reasonable
that maximize various societal benefits. The real-time cost.
information will be incorporated into advanced system control Reduce, reuse, recycle: many materials or aging infrastructure
strategies to improve safety and mobility, and to reduce energy can be incorporated into new construction. In many cases the
consumption and emissions. It also improves the resilience and mechanical behavior of the materials is well understood and can
reliability of the infrastructures through online monitoring and be used with confidence. Carbon (and other waste) accounting
real-time management, while providing data for both on-line may provide the economic impetus to accelerate the use of
operations and off-line planning applications. recycled materials and being able to quantify the cost benefit in
The use of marginal and waste materials for base and sub-base a transparent and believable framework will be required.
will be utilized increasingly as the availability of natural Bio-grouting is an area that promise to contribute to
materials is becoming scarce and the capacity for expanding sustainable and environmental-friendly solutions in soil
quarrying is becoming more challenging in the future. The use of improvement applications.
intelligent compaction and sensor technologies will become
more useful and reduce costs through optimal design and post- 5.8 Digital opportunities
construction quality assurance. Roads will also benefit from the
use of rubber intermixed aggregates and optimal use of The rapid growing field of smart mobility has brought
geosynthetics for greater internal stability. unprecedented challenges for people’s demand of building smart,
Future construction of roads will be advanced in terms of new autonomous and multimodal transportation infrastructures. Some
materials including smart aggregates and the use of sensor of emerging topics, such as connected and autonomous vehicles,
technologies to monitor performance of new design. The use of smart sensors and connected road infrastructure, automatic road
synthetic inclusions including energy absorbing damping condition monitoring, and shared mobility are bringing both
materials for increased longevity (including earthquake opportunities and challenges to practitioners.
resistance) will become more popular as research in these areas Smart sensing technologies need to be developed to improve
starts to grow. the rail industry’s ability to answer critical safety and
The Mechanistic Empirical Pavement Design Guide maintenance related questions related to the track infrastructure
(MEPDG) by the American Association of State Highway and systems by monitoring and predicting track health.
Transportation Officials (AASHTO) probably represents current Instrumentation and interpretation of big and little data sets is
best practice and has incorporated a design methodology for likely to become more prevalent in a digital world.
82
Understanding how the data analytics works will enable 7 REFERENCES
engineers to specify types and locations of instruments to
maximise their utilization as well as optimizing design and Abadi T.C. Le Pen L. Zervos A. and Powrie W. 2016. Improving the
construction. performance of railway tracks through ballast interventions. Proc
Inst Mech Eng, Part F: J Rail Rapid Transit; 232(2), 337–355
Use of data to optimize construction and reduce costs is Abadi T.C. Le Pen L. Zervos A. and Powrie W. 2019. Effect of sleeper
occurring now. interventions on railwaytrack performance. ASCE J Geotech
Predictive maintenance for rail and road pavements, driven by Geoenviron Eng;145(4).
safety and operational needs, is likely to become prevalent. Abadi T.C. Madhusudhan Li. and Le Pen LM. 2021. Re-using “life
However, predictive maintenance is an emerging field with many expired” railway ballast: laboratory testing, shape analysis and
unresolved issues. Some of these include identification of the petrographic evaluation. Submitted to the ASCE Journal of
most accurate or most practical methods to assess maintenance Geotechnical and Geo-environmental Engineering
requirements, reducing the rate of false positives and negatives, Abedi M. Fangueiro R. Correia A.G. 2020. Ultra-sensitive affordable
interpretation of black box methods, pre-processing costs of cementitious composite with high mechanical and microstructural
performances by hybrid CNT/GNP, Materials, 13 (16), art. no. 3484,
noisy input data and incorrect records and computational power DOI: 10.3390/MA13163484
and the time required to analyse big data sets Ahmed S. Harkness J. Le Pen L. Powrie W. and Zervos A. 2016.
A future development could be aircraft-to-pavement (A2P) Numerical modelling of railway ballast at the particle scale.
connectivity. The pavement would instruct the pilot/aircraft to International Journal for Numerical and Analytical Methods in
deliberately wander if required. This purposely-induced aircraft Geomechanics, 40(5), 713-737. https://doi.org/10.1002/nag.2424
wander helps spread the traffic loading on the full width of the Allou F. Chazallon C. and Hornych P. 2007. A numerical model for
pavement. This has the potential to ultimately optimize the flexible pavements rut depth evolution with time. International
pavement use, extend the existing pavement life, or reduce the Journal for Numerical and Analytical Methods in Geomechanics,
thicknesses of new pavements (while taking the negative impacts 31(1), 1-22.
Almeida J. Júlio S. Reis, M. and Carrondo M. 1995. Nitrite Inhibition of
of wander, discussed above, into account). Development of Denitrification by Pseudomonas fluorescens. Biotechnol. Bioeng.,
pavement instrumentation is required for this vision to be 46(3), 194-201.
realized. Alonso E.E. Gens A. and Josa A. 1990. Constitutive Model for Partially
New opportunities require broad research efforts in the Saturated Soils. Géotechnique. 40. 10.1680/geot.1990.40.3.405.
technology, engineering, planning, regulatory, legal, and social Alonso E.E. Pinyol N.M. and Gens A. 2012. Compacted soil behaviour:
spheres of next-generation transportation systems. Research is initial state, structure and constitutive modelling, Geotechnique
needed on (i) new technologies and innovations beyond the symposium in print, Gallipoli ed, 3-18
immediate scope of vehicle and transportation infrastructure Alves Costa P. Colaco A. Calcada R. and Silva Cardoso A. 2015. Critical
systems; and (ii) partnership with technology companies speed of railway tracks. Detailed and simplified approaches.
Transportation Geotechnics, 2, 30-46
focusing on innovative transportation ideas and coordination of Aingaran S. Le Pen L. Zervos A. and Powrie W. 2018. Modelling the
machines, objects, people, and vehicles for multiple effects of trafficking and tamping on scaled railway ballast in triaxial
transportation modes. This will need to be accomplished through tests. Transportation Geotechnics 15, 84-90. https://doi.org/
the development and implementation of embedded, wireless, 10.1016/j.trgeo.2018.04.004
“smart” infrastructure technologies capable of autonomously Ajayi O.O. Le Pen L. Zervos A. and Powrie W. 2017 A behavioural
transmitting load-response data and “state-of-repair” exception framework for fibre reinforced gravel. Géotechnique 67(1), 56-68,
reports. Wireless continuous monitoring for infrastructure doi:10.1680/jgeot.16.P.023.
sensing will improve safety and reliability of roads, bridges and Al-Shamrani M.A. and Al-Mhaidib A.I. 1999. Prediction of potential
shared mobility and establish big data for digital twin virtual vertical swell of expansive soils using a triaxial stress path cell.
Quarterly Journal of Engineering Geology & Hydrogeology, 32(1).
models, which can in turn provide new methods for multi- 45-54.
scale/multi-physics modeling and simulation of transportation Anderegg R. Von Felten D.A. and Kaufmann K. 2006. Compaction
resiliency. monitoring using intelligent soil compactors. GeoCongress 2006:
Geotechnical Engineering in the Information Technology Age (pp.
1-6).
6 ACKNOWLEDGEMENTS Asadi M. Mallick R. and Nazarian S. 2021, Numerical modeling of post-
flood water flow in pavement structures, Transportation
The authors acknowledge and thank the contribution of the Geotechnics, 27, 100468, ISSN 2214-3912, https://doi.org/10.1016/
following people to this paper: j.trgeo.2020.100468.
Aubeny C. and Long X. 2007. Moisture diffusion in shallow clay masses,
• Prof. David Airey to intelligent compaction. Journal of Geotechnical and Geoenvironmental Engineering,
• Dr Alice Duley, Dr Ann Mamou, Dr John Harkness, Dr 133(10), 1241-1248.
Louis Le Pen, Dr Anna Mamou, Dr David Milne, Australian Government Department of Infrastructure and Regional
Development. 2016. Trends Transport and Australia’s Development
Tracey Najafpour-Navaei, Dr Joel Smethurst and to 2040 and beyond
Geoff Watson and Dr Trung Ngo for railways Balay L. Gomes Correia A. Jouve P. Hornych P. and Paute J.L. 1997.
• Prof Yoichi Watabe for ground improvement of Mechanical behaviour of soils and unbound granular materials,
Japanese airports. modelling flexible pavement – Recent advances. Proc. 8th ICAP,
• Prof Masaki Kitazume for ground improvement of the Seattle, Washington, USA, 1, 823-842
Tokyo Haneda airport. Barnett I.C. and Kingsland R.I. 1999. Assignment of AS2870 Soil
• Prof Cholachat Rujikiatkamjorn for his valuable Suction Profile Parameters to TMI Derived Climatic Zones for
comments for roads and railway sections of the paper. NSW. Proceedings, 8th Australia – New Zealand Conference on
• Prof. Bill Yu from Case Western Reserve University on Geomechanics (Hobart), Expansive Soils. 1: 149-155.
Basu D. and Puppala A.J. 2015. Principles of sustainability and their
the frost heave application in geotechnical engineering, Proc. 15th PanAmerican
• Graduate student Austin Olaiz from Arizona State conference in geotechnical engineering, Buenos Aires, Argentina,
University on volume change. 162-183
Basma A.A. Al-Homoud A.S. and Malkawi A.I. 2000. Swelling-
A significant number of figures have been reproduced with kind shrinkage behaviour of natural expansive clays. Appl. Clay Sci., 11:
permission from various scholarly journals such as ASCE-JGGE, 211-227.
CGJ, Geotechnique etc.
83
Basma A.A. Al-Suleiman T.I. 1991. Climatic Consideration in New Cary C. and Zapata C.E. 2011. Resilient Modulus for Unsaturated
AASHTO Flexible Pavement Design. Journal of Transportation Unbound Materials. Int. J. of Roads Materials and Pavement
Engineering. 117(2), 210 - 223 Design. 12(3), 617-640.
BBC World Service. 2013. How much bigger can container ships get? Cary C. and Zapata C.E. 2016. Pore Water Pressure Response of Soil
https://www.bbc.com/news/magazine-21432226 Subjected to Dynamic Loading under Saturated and Unsaturated
Bian X. Jiang H. Cheng C. Chen Y. Chen R. and Jiang, J. 2014. Full- Conditions, International Journal of Geomechanics, 16(6),
scale model testing on a ballastless high-speed railway under doi:10.1061/(ASCE)GM.1943-5622.0000642
simulated train moving loads. Soil Dynamics and Earthquake Chan I. and Mostyn G. 2008. Climate Factors for AS2870 for the
Engineering, 66, 368-384. metropolitan Sydney area, Australian Geomechanics, 43(1), 17-28
Bian X. Hu J. Thompson D. and Powrie W. 2019. Pore pressure Chazallon C. Hornych P and Mouhoubi S. 2006. Elastoplastic model for
generation in a poro-elastic soil under moving train loads. Soil the long term behaviour modelling of unbound granular materials in
Dynamics and Earthquake Engineering, 125, 105711. flexible pavements. Int J. of Geomechanics. 6(4), 279-289
Bian X. Li W. Qian Y. and Tutumluer E. 2020. Analysing the effect of Chen C. Indraratna B. McDowell G. and Rujikiatkamjorn C. 2015.
principal stress rotation on railway track settlement by discrete Discrete element modelling of lateral displacement of a granular
element method. Géotechnique, 70, 803-821. assembly under cyclic loading. Computers and Geotechnics, 69:
Bjureland W. Spross J. Johansson F. Prästings A and Larsson S. 2017. 474-484
Reliability aspects of rock tunnel design with the observational Chen F.H. 1975. Foundations on Expansive Soils. Elsevier Scientific
method. International Journal of Rock Mechanics and Mining Pub. Co., Amsterdam, New York, NY.
Sciences. 98. 102-110. 10.1016/j.ijrmms.2017.07.004. Cheng L. Cord-Ruwisch R. and Shahin M. 2013. Cementation of sand
Blackmore L. Clayton C.R. Powrie W. Priest J.A. and Otter L. 2020. soil by microbially induced calcite precipitation at various degrees
Saturation and its effect on the resilient modulus of a pavement of saturation. Canadian Geotechnical Journal, 50(1), 81-90
formation material. Géotechnique, 70(4), 292-302. Collins R.J. and Ciesielski S.K. 1994. Recycling and use of waste
BRE (1998). Low rise buildings on fill. Part 3, Digest 427. materials and by-products in highway construction. Synthesis of
Briaud J.L. and Seo J., 2003. Intelligent compaction: overview and Highway Practice 199. National Cooperative Research Program
research needs. Report to the Federal Highway Administration. (NCHRP), Transportation Research Board, Washington, DC, USA.
Briaud J. Zhang X. and Moon S. 2003. Shrink Test–Water Content Connolly D. Giannopoulos A. and Forde M. 2013. Numerical modelling
Method for Shrink & Swell Predictions. Journal of Geotechnical & of ground borne vibrations from high speed rail lines on
Geoenvironmental Engineering, ASCE, 129(7), 590–600. embankments. Soil Dynamics and Earthquake Engineering, 46, 13-
Briggs K.M. Smethurst J.A. Powrie W. O'Brien A.S. and Butcher, D. 19.
2013a. Managing the extent of tree removal from railway earthwork Connolly D.P. Alves Costa P. Kouroussis G. Galvin P. Woodwards P.K.
slopes. Ecological Engineering 61(Part C), 690-696. https://doi.org/ and Laghrouche O. 2015. Large scale international testing of railway
10.1016/j.ecoleng.2012.12.076 ground vibrations across Europe. Soil Dynamics and Earthquake
Briggs K.M. Smethurst J.A. Powrie W. and O'Brien A.S. 2013b. Wet Engineering, 71, 1-12
winter pore pressures in railway embankments. Proceedings of the Coronado O. Caicedo B. Taibi S. Correia A.G. and Fleureau J.M. 2011.
Institution of Civil Engineers - Geotechnical Engineering, A macro geomechanical approach to rank non-standard unbound
166(GE5), 451-465. https://doi.org/10.1680/geng.11.00106 granular materials for pavements, Engineering Geology, 119(1-2),
Briggs K. Smethurst J.A. Powrie W. and O'Brien A.S. 2016. The 64-73.
influence of tree root water uptake on the long term hydrology of a Correia A.G. and Cunha J. 2014. Analysis of nonlinear soil modelling in
clay fill railway embankment. Transportation Geotechnics, 9, 31-48. the subgrade and rail track responses under HST. Transportation
https://doi.org/10.1016/j.trgeo.2016.06.001 Geotechnics 1, 147-156.
Brown S.F. 1996. Soil mechanics in pavement engineering. Costa P.A. Calçada R. and Cardoso A.S. 2012a. Ballast mats for the
Géotechnique, 46(3), 383–426. reduction of railway traffic vibrations. Numerical study. Soil
Brownell D.H. Garg S.K. and Pritchett, J.W. 1977. Governing equations Dynamics and Earthquake Engineering, 42(0), 137-150.
for geothermal reservoirs. Water Resources Research, 13(6), 929- Costa P.A. Calçada R. and Cardoso A.S. 2012b. Track–ground vibrations
934. induced by railway traffic: In-situ measurements and validation of a
Bull D. Smethurst J. Sinclair I. Pierron F. Roose T. Powrie W. and 2.5D FEM-BEM model. Soil Dynamics and Earthquake
Bengough A.G. 2020. Mechanisms of root-reinforcement in soils: an Engineering, 32, 111-128.
experimental methodology using four-dimensional X-ray computed Costa P.A. Colaço A. Calçada R. and Cardoso A.S. 2015. Critical speed
tomography and digital volume correlation. Proceedings of the of railway tracks. Detailed and simplified approaches.
Royal Society A: Mathematical, Physical and Engineering Sciences, Transportation Geotechnics, 2, 30-46
476(2237), 1-23. [20190838]. https://doi.org/10.1098/ Coussy O. 2005. Poromechanics of freezing materials. J. Mech. Phys.
rspa.2019.0838 Solids, 53, 1689–1718. doi:10.1016/j.jmps.2005.04.001
Burbank M. Weaver T. Green T. Williams B. and Crawford R. (2011). Coussy O. and Monteiro P. 2007. Unsaturated poroelasticity for
Precipitation of Calcite by Indigenous Microorganisms to crystallization in pores. Comput. Geotech. 34, 279–290.
Strengthen Liquefiable Soils. Geomicrobiology Journal, 28(4), 301- doi:10.1016/j. compgeo.2007.02.007
312. Dareeju B. Gallage C. Ishikawa T. and Dhanasekar M. 2017. Effects of
Burland J.B. and Jennings J.E.B. 1962. Limitations to the Use of principal stress axis rotation on cyclic deformation characteristics of
Effective Stresses in Partly Saturated Soils, Geotechnique 12(2): rail track subgrade materials. Soils and Foundations, 57, 423-438.
125-144. DOI: 10.1680/geot.1962.12.2.125 De Bruijin C.M.A. 1961. Swelling characteristics of a transported soil
Burmister D.M. 1943. The Theory of Stresses and Displacements in profile at Leeuh of Vereeniging (Transvaal). Proc. 5th Int. Conf. Soil
Layered Systems and Applications to the Design of Airport Mech. Found. Eng. 1, 43-49.
Runways. HRB Proc, 23(126) De Bruijin C.M.A. 1965. Some observations on soil moisture conditions
Cai Y. Xie Z. Wang J. Wang P. and Geng X. 2015. A new approach of beneath & adjacent to tarred roads & other surface treatments in
vacuum preloading with booster PVDs to improve deep marine clay South Africa. Moisture Equilibrium & Moisture Changes Beneath
strata. Can. Geotech. J. 55, 1359–1371. Covered Areas. Butterworths, Australia. 135-142.
Cai Y. Wu T. Guo L. and Wang J. 2018. Stiffness degradation and plastic DeJong J. Mortensen B. Martinez B. and Nelson D. 2010. Bio-mediated
strain accumulation of clay under cyclic load with principal stress Soil Improvement. Ecological Engineering, 36(2), 197-210.
rotation and deviatoric stress variation. Journal of Geotechnical and Dhowian A.W. 1990. Field performance of expansive shale formation,
Geoenvironmental Engineering, 144, 04018021. Journal of King Abdulaziz University (Engineering Sciences), 2:
Caicedo B. 2019. Geotechnics of roads: fundamentals, Taylor and 165-82.
Francis Group, London Dong K. Connolly D.P. Laghrouche O. Woodward P.K. and Costa, P.A.
Cameron D. 2001. The extent of soil desiccation near trees in a semi-arid 2018. The stiffening of soft soils on railway lines. Transportation
environment. Geotechnical & Geological Engineering, 19(3), 357- Geotechnics, 17, 178-191.
370 Dong K. Connolly D. Laghrouche O. Woodward P. and Costa P.A. 2019.
Cardoso R. Maranha Des Neves E. and Alonso E.E. 2012. Experimental Non-linear soil behaviour on high speed rail lines. Computers and
behaviour of compacted marls, Geotechnique, 62(11), 999-1012. Geotechnics, 112, 302-318.
84
Dong S.Y. and Yu X. 2018. Microstructure-based random finite element Gibbs H.J. 1973. Use of a consolidometer for measuring expansion
method simulation of frost heave: theory and implementation, potential of soils. Proc. Workshop Expansive Clays & Shales in
Transportation Research Record, 2672(52), 347-357 Highway Design & Construction. Univ. Wyoming, Laramie, May,
Donovan P. Sarker P. and Tutumluer E. 2016. Rutting prediction in 1, 206-213.
airport pavement granular base/subbase: A stress history based Gomes Correia A. 1996a. Prediction of subgrade moisture conditions for
approach, Transportation Geotechnics, 9, 139-160 purpose of pavement design. Flexible Pavement (A. Gomes Correia,
Doré G. Konrad J.M. and Roy M. 1997. Role of deicing salt in pavement ed.). A.A. Balkema, Rotterdam, 99.104
deterioration by frost action. Transportation Research Record: Gomes Correia A. (ed) 1996b. Flexible pavements. A.A. Balkema,
Journal of the Transportation Research Board, 1596, 70-75. Rotterdam
Doré G. Konrad J.M. and Roy M. 1999. Deterioration model for Gomes Correia A. (ed) 1999. Unbound Granular Materials. Laboratory
pavements in frost conditions. Transportation Research Record: testing, in-situ testing and modelling. A.A. Balkema, Rotterdam,
Journal of the Transportation Research Board, 1999 (1655), 110- Brookfield
117. Gomes Correia A. 2001. Soil mechanics in routine and advanced
Doré G. 2002. Cold region pavement. Journal of Glaciology and pavement and rail track rational design. Geotechnics for roads, rail
Geocryology, 2002, 24(5), 593-600. tracks and earth structures (A. Gomes Correia and H. Brandl, eds.),
Duley A. 2018. Soil parameters for modelling critical velocity effects of Balkema, 165-187
railways. PhD dissertation, University of Southampton, UK Gomes Correia A. and Lacasse S. 2006. Marine and Transportation
Duong T.V. Cui Y.-J. Tang A.M. Dupla, J.-C. Canou, J. Calon N. and Geotechnical Engineering. General report – session 2e. Proc. XVI
Robinet A. 2014. Investigating the mud pumping and interlayer ECSMGE, Millpress, Rotterdam, The Netherlands, 5, 3045-3069.
creation phenomena in railway sub-structure. Engineering Geology, Gomes Correia A. and Ramos A. 2021. A geomechanics classification
171, 45-58. for the rating of railroad subgrade performance. Rail. Eng. Science.
Duquennoi C. Fremond M. and Levy M. 1989. Modelling of thermal soil https://doi.org/10.1007/s40534-021-00260-z
behavior. In H. Rathmayer (ed.) Frost in geotechnical engineering, Gomez M. Anderson C. DeJong J. T. Nelson D. and Lau X. 2014.
Int. Symp., Saariselk, Finland. 13–15 Mar. 1989. VTT Symp. 94. Stimulating In situ Soil Bacteria for Bio-Cementation of Sands. Geo-
Valtion Teknillinen Tutkimuskeskus, Espoo, Finland, 895–915. Congress 2014 Technical Papers: Geo-characterization and
Duriez J. and Galusinski C. 2021. A Level Set-Discrete Element Method Modeling for Sustainability, ASCE, 1674-1682.
in YADE for numerical, micro-scale, geomechanics with refined Gräbe P.J. and Clayton C.R. 2009. Effects of principal stress rotation on
grain shapes. Computers & Geosciences 157, 104936. permanent deformation in rail track foundations. Journal of
El Kacimi A. Woodward P.K. Laghrouche O. and Medero G. 2013. Time Geotechnical and Geoenvironmental Engineering, 135(4), 555-565.
domain 3D finite element modelling of train-induced vibration at Gräbe P.J. and Clayton C.R.I. 2014. Effects of principal stress rotation
high speed. Computers & Structures, 118, 66-73. on resilient behavior in rail track foundations. Journal of
Elshaer M. and Daniel JS. 2018. Impact of pavement layer properties on Geotechnical and Geoenvironmental Engineering, 140(2), 1–10.
the structural performance of inundated flexible pavements, Grossoni I. Powrie W. Zervos A. Bezin Y. and Le Pen L. 2020.
Transportation Geotechnics, 16, 11-20, ISSN 2214-3912, Modelling railway ballasted track settlement in vehicle-track
https://doi.org/10.1016/j.trgeo.2018.06.002. interaction analysis. Transportation Geotechnics, 26, [100433].
Erol A.O. Dhowian A. and Youssef A. 1987. Assessment of oedometer https://doi.org/10.1016/j.trgeo.2020.100433
methods for heave prediction. Proceedings of 6th International Gu C. Wang Y. Cai Y. and Wang J. 2019. Deformation characteristics of
Conference on Expansive Soils, Technical Session III. 99-105. saturated clay in three-dimensional cyclic stress state. Canadian
Esmaeili M. Aela P. and Hosseini A. 2017. Experimental assessment of Geotechnical Journal, 56, 1789-1802.
cyclic behavior of sand-fouled ballast mixed with tire derived Guo L. Chen J. Wang J. Cai Y. and Deng P. 2016. Influences of stress
aggregates. Soil Dynamics and Earthquake Engineering, 98: 1-11. magnitude and loading frequency on cyclic behavior of K0-
Esveld C. 2001. Modern railway track, MRT Press, The Netherlands. consolidated marine clay involving principal stress rotation. Soil
Fabre C. 2021. Effect of COVID-19 pandemic on air travel, pavement Dynamics and Earthquake Engineering, 84, 94-107.
life and future outlook, ASCE T&DI Airfield Pavement Committee Guo L. Cai Y. Jardine R.J. Yang Z. and Wang, J. 2018. Undrained
(APC) Lecture Series behaviour of intact soft clay under cyclic paths that match vehicle
Fedakar H.I. Cetin B. and Rutherford C.J. 2021. Deformation loading conditions. Canadian Geotechnical Journal, 55(1), 90-106.
characteristics of medium-dense sand-clay mixtures under a Guo Y. Markine V. and Jing G. 2021. Review of ballast track tamping:
principal stress rotation. Transportation Geotechnics, 30: 100616. Mechanism, challenges and solutions. Construction and Building
Feldman F. and Nissen D. 2002. Alternative testing method for the Materials, 300, 123940
measurement of ballast fouling. Conference on Railway Gupta A. Madhusudhan Zervos A. Powrie W. Harkness J. and Le Pen
Engineering, Wollongong, RTSA. L.M. 2021. Grain characterisation of fresh and used railway ballast.
Fityus S.G. Walsh P.F. and Kleeman P.W. 1998. The influence of climate Submitted to Granular Matter.
as expressed by the Thornthwaite index on the design of depth of Guymon G.L. Berg R.L. and Johnston T.C. 1986. Mathematical Model
moisture change of clay soils in the Hunter Valley. Conference on of Frost Heave and Thaw Settlement in Pavements, Report: U.S.
Geotechnical Engineering & Engineering Geology in the Hunter Army Cold Regions Research and Engineering Laboratory
Valley. Springwood, Australia, 251-265 Hamdan N. 2013. Carbonate Mineral Precipitation for Soil Improvement
Fityus S. Smith D.W. and Allman M.A. 2004. Expansive Soil Test Site through Microbial Denitrification. Master's Thesis, Arizona State
Near Newcastle. Journal of Geotechnical and Geoenvironmental University.
Engineering, ASCE, 130(7), 686-695. Hamdan N. Kavazanjian E. and O’Donnell S. 2013. Carbonate
Fredlund D.G. Hasan J.U. and Filson H. 1980. The prediction of total Cementation via Plant Derived Urease. 18th Int. Conf. on Soil
heave. Proceedings 4th International Conference on Expansive Mechanics and Geotechnical Eng., Paris, France.
Soils. Denver, Colorado. 1-11 Hamdan N. 2015. Applications of Enzyme Induced Carbonate
Fredlund D.G. Rahardjo H. and Fredlund M.D. 2012. Unsaturated Soil Precipitation (EICP) for Soil Improvement. Doctoral Dissertation,
Mechanics in Engineering Practice. John Wiley & Sons, Inc. New Arizona State University.
Jersey. Hansson K. Šimůnek J. Mizoguchi M. Lundin L.C. and van Genuchten
Fremond M. and Mikkola M. 1991. Thermomechanical modeling of M.Th. 2004. Water flow and heat transport in frozen soil: Numerical
freezing soil. In X. Yu and C. Wang (ed.) Ground freezing ’91: Proc. solution and freeze–thaw applications. Vadose Zone J. 3, 693–704.
Int. Symposium on Ground Freezing, 6th, Beijing. 10–12 Sept. 1991. Harkness J. 2009. Potential particles for the modelling of interlocking
A.A. Balkema, Rotterdam, the Netherlands, 17–24. media in three dimensions. International Journal for Numerical
Furudoi T. 2010. The second phase construction of Kansai International Methods in Engineering, 80(12), 1573–1594
Airport considering the large and long-term settlement of the clay Harkness J. Zervos A. Le Pen L. Aingaran S. and Powrie W. 2016.
deposits, Soils and Foundations, 50 (6), 805–816 Discrete element simulation of railway ballast: modelling cell
Gholikhani M. Nasouri R. Tahami S.A. Legette S. Dessouky S. and pressure effects in triaxial tests. Granular Matter 18(3), 65.
Montoya A. 2019. Harvesting kinetic energy from roadway Harkness J. and Zervos A. 2019. Some effects of particle shape on the
pavement through an electromagnetic speed bump, Applied Energy, mechanical behaviour of granular materials. DEM8, Enschede,
250, 503-511, DOI: 10.1016/j.apenergy.2019.05.060 Netherlands.
85
Harlan R.L. 1973. Analysis of coupled heat–fluid transport in partially conditions. Journal of Materials in Civil Engineering, 30(3):
frozen soil. Water Resour. Res. 9, 1314–1323. 06017021.
doi:10.1029/WR009i005p01314 Indraratna B. and Ngo N.T. 2018. Ballast Railroad Design: Smart-Uow
Hasnayn M.M. McCarter W.J. Woodward P.K. and Connolly D.P. 2021 Approach: CRC Press.
Railway subgrade performance after repeated flooding – Large-scale Indraratna B. Israr J. and Li M. 2018. Inception of geohydraulic failures
laboratory testing, Transportation Geotechnics, 23, 100329, ISSN in granular soils-an experimental and theoretical treatment.
2214-3912, https://doi.org/10.1016/j.trgeo.2020.100329. Geotechnique, 68(30), 233-248
Hayward T. Lees A. Powrie W. Richards D.J. and Smethurst J A. 2000. Indraratna B. Sajjad MB. Ngo T. Correia AG and Kelly R. 2019a.
Centrifuge modelling of a cutting slope stabilised by discrete piles. Improved Performance of Ballasted Tracks at Transition Zones: A
TRL Report 471. Crowthorne, Berkshire: Transport Research Review of Experimental and Modelling Approaches. Transportation
Laboratory. 1–31. Geotechnics, 2019. 21: p. 100260.
He J. Chu J. and Ivanov V. 2013. Mitigation of liquefaction of saturated Indraratna B. Ngo T.N. and Rujikiatkamjorn C. 2019b. Performance of
sand using biogas. Geotechnique, 63(4), 267. Ballast Influenced by Deformation and Degradation: Laboratory
Hendry M. Hughes D.A. and Barbour L. 2010. Track displacement and Testing and Numerical Modeling. International Journal of
energy loss in a railway embankment. Proceedings of the Institution Geomechanics, 20(1), 04019138.
of Civil Engineers-Geotechnical Engineering, 163, 3-12. Indraratna B. Qi Y. Ngo N. T. Rujikiatkamjorn C. Neville T. Ferreira F.
Hopkins T.C. and Beckham T.L. 1998. Embankment construction using B. and Shahkolahi A. 2019c. Use of Geogrids and Recycled Rubber
Shale. Publication of Kentucky Transportation Centre in Railroad Infrastructure for Enhanced Performance. Geosciences
Houston S. and Houston W. 2017. A suction-oedometer method for 9(1): 30.
computation of heave & remaining heave. Proc. of the 2nd PanAm Indraratna B. Rujikiatkamjorn C. Tawk M. and Heitor A. 2019d.
Conf. on Unsaturated Soils, Dallas, TX, 1, ASCE. Compaction, degradation and deformation characteristics of an
Houston S. and Zhang X. 2021. Review of Expansive and Collapsible energy absorbing matrix. Transportation Geotechnics, 19: 74-83.
Soil Volume Change Models within a Unified Elastoplastic Indraratna B. Singh M. Nguyen T.T. Leroueil S. Abeywickrama A. Kelly
Framework. Soils and Rocks, 44(3). DOI: R. and Neville T. 2020a. Laboratory study on subgrade fluidization
10.28927/SR.2021.064321. under undrained cyclic triaxial loading. Canadian Geotechnical
Hromadka T.V. and Yen C. 1986. A diffusion hydrodynamic model Journal, 57(11), 1767-1779. doi: 10.1139/cgj-2019-0350
(DHM). Adv. Water Resour. 9, 118–170. doi:10.1016/0309- Indraratna B. Korkitsuntornsan W. and Nguyen T. 2020b. Influence of
1708(86)90031-X kaolin content on the cyclic loading response of railway subgrade.
Hryciw RD. Athanasopoulos-Zekkos A. and Yesiller N. (eds) 2012 Transportation Geotechnics: 100319. doi: https://doi.org/10.1016/
Geocongress 2012 State Of the Art and Practice In Geotechnical j.trgeo.2020.100319.
Engineering. ASCE, Reston, VA, USA Institution of Civil Engineers UK 2019 What should be in the National
HSBC 2019 Global Infrastructure Trends, Surbana Jurong SMEC Infrastructure Strategy? https://www.ice.org.uk/ getattachment
internal presentation /news-and- insight/policy/what-should-national-infrastructure-
Hsein Juang C. Luo Z. Atamturktur S. and Huang H. 2013. Bayesian strategy/What-should-be-in-the-National-Infrastructure-Strategy-
Updating of Soil Parameters for Braced Excavations Using Field ICE-July-2019.pdf.aspx
Observations, ASCE J. Geotech. Geoenviron. Eng., 139:395-406. IPCC (Intergovernmental Panel on Climate Change) (2007) Climate
Huang H. and Tutumluer E. 2011. Discrete Element Modeling for fouled Change 2007: Synthesis Report. Contribution of Working Groups I,
railroad ballast. Construction and Building Materials, 25, 3306- II and III to the Fourth Assessment Report of the Intergovernmental
3312. Panel on Climate Change (Pachauri RK and Reisinger A (eds)).
Indraratna B. Nimbalkar S. Christie D. Rujikiatkamjorn C. and Vinod J. IPCC, Geneva, Switzerland
2010. Field assessment of the performance of a ballasted rail track Israr J. and Indraratna B. 2019. Study of Critical Hydraulic Gradients for
with and without geosynthetics. Journal of Geotechnical and Seepage-Induced Failures in Granular Soils. J. Geotech.
Geoenvironmental Engineering, 136, 907-917. Geoenviron. Eng., 145 (7): 04019025.
Indraratna B. Rujikiatkamjorn C. Ameratunga J. and Boyle P. 2011a Ito T. and Matsui T. (1975). Methods to estimate lateral force acting on
Performance and Prediction of Vacuum Combined Surcharge stabilizing piles. Soils and Foundations, 15(4), 43–59.
Consolidation at Port of Brisbane. J. of Geotechnical & Jaksa M.B. Kaggwa W.S. Woodburn J.A. and Sinclair R. 2002g.
Geoenvironmental Engineering, ASCE, 137 (11), 1009-1018. Influence of large gum trees on the soil suction profile in expansive
Indraratna B. Ngo N.T. and Rujikiatkamjorn C. 2011b. Behavior of soils. Australian Geomechanics, 37(1), 23-33
geogrid-reinforced ballast under various levels of fouling. Jame Y.W. and Norum D.I. 1980. Heat and mass transfer in a freezing
Geotextiles and Geomembranes, 29(3), 313-322. unsaturated porous medium. Water Resources Research, 16(4), 811-
Indraratna B. Ngo N.T. and Rujikiatkamjorn C. 2013a. Deformation of 819.
coal fouled ballast stabilized with geogrid under cyclic load. Journal Janoo V.C. and Berg R.L. 1990. Thaw weakening of pavement structures
of Geotechnical and Geoenvironmental Engineering, 139(8), 1275- in seasonal frost areas. Transportation Research Record: Journal of
1289. the Transportation Research Board, 1286, 217-233.
Indraratna, B., Balasubramaniam, A. S., Poulos, H. G., Rujikiatkamjorn, Jayatilaka R. Gay D. Lytton R. and Wray W. 1992. Effectiveness of
C. and Ameratunga, J. 2013b. Performance and prediction of marine controlling pavement roughness due to expansive clays with vertical
clay treated with vacuum and surcharge consolidation at Port of moisture barriers. Research Study No. 2/11-8-88-1165. Texas Depart
Brisbane, Australian Geomechanics, 48(4), 161-180. of Transportation, Texas Transportation Institute, & Texas Tech
Indraratna B. Ngo NT. Rujikiatkamjorn C. and Vinod JS. 2014a. University, 230
Behavior of Fresh and Fouled Railway Ballast Subjected to Direct Jefferies M. Shuttle D. and Been K. 2015. Principal stress rotation as
Shear Testing: Discrete Element Simulation. International Journal cause of cyclic mobility. Geotechnical Research, 2, 66-96.
of Geomechanics, 14(1), 34-44 Jennings J.E.B and Knight K. 1957. The prediction of total heave from
Indraratna B. Nimbalkar S. and Neville T. 2014b. Performance double oedometer test, In Proceedings of Symposium on Expansive
assessment of reinforced ballasted rail track. Proceedings of the Clays. South African Institution of Civil Engineers, Johannesburg,
Institution of Civil Engineers - Ground Improvement, 167(1), 24-34. 13-19.
Indraratna B. Nimbalkar S. and Rujikiatkamjorn C. 2014c. From theory Jennings J.E.B. Firtu R.A. Ralph T.K. and Nagar N. 1973. An improved
to practice in track geomechanics–Australian perspective for method for predicting heave using the oedometer test. Proc. 3rd Int.
synthetic inclusions. Transportation Geotechnics 1(4), 171-187. Conf. Expansive Soils, Haifa, Israel, 2, 149-154.
Indraratna B. Qi Y. and Heitor A. 2017a. Evaluating the properties of Jia W. Markine V. Guo Y. and Jing G. 2019. Experimental and numerical
mixtures of steel furnace slag, coal wash, and rubber crumbs used as investigations on the shear behaviour of recycled railway ballast.
subballast. Journal of Materials in Civil Engineering, 30(1): Construction and Building Materials, 217, 310-320
04017251. Jing P. Nowamooz H. and Chazallon C. 2018. Permanent deformation
Indraratna B. Sun Q. and Grant J. 2017b. Behaviour of subballast behaviour of a granular material used in low-traffic pavements, Road
reinforced with used tyre and potential application in rail tracks. Materials and Pavement Design, 19(2), 289-314.
Transportation Geotechnics, 12: 26-36. Johnson L.D. and Snethen D.R. 1978. Prediction of potential heave of
Indraratna B. Sun Q. Heitor A. and Grant J. 2017c. Performance of rubber swelling soils. Geotechnical Testing Journal, 1(3), 117-124.
tire-confined capping layer under cyclic loading for railroad
86
Jones C.W. 1954. The permeability and settlement of laboratory Lee K.H. and Pande G.N. 2004. Development of a two-surface model in
specimens of sand and sand-gravel mixtures. ASTM: Special the Multi-laminate framework. Proceedings of the 11th Conference
Technical Publication, 163, 68-78. on Numerical Models in Geomechanics, Ottawa, 139-144.
Kang M. Qamhia IIA. Tutumluer E. Garg N. and Villafane W. 2022. Le Pen L. Powrie W. Zervos A. Ahmed S. and Aingaran S. 2013.
Airport Pavement Stiffness Monitoring and Assessment of Dependence of shape on particle size for a crushed rock railway
Mechanical Stabilization using Bender Element Field Sensor, ballast. Granular Matter, 15(6), 849-861. https://doi.org/
submitted Transport Research Record 10.1007/s10035-013-0437-5
Kang YS. Yang SC. Lee H.S. Kim Y.B. Jang S.Y. and Kim E. 2008. A Le Pen L. Ahmed S. Zervos A. Harkness J. and Powrie W. 2014. Resin
study of track and train dynamic behavior of transition zone between recovery and the use of computed tomography for quantitative image
concrete slab track and ballasted track, in SPARK. 2008, Korea analysis of railway ballast. Proceedings of the Second International
Railroad Research Institute, Uiwang, South Korea. p. 1-7. Conference on Railway Technology: Research, Development and
Karatas I. Kavazanjian E. and Rittmann B. (2008). Microbially induced Maintenance, France Ed. J. Pombo. https://doi.org/10.4203/
precipitation of calcite using Pseudomonas denitrificants. Proc., 1st ccp.104.130
International Conference on Bio-Geo-Civil Engineering, 58-66. Le Pen L. Milne D. Watson G. Harkness J. and Powrie W. (2020). A
Karunarathne A. Rajeev P. Chan D. and Kodikara J. 2012. Ground- model for the stochastic prediction of track support stiffness.
atmosphere Interaction Modelling for Long-Term Prediction of Soil Proceedings of the Institution of Mechanical Engineers, Part F:
Moisture and Temperature. Canadian Geotechnical Journal. 49(9), Journal of Rail and Rapid Transit, 234, 468-481
DOI: 10.1139/T2012-068 Li D. Hyslip J. Sussmann T. and Chrismer S. 2015. Railway geotechnics:
Kausel E. Estaire J. and Crespo-Chacón I. 2020. Proof of critical speed CRC Press.
of high-speed rail underlain by stratified media. Proc. R. Soc. A 476: Li L. Nimbalkar S. and Zhong, R. 2018. Finite element model of ballasted
20200083. http://dx.doi.org/10.1098/rspa.2020.0083 railway with infinite boundaries considering effects of moving train
Kavazanjian E. and Hamdan N. 2015. Enzyme Induced Carbonate loads and Rayleigh waves. Soil Dynamics and Earthquake
Precipitation (EICP) Columns for Ground Improvement. IFCEE, Engineering, 114, 147-153.
ASCE, San Antonio, TX. Li H. and McDowell G. 2020. Discrete element modelling of two-layered
Kavazanjian E. O’Donnell S. T. and Hamdan N. 2015. Biogeotechnical ballast in a box test. Granular Matter, 22(4), 76
mitigation of earthquake-induced soil liquefaction by denitrification: Li J. Cameron D. and Ren G. 2013. Case study and back analysis of a
a two-stage process. In Proceedings of 6th International Conference residential building damaged by expansive soils, Computers and
on Earthquake Geotechnical Engineering, Christchurch, New Geotechnics, 56, 89-99.
Zealand, 20-28 Li N. Chen B. Chen F. and Xu X. 2000. The coupled heat-moisture-
Kelly R. and Huang J. 2015. Bayesian updating for one-dimensional mechanic model of the frozen soil. Cold Regions Science and
consolidation measurements, Canadian Geotechnical Journal, 52: Technology, 31(3), 199-205.
1318–1330 (2015) dx.doi.org/10.1139/cgj-2014-0338 Li, N., Chen, F., Su, B., & Cheng, G. (2002). Theoretical frame of the
Khodadadi Tirkolaei H. Kavazanjian E. van Paassen L. and DeJong, J. saturated freezing soil. Cold Regions Science and Technology,
2017. Bio-grout materials: A review. Grouting, 1-12. 35(2), 73-80.
Kianfar K. Indraratna B. and Leroueil S. (2015) Radial consolidation Li T. Su Q. Kaewunruen S. 2019. Saturated ground vibration analysis
response upon the application and removal of vacuum and fill based on a three-dimensional coupled train-track-soil interaction
loading, Canadian Geotechnical Journal, 52 (12), 2156 - 2162 model. Applied Sciences, 9, 4991.
Kitazume M. and Terashi M. 2013. The Deep Mixing Method, CRC Li X. Ekh M. and Nielsen J.C. 2016. Three‐dimensional modelling of
Press, Taylor & Francis Group, 410p. differential railway track settlement using a cycle domain
Komornik A. and David D. 1969. Prediction of Swelling Pressure of constitutive model. International Journal for Numerical and
Clays, Journal of the Soil Mechanics and Foundations Division, 95, Analytical Methods in Geomechanics, 40, 1758-1770.
209-226 Liang T. Bengough A.G. Knappett J.A. Muir-Wood D. Loades K.W.
Konrad J.M. and Morgenstern N.R. 1980. A mechanistic theory of ice Hallett P.D. Boldrin D. Leung A.K. and Meijer G.J. 2017. Scaling
lens formation in fine-grained soils. Canadian Geotechnical of the reinforcement of soil slopes by living plants in a geotechnical
Journal, 17(4), 473-486. centrifuge. Ecological Engineering, 109(Part B), 207-227.
Konrad J.M. and Morgenstern N.R. 1981. The segregation potential of a https://doi.org/10.1016/j.ecoleng.2017.06.067
freezing soil. Canadian Geotechnical Journal, 18(4), 482-491. Liu J.K. and Xiao J.H. 2010. Experimental study on the stability of
Konrad J.M. and Morgenstern N.R. 1982a. Prediction of frost heave in railroad subgrade with increasing train speed. J. Geotech.
the laboratory during transient freezing. Canadian Geotechnical Geoenviron. ASCE, 136(6), 833-841.
Journal, 19(3), 250-259. Liu S. Huang H. Qiu T. and Gao L. 2016a. Comparison of laboratory
Konrad J.M. and Morgenstern N.R. 1982b. Effects of applied pressure testing using SmartRock and discrete element modeling of ballast
on freezing soils. Canadian Geotechnical Journal, 19(4), 494-505. particle movement, Journal of Materials in Civil Engineering, 29(3),
Kouroussis G. Van Parys L. Conti C. and Verlinden O. 2014. Using D6016001.
three-dimensional finite element analysis in time domain to model Liu Z. Yu X.B. Tao J.L. and Sun Y. 2012. Multiphysics extension to
railway-induced ground vibrations. Advances in Engineering physically based analyses of pipes with emphasis on frost actions.
Software, 70, 63-76. Journal of Zhejiang University-Science A, 13(11), 877-887.
Kouroussis G. and Verlinden O. 2015. Prediction of railway ground Lopes D. and Karunarathne A. 2017. New Conditioned Soil Index Test
vibrations: accuracy of a coupled lumped mass model for and Characteristic Ground Movement Calculation Model.
representing the track/soil interaction. Soil Dynamics and Geotechnical Special Publication: Proceedings of the Second Pan-
Earthquake Engineering 69, 220-226. American Conference on Unsaturated Soils (PanAm Unsaturated
Kujala K. 1997. Estimation of frost heave and thaw weakening by Soils 2017), Dallas, Texas, GSP 303, 81-89
statistical analyses and physical models. In: Knutsson, S. Ž . Ed. , López-Acosta N.P. Espinosa-Santiago A.L. Pineda-Núñez V.M. Ossa A.
Ground Freezing ’97, Proceedings of the International Symposium Mendoza M.J. Ovando-Shelley E. and Botero E. 2019. Performance
on Ground Freezing and Frost Action in Soils, Lulea, Sweden. A.A. of a test embankment on very soft clayey soil improved with drain-
Balkema, The Netherlands, 31–42 to-drain vacuum preloading technology, Geotextiles and
Kumara JJ. and Hayano K. 2016. Deformation characteristics of fresh Geomembranes, 47 (5), 618-631.
and fouled ballasts subjected to tamping maintenance, Soils and Lu N. and Likos W.J. 2006. Suction stress characteristic curve for
Foundations, 56(4), 652-63. unsaturated soil. J. Geotech. Geoenviron. Eng. 132, 131–142.
Lackenby J. Indraratna B. McDowell G.R. and Christie D. 2007. Effect doi:10.1061/(ASCE)1090-0241(2006)132:2(131)
of confining pressure on ballast degradation and deformation under Lytton R. Aubeny C. and Bulut R. 2005. Design Procedures for Soils on
cyclic triaxial loading. Geotechnique, 57(6), 527–536. Expansive Soils: Volume 1. FHWA/TX-05/0-4518. Texas
Lakuši S. Ahac M. and Haladin I. (2010) Experimental investigation of Department of Transportation.
railway track with under sleeper pad. Proceedings of the 10th Lytton R.L. Pufahl D.E. Michalak C.H. Liang H.S. and Dempsey K.J.
Slovenian Road and Transportation Congress, 386-393. 1990. An Integrated Model of the Climatic Effects on Pavement.
Lamas-Lopez F. Cui Y.-J. Calon N. D’Aguiar S.C. De Oliveira M.P. and Report No. FHWA‐RD‐90‐033. Federal Highway Administration,
Zhang T. 2016. Track-bed mechanical behaviour under the impact Texas Transportation Institute, Texas A&M University McLean,
of train at different speeds. Soils and Foundations, 56, 627-639. VA.
87
MacKay M.H. Hein D.K. and Emery J.J. 1992. Evaluation of frost action Navaratnarajah S.K. Indraratna B. and Ngo, N.T. 2018. Influence of
mitigation procedures for highly frost-susceptible soils. Under Sleeper Pads on Ballast Behavior Under Cyclic Loading:
Transportation Research Record: Journal of the Transportation Experimental and Numerical Studies. Journal of Geotechnical and
Research Board, 1362, 79-89. Geoenvironmental Engineering, 144(9), 04018068.
Malisetty R.S. Indraratna B. and Vinod, J. 2020a. Behaviour of ballast Navy, Dept. of, Naval Facilities Engineering Command. 1971. Design
under principal stress rotation: Multi-laminate approach for moving Manual-Soil Mechanics, Foundations & Earth Structures. U.S.
loads. Computers and Geotechnics, 125, 103655. Naval Publications & Forms Center, NAVFAC DM-7.
Malisetty R.S. Indraratna B. and Vinod J.S. 2020b. Multi-laminate Nayak N.V. and Christensen R.W. 1971. Swell characteristics of
Mathematical Framework for Analyzing the Deformation of Coarse compacted expansive soils. Clay & Clay Minerals. 19(4), 251-261.
Granular Materials. International Journal of Geomechanics, 20(6): Neaupane K.M. and Yamabe T. 2001. A fully coupled thermo-hydro-
06020004. mechanical nonlinear model for a frozen medium. Computers and
Mamou A. Powrie W. Priest J.A. and Clayton C.R.I. 2017. The effects of Geotechnics, 28(8), 613-637.
drainage on the behaviour of railway track foundation materials Nelson J.D. and Miller D.J. 1992. Expansive Soils: Problems and
during cyclic loading. Géotechnique, 67(10) 845–854. Practice in Foundation and Pavement Engineering. John Wiley &
Mamou A. Priest J.A. Clayton C.R.I. and Powrie W. 2018. Behaviour of Sons, Inc.
saturated railway track foundation materials during undrained cyclic Nelson J.D. Overton D.D. and Durkee D.B. 2001. Depth of Wetting &
loading. Canadian Geotechnical Journal, 55(5), 689 697. the Active Zone. Proceedings of Expansive Clay Soils &
Mamou A. Clayton C. Powrie W. and Priest J. 2019a. The role of clay Vegetative Influence on Shallow Foundations, Houston, Texas,
content on the response of railway track foundations during free-to- USA, ASCE, 95-109.
drain cyclic changes in principal stress rotation. Transportation Neupane D. Yasuhara H. Kinoshita N. and Ando Y. 2015. Distribution
Geotechnics, 20: 100246. of Mineralized Carbonate and Its Quantification Method in Enzyme
Mamou A. Powrie W. Priest J.A. and Clayton, C.R.I. 2019b. The use of Mediated Calcite Precipitation Technique. Soils and Foundations,
the hollow cylinder apparatus to study stress paths relevant to 55(2), 447-457.
railway track foundations. In Proceedings of the 7th International Newman G.P. and Wilson G.W. 1997. Heat and mass transfer in
Symposium on Deformation Characteristics of Geomaterials, unsaturated soils during freezing. Can. Geotech. J. 34, 63–70
Glasgow, UK. Ngo N.T. Indraratna B. and Rujikiatkamjorn C. 2014. DEM simulation
Mamou A. Powrie, W. Clayton C.R.I. Priest J.A. 2021. Suitability of of the behaviour of geogrid stabilised ballast fouled with coal.
empirical equations for estimating permanent settlement of railway Computers and Geotechnics, 55, 224-231
foundation materials subjected to cyclic loading with principal stress Ngo N.T. Indraratna B. and Rujikiatkamjorn C. 2016. Modelling
rotation. Canadian Geotechnical Journal. 58(10), 1603-1610. geogrid-reinforced railway ballast using the discrete element
Mashiri M.S. Vinod J.S. Sheikh M.N. and Tsang H.H. 2015. Shear method. Transportation Geotechnics, 8, 86-102
strength and dilatancy behaviour of sand–tyre chip mixtures. Soils Ngo N.T. Indraratna B. and Rujikiatkamjorn C. 2017. Simulation
and Foundations, 55(3), 517-528. Ballasted Track Behavior: Numerical Treatment and Field
McDowell G.R. and Li H. 2016. Discrete element modelling of scaled Application. International Journal of Geomechanics, 17(6),
railway ballast under triaxial conditions. Granular Matter 18(3), 66 04016130.
McKeen R.G. and Johnson L.D. 1990. Climate-Controlled Soil Design Ngo N.T. Indraratna B. and Rujikiatkamjorn C. 2019. Improved
Parameters for Mat Foundations. J. of Geotechnical Eng., 116(7), performance of ballasted tracks under impact loading by recycled
1073-1094. rubber mats. Transportation Geotechnics, 20, 100239.
McKenna G. Argyroudis S.A. Winter M.G. and Mitoulis S.A. 2021, Ngo, N.T. and Indraratna B. 2020. Mitigating ballast degradation with
Multiple hazard fragility analysis for granular highway under-sleeper rubber pads: Experimental and numerical
embankments: Moisture ingress and scour, Transportation perspectives. Computers and Geotechnics, 122, 103540
Geotechnics, 26, 100431, ISSN 2214-3912, https://doi.org/10.1016/ Nguyen K. Villalmanzo D. Goicolea J. and Gabaldon F. 2016. A
j.trgeo.2020.100431. computational procedure for prediction of ballasted track profile
Meijer G. Muir-Wood D. Knappett J.A. Bengough G.A. and Liang T. degradation under railway traffic loading. Proceedings of the
(Accepted/In press). Root reinforcement: continuum framework for Institution of Mechanical Engineers, Part F: Journal of Rail and
constitutive modelling. Geotechnique. Rapid Transit, 230, 1812-1827.
Miller R.D. 1972. Freezing and heaving of saturated and unsaturated Nguyen T.T. and Indraratna B. 2020. A coupled CFD-DEM approach to
soils. Highway Research Record, In Proc. Int. Conf. Permafrost, examine the hydraulic critical state of soil under increasing hydraulic
2nd, Yakutsk, Siberia. 13–28 July 1973. Natl. Acad. Sci., gradient. ASCE International Journal of Geomechanics, 20(9):
Washington, DC.1972, 344-452(393). 04020138-1:15. doi: https://doi.org/10.1061/ (ASCE)GM.1943-
Miller R.D. 1978. Frost heaving in non-colloidal soils. In Proc. Int. Conf. 5622.0001782.
on Permafrost, 3rd, Edmonton, AB, Canada. Natl. Res. Counc. of Nimbalkar S. Indraratna B. Dash S.K. and Christie D. 2012. Improved
Canada, Ottawa, ON, 708–713. Performance of Railway Ballast under Impact Loads Using Shock
Milne D. Le Pen L. Thompson D. and Powrie W. (2017). Properties of Mats. Journal of Geotechnical and Geoenvironmental Engineering,
train load frequencies and their applications. Journal of Sound and 138(3), 281-294.
Vibration 397, 123-140. https://doi.org/10.1016/j.jsv.2017.03.006 Nimbalkar S. and Indraratna B. 2016. Improved performance of ballasted
Milne D. Harkness J. Le Pen L. and Powrie W. (2019). The influence of rail track using geosynthetics and rubber shockmat. Journal of
variation in track level and support system stiffness over longer Geotechnical and Geoenvironmental Engineering, 142(8): 4016031.
lengths of track for track performance and vehicle track interaction. Nishimura S. Gens A. Olivella S. and Jardine R.J. 2009. THM-coupled
Vehicle System Dynamics. https://doi.org/10.1080/ 00423114. finite element analysis of frozen soil: Formulation and application.
2019.1677920 Geotechnique 59, 159–171. doi:10.1680/geot.2009.59.3.159
Mitchell J. and Kelly R. 2013. Addressing some current challenges in Nobel C.A. 1966. Swelling measurements & prediction of heave for a
ground improvement, Proc ICE Ground Improvement, 1, 1-11 lacustrine clay. Can. Geotechnical Journal, 3(1), 32-41.
Mitchell P.W. 1980. The Concepts Defining the Rate of Swell of Noborio K. McInnes K.J. and Heilman J.L. 1996. Two-dimensional
Expansive Soils. Proceedings of the 4th International Conference on model for water, heat and solute transport in furrow-irrigated soil: I.
Expansive Soils. Denver, USA. 1, 106-116. Theory. Soil Sci. Soc. Am. J. 60, 1001–1009.
Mitchell P.W. 2008. Footing Design for Residential Type Structures in doi:10.2136/sssaj1996.03615995006000040007x
Arid Climates. Australian Geomechanics, 41(4). 51-68. OECD Futures Project on Transcontinental Infrastructure Needs to
Mitchell P.W. 2013. Climate Change Effects on Expansive Soil 2030/50. 2011 Strategic Transport Infrastructure Needs to 2030:
Movements. Proceedings of the 18th International Conference on Main Findings
Soil Mechanics and Geotechnical Engineering. Paris, 1159-1162. O'Donnell S.T. 2016. Mitigation of Earthquake-Induced Soil
Muttuvel T. Kelly R. Malorey D. Litvin E. and Pineda J. 2020. Internal Liquefaction via Microbial Denitrification: A Two-Stage Process.
compression of fill materials originating from Bringelly Shale. Proc. Doctoral Dissertation, Arizona State University.
Australian Geomechanics Society Sydney Symposium O’Donnell S.T. Rittmann B.E. and Kavazanjian E. 2017a. MIDP:
National Academies of Sciences, Engineering, and Medicine 2019. Liquefaction Mitigation via Microbial Denitrification as a Two-
Critical Issues in Transportation 2019. Washington, DC: The Stage Process. I: Desaturation. Journal of Geotechnical and
National Academies Press. https://doi.org/10.17226/25314 Geoenvironmental Engineering, 143(12).
88
O’Donnell S.T. Kavazanjian E. and Rittmann B.E. 2017b. MIDP: Geotechnics, 21, 1-12. [100261]. https://doi.org/10.1016/
Liquefaction Mitigation via Microbial Denitrification as a Two- j.trgeo.2019.100261
Stage Process. II: MICP. Journal of Geotechnical and Priest J.A. Powrie W. Yang, L. Grabe, P.J. and Clayton, I. (2010).
Geoenvironmental Engineering, 143(12). Measurements of transient ground movements below a ballasted
Olaiz A.H. Singhar S.H. Vann J.D. and Houston S.L. 2016. Comparison railway line. Géotechnique, 60(9), 667–677.
& applications of the Thornthwaite moisture index using GIS. Proc. Putra H. Yasuhara H. Kinoshita N. Neupane D. and Lu C. 2016. Effect
of the 2nd PanAm Conf. on Unsaturated Soils, Dallas, TX, 1, ASCE of Magnesium as Substitute Material in Enzyme-Mediated Calcite
Olaiz A.H. 2017. Evaluation of Testing Methods for Suction-Volume Precipitation for Soil-Improvement Technique. Front. Bioeng.
Change of Natural Clay Soils. Arizona State University. Biotechnol., 4.
Olaiz A.H. Mosawi M. and Zapata C.E. 2021. An improved framework Qi Y. Indraratna B. and Vinod J S. 2018. Behavior of Steel Furnace Slag,
for volume change of shrink/swell soils subjected to time-varying Coal Wash, and Rubber Crumb Mixtures with Special Relevance to
climatic effects. Soils and Rocks, 44(3), [e2021065621]. Stress-Dilatancy Relation. Journal of Materials in Civil Engineering
https://doi.org/10.28927/SR.2021.065621 30(11): 04018276.
O’Neill K. and Miller R.D. 1985. Exploration of a rigid ice model of frost Qi Y. Indraratna B. Heitor A. and Vinod J. 2019a. The influence of rubber
heave. Water Resour. Res. 21, 281–296. doi:10.1029/ crumbs on the energy absorbing property of waste mixtures.
WR021i003p00281 Geotechnics for Transportation Infrastructure. Sundaram R., S. J.,
Oswell J.M. 2011. Pipelines in permafrost: geotechnical issues and Havanagi V. Singapore, Springer. 29: 271-281.
lessons. Canadian Geotechnical Journal, 48(9), 1412-1431. Qi Y. Indraratna B. and Coop M.R. 2019b. Predicted Behavior of
Paixão A. Alves Ribeiro C. Pinto N. Fortunato E. and Calçada R. (2014). Saturated Granular Waste Blended with Rubber Crumbs.
On the use of under sleeper pads in transition zones at railway International Journal of Geomechanics 19(8): 04019079.
underpasses: experimental field testing. Structure and Infrastructure Qian Y. Mishra D. Tutumluer E. and Kazmee HA. 2015. Characterization
Engineering, 1-17. of geogrid reinforced ballast behavior at different levels of
Parente M. Correia A.G. Cortez P. 2016. A Novel Integrated degradation through triaxial shear strength test and discrete element
Optimization System for Earthwork Tasks. Transportation Research modeling. Geotextiles and Geomembranes 43(5), 393-402
Procedia, 14, pp. 3601-3610, DOI: 10.1016/j.trpro.2016.05.428 Qian J.G. Wang Y.G. Yin Z.Y. and Huang M.S. 2016. Experimental
Perera Y.Y. Zapata C.E. Houston W.N. and Houston S.L. 2004. Moisture identification of plastic shakedown behavior of saturated clay
Equilibria beneath Highway Pavements, Transportation Research subjected to traffic loading with principal stress rotation.
Board 83rd Annual Meeting CD-ROM, National Research Council, Engineering Geology, 214, 29-42.
Washington D.C. Rabin Y. and Steif P.S. 1998. Thermal stresses in a freezing sphere and
Pham V.P. van Paassen L.A. Star W.R.L.v.d. and Heimovaara T.J. 2018. its application to cryobiology. Journal of Applied Mechanics, 65(2),
Evaluating Strategies to Improve Process Efficiency of 328-333.
Denitrification-Based MICP. Journal of Geotechnical and Ramos A. Gomes Correia A. Indraratna B. Ngo T. Calçada R. and Alves
Geoenvironmental Engineering, 144(8). Costa P. 2020. Mechanistic-empirical permanent deformation
Picornell M. and Lytton R.L. 1984. Modeling the heave of a heavily models: Laboratory testing, modelling and ranking. Transportation
loaded foundation. Proceeding of 5th International Conference on Geotechnics. 23. 100326. 10.1016/j.trgeo.2020.100326.
Expansive Soils. Adelaide, Australia. 104-108 Rana S. Subramani P. Fangueiro R. and Correia A.G. 2016. A review on
Pineda J.A. Alonso E.E and Romero E. 2014. Environmental degradation smart self-sensing composite materials for civil engineering
of claystones, Geotechnique, 64(1), 64-821 applications, AIMS Materials Science, Vol 3(2), 357-379
Porter A.A. and Nelson J.D. 1980. Strain controlled testing of soils. Proc. Ranganathan B.V. and Satyanarayana B. 1965. A rational method of
4th Int. Conf Expansive Soils, ASCE & Int. Soc. Soil Mech. Found. predicting swelling potential for compacted expansive clays.
Eng. Denver, CO. 34-44 Proceedings of the 6th International Conference on Soil Mechanics
Post-Tensioning Institute. 2004. Design of Post-Tensioned Slabs-on- & Foundation Engineering. International Society for Soil Mechanics
Ground. Phoenix: 3rd Edition, Post-Tensioning Institute. & Geotechnical Engineering, London, 1, 92-96.
Post-Tensioning Institute, 2008. Design & construction of post-tensioned Rebata-Landa V. and Santamarina J.C. 2012. Mechanical effects of
slabs-on-ground, 3rd edition. Post Tensioning Institute, Phoenix. biogenic nitrogen gas bubbles in soils. Journal of Geotechnical and
Potticary M. Zervos A. and Harkness J. 2014. A numerical investigation Geoenvironmental Engineering, 138(2), 128-137.
into the effect of particle form on the strength of granular materials. Rosenbalm D.C. and Zapata C.E. 2013. Incorporation of the Soil-Water
22nd UK National Conference of the Association for Computational Characteristic Curve Hysteresis in Pavement Design. Advances in
Mechanics in Engineering, United Kingdom. Unsaturated Soils. Caycedo et al. (eds). Proc. of the First Pan-
Potticary M. Zervos A. and Harkness J. 2015. An investigation into the American Conf. on Unsaturated Soils, ASCE, Cartagena, Colombia,
effect of particle platyness on the strength of granular materials using Taylor & Francis Group, 461-467.
the discrete element method. IV International Conference on Rosenbalm D.C. and Zapata C.E. 2017. Effect of wetting and drying
Particle-based Methods. Fundamentals and Applications cycles on the behavior of compacted expansive soils, Journal of
(PARTICLES 2015), Spain. 767-778. Materials in Civil Engineering, 29(1), 04016191.
Potticary M. Zervos A. and Harkness J. 2016. The effect of particle https://doi.org/10.1061/(ASCE)MT.1943-5533.0001689
elongation on the strength of granular materials. 24th Conference on Rouainia M. Helm P. and Davies O. 2020. Deterioration of an
Computational Mechanics, United Kingdom, 239-242. infrastructure cutting subjected to climate change. Acta Geotechnica
Potticary M. 2018. A Numerical Investigation of the Effect of Particle 15, 2997–3016. https://doi.org/10.1007/s11440-020-00965-1
Shape on the Strength of Coarse Granular Materials. PhD Thesis, Roy M. Crispin J. Konrad J.M. and Larose G. 1992. Field study of two
University of Southampton, UK. road sections during a freeze-thaw cycle. Transportation Research
Powrie W. Yang L.A. and Clayton C.R.I. 2007. Stress changes in the Record: Journal of the Transportation Research Board, 1992,
ground below ballasted railway track during train passage. (1362), 71-78
Proceedings of the Institution of Mechanical Engineers: Part F: Rujikiatkamjorn C. Indraratna B. and Chiaro G. 2013. Compaction of
Journal of Rail and Rapid Transit, 247-261. coal wash to optimise its utilisation as water-front reclamation fill.
Powrie W. 2014. On track: the future for rail infrastructure systems. Geomechanics and Geoengineering, 8, 36-45.
Proceedings of the Proceedings of the Institution of Civil Engineers- Salour F. Erlingsson S. and Zapata C.E. 2014. Modelling Resilient
Civil Engineering. Thomas Telford Ltd, 167, 177-185. Modulus Seasonal Variation of Silty Sand Subgrade Soils with
Powrie W. and Smethurst J.A. 2015. Discrete passive piles for Matric Suction Control. Canadian Geotech. J. 51(12), 1413-1422.
infrastructure slope stabilization. International Conference on Soft Salour F. Erlingsson S. and Zapata C.E. 2015. Model for Seasonal
Ground Engineering (ICSGE 2015), Singapore. Variation of Resilient Modulus in Silty Sand Subgrade Soil:
Powrie, W., Le Pen, L., Milne, D., Watson, G., & Harkness, J. (2019a). Evaluation with Falling Weight Deflectometer. Transportation
Behaviour of under-track crossings on ballasted railways. Research Record: J. of the Transportation Research Board, No.
Transportation Geotechnics, 21, [100258]. https://doi.org/10.1016 2510. Transportation Research Board of the National Academies,
/j.trgeo.2019.100258 Washington, D.C, 65-73.
Powrie W. Le Pen L. Milne D. and Thompson D. (2019b). Train loading Sampson E. Schuster R.L. and Budge W.D. 1965. A method of
effects in railway geotechnical engineering: ground response, determining swell potential of an expansive clay. Proc. Engineering
analysis, measurement and interpretation. Transportation
89
Effects of Moisture Changes in Soils. Int. Res. Eng. Conf Expansive Conference, 195, 05003 (2020), October 19-21st, 2020, DOI:
Clay Soils. Texas A & M Univ. Press, College Station, TX, 255-275. https://doi.org/10.1051/e3sconf/202019505003
Sarker P. and Tutumluer E. 2020. A Stress-history-based Approach for Stallings Young E.G. 2021. Fluid Flow through Granular Soils Treated
Predicting Deformation Potentials of Granular Base and Subbase with Microbial Induced Desaturation and Precipitation. Doctor of
Layers in Airport Pavements (NAPTF CC5 data analysis), FAA Philosophy Dissertation, Arizona State University, ProQuest.
Final Report FAA_UILU-ENG-2020-2013 Stallings Young E.G. Mahabadi N. Zapata C.E. and Paassen L.A.v. 2021.
Sayeed M.A. and Shahin M.A. 2016. Three-dimensional numerical Microbial-Induced Desaturation in Stratified Soil Conditions.
modelling of ballasted railway track foundations for high-speed International Journal of Geosynthetics and Ground Engineering,
trains with special reference to critical speed. Transportation 7(37), 1-17.
Geotechnics, 6, 55-65. Style R.W. Peppin S.S. Cocks A.C. and Wettlaufer, J.S. 2011. Ice-lens
Schaus L. Eng P. Popik M. and Eng M. 2011. Frost heaves: a problem formation and geometrical supercooling in soils and other colloidal
that continues to swell. In Conference and Exhibition of the materials. Physical Review E, 84(4), 041402
Transportation Association of Canada. Transportation Successes: Su Y. and Cui Y-J. 2022. Modelling the suction-dependent plastic
Let's Build on Them. shakedown limit of unsaturated fine/coarse soil mixture with
Schneider G.L. and Poor A.R. 1974. The prediction of soil heave & swell consideration of soil-water retention curve. Transportation
pressures developed by an expansive clay. Research Report, No: Geotechnics, 32, art. no. 100698
TR-9-74. Construction Research Center, Univ. of Texas. Subba Rao K.S. and Tripathy S. 2003. Effect of aging on swelling &
Seed H.B. Woodward R.J. and Lundgren R. 1962. Prediction of Swelling swell-shrink behavior of a compacted expansive soil. ASTM
Potential for Compacted Clays. Journal of the Soil Mech. Found. Geotechnical Testing Journal, 26(1), 36-46.
Div., ASCE, 88, SM4, 57-59. Suhr B. Marschnig S. and Six K. 2018. Comparison of two different types
Selig E.T. and Waters J.M. 1994. Track geotechnology and substructure of railway ballast in compression and direct shear tests: experimental
management, Thomas Telford, London. results and DEM model validation. Granular Matter, 20(4), 70
Senetakis K. Anastasiadis A. and Pitilakis K. 2012. Dynamic properties Sullivan R.A. and McClelland B. 1969. Predicting heave of buildings on
of dry sand/rubber (SRM) and gravel/rubber (GRM) mixtures in a unsaturated clay. Proc. 2nd Int. Res. Eng. Conf. Expansive Soils.
wide range of shearing strain amplitudes. Soil Dynamics and Texas A & M Univ. Press, College Station, TX, 404-420.
Earthquake Engineering, 33(1): 38-53. Summersgill F.C. Kontoe S. and Potts D.M. 2018. Stabilisation of
Shanker N.B. Ratnam M.V. and Rao A.S. 1987. Multi-dimensional swell excavated slopes in strain-softening materials with piles.
behaviour of expansive clays. Proc. 6th Int. Conf. Expansive Soils. Géotechnique 68, No. 7, 626–639, https://doi.org/10.1680/
New Delhi, India. jgeot.17.P.096.
Sheng X. Jones C.J.C. and Thompson D.J. 2004. A theoretical study on Sun Q. Indraratna B. and Ngo T.N. 2019. Effect of increase in load and
the influence of the track on train-induced ground vibration. Journal frequency on the resilience of railway ballast. Géotechnique, 69(9),
of Sound and Vibration, 272, 909-936 833-840.
Sheng D. Fredlund D.G. and Gens A. 2008. A new modelling approach Sun. Xi. Li. Jie. and Zhou A. 2017. Evaluation and Comparison of
for unsaturated soils using independent stress variables. Canadian Methods for Calculating Thornthwaite Moisture Index. Australian
Geotechnical Journal, 45(4): 511–534. doi:10. 1139/T07-112. Geomechanics Journal. 52(2), 61-75
Shi S. Gao L. Cai X. Yin H. and Wang X. 2020. Effect of tamping Svalova A. Prangle D. Wilkinson D.J. Helm P. Rouainia M. and
operation on mechanical qualities of ballast bed based on DEM- Glendinning S. (2021). Emulating computer experiments of
MBD coupling method. Computers and Geotechnics, 124, 103574 infrastructure slope stability in the London-Bristol rail corridor using
Shih J.-Y. Thompson D. and Zervos A. 2016. The effect of boundary Gaussian processes and Bayesian inference, Data-Centric
conditions, model size and damping models in the finite element Engineering, 2:e12.
modelling of a moving load on a track/ground system. Soil Taber, S. The mechanics of frost heaving. The Journal of Geology, 1930,
Dynamics and Earthquake Engineering, 89, 12-27. 38(4), pp. 303-317.
Signes C.H. Fernández P.M. Perallón E.M. and Franco R.I. 2015. Tang Y. Xiao S. and Yang Q. 2019a. Numerical study of dynamic stress
Characterisation of an unbound granular mixture with waste tyre developed in the high speed rail foundation under train loads. Soil
rubber for subballast layers. Materials and Structures, 48(12): 3847- Dynamics and Earthquake Engineering, 123, 36-47.
3861. Tang Y. Yang Q. Ren X. and Xiao S. 2019b. Dynamic response of soft
Simpson B. and Tatsuoka F. 2008. Geotechnics: the next 60 years. The soils in high-speed rail foundation: in situ measurements and time
Essence of Geotechnical Engineering: 60 Years of Geotechnique. domain finite element method model. Canadian Geotechnical
Thomas Telford, London, UK, 49–60 Journal, 56, 1832-1848.
Singhal S. 2010. Expansive Soil Behavior: Property Measurement Tasalloti S.M.A. Indraratna B. Rujikiatkamjorn C. Heitor A. and Chiaro
Techniques & Heave Prediction Methods. Ph.D. Dissertation, G. 2015a. A laboratory study on the shear behavior of mixtures of
Arizona State University, Tempe, AZ, USA coal wash and steel furnace slag as potential structural fill,
Smethurst J.A. Clarke D. and Powrie W. 2012. Factors controlling the Geotechnical Testing Journal, 38(4), 1-12.
seasonal variation in soil water content and pore water pressures Tasalloti S.M.A. Indraratna B. Chiaro G. and Heitor A. 2015b, Field
within a lightly vegetated clay slope. Géotechnique, 62(5), 429-446. investigation on compaction and strength performance of two coal
https://doi.org/10.1680/geot.10.p.097 wash-BOS slag mixtures, ASCE Geotechnical Special Publication,
Smethurst J.A. Briggs K.M. Powrie W. Ridley A. and Butcher D.J.E. 256, 2359-2368 (Proc. of the 2015 International Foundations
2015. Mechanical and hydrological impacts of tree removal on a clay Congress. & Equipment Expo,, 17-21 March 2015, San Antonio,
fill railway embankment. Géotechnique, 65(11), 869-882. Texas, USA).
https://doi.org/10.1680/jgeot.14.P.010 Tatsuoka F. Hashimoto T. and Tateyama K. 2021, Soil stiffness as a
Smethurst J. Bicocchi N. Powrie W. and O'Brien. A. S. 2020. Behaviour function of dry density and the degree of saturation for compaction
of discrete piles used to stabilise a tree-covered railway control, Soils and Foundations, 61, 989-1002
embankment. Géotechnique, 70(9), 774-790. [18-P-150R1]. Taylor G.S. and Luthin J.N. 1978. A model for coupled heat and moisture
https://doi.org/10.1680/jgeot.18.P.150 transfer during soil freezing. Can. Geotech. J. 15, 548–555.
Smith A.W. 1973. Method for determining the potential vertical rise, doi:10.1139/t78-058
PVR. Proc. Workshop Expansive Clays & Shales in Highway Design Teng T.C.P. and Clisby M.B. 1975. Experimental work for active clays
& Construction. Univ. of Wyoming, Denver, CO, 245-249. in Mississippi. Transport. Eng. J. ASCE 101 (TEI): 77-95.
Sol-Sánchez M. Moreno-Navarro F. and Rubio-Gámez M.C. 2015. The Teng T.C.P. Mattox R.M. and Clisby M.B. 1972. A study of active clays
use of elastic elements in railway tracks: A state of the art review. as related to highway design. Research & Development Division,
Construction and Building Materials, 75, 293-305. Mississippi State Highway Dept., Engineering & Industrial
Sridharan A. Sreepada Rao A. and SivapuUaiah, P.V. 1986. Swelling Research Station, Mississippi State University, MSHD-RD-72-045,
pressure of clays. Geotechnical Testing Journal, GTJODJ, 9(1), 24- 134.
33. Teng T.C.P. Mattox R.M. and Clisby M.B. 1973. Mississippi's
Stallings Young E.G. Zapata C.E. and van Paassen L. 2020. Unsaturated experimental work on active clays. Proc. Workshop on Expansive
Fluid Flow through Granular Soils Treated with Microbial Induced Clays & Shales in Highway design & Construction. Univ. of
Desaturation and Precipitation. In Proceedings of the 4th European Wyoming, Laramie. 1-17
Conference on Unsaturated Soils (E-UNSAT 2020) Online
90
Tennakoon N. Indraratna B. Rujikiatkamjorn C. Nimbalkar S. and with different frequencies. Soil Dynamics and Earthquake
Neville T. 2012. The role of ballast-fouling characteristics on the Engineering, 118: 75-85.
drainage capacity of rail substructure. Geotechnical Testing Journal, Wang L. Van Paassen L. Gao Y. He J. Gao Y. and Kim, D. 2020.
35(4), 1-11. Laboratory tests on mitigation of soil liquefaction using microbial
The Communist Party of China Central Committee and the State Council induced desaturation and precipitation. Geotechnical Testing
2019. Program of Building National Strength in Transportation. Journal, 44(2).
http://www.gov.cn/xinwen/2019-09/19/content_5431432.htm (in Wang L. Gao Y. He J. Gao Y. and van Paassen L.A. 2021. Effect of
Chinese) Biogenic Gas Formation through Microbial Induced Desaturation
Thomas H.R. Cleall P. Li Y.C. Harris C. and Kern-Luetschg M. 2009. and Precipitation on the Static Response of Sands with Varied
Modelling of cryogenic processes in permafrost and seasonally Relative Density. Journal of Geotechnical and Geoenvironmental
frozen soils. Geotechnique 59, 173–184. doi:10.1680/geot. Engineering, 147(8), 04021071.
2009.59.3.173 Wang W.L. and Yen B.C. 1974. Soil arching in slopes. Journal of the
Thornthwaite C. 1948. An Approach Toward A Rational Classification geotechnical Engineering division ASCE, 100(GT1), 61–78.
of Climate, Geographical Review, 38(1), 55-94. Ward Wilson, G. 1990. Soil Evaporative Fluxes for Geotechnical
Thornthwaite C. and Mather J. 1955. The Water Balance, Publications in Engineering Problems. PhD Thesis, University of Sasketchewan,
Climatology, 8(1) Canada.
Torres-Hernandez G. 2011. Estimating the Soil–Water Characteristic Weston D.J. 1980. Expansive roadbed, treatment for Southern Africa.
Curve using Grain Size Analysis and Plasticity Index. Master’s Proceeding of the 4th International Conference on Expansive Soils,
Thesis. Arizona State University, Tempe, AZ. 1, 339-360.
Track Stiffness Working Group (2016). A Guide to Track Stiffness, Witczak M.W. Zapata C.E. and Houston W.N. 2006. Models
Edited by W. Powrie & L Le Pen on behalf of the cross industry Incorporated into the Current Enhanced Integrated Climatic Model:
track stiffness working group (TSWG). Southampton, UK; NCHRP 9-23 Project Findings and Additional Changes after
University of Southampton. ISBN: 9780854329946. Version 0.7. Final Report. Project NCHRP 1-40D.
Trani L.D.O. and Indraratna B. 2012. A New Method of Filtration Design Woldringh R. and New B. 1999. Embankment design for high speed
on Rail Track Subballast. J. Geotech. Geoenviron. Eng, 136(11), trains on soft soils. Twelfth European Conference on Soil Mechanics
1519-1528. and Geotechnical Engineering, Geotechnical Engineering for
Tutumluer E. Huang H. and Bian X. 2012. Geogrid-Aggregate Interlock Transport Infrastructure. Amsterdam. Rotterdam: Balkema
Mechanism Investigated through Aggregate Imaging-Based Woodman N. Smethurst J. Roose T. Powrie W. Meijer G. Knappett J.
Discrete Element Modeling Approach. International Journal of and Dias T. 2020. Mathematical and computational modelling of
Geomechanics, 12(4), 391-398. vegetated soil incorporating hydraulically-driven finite strain
Tutumluer E. Qian Y. Hashash YMA. Ghaboussi J. and Davis DD. 2013. deformation. Computers and Geotechnics, 127, [103754].
Discrete element modelling of ballasted track deformation https://doi.org/10.1016/j.compgeo.2020.103754
behaviour. International Journal of Rail Transportation, 1(1-2), 57- Wong H.Y and Yong R.M. 1973. A study of swelling & swelling force
73. during unsaturated flow in expansive soils. Proc. 3rd Int. Conf.
University of Southampton (2013). Confidential report on Sand Expansive Soils, Haifa, Israel, 1, 143-151.
mitigation measures efficiency track resilience test carried out for Wu, T. Cai Y. Guo L. Ling D. and Wang J. 2017. Influence of shear stress
Etihad Rail for the Shah-Habshan-Ruwais Railway Project level on cyclic deformation behaviour of intact Wenzhou soft clay
USA Army Corps of Engineers. 1948. Field Moisture Content under traffic loading. Engineering Geology, 228, 61-70.
Investigation. T.M. 3-401, Interim Report No. 1, Waterways Xiao J. Juang C.H. Wei K. and Xu S. 2014. Effects of principal stress
Experiment Station rotation on the cumulative deformation of normally consolidated soft
USA Army Corps of Engineers. 1955. Field Moisture Content clay under subway traffic loading. Journal of Geotechnical and
Investigation. T. M. 3-401, Report No. 2, October 1945 -November Geoenvironmental Engineering, 140(4), 04013046.
1952 Phase, Waterways Experiment Station Xie J Huang J Zeng C Jiang S-H and Podlich N. 2020. Systematic
Van Der Merwe D.H. 1964. The prediction of heave from the plasticity Literature Review on Data-Driven Models for Predictive
index & percentage clay fraction of soils. Civil Engineers in South Maintenance of Railway Track: Implications in Geotechnical
Africa, 6, 337-42. Engineering, Geosciences, 10, 425; doi:10.3390/geosciences
van Paassen L. Daza C. Staal M. Sorokin D. van der Zon W. and van 10110425
Loosdrecht M. 2010. Potential Soil Reinforcement by Biological Xu S. Liang R. and Edil T. (2015) Innovative Soft Soil Improvement
Denitrification. Ecological Engineering, 36(2), 168-175. Method through Intelligent Use of Vacuum De-Watering and
Vanapalli S. and Lu L. 2012. A State-of-the Art Review of 1-D Heave Dynamic Compaction Techniques, Geotechnical Engineering 46(4),
Prediction Methods for Expansive Soils. International Journal of 57-67.
Geotechnical Engineering, 6(1), 15-41. Yang L. Powrie W. Priest J. 2009. Dynamic stress analysis of a ballasted
Vann J. Houston S. Houston W. Singhar S. Cuzme A. and Olaiz A. 2018. railway track bed during train passage. Journal of Geotechnical and
A soil suction surrogate & its use in the suction-oedometer method Geoenvironmental Engineering 135, 680-689.
for computation of volume change of expansive soils. Proc. of the Yang Y.B. Liu S.J. Li Q.M. and Ge P.B. 2019. Stress waves in half-space
7th International Conference on Unsaturated Soil. Hong Kong. due to moving train loads by 2.5D finite/infinite element approach.
1205-1210 Soil Dynamics and Earthquake Engineering, 125, 105714.
Vann J.D. 2019. A Soil Suction-Oedometer Method and Design Soil Yasuhara H. Neupane D. Hayashi K. and Okamura M. 2012.
Suction Profile Recommendations for Estimation of Volume Change Experiments and Predictions of Physical Properties of Sand
of Expansive Soils, Doctoral Dissertation, Arizona State University, Cemented by Enzymatically-induced Carbonate Precipitation. Soils
Tempe, AZ. and Foundations, 52(3), 539-549.
Vann J.D. and Houston S. 2021. Field Suction Profiles for Expansive Youwai S. and Bergado D.T. (2003). Strength and deformation
Soil, J. of Geotechnical and Geoenvironmental Engineering, 147(9), characteristics of shredded rubber tire-sand mixtures. Canadian
doi:10.1061/(ASCE)GT.1943-5606.0002570. Geotechnical Journal, 40(2), 254-264.
Varandas J.N. Paixão A. Fortunato E. and Hölscher P. 2016. A numerical Yu X. and Kwon E. 2012. Carbon nanotube based self-sensing concrete
study on the stress changes in the ballast due to train passages. for pavement structural health monitoring, Research report
Procedia engineering, 143, 1169-1176. University of Minnesota, Duluth, Contract Number: US DOT:
Varandas J. Paixão A. Fortunato E. Coelho B.Z. and Hölscher, P. 2020. DTFH161-10-C-00911
Long-term deformation of railway tracks considering train-track Zapata C.E. Andrei D. Witczak M.W. and Houston W.N. 2007.
interaction and non-linear resilient behaviour of aggregates–a 3D Incorporation of Environmental Effects in Pavement Design. Int. J.
FEM implementation. Computers and Geotechnics, 126, 103712. of Road Materials and Pavement Design, 8(4). 667-693
Walker R.T. Indraratna B. 2018. Moving loads on a viscoelastic Zapata C.E. Perera Y.Y. and Houston W.N. 2009. Matric Suction
foundation with special reference to railway transition zones. Prediction Model Used in the New AASHTO Mechanistic-
International Journal of Geomechanics, 18, 04018145. Empirical Pavement Design Guide. Transportation Research
Wang J. Zhou Z. Hu X. Guo L. and Cai Y. 2019. Effects of principal Record: J. of the Transportation Research Board, No. 2101, Geology
stress rotation and cyclic confining pressure on behavior of soft clay and Properties of Earth Materials, 53-62.
91
Zapata, C.E. 2018. Empirical approach for the use of unsaturated soil
mechanics in pavement design. Geotechnical Special Publication,
2017-November (GSP 300), 149-173. https://doi.org
/10.1061/9780784481677.008
Zeng K. Qiu T. Bian X. Xiao M. and Huang H. 2019. Identification of
ballast condition using SmartRock and pattern recognition,
Construction and Building Materials, 221, 50-59.
Zhao H.D. Ling J.M. Fu P.C. 2013. A review of harvesting green energy
from road. Advanced Materials Research, 723, 559-566, DOI:
10.4028/www.scientific.net/AMR.723.559
Zhao HY. Indraratna B. and Ngo N.T. 2021. Numerical simulation of the
effect of moving loads on saturated subgrade soil, Computers and
Geotechnics https://doi.org/10.1016/j.compgeo.2020.103930.
Zhang X. and Lytton R. 2009. Modified State-Surface Approach to the
Study of Unsaturated Soil Behaviour. Part I: Basic Concept.
Canadian Geotechnical Journal, 46, 1-17.
Zhang X. Zhao C. and Zhai W. 2017. Dynamic behavior analysis of high-
speed railway ballast under moving vehicle loads using discrete
element method. International Journal of Geomechanics, 17,
04016157.
Zhang Y. Gu C. Wang J. and Cai Y. 2021. Three-dimensional cyclic
behavior of saturated clays: comparison between undrained and
partly drained conditions. Canadian Geotechnical Journal, ahead of
print
Zhang Y. and Michalowski R.L. 2015. Thermal-hydro-mechanical
analysis of frost heave and thaw settlement. Journal of Geotechnical
and Geoenvironmental Engineering, 141(7), 04015027.
Zhai W. Wang K. and Lin J. 2004. Modelling and experiment of railway
ballast vibrations, Journal of Sound and Vibration, 270(4-5), 673-
683.
Zheng D. Huang J. Li D-Q. Kelly R. and Sloan S.W. 2018, Embankment
prediction using testing data and monitored behaviour: A Bayesian
updating approach, Computers and Geotechnics, 93, 150–162,
Zhou M.M. and Meschke G. 2013. A three‐phase thermo‐hydro‐
mechanical finite element model for freezing soils. International
Journal for Numerical and Analytical Methods in Geomechanics,
37(18), 3173-3193
92
State of the Art 2: Energy Geotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Energy geo-engineering
Geo-ingenierie pour l’energie
J. Carlos Santamarina
Energy Resources and Petroleum Engineering, KAUST, Saudi Arabia, [email protected]
Adnan Aftab
WASM: Minerals, Energy, and Chemical Engineering, Curtin University, Australia
D. Nicolas Espinoza
Petroleum and Geosystems Engineering, U. Texas at Austin, USA
Maurice Dusseault
Earth and Environmental Sciences, U. Waterloo, Canada
Antonio Gens
Civil and Environmental Engineering, U. Politecnica de Catalunya - CIMNE, Spain
Hussein Hoteit
Energy Resources and Petroleum Engineering, KAUST, Saudi Arabia
Seunghee Kim
Civil and Environmental Engineering, U. Nebraska-Lincoln, USA
Joo Yong Lee
Climate Change Response Division, Korea Institute of Geoscience and Mineral Resources, Republic of Korea
Liang Lei
School of Engineering, Westlake U., China
Guillermo Narsilio
Department Infrastructure Eng., U. Melbourne, Australia
Jean-Michel Pereira
Navier, Ecole des Ponts, Univ Gustave Eiffel, CNRS, France
Marcelo Sanchez
Civil and Environmental Engineering, Texas A&M, USA
Kenichi Soga
Civil and Environmental Engineering, UC Berkeley, USA
María Victoria Villar
CIEMAT, Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas, Spain
Marie Violay
Lab Experimental Rock Mechanics, EPFL, Switzerland
ABSTRACT: The geo-science and engineering fields have critical roles to play towards a sustainable energy future. This state-of-
the-art review focusses on five areas where the geotechnical community has been involved the most: the oil and gas sector with
emphasis on methane hydrates, carbon geological storage, geothermal, energy geo-storage, and nuclear waste storage. Extensive
tables in the appendix identify potential geotechnical contributions to all energy resources (mining is critically needed to supply raw
materials for the energy sector, yet is not addressed in this review). Energy-related applications involve a wide range of time and
length scales, coupled thermo-hydro-chemo-mechanical processes, multi-phase fluids, high pressure-temperature-stress conditions,
fines migration, reactive fluid transport and phase transformations, and multi-physics repetitive loads. Analysis and design require
careful experimentation (including field-scale studies) and advanced numerical simulations. Educational programs must address the
knowledge needs in energy geo-engineering.
Les domaines des géosciences et de l’ingénierie ont un rôle essentiel à jouer pour un avenir énergétique durable. Cette
revue de l’état de l’art se concentre sur cinq domaines dans lesquels la communauté géotechnique a été le plus impliquée : le secteur
pétrolier avec un focus sur les hydrates de méthane, le stockage géologique de dioxyde de carbone, la géothermie, le géostockage
d’énergie et le stockage des déchets radioactifs. Des tableaux détaillés donnés en annexe identifient les potentielles contributions
95
95
géotechniques à toutes les ressources énergétiques (l’exploitation minière est indispensable pour fournir des matières premières au
secteur de l’énergie, mais n’est pas abordée dans cette revue). Les applications de la géo-ingénierie pour l’énergie impliquent une
large gamme d’échelles de temps et de longueur, des processus couplés thermo-hydro-chimio-mécaniques, des fluides multiphases,
des conditions de pression-température-contrainte élevées, la migration de fines, le transport de fluide réactif et des changements de
phase, ainsi que des sollicitations multiphysiques répétitives. L’analyse et la conception nécessitent des expérimentations minutieuses
(y compris des essais de terrain) et des simulations numériques avancées. Les programmes d'enseignement doivent répondre aux
besoins de connaissances en matière de géo-ingénierie pour l’énergie.
KEYWORDS: geothermal, hydrates, geological storage, nuclear waste, energy storage, carbon geological storage
1. INTRODUCTION: ENERGY GEOENGINEERING In all cases, designs must take into consideration site
conditions, construction, maintenance and retrofitting needs, and
The complex energy situation is intimately linked to quality of anticipate decommissioning difficulties (this applies to all
life, water, food security and the environment. The current global developments, from hydroelectric dams, to oil and gas wells and
power consumption is about 18.5 TW, and a staggering increase nuclear power plants). Design optimization must consider the
of 2%-per-year in capacity will be needed to support current embodied energy associated with the project throughout its life
trends including growth in the developing world. Eighty percent and decommissioning.
80% of all primary energy comes from fossil fuels and the Tables in the appendix provide details and engineering needs
ensuing release of CO2 into the atmosphere is the main associated with: (A) energy resources, (B) storage, and (C)
contributor to climate change. In particular, oil satisfies 33% of conservation and efficiency. Fossil fuels have fueled the
the total energy needs and most of it is used for transportation industrial revolution, satisfy about 85% of all energy needs
(today’s transportation is the most inefficient component in the today, and will fuel the transition out of fossil fuels, therefore
energy system). In addition, consider the geographic mismatch they will remain the main energy source until 2050 and beyond
between energy resources and consumers, dramatic inequality in (70% by 2050). Fossil fuels are summarized in Tables A1: oil,
power consumption between “energy-poor” countries (<100 A2: gas and A3: coal. The oil and gas sector is well advanced,
W/person) and “energy-affluent” countries (>5,000 W/person), yet, many classical problems remain (Tables A1 & A2). The case
the lack of nuclear waste disposal facilities for commercial of methane hydrate is of particular interest given the large
applications, and rapidly evolving mining needs to satisfy the amount of carbon trapped in hydrate bearing sediments around
energy sector. the world and the challenges associated with gas production from
Achieving net zero targets will be costly and energy intensive this resource (Table A2). While there is a general aversion to
(EU 2021, example in Meys et al., 2021). Furthermore, given coal, the vast world reserves are likely to ensure that coal remains
today’s energy mix, the transition out of fossil fuels will be a key component of the energy mix (Table A3), yet, coal burning
fueled by fossil fuels. Consequently, efforts to decarbonize the calls for pressing improvement in recovery with reduced
economy have had limited impact so far, and plans to scale up environmental impact and efficient C-capture and storage.
carbon neutral energy systems are progressing slowly in Solar and wind sources require limited geotechnical
comparison to increasing demands. engineering apart from foundation systems and transient energy
On the other hand, there are positive indicators, such as the storage in the subsurface (Tables A4 & A5). The developing
rapid growth in distributed solar energy that does not require a world uses biofuels extensively and this is an ‘almost’ carbon
power grid to address the need of underdeveloped nations that neutral energy resource (Table A6). The estimated geothermal
are often located in regions with high solar radiation. Overall, a potential is three orders of magnitude higher than the installed
sustainable energy solution must promote increased quality of capacity (Table A7). Geotechnology has played a central role in
life in the developing world, reduced power consumption in the design and construction of hydroelectric projects throughout
developed nations, and minimize the environmental impact of the 20th century, yet, the global fluvial hydroelectric capacity is
energy consumption, in particular CO2 emissions. almost saturated (Table A8 - it includes tidal energy). Nuclear
The geo-sciences and engineering fields have critical roles to accidents and the challenges associated with nuclear waste
play towards a sustainable energy transition, in line with IPCC storage facilities have hindered the development of nuclear
recommendations to address climate change. Our community energy (Table A9); still, the extreme energy density in nuclear
understands geological processes, comprehends the underlying fuel allows for exceptional compact designs and alternative
physical foundations and theoretical framework, and has the nuclear power plant designs are in progress (Gen IV).
analytical, numerical and experimental expertise to advance Short-term energy storage is needed as part of intermittent
robust engineering solutions. Figure 1 presents a summary of the renewable sources and base-load nuclear power generation
main energy resources along with high-level geotechnical (Table B1). On the other hand, we need to accommodate waste
themes. from energy production/consumption, such as fly ash (coal-
Fossil Fuels Renewables
burning), CO2 (all fossil fuels), produced water (9-to-15 barrels
Critical Minerals for
Energy Systems
of water are produced for every 1 barrel of oil), and nuclear waste
Hydroelectric
Geo-Thermal
Nuclear
(there is no operating disposal site for commercial waste
Gas *
Solar
Wind
Coal
anywhere in the world). Table B2 reviews CO2 geo-storage.
Oil
Onshore
A
Resources -
Assessment
Onshore & offshore Onshore & offshore
Mines &
seawater
& Finally, energy geo-engineering must also address
offshore
Hydroelectric
environmental remediation, the design of new facilities in view
Includes strategic fuel Molten
B Energy Storage reserves
Heat
Compressed air energy storage
salt, etc.
-
of life-cycle needs and decommissioning, as well as the
CO2 Low & development of geotechnical construction methods that can
C Waste Produced water Obsolete hardware disposal high Tailings
Fly ash waste reduce the embodied energy in infrastructure projects (See Table
D
Depleted sites
Abandoned wells & mines - Dams - -
Mines &
Sites
Mines C1 for significant examples). Besides the recovery of mineral
Legacy
resources such as coal, oil sands and uranium, mining is critically
Conservation & Efficiency needed to provide the ingredients for batteries (e.g., Li, Co),
magnets in wind turbines (e.g., rare earth minerals), and the
Figure 1. Energy Resources and Energy GeoEngineering power grid (e.g., Cu and Al).
96
96
The rest of this state-of-the-art review takes a closer look at Often, three fluids coexist and their interfacial tensions
five areas where our geotechnical community is involved the combine to define the spreading of the intermediate fluid
most: the oil and gas sector with emphasis on methane hydrates, between the other two (Figure 2). Multiphase fluids interact with
carbon geological storage, geothermal, energy geo-storage, and the mineral walls along the interconnected porosity in soils, rocks
nuclear waste storage. and fractures. The contact angle between the fluid interface and
the mineral surface plays a central role in capillary phenomena.
In particular, differences between the advancing and receding
2. OILS AND GAS - METHANE HYDRATES (Appendix contact angles combine with fluid viscosities, pore topology and
Tables A1 and A2) spatial variability to define storativity, saturation patchiness,
Typical downhole geotechnical challenges in the oil and gas instabilities (Haines jumps, viscous fingering), displacement and
sector include fines migration and clogging (akin to filter recovery efficiencies, invasion patterns, bypassed oil and ganglia
criteria), sand production, stimulation, enhanced oil recovery stability (Wang et al. 2020; Liu and Santamarina, 2022).
(akin to unsaturated soils), and various needs related to the
CO2liq oil CH4gas
production of heavy oil and tar sands. 30 34-46 64
There are more than 4,500,000 wells drilled around the world (a) [mN/m]
30 40 50 60 70
for oil and gas production; about half of them remain in operation
and the rest are either abandoned or decommissioned. Some 31 1
Ice CH4hyd
72
water vapor
estimates indicate that 60% of all wells leak. Clearly, drilling is CO2hyd CO2gas
a critical component and advances to tackle some of the
following drilling challenges will have a major impact across the (b) (c)
energy sector: high-pressure and temperature deep wells;
Tao
efficient, low cost and low environmental impact; safe, lost Air
oil
Taw
circulation control; reliable completion, cake formation and Tow
Water
cementing; THCM coupling and shale instability.
The scientific foundations for soil behavior developed in the
classical geotechnical field are readily extended to the oil and gas CS Taw Tow Tao
sector, but properly adjusted for prevailing field conditions, for
example: lightly cemented sandstones; clay geochemistry and Figure 2. (a) Interfacial tension. (b) Spreading. (c) Contact angle hyste-
shales instability during drilling and seal layer performance; resis
stable suspensions used for drilling muds; particle crushing and
permeability changes in the design of proppants used for 2.2 Fines Migration, Clogging and Sand Production
reservoir stimulation during hydraulic fracturing; the prevalent Fine particles can become detached and displaced during fluid
role of fines in resource recovery; filters in fines migration, lost flow. Fines migration, clogging and sand production are common
circulation materials for drilling, and sand production during oil in oil and gas recovery. In particular, sanding is the internal
recovery. erosion and transport of the sand particles as reservoir fluids flow
Similarly, the extensive study of rocks during the last century through the formation into the wellbore, and it is one of the major
is critically important to the energy sector, with renewed issues for maintaining safe and sustainable gas production from
emphasis on carbonates (prevalent oil reservoirs) as well as gas hydrate deposits by depressurization (Collett et al. 1999,
shales, evaporites (oil and gas traps and seal layers for CO2 Yamamoto et al. 2014, Yamamoto et al., 2017). Recent
geological storage). Field conditions in these applications experimental studies on fine migration identified the following
demand characterization and modeling at high pressure, high underlying mechanisms (Valdes and Santamarina 2007; Liu et
effective stress (brittle-ductile transition) and high temperature al., 2019; Jung et al.2019; Han et al. 2020; Jang et al. 2020;
(thermo-mechanical coupling), and the analysis of long-term Murphy et al. 2020; Hafez et al., 2021):
time-dependent response (creep, diagenesis, and reactive Migrating fines concentrate near the well-bore during gas
transport). While the rock matrix is responsible for storativity, production and can form a clogging annulus
fractures control conduction and define the reservoir Clogging depends on the pore size distribution of the host
“geoplumbing”. For example, a 1 mm aperture fracture sediment, the fines characteristics and the flow regime
transports the same flow rate as an 80 km wide intact carbonate Heterogeneity and layering can cause a slab effect that
rock with a normal 1 mD permeability. Furthermore, fractures facilitates the development of high permeability pathways
determine all mechanical properties, and render the rock mass below
behavior stress-dependent.
Multiphase flow conditions tend to facilitate clogging
whereby fines cluster at the gas-liquid interface and become
2.1 Multiphase Fluids
stabilized by capillary forces
Multiphase fluid conditions prevail in oil and gas recovery; Changes in pore-fluid chemistry can trigger fines detachment
processes resemble those explored in unsaturated soils, but the and/or aggregation.
emphasis rests on fluid flow rather than changes in effective
stress and mechanical coupling. The interaction between 2.3 Hydrates: Hydrate Morphology and Physical Properties
different fluids is characterized by their interfacial tension fl and
Hydrate growth in sediments reflects the competition between
mutual solubilities (e.g., very limited for CH4 in water, but high
capillarity and skeletal forces (Clennell et al., 1999; Dai et al.,
for CO2 in oil). When one of the interacting fluids is a gas such
2012; Lei and Santamarina, 2018; Terzariol et al., 2020). Hydrate
as CO2 or CH4, both interfacial tension and solubility are pressure
growth is grain-displacive in fine-grained sediments (and low
dependent. Fluid properties can change significantly following
effective stress), resulting in segregated lenses, veins, and chunks
dissolution, e.g., water acidifies while oil swells and changes
(Figure 3). On the other hand, hydrate growth is pore-invasive in
viscosity after CO2 dissolution. Surfactants and some
coarse-grained sediments at high effective stress where the pore
nanoparticles preferentially accumulate at the interface and
habit becomes cementing, grain-coating, pore filling and patchy
profoundly change the interaction between neighboring fluids.
saturation (Kleinberg et al., 2003). Micro-CT images show that
97
97
cementing and grain-coating pore habits prevail in excess-gas 2014, Ryu et al. 2013, Zhao et al. 2018; Kim et al. 2021; White
conditions while pore-filling is most common in excess-water et al. 2020; Zhou et al. 2020a&b).
(Lei et al., 2019a; Le et al., 2020). “Load-bearing hydrate”
evolves as hydrate creeps and the effective stress field Sand Production
redistributes within the sediment (Lei et al., 2020). • Production potential
Cambridge MH Code – FLAC2D FD code (2005-)
• Volumetric strain
• Effect stress relaxation
(a) (b) • Permeability
S
W
Figure 4. Hydrate bearing sediments - Coupled processes
There are several possible production strategies, including:
(c) (d)
depressurization, heating, inhibitors and CO2-CH4 replacement
(Jung et al., 2010; Moridis et al., 2011). Production by
depressurization is most frequently explored, however,
S production is inherently limited by the extent of the
H
H depressurized field which decays rapidly away from the well
S
(Terzariol et al., 2017 - see multi well strategy in Terzariol and
Santamarina, 2020). Geomechanical implications associated
2 cm
5mm with depressurization include soil structure collapse associated
with the loss of hydrate cementation at particle contacts, sand
S=Sediments/Sand H=Hydrate W=Water G=Gas production, sediment compaction due to the increase in effective
stress, and increase in shear stress along the well. Field tests have
Figure 3. Hydrate pore habits – Micro CT images. Pore-invasive hydrate highlighted issues related to wellbore instability caused by
in coarse sediments,: (a) lab-formed hydrate under excess-gas condition
(Lei et al., 2019a); (b) Natural hydrate-bearing silt pressure cores re-
formation compaction and shear loading of the well which can
trieved from Gulf of Mexico (Lei et al., 2022). Grain-displacive hydrate cause buckling collapse within the production layer as well as
in fine-grained sediments: (c) lab-formed CO2 hydrate in a gas-driven axial tensile failure above (Moridis et al. 2013; Shin and
fracture across a water-saturated kaolinite specimen (Lei and Santama- Santamarina 2016; Kim et al. 2018; Yoon et al. 2021; Sasaki et
rina, 2018); (d) Natural hydrate lenses in a pressure core retrieved from al. 2021). Finally, gas recovery by depressurization from fine-
NGHP 01 project (National Energy Technology Laboratory - See also grained hydrate bearing sediments faces difficulties related with
Yun et al., 2011). high entry pressure and compressibility; in fact, mud production
is needed to initiate gas production.
Reproducing methane-limited hydrate formation in nature is
particularly challenging within typical laboratory time scales,
3. CARBON GEO-STORAGE (Appendix Table B2)
i.e., weeks to months. Because of high advection, laboratory
studies of hydrate formation are much simpler in coarse-grained Anthropogenic activity emits 35 109 ton/yr of carbon dioxide
sediments than in fine-grained deposits (Spangenberg et al., CO2 into the relatively thin atmosphere layer that controls the
2005; Lei and Santamarina, 2018). Hydrate can form by forcing Earth’s thermal balance. Clearly, the geological storage of CO2
free gas invasion into water-saturated sediment within the emerges as a critical component of a sustainable energy system
stability field; in this case, a hydrate film/crust forms around the which will remain strongly fossil fuel dependent for the next
invading gas body (Lei et al., 2019a&b). decades as we transition out of fossil fuels. Storage systems do
At the layer-scale, hydrate distribution reflects inherent not need to be permanent, but must delay the release of CO2 into
sediment layering, and both fabric and stress anisotropy; these the atmosphere for hundreds of years until natural carbon sinks
layer-scale effects can be more important than pore-scale hydrate can adequately trap emitted and leaked CO2.
habit on macro-scale properties. Together, pore-scale and layer- While there are various attempts at capturing carbon from air,
scale hydrate formation patterns affect all the physical properties most significant efforts focus on carbon capture at focal emitting
of hydrate-bearing sediments, e.g., mechanical, acoustic, facilities e.g. fossil fuel-based power plants and cement factories.
hydraulic and thermal. (Extensive data compilation in Waite et Currently, there are ~60 large-scale C-storage projects in
al., 2009. For more recent insights, see: Dai et al., 2012; Lei and operation or under development (GCCSI, 2020; IEA, 2020);
Santamarina, 2019; Atig et al., 2020; Pan et al., 2021). implemented and potential geological storage methods include
(Figure 5 - DOE, 2017; IEA, 2020):
2.4 Gas Production from Hydrate-Bearing Sediments
Storage in depleted oil and gas reservoirs (possibly as part of
Gas hydrate accumulates mostly in unconsolidated deposits. enhanced oil recovery) or in saline aquifers (Birkholzer et al.,
During hydrate dissociation, sediments experience phase 2009). In these cases, supercritical CO2 would remain trapped
transition (the solid hydrate phase converts to gas and water and in pores by capillary action and/or dissolved in the pore fluid.
experiences high volume expansion), thermal changes and heat Storage in un-minable coal seams (trapping by CO2 adsorbed
flow (dissociation is endothermic), multiphase fluid (water and on mineral surfaces).
gas), changes in effective stress and associated instabilities. Fully Mineralized storage e.g., mafic igneous rocks (Iceland
coupled THM simulators with history matching help to gain CARBFIX project; White et al., 2020).
insight into the evolution of underlying processes within the Storage in the form of hydrates, either at new locations where
formation; reliable reservoir properties are as crucial as robust P&T conditions support CO2 hydrate formation (House et al.,
numerical simulators (Figure 4 - Collett et al. 2011, Collet et al. 2006), or by CO2 CH4 substitution in hydrate bearing
98
98
sediments (Komai et al., 2000; Espinoza and Santamarina 4. GEOTHERMAL ENERGY (Appendix Table A7)
2011; Zhang et al., 2017).
The critical point for CO2 is at T=31 C and P=7.4 MPa which Average values of geothermal gradient iT 25 C/km, rock
corresponds to onshore reservoirs deeper than ~800 m. thermal conductivity kT 3 W m-1C-1 and heat capacity 2 kJ kg-
1 C-1 allow us to obtain the following first-order estimates: (1)
~98% of the Earth is hotter than 1000 C, (2) the heat flux from
the Earth’s interior to the ground surface is ~40×1012 W (see
Pollack et al., 1993), and (3) a 3x3x0.1 km3 geothermal field can
store enough energy to deliver ~500 MW of average power over
30 years assuming a temperature drawdown of 100°C (see Jaeger
et al., 2009; Blackwell et al., 2011). Most geothermal fields are
located around the Ring of Fire (Indonesia, Japan, Philippines,
Russia, Tibet, New Zealand, western USA, and some Latin
American countries), Turkey, western Europe, and some African
Figure 5. Carbon geological storage countries. Geothermal sources can be used for power generation
(typically deep and hot sources) or space heating (typically near-
CO2 injection in geological reservoirs disturbs the surface).
thermodynamic equilibrium and affects the short- and long-term
reservoir and caprock responses. Complex hydro-chemo-thermo- 4.1 Subsurface Thermal Energy Storage Systems (Shallow)
mechanically coupled phenomena include:
CO2 injection can increase the total stresses in both vertical The thermal energy demand for buildings and industries is a
and horizontal directions away from the injection well significant portion of the global energy demand. At present, most
(Altmann et al., 2010; Kim and Hosseini, 2016), and alter the of the heating energy demands are met by natural gas and oil;
state of stresses in layers above and below the storage reservoir energy efficient construction and electrification have helped
(poroelastic effects - Chang and Segall, 2016&2017). reduce CO2 emissions for space heating. On the other hand, space
Rapid pressure decay radially away from the well during cooling demand is increasing and related CO2 emissions are
injection causes cooling (and even water freezing) and stress expected to increase between 58 and 82% (Mastrucci et al. 2021).
changes (Vilarrasa et al., 2014; Kim and Hosseini, 2015, As the temperature variation in the subsurface is small over
2016). the year, the ground can become both a heat source and sink,
Drying associated with dry CO2 injection triggers salt providing renewable heating and cooling to buildings.
precipitation and alters the porosity particularly in saline Subsurface thermal energy storage systems can help balance
aquifers (Peysson et al., 2014). intermittent renewable energy sources such as wind and solar.
Injection in coal seams leads to significant swelling and Heating and cooling demands vary seasonally and daily; to make
closure of the reservoir cleats (Espinoza et al., 2014). subsurface geothermal systems economically feasible, the
Similarly, CO2 can adsorb in the shale matrix, cause swelling variable heating and cooling demands must be optimized with
of organic matter and a decrease in permeability (Song and respect to heat/cold supply from the subsurface and heat storage
Zhang, 2013). On the other hand, adsorption can decrease the both at the building-scale (Garber et al. 2013) and the city-scale
frictional strength of the caprock (Spokas et al., 2019). (Zhang et al. 2015).
There are open and closed-loop systems (Figure 6 - Florides
The solubility of supercritical CO2 in water can reach 1~2
and Kalogirou, 2006, Johnston et al., 2011, Narsilio and Aye
moles/L, water acidifies to pH~3 and prompts mineral
2018). In open-loop systems, groundwater is passed through the
dissolution in the reservoir rock, the caprock and the cement
ground source heat pump GSHPs and returns back to the porous
around casings (Vallin et al., 2013; Selvadurai et al., 2017;
ground; challenges are related to corrosion, system maintenance,
Cha and Santamarina 2019; Barria et al., 2021). The evolution
clogging, and dissolution/precipitation.
of reactive fluid flow depends on mineral reactivity, advection
and diffusion rates.
Water with dissolved CO2 and minerals becomes heavier and
creates convective currents that sustain further dissolution.
Adsorption, mineralization, and dissolution in water lowers
the excess pressure over time in favor of long term stability
(similarly, leaks help by pressure bleed-off - Birkholzer et al.,
2015)
Simple scaling relationships facilitate first-order spatial and
temporal analyses of coupled phenomena during geologic CO2
sequestration (Kim et al. 2019).
Large-scale CO2 injection projects may induce seismic events
(Ellsworth, 2013; Yarushina & Bercovici, 2013; Vilarrasa et al.,
2014; Elsworth et al., 2016). In general, sedimentary formations
are softer than the crystalline basement and facilitate pressure
buildup around faults (Vilarrasa and Carrera, 2015). In addition, Figure 6: Shallow geothermal systems - Open and closed loops.
CO2 leakage may take place preferentially along pre-existing
fractures and abandoned wells (Hawkes et al., 2005). Injection In traditional closed-loop shallow geothermal systems, high-
strategies must be designed to maximize the use of the pore space density polyethylene or cross-linked polyethylene pipes are
in the reservoir while preserving seal integrity. Surface and embedded in vertical boreholes down to ~100 to 200 m in depth
subsurface monitoring programs should be used to anticipate or in horizontal trenches approximately ~1 to 1.5m in depth to
these risks; their design must be based on the thorough form ground heat exchangers. The loop involves a ground source
understanding of such prevalent regimes. heat pump and the circulating carrier fluid (water or water-
antifreeze solution) and there is no mass exchange with the
99
99
ground but heat transfer only. The system performance compares 4.2 Geothermal Power (Typically Deep)
the thermal kWhth delivered by the system for every kWhe of
Geothermal energy is a promising non-carbon energy resource,
electricity used by the heat pump. A good year-round
yet, the installed capacity remains low; in fact, the 3.6 GW online
performance corresponds to a ratio kWhth/kWhe 5. Open-loop
geothermal power in the USA is a quarter of the world’s total
systems typically show higher performance than closed-loop
capacity (Huttrer, 2020). The Geyser field in California has over
ones.
300 wells and a total installed capacity of 2 GW (Calpine
Shallow geothermal systems may benefit from other
Corporation, 2019; Sanyal and Enedy, 2011).
geotechnical geostructures such as piles, retaining walls, and
Most geothermal power plants target hot hydrothermal
tunnel linings to significantly reduce capital costs (Brandl 2006,
systems with a single or double vertical wellbore completion
Laloui et al., 2006; Barla et al., 2016; Bourne-Webb et al., 2016,
(Hamm et al., 2009). In doublets, injection and production well
Brandl 2016, Soga and Rui 2016, Loveridge et al., 2020). Energy
spacing ranges between 300 and 1500m, and production rates
piles have attracted most of the attention by the geotechnical
vary from 10 to 140 liters/second (Figure 7). New generation
scientific community to date (Laloui and Di Donna 2011,
deep geothermal systems aim at increasing reservoir size and
Amatya et al., 2012, Loveridge and Powrie 2013, Fadejev et al.,
temperature to increase power output and involve deviated
2017, Gawecka, Taborda et al. 2017, Makasis et al., 2018,
wellbores, hydraulic fractures (enhanced geothermal systems
Alberdi-Pagola et al., 2018). Experimental studies have been
EGS) and deep closed-loop horizontal wellbores.
conducted in situ or in scaled models both at 1-g and N-g
(Kalantidou et al., 2012; Stewart & McCartney, 2013). In recent
years, studies have expanded to other geostructures where more
roduction
n ection
complex geometrical and constructability arrangements
dominate, including works on energy tunnels (Sadokierski and
Thiffeault 2008; Nicholson et al., 2014; Barla and Di Donna o surface
2018; Bidarmaghz and Narsilio 2018; Dai et al., 2022) and
energy retaining walls, mainly on diaphragm walls (Dong et al.,
2019, Di Donna et al., 2017, Sterpi et al., 2018, Rammal et al.,
2018, Sun et al., 2013). These structures have an exposed face 1000 m
resulting in lower thermal storage capacity and the potential need
for thermal insulation.
Important coupling processes relevant to energy
geostructures include (Soga and Rui, 2016): (1) heat-fluid
coupling inside heat exchanger pipes, (2) a complex thermo-
hydro-mechanical coupled soil response which is aggravated in
lane 1 250 m
low-permeability high-compressibility fine-grained soils
lane 2
(Bourne-Webb et al., 2009; Rui and Soga, 2019), and (3) thermo-
mechanical coupling in the substructure including the impact of
temperature change on concrete structures and the long-term Figure 7: Stimulated horizontal wellbores - A two-fracture EGS system
soil-structure interaction response for systems subjected to a (Gee, Gracie and Dusseault, U. Waterloo)
large number of thermal cycles and potential ratcheting effects
(Pasten et al., 2014& 2019; Nguyen et al., 2020).
The performance of subsurface energy storage is significantly All deep geothermal systems face drilling and engineering
affected by the spatial and temporal heterogeneity of the challenges (Kocis et al., 2017; Rossi et al., 2020): bit-steering at
subsurface thermo-hydrological thermal regime. The extent of high temperatures (>300°C), long and costly laterals for closed
subsurface storage capacities can vary locally. Subsurface wellbore systems, casing and cementing difficulties, alternatives
temperatures are affected by surface warming and the heat island to steel casing, fluid losses in uncased wellbores and the potential
effect due to urbanization, which creates subsurface thermal for early depletion, hydraulic fracture control (Ugueto et al.,
anomalies in many cities. The depth of deviation from the 2021; Adachi et al., 2007), natural fracture network
regional geothermal gradient due to surface warming can reach characterization, local stress field and implications (Forbes et al.,
100 m (Tokyo - Taniguchi et al. 2007). Anthropogenic heat flow 2019; Xing et al., 2020).
from building basements and subways influences subsurface The efficiency of thermal extraction often remains as low as
temperatures, especially when in proximity of urban aquifers ~16%, and it is affected both by extraction parameters (e.g.,
(Bidarmaghz et al., 2019). The combined effects of urban heat production/injection rate, well spacing) as well as reservoir
islands and underground heat islands can change the subsurface characteristics: rock types (igneous, sedimentary, and
temperature by several degrees, i.e., “underground climate metamorphic), the subsurface natural geoplumbing and the
change”, and affect the performance of shallow geothermal morphology of induced hydraulic fractures (matrix hydraulic
systems (Bidarmaghz et al., 2019). conductivity, fracture network and transmissivity - Cardona et
There are various engineering design charts, codes and al., 2021), and thermal properties (Garcia and Santamarina
guidelines around the world for the more common shallow 2021). Subsurface properties are difficult to characterize, thus
geothermal systems. Furthermore, considerable advances in associated uncertainties lead to inaccurate predictions and hinder
numerical modelling facilitate the analysis of more complex optimization towards higher efficiency. Data collected for 136
systems. Still, experimental data are scarce for thermal- geothermal fields show that permeability varies in 3-orders of
geostructures such as soldier pile retaining walls (Makasis et al., magnitude between 20 and 1600 mD, porosity between 0.02 and
2021; Zhong et al. 2022), pavements (Muñoz-Criollo et al., 2016, 0.16, and matrix thermal conductivity between 1.6 to 3W/(m.K).
Gu et al. 2021), underground utilities (Shafagh et al. 2022), and The output power of a geothermal system is proportional to
decommissioned mines. the flow rate and the output temperature. The underlying
equations are mass, momentum, and energy conservation. Useful
dimensionless ratios capture the interplay between underlying
parameters (Hoteit et al., 2021; Juliusson and Bjornsson 2021):
100
100
For example, heat depletion at the Geysers, California, has
resulted in continued seismicity over decades with magnitude
The dimensionless time represents the cumulative volume Q.t events as large as MW = 4.5 (Martinez-Garzon et al., 2014). In
of water injected into the reservoir at reservoir conditions relative fact, it has been proposed that a gradual release of strain energy
to the reservoir's pore-volume Pv. The thermal recovery factor triggered by thermal destressing can help mitigate large
FTR quantifies the energy production efficiency in terms of the earthquake risks (Im et al., 2021). Yet, other geothermal projects
extracted energy relative to the original energy in place . appear to have induced large events of induced seismicity (MW
= 5.5) caused by thermal destressing, such as in Pohang, South
The enthalpy production factor FEP in terms of the instantaneous
Korea after 2 years of operation (Grigoli et al. 2018). Designing
and the initial enthalpies represents the sustainability
the operation to prevent fracture reactivation can decrease the
of constant enthalpy (or temperature) at the production well
recoverable heat to 10-50% of the available heat (McLean and
during the extraction process, i.e., it declines as a result of
Espinoza, 2021, 2022).
reservoir cooling and thermal front breakthrough. Thermal
recovery and enthalpy production efficiency are significantly
influenced by permeability, rate, porosity, and well spacing. The 5. ENERGY GEO-STORAGE (Appendix Table B1)
optimal geothermal field development requires the simultaneous
maximization of both FEP and FTR; however, they often exhibit Energy storage is a critical component for a sustainable energy
opposite trends, for instance, a tighter well spacing can help system: it allows the operation of power plants at their highest
increase the recovery factor but may lead to an early efficiency throughout the year and optimizes the use of
breakthrough of recycled water. Simulation studies are crucial renewable energy sources by addressing the intrinsic fluctuations
for optimal geothermal field development. and time mismatch between energy demand and renewable
sources (See review in Pasten and Santamarina 2011). The
4.3 Hydraulic Stimulation and Induced Seismicity. selected system must satisfy the transient capacity gap and
needed power, and may combine hourly, daily, and seasonal
Hydraulic stimulation is commonly used in the oil and gas sector
storage components.
to improve flow in reservoirs (Hubbert and Wallis, 1957;
Energy storage depends on energy quality: high-grade
Zemlak, et al., 2002). Similarly, engineered geothermal system
mechanical, electrical, magnetic, and chemical energy can be
use hydraulic fractures to improve the transmissivity between
readily converted into other forms of energy with high efficiency
wells and to create a fracture network that enables sustainable
(greater than 85%); however, thermal energy is low-grade and its
heat extraction (Majer et al., 2007). In practice, the operator
transformation into a higher form of energy is limited by the
increases the injection rate and surface injection pressure until
Carnot efficiency (efficiency can be lower than 10% for shallow
the fluid pressure at the formation face is large enough to activate
geothermal systems).
the propagation and growth of existing fractures and open new
In terms of power and energy capacity, mechanical energy
fractures (Maxwell et al., 2010). storage systems such as compressed air and pumped hydro are
When large volumes of fluid are injected into the reservoir, most suitable for large-scale power systems since they have a
fluid may travel into faults (Rutqvist and Stephanson, 2003) and
lower levelized annual cost for high discharge times compared to
generate micro-seismic events, which are commonly recorded
batteries, flywheels, and electric storage systems (Groenenberg
(Shapiro and Dinske, 2009). Indeed, many faults throughout the et al., 2020). Thermal energy storage either as sensible heat
Earth’s crust are in a state of critical failure equilibrium (Zoback (change in temperature) or latent heat (change in phase) is
and Zoback, 1980, Zoback et al., 2002) and fluid injection can
preferred when the available surplus energy is heat (see shallow
induce small stress, chemical or thermal perturbations in the
geothermal systems discussed above. Note: latent heat has a
underground and lead to fault reactivation and enhanced higher energy density within the range of working temperatures).
seismicity. The magnitudes of these events vary depending on Geo-storage in chemical form includes fuel storage in
the site conditions and hydraulic energy (McGarr, 2014). Various
underground caverns (e.g., strategic petroleum reserves
geothermal projects have exhibited induced seismicity during the
maintained by the Department of Energy in the USA and the
process of hydraulic fracturing, ranging from microseismicity to Federal Oil Reserve in Germany), and the commercial storage of
MW = 3.2 in Basel, Switzerland (Moeck et al., 2009; Mignan et natural gas, butane, propane, propylene, and gasoline in salt
al., 2015). Many different operational protocols and stimulation
caverns, aquifers or depleted reservoirs. There is increased
designs (dealing with well strings and trajectories, well
interest in chemical energy storage in the form of hydrogen and
completion, interval isolation and sequences, shut-in ammonia.
management, fluid pressure and volume, fluid properties, All energy geo-storage systems impose coupled stress,
proppants, and responses to incidents) have been proposed to temperature, wet-dry, and freeze-thaw cycles; while these
mitigate the risk that is associated with fluid-induced
excitations may not cause monotonic failure, ratcheting may
earthquakes, (e.g., Majer et al., 2012), but engineering a fracture affect the long-term cyclic response of energy geo-storage
topology that suits the needs of optimized fluid and heat recovery systems. The rest of this section focusses on three potential grid-
without generating induced seismicity is still difficult (the same scale energy storage systems that are currently gaining increased
as in oil and gas). Difficulties arise from the complex thermo-
attention: compressed air, hydrogen and ammonia.
hydro-chemo-mechanical reservoir response and difficulties in
mapping and characterizing fracture networks. Current research 5.1 Compressed Air Energy Storage CAES
efforts range from the laboratory-scale (e.g. Passelègue et al,
2020, Acosta et al., 2020, Violay et al., 2019) to the crustal fault- Compressed air energy storage can achieve higher energy density
scale (Eaton et al., 2018, Eyre et al., 2019). than pumped hydro, hence it uses less volume, and the
Heat depletion from geothermal fields can also result in stress underground space for compressed air has a lower environmental
changes that lead to fault/fracture reactivation, as can be inferred impact than surface water impoundment (furthermore,
from coupled thermo-hydro-mechanical analyses. Similar to oil compressed air can be stored underwater in offshore facilities).
and gas depletion, the magnitude of stress/strain changes depends Potential onshore geologic formations include hard rock caverns,
on reservoir geometry, boundary conditions, and transient pore salt caverns, bedded salts, porous aquifers and depleted
pressure and temperature changes (Segall and Fitzgerald, 1998). reservoirs; cased wellbores can also be used for underground
101
101
CAES (Figure 8). Solution-mined salt caverns are advantageous storage conditions (0.01 GJ/m3 at 1 atm and 25°C - for
given the low salt permeability and self-healing characteristics comparison, gasoline has 34 GJ/m3).
(Kushnir et al., 2012; Kim et al., 2012). On the other hand, porous Hydrogen production methods and ensuing emissions define
rock formations with impermeable upper and lower layers (e.g., three hydrogen categories (Figure 9 - production pathways in
depleted gas reservoirs or porous aquifers) are more frequently Nikolaidis and Poullikkas 2017): (1) grey hydrogen is obtained
found than salt formations and offer larger storage capacities from fossil fuels such as methane and emits 10 kg of CO2 for
(Wang and Bauer, 2017). each kg or H2 produced, (2) blue hydrogen is produced the same
way but CO2 is captured and stored, and (3) green hydrogen
obtained by electrolysis using renewable energy sources. There
is a trade-off between sustainability and affordability: current
Motor Compressor High Low Generator production costs are ~2 US$/kg for blue hydrogen and
Pressure Pressure
Turbine Turbine
4.60 US$/kg for green hydrogen. Geothermal-assisted hydrogen
production exhibits higher energy and exergy efficiency than
solar-based technologies (Balta et al., 2009&2010; Dincer and
Acar 2017; Ghazvini et al., 2019; Mahmoud et al., 2021).
Air Hydrogen storage in porous media and salt caverns can
overcome the intermittency of renewable energy production
(Mostafaeipour et al., 2016; Notton et al., 2018; Hashemi et al.,
Salt Dome 2021; Hassanpouryouzband et al., 2021; Heinemann et al., 2021).
Figure 8. Compressed air energy storage CAES Anticipated hydro-bio-chemo-thermo-mechanically coupled
phenomena in H2 geo-storage are quite different from either CH4
or CO2 storage
CAES can be categorized as diabatic, adiabatic, and -Bio-photolysis
isothermal based on the heat management during air compression Biological -Dark fermentation
-Photo fermentation
and expansion (Venkataramani et al., 2018). The two existing Biomass -Gasification
large-scale CAES plants are diabatic and use natural gas to -Pyrolysis
Thermochemic
provide heat during air expansion. Adiabatic/isothermal CAES Renewable
-Combustion
-Liquefaction
reuse locally stored heat captured during air compression (Budt sources
et al., 2016; Eckard, 2010). Heat recovery can increase the Thermolysis
operational efficiency from 50–55% to as high as 70–75% (fully -Alkaline
Water splitting Electrolysis -Solid oxide
adiabatic). -PEM
Small-scale CAES systems can use steel containers and Photolysis
wellbores, underground and submerged bags (Pimm et al., 2014), Pyrolysis
or building foundations (Ko et al., 2020; Zhang et al., 2020) as Fossil fuels
-Steam reforming
storage media. Currently, CAES requires sufficiently large Reforming -Partial oxidation
storage volumes or high pressures for economic benefit. -Autothermal reforming
The air pressure must remain below the fracture pressure Figure 9. Hydrogen production
when the compressed air acts directly against the host rock (e.g.,
salt caverns, porous aquifers or depleted reservoirs); on the other
hand, a minimum pressure must be kept to limit the stress ratio Hydrogen serves as an electron donor or energy source for
on cavern walls during depressurization. In the case of storage in many microorganisms. Expected bio-mediated reactions during
porous aquifers, air injection causes water drainage followed by underground hydrogen storage include methanogenesis,
imbibition during air discharge; associated processes involve acetogenesis, and the reduction of sulfates, iron and inorganic
two-phase fluid flow under the desired air injection/extraction carbon which are associated with common subsurface minerals
rates, flow instabilities, fines migration, and/or freezing-drying. e.g., calcite, anhydrite, hematite (Hagemann et al., 2016; Berta et
Pore-size heterogeneity and flow rate have a pronounced effect al. 2018). Consequently, microbial activity can lead to mineral
on flow, sweep efficiency and the evolution of residual saturation dissolution/precipitation, altered hydraulic conductivity and
during repetitive drainage-imbibition cycles (Zhang, 2021). changes in geomechanical properties of the rock mass.
Associated heating/cooling during pressurization/ Under most geologic storage conditions, hydrogen will be
depressurization may also negatively affect the long-term stored in the gaseous state. Ensuing questions relate to pressure-
stability of the formation. Numerical simulations must capture dependent interfacial tension, cyclic injection-extraction effects
such coupled thermo-hydro-mechanical effects and the on wettability and entrapment, hysteretic capillary properties and
consequences of repetitive cycles to anticipate the system relative permeability among others (Heinemann et al., 2021).
performance. Hydrogen is much lighter than CO2 under storage conditions; for
example at 10 MPa and 40 C, their mass densities are CO2= 630
5.2 Hydrogen: Production and Storage kg/m3 and H2= 7.8 kg/m3. Consequently, the capillary pressure
Hydrogen combines with oxygen to produce water and releases u exerted against the seal will be significantly higher in H2
~120 kJ/gr of H2 (Park et al., 2000 - For comparison, the storage: u=H g ( H2O- fluid).
gravimetric energy density is 54 kJ/gr in methane and 46 kJ/gr in The solubility of H2 in water is very small (e.g., 0.14 mol/L
gasoline). As an energy carrier, hydrogen can be employed in a at 65 C and 20MPa) and dissolved H2 in water has no impact on
variety of applications from transportation (Cipriani et al., 2014) pH; as such, framework minerals such as quartz and feldspar
to heating/cooling (Maroufmashat et al., 2016; Valente et al., appear to be unaffected by hydrogen exposure. However,
2018). Therefore, hydrogen emerges as an alternative to fossil dissolved hydrogen may react with chemical components that are
fuels, albeit with its own challenges (Brumfiel 2003; Winter originally present in the formation water, driving the dissolution
2009; van Renssen 2020; Aftab et al., 2022). In particular, the of carbonate minerals. Moreover, the hydrogen-driven redox
low volumetric energy density in hydrogen gas challenges reactions with iron-bearing minerals and sulfate minerals could
lead to additional mineral dissolution resulting in mechanical
102
102
weakening and volumetric changes in the reservoir and caprock. TWh of electricity in 2020, equivalent to 4.3% of the world
Furthermore, mineral dissolution may lead to positive feedback energy consumption and about 10.1% of the global gross
with fluid flow and further compromise the containment of electricity supply (IAEA 2021, Schneider and Froggatt 2021).
hydrogen: for example, a clear increase in the porosity and According to the Intergovernmental Panel on Climate Change
permeability were observed for sandstone samples that contained IPCC report, the carbon intensity of nuclear power per kWh
anhydrite CaSO4 and carbonate caused by mineral dissolution by produced is more than an order of magnitude lower than that of
H2-enriched water (Flesch et al., 2018). any fossil fuel and comparable to the carbon intensity of
renewable sources such as hydroelectric, wind or solar. For this
5.3 Ammonia reason, the European Commission has recently proposed that
investments in nuclear power plants could be labelled as
A comparison between hydrogen and ammonia properties helps
sustainable, under certain conditions. Nevertheless, because of
us to identify their relative advantages and potential storage
the potential risks associated with accidents, as exemplified in
possibilities (Clifford and Hunter 2002; The Engineering
the cases of Three Mile Island in 1979, Chernobyl in 1986 and
ToolBox):
Fukushima in 2011, nuclear power remains controversial and
Gravimetric energy density: ammonia is much lower than faces considerable opposition, making its future quite uncertain.
hydrogen (NH3: 22.5 MJ/kg; H2: 120 MJ/kg) The safe storage and disposal of high-level nuclear waste
Volumetric energy density: greater for liquid ammonia than HLW remains a major issue. HLW is generated either in nuclear
liquid hydrogen (liq-NH3: 15.6 MJ/L; liq-H2: 9.1 MJ/L) power plants as spent fuel or in nuclear fuel reprocessing
Liquefaction: ammonia at 1atm & -33°C, or at 750kPa & facilities in the form of vitrified waste. HLW represents 95% of
20°C; in comparison, hydrogen liquefies at -253°C at 1atm the total radioactivity generated by civilian use, contains long-
Solubility in water at 20 C: ammonia is highly soluble ~530g lived isotopes and is heat emitting. The main safety objective is
NH3/liter H2O, as compared to 0.0016g H2/liter H2O to protect human health and the environment without imposing
Diffusion in water at 20 C: three times higher for hydrogen undue burdens on future generations (IAEA, 2005). The general
than ammonia (H2: 4.6 10-9m2/s; NH3: 1.5 10-9m2/s) consensus is to place the HLW in deep geological repositories
Liquid viscosity: liq-NH3: 0.27 cp much higher than liq-H2: using a combination of natural and engineered barriers
0.009 cp (Chapman & Mc Kinley 1987). There are several repositories in
Consequently, it is much easier to liquefy ammonia than operation for the disposal of low- and medium-level radioactive
hydrogen, to attain higher energy density and to store it due to waste. Finland’s Onkalo facility will be the first deep geological
lower diffusion and higher viscosity. Therefore, ammonia is a repository for spent fuel (will start operation in 2023), and a
promising hydrogen carrier for transportation and storage (Figure repository in Sweden is expected in the 2030’s.
10 - Wentorf and Hanneman 1975; Koike et al., 2012; Hayashi The multi-barrier concept should provide enough waste
et al., 2013; Kato et al., 2016; Nagaoka et al., 2017; Kojima and isolation to reduce the migration of radionuclides to the
Yamaguchi 2020; Caglayan et al., 2020). In addition, ammonia biosphere to acceptable levels over extremely long periods of
NH3 has a low production cost and there is extensive experience time - a million years in some regulations. The system comprises
in ammonia production and transportation developed by various the treated waste in a canister, an expansive clay engineered
industries, in particular the fertilizer sector. barrier and the geological host rock. The role of geotechnical
engineering in deep geological repositories ranges from the
Carbon
Intensive
Hydrogen
generation
Ammonia synthesis
/processing
Transportation
/storage
Applications construction of the repository itself to understanding and
energy
predicting the underlying thermo-hydro-chemo-mechanical
Natural gas
Reforming
H2
H 2 NH3 H 2 coupled response of the engineered barrier and the host rock
H2 Cracking 90000 to
Mobile
Coal Gasification
H2
Liquefied /processing 70000 KPa
mass (Gens 2003; Gens 2010).
N 2
Catalyst
773.1 K
298.15 K
1000 KPa
20000 to Fuel cell vehicles
Capturing CO2 40000 KPa
N 2
6.1 Engineered Barrier
Air
Storing Depleted reservoirs
Saline aquifers
Refrigerated
240.1 K Chemical Power
The engineered barrier around the waste containers provides
CO2
Salt caverns industry Generation mechanical stability to the gallery and canister, delays water
Carbon- access to the waste package and retards the migration of
neutral Electrolysis
radionuclides that could eventually leak from a deteriorating
energy
Wind H 2
Long-term
NH3 Depleted reservoirs
Saline aquifers
Direct
NH 3
canister. Bentonite-based mixtures have been proposed to fulfil
Solar
H2 O2
N 2
Catalyst
773.1 K
20000 to
Salt caverns
Internal
this function for their low permeability, high swelling, high
40000 KPa
Agriculture combustion
engine
specific surface area and high retention capacity. High density
fertilizer
Hydel
Air enhances its response, thus the target barrier dry density ranges
between 1.45 and 1.65 g/cm3 (Sellin and Leupin 2013) and
Figure 10. Hydrogen and ammonia. From production to utilization. requires the use of pre-compacted blocks of bentonite-based
materials (SKB 2010) or high-density bentonite pellets prepared
However, ammonia must be stored in caverns as it readily with pre-heated bentonite powder to obtain a more compressible
dissolves in water aquifers. The interaction with brine, reservoir material. Block placement combined with pellet backfilling add
gases, cushion fluids and bio-activity requires detailed complexity to modelling and understanding of system
consideration. For example, when liquid NH3 mixes with water, performance (NAGRA 2019). Although the bentonite buffer
it forms (aq) ions and OH-(aq) ions to establish an alkaline shows a tendency towards homogenization during imbibition,
solution that will react with storage rocks causing both long-term observations show memory of initial conditions on
dissolution and precipitation. density and moisture content evolution (Sellin et al. 2020).
Coupled thermo-hydro-chemo-mechanical processes take
place during the transient period of the repository life. The
6. NUCLEAR WASTE STORAGE (Appendix Table A9) bentonite is installed unsaturated and experiences a thermal
gradient driven by radioactive decay (Gens 2010). Concurrent
The 442 operating nuclear power reactors in the world have an chemical changes affect the barrier as well, such as
installed capacity of 393 GW and produced a net total of 2553 dissolution/precipitation of mineral species, cation exchange
103
103
processes, corrosion of metallic components– triggered by the argillaceous rocks may cause hydraulic fracturing (see Gens et
interaction of groundwater with the clay barrier and other al. 2007 for temperature and pore pressure data gathered during
repository elements (the metal canisters, concrete plugs or an in-situ heating test on Opalinus clay in the Mont Terri URL,
reinforcement). Corrosion in turn, along with water hydrolysis Switzerland). The phenomenon is triggered by the water thermal
and microbial activity, can generate gases at the canister-EBS expansion and the response depends on the relative heat and
interface which, depending on the production rate, can be pressure diffusion rates (Note: temperature and pressure maxima
dissolved or will have to be transported away as a separate phase, do not coincide). Furthermore, sudden volumetric contraction
either by visco-capillary two-phase flow, microscopic pathway may take place when the temperature exceeds the maximum
dilation or breakthrough, if the gas pressure exceeds the total temperature experienced by the rock during its geological past
stress experienced by the clay (Levasseur et al. 2021). (e.g. Mohajerani et al. 2012). Such thermally-induced processes
Most repository concepts consider that the temperature at the will become more relevant as the design of repositories evolves
surface of the canisters remains below 100°C. However, the towards higher allowable temperatures.
ability to tolerate higher temperatures whilst still ensuring
appropriate performance has significant advantages, e.g., shorter 6.3 From Laboratory Tests to Underground Research Labs
above-ground cooling times, more efficient packaging, fewer
Laboratory element-tests and large-scale tests carried out in
disposal containers, fewer transport operations, and smaller
surface facilities allow for well-controlled configurations and
facility footprints. Hence, current research focusses on higher
boundary conditions (Martín et al. 2014; Garitte et al. 2017).
temperature - up to 200°C - effects on buffer safety functions
These have been instrumental for gaining a better understanding
(García-Siñeriz et al., 2015&2020; Gens et al. 2022).
of the barriers behavior but also for developing and validating
6.2 Host Rock coupled-process constitutive models and numerical simulators
(e.g. Lloret et al., 2003; Villar et al., 2008; Tang et al., 2008;
Current developments focus on hard crystalline rocks and Gens et al., 2009, 2011; Sánchez et al., 2012; Zheng et al., 2020).
argillaceous rocks; other host rocks considered in the past Twelve underground research laboratories are currently in
included rock salt and unsaturated tuff. Crystalline rocks operation in Europe and Asia to conduct field and full-scale tests
generally exhibit low permeability in the absence of fractures, under conditions similar to those of an underground repository
high chemical stability and low economic value. In addition, their (NEA 2013; Gaus et al. 2014; NAGRA 2019; Villar et al. 2020).
high strength requires limited excavation support. However, Field studies allow testing and demonstrating various barrier
crystalline rocks are brittle and they sustain fractures that concepts, manufacturing methods, handling and assembling the
constitute the main pathways for water flow and potential engineered barriers. Typically, heaters simulate the waste
radionuclide transport. The characterization of the initial fracture canister and hydration is either natural or augmented. Their
network and its interconnectivity remains very challenging; design reflects the particular national repository concepts, e.g.,
furthermore, excavation may cause additional fractures within horizontal or vertical configuration, heater temperature, and
the excavation damaged zone EDZ that have no capacity for self- barrier composition. Some large-scale experiments have run for
sealing. On the other hand, argillaceous rocks have very low long testing periods; examples include the Prototype Repository
permeability, significant retardation properties for radionuclide test performed since 2001 at the Aspö Rock Laboratory, (host:
transport, and some self-sealing capacity; in fact, bentonite granite - Johannesson et al., 2007) and the Full-scale Engineered
barriers may not be needed around canisters. However, Barriers Experiment FEBEX conducted at the Grimsel Test Site
significant support is required due to their low strength and they Switzerland since 1997 (host: granite - Gens et al., 2009; Sanchez
are more sensitive to chemical changes. et al., 2022). Figure 11 and 12 show salient results.
The repository depth will depend on the stratigraphic
sequence (typically several hundred meters deep). The selected
layer must have favorable properties for radionuclide
containment and sufficient thickness to host the repository. Site
investigations must pay close attention to rock structure and
permeability; mechanical properties, though significant, have a
lower priority. The construction of the repository will involve
various access shafts or ramps, a network of horizontal access
drifts, and vertical excavation or additional horizontal drifts for
waste emplacement.
The required underground construction is well within the
capabilities of conventional geotechnical and mining
engineering. However, the design and construction of a HLW
repository involves specific features not usually encountered in
conventional projects. In particular, excavation must minimize
the damaged zone that invariably appears around openings and
becomes a preferential pathway for water migration and Figure 11. Evolution of temperature and pore pressures observed in a
heating experiment on Opalinus clay at the Mont Tetti underground
radionuclide transport; therefore, excavations use tunnel boring
laboratory (Gens et al. 2007).
machines or road headers rather than blasting, prevent rock
relaxation by the early emplacement of rock support systems, and 6.4 Numerical Studies
repressurize the damaged zone using expansive backfill or by
mechanical means. In general, all openings created during site The engineered barrier and the host rock will experience complex
investigation and construction are preferential pathways for thermo-hydro-chemo-mechanical coupled processes triggered by
water, gas and radionuclide migration; therefore, boreholes, the heat released by the spent fuel, hydration of the barrier driven
access galleries and shafts must be carefully sealed and by the strong water potential difference between the unsaturated
backfilled. clay and the surrounding rock, stress changes induced by the
Heating due to radioactive decay may not affect crystalline progressive wetting and swelling of the compacted bentonite
rocks, however, the associated pore pressure increase in under highly confined conditions, and the chemical
104
104
interactions/reactions between the barrier materials and the experimental observations gathered from the 18.3 year long
chemical species dissolved in the host-rock water. These are FEBEX in-situ experiment: the dry density decreases close to the
concurrent and mutually coupled processes. rock because of hydration-induced swelling of the clay, and it
increases close to the heater because of the drying of the
bentonite and the compression caused by the gradual increase of
swelling stresses (Sanchez et al., 2022). Experimental results
show memory of the initial density heterogeneity.
7. CLOSING: GEOSCIENCE AND ENGINEERING
The strong causal link between energy and quality of life
demands a sustainable energy solution that addresses the
evolving energy demands in developing and developed nations.
While renewables are gradually contributing an increasingly
larger fraction of the energy mix, the actual consumption of fossil
fuels will continue to rise during the next 25 years.
Our field has a central role to play in the energy challenge,
from resource recovery to waste management. The important
problems our field solved in the 20th century will remain relevant
to engineering practice in this century (Figure 14). However, the
energy challenge and associated environmental-livability
implications bring new opportunities for our field to embrace and
advance. But, it will require the understanding of new
Figure 12. Persistence of heterogeneities in the bentonite barrier. phenomena.
Contour map of dry density (g/cm3) after 18 years of operation of an
Soil Foundation Sand GeoEnviron.
originally homogeneous barrier (FEBEX in situ test, Villar et al. 2020) Mechanics Engineering Liquefaction Engineering
Therefore, phenomena considered THCM formulations must
include: (i) heat transport (including heat conduction, heat
convection of both liquid water and vapor) and phase changes;
(ii) water flow, liquid phase properties and water vapor diffusion;
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
(iii) air flow, gas phase properties, dissolved air diffusion; (iv)
mechanical behavior of barrier materials, including the effect of Climate
Change
stress, fluid pressure, and temperature; and (v) chemical
C-
reactions in the liquid phase, cation exchange, storage
dissolution/precipitation and adsorption. These processes are
cast in mathematical form through balance equations (mass
balance for the various species, energy, and momentum balance 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
equations), constitutive equations (for thermal, hydraulic,
ACOPLAMIENTO Figure 14. The evolution of frequently used geotechnical terms
mechanical and chemical processes), and equilibrium restrictions
(Figure 13THM- Thomas and He, 1995; Rutqvist
x TRANSPORTE et al., 2005; Olivella
REACTIVO Classical assumptions in geotechnical engineering need to be
et al., 1996). reassessed to address the needs in the energy sector. In contrast
to classical geotechnical engineering, energy geoengineering
covers wide-ranging scales for almost all physical parameters
REACTION KINETICS
including length (meter-scale foundations to hundreds of
kilometers long oil reservoirs), time (the 30 year design life for
civil constructions to the need to predict performance for 100,000
to 1,000,000 years in the case of nuclear waste storage), depth
REACTIVE (the upper 50 m for infrastructure projects to wells drilled deeper
THM TRANSPORT than 10 km), effective stress (500 kPa to greater than 100 MPa),
n interacting species
temperature (from normal ambient conditions to temperatures
that can range from -50 C in the artic to higher than 300 C for
molten salt energy storage). Consequently, phenomena that may
play a secondary role in the near surface gain relevance at depth
CHEMICAL because of high fluid pressure, high effective stress, small pores,
EQUILIBRIUM MODEL and high temperature. Furthermore, the long timescales
associated with energy-related processes require further
Figure 13. THCM coupling consideration of time-dependent phenomena. Notable examples
follow:
Effective Stress: erzaghi’s effective stress assumes
Laboratory, surface and underground tests have played a incompressible grains. This assumption no longer holds at
critical role in validating coupled THCM formulations, high effective stress and pressure. Biot’s effective stress B’=
constitutive models and their numerical implementation in - B p considers the relative bulk stiffness between the dry
computer codes (Sánchez et al. 2022). Results highlight the solid skeleton Ksk and the mineral Km through the Biot-Willis
critical role of the bentonite dry density on the evolution of the parameter B=1-Ksk/Km. High effective stress conditions give
liner system. Furthermore, numerical predictions agree with
105
105
rise to crushing-mediated compaction, and diagenetic geotechnical engineering analyses are diffusive, advective and
processes such as pressure solution-precipitation. reaction rates, mechanical creep (significant in salt), and
Compressibility: The linear approximation in log scale changes in boundary conditions, from climate to tectonism.
e=Cc log[( o+ )/ o] is locally applicable, yet predictions
are physically incorrect at small and large stress levels. Educational programs must evolve to address the knowledge
Instead, asymptotically correct power or exponential models needs in energy geoscience and engineering (Editorial 2021).
are more adequate for wide stress range conditions from the Let’s prune obsolete concepts and approaches, place renewed
depressurization of hydrate-bearing sediments to mud cake emphasis on fundamentals, and extend the curriculum to
formation during drilling (Chong and Santamarina 2016). adequately cover the knowledge and tools relevant to energy-
Diffusion: The mean free path is the average distance a environment-livability demands. Highly qualified new graduates
molecule travels between collisions. The mean free path in will facilitate the rapid transition in engineering practice needed
liquids is shorter than the molecular size and Fick’s diffusion to address the challenges and opportunities in the energy sector.
applies in most cases. However, the mean free path in gases
depends on density, i.e., pressure, and may approach the pore
8. ACKNOWLEDGEMENTS & OTHER SOURCES
size dpore in small pores such as in many shales; then diffusion
will involve collisions against pore walls (Knudsen diffusion The authors acknowledge the support from various national
– The Knudsen number is Kn= /dpore). agencies and industry partners. Compilation efforts were
Flow: Fluids may adhere to the boundary (non-slip flow – supported by the KAUST Endowment at King Abdullah
liquids) or slide along the pore as free molecular flow (e.g., University of Science and Technology. G. E. Abelskamp edited
gases in small pores). In general, Navier-Stokes and Darcian the manuscript.
flow apply when there is no-slip at boundaries (Kn<10-2). Energy Geotechnology was the focus of a Special Issue of the
Navier-Stokes is extended by considering slip boundary KSCE Journal of Civil Engineering (2011, volume 15, issue 4 -
conditions when 10-2<Kn<10-1. The free molecular flow Open access) and the 50th Terzaghi Lecture (Santamarina 2014
regime takes place when molecular interaction with pore walls available on YouTube). The comprehensive list of references
vanishes at Kn>10. Gas permeability tests require proper captures the evolution of energy geoengineering within the fields
correction for gas interaction with pore walls (Klinkenberg covered in this report, and important contributions.
correction).
Mixed Fluids - Capillarity: van Genuchten (geomechanics)
and Brooks-Corey (oil) models capture capillarity-saturation-
relative permeability trends.
Reactive transport: Dissolution and precipitation. Reactive
fluid flow may trigger positive feedback effects whereby the
most conductive pathways dissolve more actively and increase
their flow rate. Dissolution and the evolving pore structure
reflect the interaction between diffusion, advection and
reactive transport, as captured in the Damkohler and Peclet
numbers. Precipitation gives rise to self-stabilizing processes.
(Caution: carbonates exhibit retrograde solubility, i.e.,
solubility decreases with temperature).
Phase transformation and gas exsolution: Common phase
transformations are encountered in energy applications,
including hydrate formation-dissociation, water-vapor, liquid-
gas CO2. These are accompanied by pronounced volume
changes, and can be either exothermic or endothermic (heat of
transformation) and lead to complex thermo-hydro-
mechanical coupling effects. Gas exsolution from oil or water
has similar effects, e.g., foamy flow.
Repetitive loads: Energy piles, compressed air energy storage,
tidal energy systems, the foundations of wind turbines, and
pump-hydro power plants are examples of energy systems
subjected to repetitive loads. Geomaterials subjected to
repetitive loads experience the accumulation of volumetric
strain (towards “terminal density”) and shear strains (arrest:
“shakedown”; or ceaseless accumulation: “ratcheting”). Strain
accumulation may eventually impair the system performance.
Repetitive loads may involve not only effective stress, but
thermal, humidity, chemical, fluid pressure, and freeze-thaw
cycles. Numerical simulators must capture the effect of
thousands of cycles even when the cyclic strain falls below the
monotonic elastic threshold strain.
Time: The sense of a static, inert geologic system vanishes in
long time scales. In fact, time emerges as a critical parameter
in some energy geotechnology challenges. For example, there
is an expectation that radioactive waste will be safely stored
for 100,000 years, and that CO2 will remain stored
underground for hundreds to thousands of years (past the fossil
fuel era). Common phenomena that bring time into
106
106
9. APPENDIX Table A1: Oil (33% of worldwide energy consumption)
The following summary tables have been updated from Santamarina Accumulation of organic matter in sedimentary
and Rached (2021). Sources of information include: basins maturation ( & in “oil window”)
Origin
World Nuclear Association (2020) migration from source to reservoir rock
(geoplumbing and traps)
AWEA, American Wind Energy Association, 2020.
World: 1.5x1012 barrels (Venezuela, Saudi Arabia,
British Petroleum Company, 2020: BP statistical review of world Reserves
Canada, Iran, Iraq, Kuwait)
energy. London: British Petroleum Co. World: 3.2x1010 barrels/yr (USA, EU, China, Japan,
Consumption
CIA World Factbook. India). Main use: transportation
Energy density: 46 MJ/kg (effective: 15 MJ/kg)
DOE, Department of Energy. Energetics
Emitted CO2 at power plants: 0.88 kg/kW.h
IAEA 2021. Nuclear power reactors in the world
New reservoirs and challenges
IGA, International Geothermal Association, 2020. Locations: deep (high temperature and pressure), arctic regions
International Energy Agency, 2019: World Energy Outlook. Formations: weak (uncemented sands), intensely fractured
IPCC, 2018: Global Warming of 1.5°C. Masson-Delmotte et al., eds. (carbonates), highly compressible (chalk, diatomite), beneath
difficult layers (e.g., salt, hydrates)
NOAA, Trends in atmospheric CO2. Very viscous or immobile oils (oil sands and oil shales)
NSRDC, The National Solar Radiation Data Base (NSRDB), 2018. Geotechnical issues
Organization of the Petroleum Exporting Countries (OPEC), 2020: Characterization:
Geo-plumbing and fracture networks (conventional and
The Annual Statistical Bulletin.
unconventional reservoirs)
Schneider, M. & Froggatt, A. (2021). The world nuclear industry Geophysical methods - Subsalt characterization (e.g., Brazil)
U.S. Energy Information Administration, 2020: Annual Energy Assessment of in situ pressure, effective stress and temperature
Outlook 2020. Physical properties: mineralogy, porosity, permeability, mechanical,
thermal
Correlations with in situ tests, sampling (pressure cores if needed),
laboratory testing at field [ ’, u, ] conditions
Drilling (drilling costs increase higher-than-linearly with depth)
Borehole wall stability (swelling smectite shale and fractured
quartz-illite shale)
Hydrate bearing sediments - Salt layers
Deepwater drilling and directional drilling
Well leak: excessive mud weight, overpressure, cementation
assurance & control
Production: inherently mixed fluid flow (water, oil, gas)
Reservoirs in carbonates (most oil reserves are in carbonates)
Sand production management (control or exploitation)
Fines migration and formation clogging (extend to asphaltenes)
Deep water / high fluid pressure (stiffness of clays and sands at
high stress?)
Reservoir stimulation:
Hydraulic fracturing (modeling and monitoring of multiple-
stage HF)
Steam assisted gravity drainage for viscous oil production
(weak sandstones and naturally fractured carbonates) - Strains
Acid injection (carbonate dissolution)
Enhanced oil recovery (water and surfactants, alcohol, CO2
injection)
Subsidence (high porosity and compressibility i.e., low seismic
velocity, thick reservoir, large depressurization during production)
Casing buckling and shear triggered by reservoir subsidence, fault
reactivation, differential thermal strains in thermal stimulation
Monitoring: deformations (surface and subsurface), microseismicity,
and fluid pressure. Integration of monitoring data into reservoir
management
Infrastructure:
Foundations (carbonates, hydrates, chalk), large size jack up rigs
and storm loads
Stability and performance of pipelines – Submarine landslides
Waste reinjection (fluids, cuttings, produced fines) in deep
sedimentary basins
Reassessment of super-basins within a comprehensive petroleum
system perspective
107
107
Table A2: Gas (24% of worldwide energy consumption) Table A3: Coal (27% of worldwide energy consumption)
Found in traps (floating above oil or in gas fields), Sedimentary rock made of carbon (anthracite:
shale gas, coal beds, gas hydrates metamorphosed). Forms seams or beds
Thermogenic: thermal cracking of petroleum (P&T Origin
Location Recovery: shaft to underground mines or open pit
above the “oil window”) surface mining
Biogenic: methanogens (e.g., high purity methane World: 1012 ton = 1000 Gt (China, USA, Pakistan,
found in hydrate bearing sediments) Reserves
Russia, India, Australia)
World: recoverable >200x1012 m3 (Russia, Iran, World: 8x109 ton/yr (China, USA, India).
Qatar, Turkmenistan, USA) Consumption*
Reserves* Primarily: electric power generation
Gas hydrates: 3×1015-to-1x1016 m3 of carbon. Pro- Energy density: 24 MJ/kg (effective: 10 MJ/kg -
ducible: probably <4% Energetics Most efficient plant: 49%)
Consump- Emitted CO2= 0.96 kg/kW.h
World: 3.3 x 1012 m3/yr (USA, Russia, EU)
tion Geotechnical issues
Energy density: 45 MJ/kg of gas. Note: for hydrates, Characterization:
5.9 MJ/kg of hydrate mass Stratigraphy and faulting
Energetics
Carbon intensity: emitted CO2 at power plants: Material properties (including gas content)
0.5 kg/kW.h Mine design and operation:
Common Geotechnical issues Global and local stability (including roof stability and the effects
Characterization: of faults)
Geophysical methods: stratigraphy, natural fractures and geo- Optimal extraction strategies
plumbing, interpretation of in situ conditions (pressure, effective Coal material handling - Flow
stress, gas content or hydrate saturation Shyd), and estimation of Environmental impact (local and regional)
physical properties. Well logging Coal combustion products - Fly ash (130 million tons/year generated
In situ testing. Sampling (pressure coring). Laboratory tests at in in the USA): Handling and reutilization
situ conditions Early diagenesis and hindered compaction
Properties (mineral, porosity, permeability, compressibility, Pond assessment. Static and dynamic stability. Pond closure
strength, brittleness) Rare earth metals
Well: Horizontal drilling (shale gas) - Casing and cementation - Ver- Mine closure and abandoned mines:
ification - Stability during drilling and production Decommissioning and closure accelerating to help reach net zero
Monitoring: emissions
Deformations (surface & subsurface), microseismicity, fluid Closure = largely an uncontrolled process to date
pressure, temperature Decommissioning = ensure long term physical and chemical
Integration of monitoring data into reservoir management stability
Infrastructure: Foundations, pipelines, erosion control around well Potential re-use. Reclamation. Backfilling
pads Geo-environmental implications: acid mine drainage, methane
Waste disposal, containment, management (HF fluids, cuttings, pro- release
duced fines) Monitoring active and abandoned mines:
Shale gas Data collection and interpretation (surface, within mine, and
Fracking: water demand (1-to-30x106 liters of water per well) borehole)
Contamination (chemical additives for fracking) Deformations (surface and subsurface, regional subsidence, creep)
Induced seismicity Microseismicity and stress redistribution
Evolution of fractures in pre-structured shales Integration of monitoring data into mine management
Early drop in production - Overworking wells - CO2 injection Note: * Atomic mas (C:12, O:16) 3.7 kg of CO2 per kg of C-burnt
CH4 leakage (0.4% for well completion - total may reach 8% of the
produced gas)
Hydrates
Understanding hydrate nucleation and growth in sediments
Physical properties: compressibility, fluid and thermal conductivities Table A4: Solar (~1% of worldwide energy consumption)*
Production (mixed fluid flow: water and gas - fully coupled THCM)
Hydrate dissociation, gas-driven fractures, gas migration Insolation: 150-to-300 W/m2 (primarily between the
Production methods (the petroleum production framework does Solar tropics)
not apply) power Total earth insolation: 6500 times our total energy
CO2-CH4 replacement consumption
Fines migration - Clogging - Sand production Heating, bio-photosynthesis, photovoltaics (power
Multi-well depressurization strategies Harnessing stations can exceed 200MW) concentrated solar
Volume change during hydrate dissociation - Subsidence predic- power plants (10 MW units)
tion Geotechnical issues
Casing buckling and shear triggered by reservoir subsidence Solar panels (1-to-3m above ground)
Environmental hazard: seafloor stability and gas release Loads: low, but consider uplift
to the environment Key: low cost and high installation rate. Types: fin piles, helical
Notes: * Volume measured at: T=60°F and P=100 kPa piles
Frequently changing due to new findings and technological Concentrated solar power - Solar towers exert large moments
developments, e.g., shale gas Solar-thermal energy systems: Near surface geothermal storage of
Assumes methane hydrates CH4. H2O where 6 heat captured using solar collectors. Geo-implications: HTM
Conventional natural gas reservoirs: refer to Table A1 on Oil coupling (see Table on Energy Storage)
Solar Ponds: pools of saltwater can be used for thermal energy col-
lectors and storage. Require geomembranes, geosynthetic or clay
liners to seal the bottom of the pond
Note: * Rapid increase - updated annually
Refer to table on Energy Storage
108
108
Table A5: Wind (~1% of worldwide energy consumption and ~2.5% of Table A7: Geothermal (0.2-to-0.3% of worldwide energy consumption)
electricity)*
Shallow (<0.1 km)*: use the subsurface thermal capacitance
Wind turbines: may be up to ~150m diameter, as large as for space heating and cooling. Requires Ground Source
Produc- 8MW Heat Pumps GSHPs
tion Wind farms: some exceed 1GW Hydrothermal systems (intermediate depths): Volcanic / tec-
Advantage: day and night tonic regions, primarily provision of space heating and/or
Ex- domestic/industrial hot water
Types
Worldwide: most estimates exceed the current total en-
tracta- Deep geothermal energy systems (~3-5 km, temperatures
ergy consumption > 17 TW
ble 150 C<T< 350 C)
In Worldwide: installed 650 GW Produce steam (direct use or to generate electricity)
place * USA: installed 105 GW Hot dry rock HDR formations: high geothermal gradi-
ents, volcanic/tectonic
Wind power P is a function of the 1
Ener- Engineered to enhance conductivity (Enhanced Geother-
area A, air mass density , and P A v3
getics 2 mal Systems EGS)
wind speed v to the third power
Average geothermal gradient iT 25 C/km. Rocks thermal
Geotechnical issues
conductivity k 1-3W/(m C). Soils thermal conductivity
Resources worldwide
Onshore and offshore foundations
Sediment characterization - Material properties k 0.4-2W/(m C)
Stress-dependent stiffness 98% of the Earth volume is above 1000 C
Repetitive load (wind and waves) – ratcheting, terminal densities Average heat flux from the Earth is ~0.08W/m2
Constitutive models Thermal yield from ground heat exchangers is ~30 to
Analyses - Numerical modeling 150 W/m & ~5 to 40 W/m2
Large horizontal loads and overturning moment (wind and wave Potential : 4x106 MW of electricity (z<3 km); same order for
action) direct heating
Cyclic loading and large number of repetitions Installed: 12x103 MW of electricity; 30x103 MW of heating
Monitoring short and long term performance (fluid pressure, defor- Geotechnical issues: Shallow
mations) Characterization: thermal conductivity, heat capacity, ground tem-
Note: * Rapid increase - updated annually perature, and groundwater regime (Note: heat ownership will be
Refer to Table on Energy Storage an issue in urban settings)
Coupled HTCM processes (e.g., pile-soil interaction, thermo-me-
chanical ratcheting)
Engineering: piping and backfill material for optimal heat ex-
change, constructability, design of energy geostructures
Optimal operation: target a constant yearly average ground tempera-
Table A6: Bio Fuels (7.7% of worldwide energy consumption; >90% ture (balanced winter/summer thermal loads), aim to reduce cap-
of heat generated from renewables) ital cost
Geotechnical issues: Deep EGS
Combustible renewables (e.g., sugar beet, corn, wood, bi- Characterization: formation, temperature, preexisting fracture net-
ogenic gas) works and geo-plumbing. Sampling and core recovery
Collateral: land use, water consumption, effect on food Rock properties/behavior at high temperature (400 C) and pressure
Sources supply (100 MPa) Transition from brittle to ductile response
Caution: subsidized development that is not energeti- Drilling to great depths (3-10 km); rock ductility, steering precision,
cally advantageous casing costs for long laterals, accelerated casing corrosion
Waste (Note: not necessarily carbon neutral) Reservoir engineering: fracturing to attain ideal conductivity in the
Various bio-mass sources eV= 15 5 MJ/kg mass size of interest (flow rates vs. thermal breakthrough and
Paper, cardboard eV= 10 MJ/kg short-circuiting, well spacing). Avoid: surface damage, induced
Energy
seismicity, pollutant migration
den- Plastic eV= 30 10 MJ/kg
Coupled HTCM processes:
sity* Municipal solid waste eV 12 MJ/kg or eV
Stress changes, tensile fracturing, shear (lithological interfaces)
14,000 MJ/m3
Mineral dissolution and reprecipitation, creep
Geotechnical issues: Landfills Constitutive models to capture HTCM response
Waste composition and time-varying physical properties Time-dependent behavior. Long-term performance
Liners (clays, geotextiles). Coupled THCBM processes. Consti- Analysis to upscale plant capacity
tutive models. Subsidence, well failure. Induced seismicity
Volume change during decomposition, subsidence Optimal operation (temporary depletion if heat recovery faster than
Monitoring: deformations, P&T, leakage recharge)
Geotechnical issues: Agriculture Waste reinjection (waste during drilling, and brine during produc-
Unsaturated soil behavior - Properties and property variation tion)
with saturation Monitoring: deformations (surface and subsurface), induced seis-
Coupled THCBM processes - Root-soil interaction micity (during hydraulic fracturing, and heat extraction and local
Erosion control. Desertification. cooling), fluid P&T
Geomechanical tool optimization, tire-soil interaction Integration of monitoring data into reservoir management
Monitoring (local, remote)
Note: * Refer to Table on Energy Storage
Note: * Nominal values.
Estimates vary by orders of magnitude
Includes biogenic and non-biogenic contributions. The mix-
ture varies
109
109
Table A8: Hydroelectric Power (2.6% of worldwide energy consump- Table A9: Nuclear Energy and Waste (4.3% of worldwide energy con-
tion (14% of global electricity) sumption, 10.1% of global electricity)
Dams: producible capacity is almost saturated with cur-
rent and planned dams U: radioactive chemical element (atomic weight 92)
Possible alternative (for energy storage): use of shafts Found in minerals (oxides – mined when >1000 ppm)
Resources and abandoned mines Uranium and seawater (3 ppb)
Installed capacity worldwide: 400 GW (China, Canada, Uranium-235 has a half-life of 7x108 years (concentra-
Brazil, USA) tion 0.7% in natural uranium)
Hydroelectric power plants: some exceed 10 GW Energet- Energy density for 235U is 80x106 MJ/kg (effective: 2-
Geotechnical issues: Dams ics to-5 105 MJ /kg)
Classical problems: piping, dispersion, toe instability, filters, frozen In minerals: 5x106 tons. In seawater: 4 x109 tons
ground, overtopping, differential settlement, uplift, seepage, intel- Annual production 2020: 47.7 x103 tons
Reserves
ligent compaction Producers: Kazakhstan, Canada, Australia, Namibia,
For storage (pump-hydro): consider repetitive load, fast drawdown Niger, Russia, Uzbekistan
Reservoir sedimentation and capacity loss – Dredging (reusable sedi- Context
ments) Electricity was first generated from a nuclear plant in 1951, electric-
Maintenance and retrofitting of existing dams (inadequate perfor- ity supplied to the grid in 1954
mance, modified land use, revised codes and standards, conse- Nuclear power plants in operation: 442. Average size: 0.89 GW
quences of long-term processes such as dissolution and repetitive Nuclear waste: (1) storage (USA, Germany), (2) reprocessing
loads) (France, UK, China, Japan, India, Pakistan, Russia), (3) direct dis-
Dynamic stability (shells and foundation) – Stability berms posal (Sweden, Finland)
Drainage – Erosion control There is no nuclear high-level waste repository in operation in the
Leaks and leak repair – Cutoff walls world (for commercial use). Finland and Sweden have been given
the go-ahead for the implementation of a deep geological reposi-
Geotechnical issues: Tidal tory.
Tidal Systems (underwater turbines, floating and fixed systems) Critical time for waste fuel storage: ~100 years
Anchoring in soft marine clays Design horizon for waste repositories: 10,000 yr to 1,000,000 yr
Cavitation and scour
Loads: repetitive dynamic load; high drag and drag-induced mo- Commercial accidents*
ments Three Mile Island 1979 (level 5 in the INES scale), Chernobyl 1986
Complemented with energy storage (refer to Table on Energy Storage) and Fukushima Daiichi 2011 (level 7 in the INES scale)
Minor accidents are more common (e.g., leaks from spent fuel
Geotechnical issues: Both pools)
Site characterization and physical properties
Tidal: stratigraphy, soft sediments, erodability and mass transport Geotechnical issues
Dams: fractures and relic river channels (geo-plumbing), abutment Common to Mining, Infrastructure, and Waste Storage:
stability, stability of slopes around reservoir Site characterization - Geo-plumbing (migration and dispersion)
Geo-sustainability: environmental assessment, life cycle assessment Local environmental conditions - Baseline conditions before de-
Monitoring: deformations (surface and subsurface), fluid pressure, sign
leaks Material properties (long-term response needed for waste stor-
Integration of monitoring data into resource management age)
Infrastructure development: power lines Monitoring: infrastructure, spent fuel pools, waste repositories
Deformations (wet-dry, freeze-thaw) - Thermal field
Leakage and accidental release - Fluid chemistry and geophys-
ical detection
Detailed monitoring program to detect changes from baseline
Long-term monitoring related to the special conditions (flood-
ing, freezing)
Integration of monitoring data into system reliability strategy
Mining: excavation and handling of tailings
Infrastructure / Foundations:
Static and seismic conditions – Natural hazards
Heat absorption/release (new generation reactors)
Design for life-cycle needs (monitoring, access, repair) and for
decommissioning
Flooding protection and mitigation (most designs place SMR’s
below grade)
Stability of engineered backfill surrounding the SMR (flood-
ing)
Regional and local subsidence: water, oil and gas extraction,
and karst processes
Nuclear waste storage (short time to 1,000,000 yr)
Potential geological formations: salt, hard rock, or clay
Requirements: self-healing, mineral HTCM stability, seismic
stability, biological shield, high heat conduction, low fluid
conductivity, and radionuclide retention
Coupled HTCM processes: understanding, properties, and
modeling
Design for retrievability
Remediation
Note: * several military accidents have been documented
110
110
Table B1: Energy Storage Table B2: Geological Storage of CO2
Satisfy: daily peaks in demand (optimize power plant Global anthropogenic CO2 emissions: 33x1012 t/yr
operation) Current CO2 concentration in atmosphere ~400 ppm.
Need
Accommodate: intermittent renewable energy sources Increase: 2ppm/yr
Storage scale: hourly, daily, and seasonal A CO2 concentration of 550 ppm could trigger severe climate
effects (IPCC, 2018)
National/commercial: chemical storage in salt caverns/domes, There is no low cost and scalable technology to capture CO2
Situation
aquifers or depleted reservoirs (e.g., US Strategic Petroleum from flue gas*
Methods
(scales)
Reserve) More than 50 CO2 geostorage pilot projects worldwide
Urban: pumped hydro storage, compressed air energy storage, CO2 injection is common practice in oil production
molten salt Supercritical: deep saline aquifers (non-usable), depleted oil
Residential: distributed thermal energy storage and gas reservoirs (enhances oil production and cap rock
Given: energy density eV [J/m3] of acts as seal), low-grade or unmineable coal seams
Methods and target
Volume*
adopted system, and stored energy E P t (promotes CH4 recovery)
CO2 Geostorage:
V Liquid CO2 pools in deep ocean (> 3000 m)
E [J], or delivered power P [W] eV eV
and supply duration [s]: Hydrate: inject into low temperature deep ocean sediments or
formations
permafrost to form CO2 hydrate. Injection into CH4
Hydrogen, methane, gasoline CH4: 56 MJ/kg hydrate-bearing layers to replace CH4 with CO2
Chemical
ammonia. Energy densities H2: 120 MJ/kg Chemically trapped: carbonation, natural sinks (trees, algae),
shown per mass(refer to text NH3: 22.5 MJ/kg coal, organic-rich shales
for volumetric data) Gasoline: 56 MJ/kg Geotechnical issues
Fluid unit weight [kN/m3] raised Identification and characterization of suitable formations
Pumped
Hydro
to elevation [m]. Continuity and long-term stability of seal layers (e.g., shale, salt
Example: for water elevated by
eV H
rock )
3
=100m, eV = 1 MJ/m Porosity (tradeoff between injectability and capillary trapping)
Cycle’s minimum and maximum Long-term physical properties evolution (e.g., 1,000 to 10,000 yr)
Compress
P
ed Air
pressures Pmin and Pmax [kPa]. Engineered injection to reduce viscous fingering, increase
Example: if Pmin=4 MPa &
eV Pmax ln max storativity, minimize dissolution, hinder leakage
Pmax=7 MPa, then eV = 4 MJ/m3
Pmin
Coupled HCM processes, such as: mixed fluids and fingered
Sensible heat : material mass invasion; acidification and mineral dissolution; convection
density [kg/m3], specific heat sustained dissolution; reactive fluid transport, piping and localized
capacity Cp [kJ/(kg.°C)], and eV Cp T fluid flow; coal swelling, change in permeability, local
change in temperature [°C]. deformations, development of strain discontinuities such as shear
Example: for water subjected to localization in contraction.
Thermal
=10°C, eV = 42 MJ/m3 Long-term performance of grouts and well plugs
Latent heat - Phase change: Monitoring: subsurface plume tracking, leak detection, deformations
material latent heat for phase (surface and subsurface), microseismicity, fluid P&T and
transformation L [kJ/kg] and chemistry
mass density [kg/m3].
eV L Integration of monitoring data into storage management
Example: ice water eV = 305 Sealing strategies
MJ/m3 Note: * Target: 90% reduction in emissions. Capture and compression
Geotechnical issues reduces power plant efficiency
Characterization: Stratigraphy, geo-plumbing (fracture networks), regional
conditions
Material properties under repetitive thermal, pressure, humidity, effective
stress cycles
Time-dependent behavior (creep, ratcheting, progressive deterioration)
Proofing existent caverns and depleted porous reservoirs for reuse as
storage vessels
Caprock integrity, roof stability and cavern integrity
Design for coupled HTCM processes. Consider working pressure and
temperature range, loading/unloading rate, cyclic multidirectional flow,
mixed fluid composition, repetitive loading, cementation loss, fines and
clogging, accelerated dissolution
Cavity formation/construction with minimal disturbance
Site pre-engineering and preconditioning to enhance subsurface conditions
Monitoring: incremental deformation (surface & subsurface),
microseismicity, fluid P&T. Integration of monitoring data into reservoir
management
Leak monitoring and repair (including bio-mediated processes)
Note: * Energy density shown as J/m3 to facilitate storage volume
computation
The transforming of heat into a higher form of energy is
limited by the Carnot efficiency C=1-(Tlow/Thigh) where the
cycle's low and high temperatures are in degrees Kelvin.
Hence, high Thigh value is sought, as in molten salt.
Refer to Table on Geothermal
111
111
Table C1: Energy Conservation and Efficiency in Geotechnical 10. REFERENCES
Engineering - Embodied Energy
Acosta, M., Maye, R. and Violay, M. 2020. Hydraulic transport through
Conservation: important effect in developed nations with calcite bearing faults with customized roughness: effects of normal
overconsumption* and shear loading. Journal of Geophysical Research: Solid Earth,
125(8), e2020JB019767.
Efficiency complements conservation
Adachi, J., Siebrits, E. M., Peirce, A. and Desroches, J. (2007). Computer
Embodied energy parallels embodied CO2 (due to fossil fuel
simulation of hydraulic fractures. International Journal of Rock
dependency)
Mechanics and Mining Sciences, 44(5), 739-757.
The use of Portland cement increases the embodied CO2 in
Aftab A., Hassanpouryouzband A., Xie, Q., Machuca, L. L. and
infrastructure Sarmadivaleh, M. 2022. Toward a Fundamental Understanding of
General
Some construction processes are very inefficient Geological Hydrogen Storage. Industrial & Engineering Chemistry
Crushing: only 2-to-5% of the consumed energy used to Research, 61(9): 3233-3253.
generate surfaces Alberdi-Pagola, M., Poulsen, S.E., Loveridge, F., Madsen, S. and Jensen,
R. L. (2018). Comparing heat flow models for interpretation of precast
Biological processes have developed optimal strategies, often
Biomimetic
quadratic pile heat exchanger thermal response tests. Energy, 145,
to minimize energy 721–733.
Roots: optimal, self-adaptive, self-sensing and self-healing Altmann, J. B., Muller, T. M., Muller, B. I., Tingay, M. R. and Heidbach,
foundation systems O. (2010). Poroelastic contribution to the reservoir stress path.
International Journal of Rock Mechanics and Mining Sciences, 47 (7),
Geotechnical issues 1104-1113.
Efficient use of natural resources (e.g., aggregates). Reduced volume Amatya, B.L., Soga, K., Bourne-Webb, P.J., Amis, T. and Laloui, L.
extraction (2012). Thermo-mechanical behaviour of energy piles. Géotechnique,
Prefer materials with low embodied energy intensity (high in 62, 503–519.
concrete and steel) Atig, D., Broseta, D., Pereira, J. M. and Brown, R. 2020. Contactless
Energy-efficient construction practices probing of polycrystalline methane hydrate at pore scale suggests
Energy return on investment EROI (considers all invested energy). weaker tensile properties than thought. Nature Communications,
Waste recycling/reutilization: engineering waste reuse for long-term 11(1): 3379.
performance Balat H. –present scenario
Observational approach: monitoring as an integral component of and future prospects. International Journal of Hydrogen Energy, vol.
energy efficient design and construction practice 35, no. 14, pp. 7416-7426.
Note: * Overconsumption is defined as power consumption rates Balta, M. T., Dincer, I. and Hepbasli, A. 2009. Thermodynamic
above values needed to attain the highest quality of life. assessment of geothermal energy use in hydrogen production.
Given today’s technology, very high quality of life can be International Journal of Hydrogen Energy, vol. 34, no. 7, pp. 2925-
attained at ~3000 W/person. 2939.
Portland cement: CO2 from heating with fossil fuels, and Balta, M. T., Dincer, I. and Hepbasli, A. 2010. Potential methods for
CO2 released from carbonates CaCO3 CaO + CO2 geothermal-based hydrogen production. International Journal of
Hydrogen Energy, vol. 35, no. 10, pp. 4949-4961.
Barla, M. and Di Donna, A. 2018. Energy tunnels: Concept and design
aspects. Undergr. Sp.
Barla, M., Di Donna, A. and Perino, A. 2016. Application of Energy
Tunnels to an Urban Environment. Geothermics, 61: 104–13.
Barría, J. C., Manzanal, D., Cerrutti, P. and Pereira, J.M. 2021. Cement
with Bacterial Nanocellulose Cured at Reservoir Temperature:
Mechanical Performance in the Context of CO2 Geological Storage.
Geomechanics for Energy and the Environment, 100267.
Berta, M., Dethlefsen, F., Ebert, M., Schäfer, D. and Dahmke, A. 2018.
Geochemical effects of millimolar hydrogen concentrations in
groundwater: an experimental study in the context of subsurface
hydrogen storage. Environmental Science & Technology, 52, 4937-
4949.
Bidarmaghz, A., and Narsilio, G. A. 2018. Heat exchange mechanisms
in energy tunnel systems. Geomech. Energy Environ. 16, 83–95.
Bidarmaghz, A., Choudhary, R., Soga, K., Kessler, H., Terrington, R. L.
and Thorpe, S. 2019. Influence of geology and hydrogeology on heat
rejection from residential basements in urban areas. Tunnelling and
Underground Space Technology, 92, p.103068.
Birkholzer, J. T., Oldenburg, C. M. and Zhou, Q. 2015. CO2 migration
and pressure evolution in deep saline aquifers. International Journal
of Greenhouse Gas Control, 40, 203-220.
Birkholzer, J. T., Zhou, Q., and Tsang, C. F. 2009. Large-scale impact of
CO2 storage in deep saline aquifers: A sensitivity study on pressure
response in stratified systems. International Journal of Greenhouse
Gas Control, 3 (2), 181-194.
Blackwell, D., Richards, M., Frone, Z., Batir, J., Ruzo, A., Dingwall, R.
and Williams, M. 2011. Temperature at depth maps for the
conterminous US and geothermal resource estimates. GRC
Transactions, 35 (GRC1029452).
Bourne-Webb, P. J., Amatya, B., Soga, K., Amis, T., Davidson, C. and
Payne, P. 2009. Energy pile test at Lambeth College, London:
geotechnical and thermodynamic aspects of pile response to heat
cycles. Géotechnique, 59(3), pp.237-248.
Bourne-Webb, P., Burlon, S., Javed, S., Kürten, S. and Loveridge, F.
2016. Analysis and design methods for energy geostructures, Renew.
Sustain. Energy Rev, 65, 402–419.
Brandl, H. 2006. Energy foundations and other thermo-active ground
structures. Géotechnique. 56 (81–122).
112
112
Brandl, H. 2016. Geothermal Geotechnics for Urban Undergrounds, 251.
Procedia Eng. 165, 747–764. Durminey, B. 2012. EGS Takes Geothermal Global.
Brumfiel G., 2003. Hydrogen cars fuel debate on basic research. Nature, Eaton, D. W., Igonin, N., Poulin, A., Weir, R., Zhang, H., Pellegrino, S.
422:104-5. and Rodriguez, G. 2018. Induced seismicity characterization during
Budt, M., Wolf, D., Span, R. and Yan, J. 2016. A review on compressed
air energy storage: Basic principles, past milestones and recent array and broadband sensors. Seismological Research Letters, 89(5),
developments. Applied Energy 170, 250-268. 1641-1651.
Caglayan D. G., Weber, N., Heinrichs, H. U., Linßen, J., Robinius, M. Eckard, R. 2010. CAES compressed air energy storage worldwide. SBI
and Kukla, P. A. et al. 2020. Technical potential of salt caverns for Energy: Rockville, MD, USA.
hydrogen storage in Europe. International Journal of Hydrogen Editorial. 2021. Geoscience on the Chopping Block. Nature Reviews
Energy. 45:6793-805. Earth & Environment 2 (9): 587–587.
Calpine Corporation. 2019. The Geysers Geothermal Field 2019 Ellsworth, W. L. 2013. Injection-induced earthquakes. Science,
Statistics. 341(6142).
Cardona, A., Finkbeiner, T., Santamarina, J.C. (2021), Natural Rock Elsworth, D., Spiers, C. J. and Niemeijer, A. R. 2016. Understanding
Fractures: From Aperture to Fluid Flow, J. Rock Mechanics and Rock induced seismicity. Science, 354(6318), 1380-1381.
Engineering, 54:5827–5844. Espinoza, D.N. and Santamarina, J.C. (2011). P-wave monitoring of
Cha, M and Santamarina, J.C. (2019). Localized Dissolution in hydrate-bearing sand during CH4-CO2 replacement, Int. J.
Sediments Under Stress, Granular Matter, vol. 21. Greenhouse Gas Control, vol. 5, pp. 1031–1038.
Chang, K. W. and Segall, P. 2016. Injection-induced seismicity on Espinoza, N. D., Vandamme, M., Pereira, J. M., Dangla, P. and Vidal-
basement faults including poroelastic stressing. Journal of Gilbert, S. 2014. Measurement and Modeling of Adsorptive–
Geophysical Research: Solid Earth, 121, 2708-2726. Poromechanical Properties of Bituminous Coal Cores Exposed to
Chang, K. W. and Segall, P. 2017. Reduction of injection-induced pore- CO2: Adsorption, Swelling Strains, Swelling Stresses and Impact on
pressure and stress in basement rocks due to basal sealing layers. Pure Fracture Permeability. International Journal of Coal Geology, 134–
and Applied Geophysics, 174(7), 2649-2661. 135: 80–95.
Chapman, N.A. & Mc Kinley, I.G. 1987. The geological disposal of nu- EU Paris Agreement, https://ec.europa.eu/clima/policies
clear waste. John Wiley: Chichester. Eyre, T. S., Eaton, D. W., Garagash, D. I., Zecevic, M., Venieri, M.,
Chong, S.H. and Santamarina J.C. (2016). Soil Compressibility Models Weir, R. and Lawton, D. C. 2019. The role of aseismic slip in
for Wide Stress Range, Journal of Geotechnical and Geoenv. hydraulic fracturing–induced seismicity. Science Advances, 5(8),
Engineering, ASCE vol. 142(6). eaav7172.
Cipriani G. et al. 2014. Perspective on hydrogen energy carrier and its Fadejev, J., Simson, R., Kurnitski, J. and Haghighat, F. 2017. A review
automotive applications. International Journal of Hydrogen Energy, on energy piles design, sizing and modelling. Energy, 122, 390–407.
vol. 39, no. 16, pp. 8482-8494. Flesch, S., Pudlo, D., Albrecht, D., Jacob, A. and Enzmann, F. 2018.
Clennell, M. B., Hovland, M., Booth, J. S., Henry, P. and Winters, W. J. Hydrogen underground storage—Petrographic and petrophysical
1999. Formation of natural gas hydrates in marine sediments: 1. variations in reservoir sandstones from laboratory experiments under
Conceptual model of gas hydrate growth conditioned by host sediment simulated reservoir conditions. International Journal of Hydrogen
properties. Journal of Geophysical Research: Solid Earth, 104 (B10). Energy, 43, 20822-20835.
Clifford, I. and Hunter, E. 2002. The System Ammonia–Water at Florides, G. and Kalogirou, S. 2006. Ground heat exchangers-A review
Temperatures up to 150° C. and at Pressures up to Twenty of systems, models and applications. Renew. Energy. 32, 2461–2478.
Atmospheres. The Journal of Physical Chemistry, 37:101-18. Forbes, B., Moore, J. N., Finnila, A., Podgorney, R., Nadimi, S. and
Collett T. S., Lewis, R. E., Dallimore, S. R., Lee, M. W., Mroz, T. H and McLennan, J. D. 2019. Natural fracture characterization at the Utah
Uchida, T. 1999. Detailed evaluation of gas hydrate reservoir FORGE EGS test site - Discrete natural fracture network, stress field,
properties using japex/jnoc/gsc mallik 2l-38 gas hydrate research well and critical stress analysis. Geothermal Characteristics of the
downhole well log displays. In: Dallimore S. R, Uchida T and Collett Roosevelt Hot Springs System and Adjacent FORGE EGS Site,
T. S, eds. Scientific results from JAPEX/JNOC/GSC Mallik 2L-38 gas Milford, Utah. Utah Geological Survey Miscellaneous Publication,
hydrate research well, mackenzie delta, northwest territories, Canada. 169.
In: Bull Geol Surv Can, vol. 544, p. 295–312. Garber, D., Choudhary, R. and Soga, K. 2013. Risk based lifetime costs
Collett, T. S., Boswell, R., Cochran, J. R., Kumar, P., Lall, M., assessment of a ground source heat pump (GSHP) system design:
Mazumdar, A., Ramana, M. V., Ramprasad, T., Riedel, M., Sain, K. Methodology and case study. Building and Environment, 60, pp.66-
and Sathe, A.V. 2014. Geologic implications of gas hydrates in the 80.
offshore of India: Results of the National Gas Hydrate Program Garcia, A.V. and Santamarina J.C. (2021), Heat Flow in Fractured
Expedition 01. Marine and Petroleum Geology, 58, 3-28. Rocks: Stress and Moisture-Dependent Thermal Contact Resistance,
Collett, T. S., Lee, M. W., Agena, W. F., Miller, J. J., Lewis, K. A., Geothermics. doi.org/10.1016/j.geothermics.2021.102113
Zyrianova, M. V., Boswell, R. and Inks, T. L. 2011. Permafrost- García-Siñeriz, J. L., Villar, M. V., Rey, M. and Palacios, B. 2015.
associated natural gas hydrate occurrences on the Alaska North Slope. Engineered barrier of bentonite pellets and compacted blocks: state
Marine and Petroleum Geology, 28 (2), 279-294. after reaching saturation. Engineering Geology, 192: 33-45.
Dai, S., Santamarina, J. C., Waite, W. F. and Kneafsey, T. J. 2012. García-Siñeriz, J.L. & Tuñon, S. 2020. HotBENT: Experiment plan.
Hydrate morphology: Physical properties of sands with patchy hydrate Nagra Arbeitsbericht. NAB 19-243. Nagra, Wettingen, Switzerland.
saturation. Journal of Geophysical Research: Solid Earth, 117 (B11), Garitte, B., Shao, H., Wang, X. R., Nguyen, T. S., Li, Z., Rutqvist, J.,
B11205. Birkholzer, J., Wang, W. Q., Kolditz, O., Pan, P. Z., Feng, X. T., Lee,
Dai, X., Bidarmaghz, A. and Narsilio, G. 2022. Energy tunnels: a review C., Graupner, B. J., Maekawa, K., Manepally, C., Dasgupta, B.,
of analysis and design approaches to harness energy from Stothoff, S., Ofoegbu, G., Fedors, R. and Barnichon, J. D. 2017.
infrastructure, Renewable and Sustainable Energy Reviews, Evaluation of the predictive capability of coupled thermo-hydro-
forthcoming. mechanical models for a heated bentonite/clay system (HE-E) in the
Di Donna, A. D., F. Cecinato, F. Loveridge and M. Barla 2017. "Energy Mont Terri Rock Laboratory. Environmental Earth Sciences, 76: 64.
performance of diaphragm walls used as heat exchangers." Gaus, I., Garitte, B., Senger, R., Gens, A., Vasconcelos, R., Garcia-
Proceedings of the Institution of Civil Engineers - Geotechnical Sineriz, J. L., Trick, T., Wieczorek, K., Czaikowski, O., Schuster, K.,
Engineering 170(3): 232-245. Mayor, J. C., Velasco, M., Kuhlmann, U. and Villar, M. V. 2014. The
Dincer I. and Acar, C. 2017. Innovation in hydrogen production. HE-E Experiment: Lay-out, Interpretation and THM Modelling
International Journal of Hydrogen Energy, vol. 42, no. 22, pp. 14843- Combining D2.2-11. Final Report on the HE-E Experiment and D3.2-
14864. 2: Modelling and Interpretation of the HE-E Experiment of the PEBS
DOE-NETL. 2017. Best Practices for Site Screening, Selection, and Project. NAGRA Working Report NAB, 14-53, Wettingen, Switzerland
Initial Characterization for Storage of CO2 in Deep Geologic and EU PEBS.
Formations. United States Department of Energy National Energy Gawecka, K. A., Taborda, D. M. G., Potts, D. M., Cui, W., Zdravkovic´,
Technology Laboratory 2017; DOE/NETL-2017/1844. L. and Haji Kasri, M. S. (2017). Numerical modelling of thermo-active
Dong, S., Li, X., Tang, A. M., Pereira, J. M., Nguyen, V.T., Che, P. and piles in London clay. Proc. Inst. Civ. Engrs – Geotech. Engng 170,
Xiong, Z. 2019. Thermo-mechanical behavior of energy diaphragm No. 3, 201–219.
wall: Physical and numerical modelling. Appl. Therm. Eng, 146 243– GCCSI (Global CCS Institute). 2020. The Global Status of CCS, Report,
113
113
Global Carbon Capture and Storage Institute. Permanent Carbon Dioxide Storage in Deep-Sea Sediments.
Gens, A. 2003. The role of geotechnical engineering in nuclear energy Proceedings of the National Academy of Sciences, 103 (33): 12291–
utilization. Proc. 13th. Eur. Conf. Soil Mech. Geotech. Engng, 95.
(Vanicek et al. eds.), CGtS, Prague. 3, 25–67. Hubbert, M. K. and Willis, D. G. 1957. Mechanics of hydraulic
Gens, A. 2010. Soil-environment interactions in geotechnical fracturing. AIME Trans., 210, 153–168.
engineering. 47th Rankine Lecture. Géotechnique, 60, 3-74. Huttrer, G. W. 2020. Geothermal power generation in the world 2015-
Gens, A., de Vasconcelos, R.B. & Vaunat, J. (2022). Towards higher 2020 update report. In Proceedings, World geothermal congress (Vol.
temperatures in nuclear waste repositories. Proc. 2nd International 2020, pp. 1-17).
Conference on Energy Geotechnics, San Diego (in press) IAEA 2005. Draft safety requirements: geological disposal of
Gens, A., Sánchez, M., Guimaraes, L., Alonso, E. E., Lloret, A., Olivella, radioactive waste. Report no. WS-R-4, International Atomic Energy
S., Villar, M. V. and Huertas, F. 2009. A full-scale in situ heating test Agency, Vienna.
for high-level nuclear waste disposal. Observations, analysis and IAEA 2021. Nuclear power reactors in the world. Reference data series
interpretation. Géotechnique, 59(4): 377-399. no. 2. 2021 Edition. International Atomic Energy Agency, Vienna.
Gens, A., Valleján, B., Sánchez, M., Imbert, C., Villar, M. V. and Van Im, K., Avouac, J. P. and Heimisson, E. R. et al. 2021. Ridgecrest
Geet, M. 2011. Hydromechanical behaviour of a heterogeneous aftershocks at Coso suppressed by thermal destressing. Nature, 595,
compacted soil: experimental observations and modelling. 70–74.
Géotechnique, 61(5), 367-386. International Energy Agency (IEA). 2020. CCUS in Clean Energy
Gens, A., Vaunat, J., Garitte, B. and Wileveau., Y. 2007. In situ Transitions, IEA, Paris.
behaviour of a stiff layered clay subject to thermal loading: Jaeger, J. C., Cook, N. G. and Zimmerman, R. 2009. Fundamentals of
observations and interpretation. Geotechnique. 57, 207-228. rock mechanics. John Wiley & Sons.
Ghazvini, M., Sadeghzadeh, M., Ahmadi, M. H., Moosavi, S. and Jang, J., Cao, S. C., Stern, L. A., Waite, W. F., Jung, J. and Lee, J. Y.
Pourfayaz, F. 2019. Geothermal energy use in hydrogen production: 2020. Potential freshening impacts on fines migration and pore-throat
A review. International Journal of Energy Research, vol. 43, no. 14, clogging during gas hydrate production: 2-D micromodel study with
pp. 7823-7851. Diatomaceous UBGH2 sediments. Marine and Petroleum Geology,
Grigoli, F., Cesca, S., Rinaldi, A. P., Manconi, A., Lopez-Comino, J. A., 116, 104244.
Clinton, J. F. and Wiemer, S. 2018. The November 2017 M w 5.5 Johannesson, L. E., Börgesson, L., Goudarzi, R., Sandén, T.,
Pohang earthquake: A possible case of induced seismicity in South Gunnarsson, D. and Svemar, C. 2007. Prototype repository: a full-
Korea. Science, 360(6392), 1003-1006. scale experiment at Äspö HRL. Physics and Chemistry of the Earth
Groenenberg, R. M., Koornnef, J. M., Sijm, J. P. M., Janssen, G. J. M., Parts A/B/C, 32(1-7): 58-76.
Morales-Espana, G. A., van Stralen, J., Hernandez-Serna, R., Johnston I.W., Narsilio, G.A. and Colls, S. 2011. Emerging geothermal
Smekens, K., JuezLarré, J., Goncalvez, C., Wasch, L., Dijkstra, H., energy technologies. KSCE J. Civ. Eng. 15, 643–653.
Wassing, B., Orlic, B., Brunner, L., van der Valk, K., van Unen, M., Juliusson, E. and Bjornsson, S. 2021. Optimizing production strategies
van der Meulen, T.H., Kranenburg-Bruinsma, K., Winters, E., Puts, for geothermal resources. Geothermics, 94, 102091.
H., van Popering-Verkerk, J. and Duijn, M. 2020. Large-Scale Energy Jung, J., Kang, H., Cao, S. C., Al-Raoush, R. I., Alshibli, K. and Lee, J.
Storage in Salt Caverns and Depleted Fields (LSES)–Project Findings. Y. 2019. Effects of Fine-Grained articles’ Migration and Clogging in
Gu, X., Makasis, N., Motamedi, Y., Narsilio, G. A., Arulrajah, A. and Porous Media on Gas Production from Hydrate-Bearing Sediments.
Horpibulsuk, S. (2021). Geothermal pavements: field observations, Geofluids, 5061216
numerical modelling and long-term performance. Géotechnique, 1-15. Jung, J.W., Espinoza, D.N. and Santamarina, J.C. (2010). Properties and
Hafez, A., Liu, Q. Finkbeiner, T., Alouhali, R.A., Moellendick, T.E., phenomena relevant to CH2-CH4 replacement in hydrate-bearing
Santamarina, J.C. (2021), The Effect of Particle Shape on Discharge sediments, Journal of Geophysical Research, vol. 115, B10102,
and Clogging, Nature Scientific Reports. DOI: 10.1038/s41598-021- DOI:10.1029/2009JB000812
82744-w. Kalantidou, A., Tang, A. M., Pereira, J. M. and Hassen, G. 2012.
Hagemann, B., Rasoulzadeh, M., Panfilov, M., Ganzer, L. and Preliminary Study on the Mechanical Behaviour of Heat Exchanger
Reitenbach, V. 2016. Hydrogenization of underground storage of Pile in Physical Model. Géotechnique 62 (11): 1047-51.
natural gas. Computational Geosciences, 20, 595-606. Kato, Y., Koyama, M., Fukushima, Y. and Nakagaki, T. 2016. Energy
Hamm, S., Hass, E., Winick, J., Tasca, C., Albayrak, F. and Augustine, Technology Roadmaps of Japan, 537-547.
C. 2019. GeoVision: Harnessing the heat beneath our feet. US Kim, A. R., Kim, J. T., Cho, G. C and Lee, J. Y. 2018. Methane
Department of Energy.
Han, G., Kwon, T. H., Lee, J. Y. and Jung, J. 2020. Fines migration and lation on production wellbore stability.
pore clogging induced by single-and two-phase fluid flows in porous Journal of Geophysical Research: Solid Earth, 123 (11), 9555-9569.
media: From the perspectives of particle detachment and particle-level Kim, J., Lee, J. Y., Ahn, T. W., Yoon, H. C., Lee, J., Yoon, S. and
forces. Geomechanics for Energy and the Environment, 23, 100131. Moridis, G. J. 2021. Validation of strongly coupled geomechanics and
Hashemi, L., Blunt, M. and Hajibeygi. H. 2021. Pore-Scale Modelling gas hydrate reservoir simulation with multiscale laboratory tests. Int.
and Sensitivity Analyses of Hydrogen-Brine Multiphase Flow in J. Rock Mech. Min. Sci., 149, 104958.
Geological Porous Media. Scientific Reports, 11 (1): 8348. Kim, S. and Hosseini, S. 2015. Hydro-thermo-mechanical analysis
Hassanpouryouzband, A., Joonaki, E., Edlmann, K. and Haszeldine, R.S. during injection of cold fluid into a geologic formation. International
2021. Offshore Geological Storage of Hydrogen: Is This Our Best Journal of Rock Mechanics and Mining Sciences, 77, 220-236.
Option to Achieve Net-Zero? ACS Energy Letters, 6, 2181-2186. Kim, S. and Hosseini, S. 2016. Study on the ratio of pore-pressure/stress
Hawkes, C. D., Bachu, S., Haug, K. and Thompson, A. 2005. Analysis changes during fluid injection and its implications for CO 2 geologic
of in-situ stress regime in the Alberta Basin, Canada, for performance storage. Journal of Petroleum Science and Engineering, 149, 138-150.
assessment of CO2 geological sequestration sites, in Proceedings Kim, S., Espinoza, D. N., Jung, J., Cha, M. and Santamarina, J. C. 2019.
Fourth Annual Conference on Carbon Capture and Sequestration, 2- Carbon geological storage: Coupled processes, engineering and
5. monitoring. In Science of carbon storage in deep saline formations
Hayashi, F., Toda, Y., Kanie, Y., Kitano, M., Inoue, Y. and Yokoyama (pp. 383-407). Elsevier.
T, et al. 2013. Ammonia decomposition by ruthenium nanoparticles Kim, Y. M., Lee, J. H., Kim, S. J. and Favrat, D. 2012. Potential and
Chemical Science. 24:3124- evolution of compressed air energy storage: energy and exergy
30. analyses. Entropy, 14, 1501-1521.
Heinemann, N., Alcalde, J., Miocic, J.M., Hangx, S.J., Kallmeyer, J., Kleinberg, R. L., Flaum, C., Griffin, D. D., Brewer, P. G., Malby, G. E.,
Ostertag-Henning, C., Hassanpouryouzband, A., Thaysen, E.M., Peltzer, E. T. and Yesinowski, J. P. 2003. Deep sea NMR: Methane
Strobel, G.J. and Schmidt-Hattenberger, C. 2021. Enabling large-scale hydrate growth habit in porous media and its relationship to hydraulic
hydrogen storage in porous media–the scientific challenges. Energy & permeability, deposit accumulation, and submarine slope stability.
Environmental Science, 14, 853-864. Journal of Geophysical Research: Solid Earth, 108 (B10), 2508.
Hoteit, H., He, X., Yan, B. and Vahrenkamp, V. 2021. Optimization and Ko, J., Kim, S., Kim, S. and Seo, H. 2020. Utilizing building foundations
uncertainty quantification model for time-continuous geothermal as micro-scale compressed air energy vessel: Numerical study for
energy extraction undergoing re-injection. arXiv preprint mechanical feasibility. Journal of Energy Storage, 28, 101225.
arXiv:2112.05544. Kocis, I., Kristofic, T., Gebura, M., Horvath, G., Gajdos, M. and
House, K. Z., Schrag, D. P., Harvey, C. F. and Lackner, K. S. 2006. Stofanik, V. 2017. Novel deep drilling technology based on electric
114
114
plasma developed in Slovakia. In XXXIInd General Assembly and Makasis, N., Narsilio, G. A., Bidarmaghz, A., Johnston, I. W. and Zhong,
Scientific Symposium of the International Union of Radio Science Y. 2021. The importance of boundary conditions on the application of
(URSI GASS) (pp. 1-4). IEEE. soldier pile retaining walls and slabs as energy geo-structures,
Koike, M., Miyagawa, H., Suzuoki, T. and Ogasawara, K. 2012. Comput. Geotech. 120.
Ammonia as a hydrogen energy carrier and its application to internal Maroufmashat, A., Fowler, M., Khavas, S. S., Elkamel, A., Roshandel,
combustion engines. Sustainable vehicle technologies: driving the R. and Hajimiragha, A. 2016. Mixed integer linear programing based
green agenda,61-70. approach for optimal planning and operation of a smart urban energy
Kojima, Y. and Yamaguchi, M. 2020. Ammonia storage materials for network to support the hydrogen economy. International Journal of
nitrogen recycling hydrogen and energy carriers. International Hydrogen Energy, vol. 41, no. 19, pp. 7700-7716.
Journal of Hydrogen Energy,45:10233-46. Martín, P. L., Barcala, J. M. and Villar, M. V. 2014. Behaviour of a
Komai, T., Y. Yamamoto, and K. Ohga (2000), Dynamics of reformation bentonite barrier under thermal and hydraulic gradients: FEBEX
and replacement of CO2 and CH4 gas hydrates, Ann. N. Y. Acad. Sci., mock-up test, 16 years of operation. In: Khalili, N., Russell, A. and
912, 272–280. Khoshghalb, A. (Eds.): Unsaturated Soils: Research and
Kushnir, R., Dayan, A. and Ullmann, A. 2012. Temperature and pressure Applications, vol. 1: 857-862. Taylor & Francis Group, London.
variations within compressed air energy storage caverns. International
Journal of Heat and Mass Transfer, 55, 5616–5630. and Hartline, C. 2014. Spatiotemporal changes, faulting regimes, and
Laloui, L, Nuth, M. and Vulliet, L. 2006. Experimental and Numerical source parameters of induced seismicity: A case study from The
Investigations of the Behaviour of a Heat Exchanger Pile. Geysers geothermal field. Journal of Geophysical Research: Solid
International Journal for Numerical and Analytical Methods in Earth, 119 (11), 8378-8396.
Geomechanics, 30 (8): 763–81. Mastrucci, A., van Ruijven, B., Byers, E., Poblete-Cazenave, M. and
Laloui, L., and Di Donna, A. 2011. Understanding the behaviour of Pachauri, S. 2021. Global scenarios of residential heating and cooling
energy geo-structures. Proc. Inst. Civ. 164(4):184-191. energy demand and CO2 emissions. Climatic Change, 168 (3), pp.1-
Le, T. X., Bornert, M., Aimedieu, P., Chabot, B., King, A. and Tang, A. 26.
M. 2020. An experimental investigation on methane hydrate Maxwell, S. C., Jones, M., Parker, R., Leaney, S., Mack, M., Dorvall, D.
morphologies and pore habits in sandy sediment using synchrotron X- and Hammermaster, K. 2010, June. Fault activation during hydraulic
ray computed tomography. Marine and Petroleum Geology, 122, fracturing. In 72nd EAGE Conference and Exhibition incorporating
104646. SPE EUROPEC 2010.
Lei, L. and Santamarina, J. C. 2019. Physical properties of fine-grained McGarr, A. 2014. Maximum magnitude earthquakes induced by fluid
sediments with segregated hydrate lenses. Marine and Petroleum injection. Journal of Geophysical Research: Solid earth, 119(2),
Geology, 109, 899-911. 1008-1019.
Lei, L., and Santamarina, J. C. 2018. Laboratory strategies for hydrate McLean, M. L. and Espinoza, D. N. 2021. Depth Dependent Thermo-
formation in fine-grained sediments. Journal of Geophysical Poro-Elastic Response of Geothermal Reservoirs During Heat
Research: Solid Earth, 123 (4), 2583-2596. Extraction. In 55th US Rock Mechanics/Geomechanics Symposium,
Lei, L., Gai, X. and Seol, Y. 2020. Load-bearing characteristic of OnePetro.
methane hydrate within coarse-grained sediments – Insights from McLean M. L. and Espinoza, D. N. 2022. Geometrical Controls on
isotropic consolidation. Marine and Petroleum Geology, 121, 104571. Thermal Short-circuiting in Multi-fracture Enhanced Geothermal
Lei, L., Seol, Y. and Myshakin, E. M. 2019b. Methane hydrate film Systems. In 56th US Rock Mechanics, Geomechanics Symposium,
thickening in porous media. Geophysical Research Letters, 46(20), Santa Fe, New Mexico, USA.
11091-11099. Meys R., Kätelhön, A., Bachmann, M., Winter, B., Zibunas, C. and Suh
Lei, L., Seol, Y., Choi, J. H. and Kneafsey, T. J. 2019a. Pore habit of S, et al. 2021. Achieving net-zero greenhouse gas emission plastics by
methane hydrate and its evolution in sediment matrix – Laboratory a circular carbon economy. Science, 374:71-6.
visualization with phase-contrast micro-CT. Marine and Petroleum Mignan, A., Landtwing, D., Kästli, P., Mena, B. and Wiemer, S. 2015.
Geology, 104, 451-467. Induced seismicity risk analysis of the 2006 Basel, Switzerland,
Levasseur S., Collin F., Daniels K., Dymitrowska M., Harrington J., Enhanced Geothermal System project: Influence of uncertainties on
Jacops E., Kolditz O., Marschall P., Norris S., Sillen X., Talandier J., risk mitigation. Geothermics, 53, 133-146.
Truche L. and Wendling, J. 2021. Initial State of the Art on Gas Moeck, I., Kwiatek, G. and Zimmermann, G. 2009. Slip tendency
Transport in Clayey Materials. Deliverable D6.1 of the HORIZON analysis, fault reactivation potential and induced seismicity in a deep
2020 project EURAD, Work Package Gas. EC Grant agreement no: geothermal reservoir. Journal of Structural Geology, 31 (10), 1174-
847593. 364 pp. 1182.
Liu Q., Zhao, B. and J.C. Santamarina J.C. (2019). Particle Migration and Mohajerani, M., Delage, P., Sulem, J., Monfared, M., Tang, A.M. &
Clogging in Porous Media: A Convergent-Flow Microfluidics Study, Gatmiri, B. 2012. A laboratory investigation of thermally induced pore
Journal of Geophysical Research - Solid Earth. vol. 124, pp. 9495- pressures in the Callovo-Oxfordian claystone. Int. J. Rock Mech. &
9504. Mining Sci. 52, 112–121.
Liu Q and Santamarina JC, 2022. Fluid-Driven Instabilities In Granular Moridis, G. J., Kim, J., Reagan, M. T. and Kim, S. 2013. Feasibility of
Media, under review in Frontiers in Mechanical Engineering. gas production from a gas hydrate accumulation at the UBGH2-6 site
Available from the uthors. of the Ulleung Basin in the Korean East Sea. J. Pet. Sci. Eng. 108,
Lloret, A., Villar, M. V., Sánchez, M., Gens, A., Pintado, X. and Alonso, 180–210.
E. E. 2003. Mechanical behavior of heavily compacted bentonite Moridis, G.J., Collett, T.S., Pooladi-Darvish, M., Santamarina, J.C.,
under high suction changes. Géotechnique, 53(1): 27-40. Boswell, R., Kneafsey, T.J., Rutqvist, J., Reagan, M.T., Sloan, E.D.,
Loveridge, F. and Powrie, W. 2013. Pile heat exchangers: thermal Sum, A., and Koh, C. (2011). Challenges, Uncertainties, and Issues
behaviour and interactions. Proc. ICE - Geotech. Eng. 166, 178–196. Facing Gas Production From Gas-Hydrate Deposits. SPE Res Eval &
Loveridge, F., McCartney, J. S., Narsilio, G. A. and Sánchez, M. 2020. Eng 14(1): 76-112. SPE-131792-PA.
Energy geostructures: a review of analysis approaches, in situ testing Mostafaeipour A. et al. 2016. Evaluating the wind energy potential for
and model scale experiments. Geomech. Energy Environ, 22. hydrogen production: a case study. International Journal of Hydrogen
Mahmoud, M., Ramadan, M., Naher, S., Pullen, K., Abdelkareem, M.A. Energy, vol. 41, no. 15, pp. 6200-6210.
and Olabi, A. G. 2021. A review of geothermal energy-driven Muñoz-Criollo, J. J., Cleall, P. J. and Rees, S. W. 2016. Factors
hydrogen production systems. Thermal Science and Engineering influencing collection performance of near surface interseasonal
Progress, p.100854. ground energy collection and storage systems. Geomech. Energy
Majer, E. L., Baria, R., Stark, M., Oates, S., Bommer, J., Smith, B. and Environ. 6, 45–57.
Asanuma, H. 2007. Induced seismicity associated with enhanced Murphy, A., Soga, K. and Yamamoto, K. 2020. Experimental
geothermal systems. Geothermics, 36(3), 185-222. investigation into sand production from turbidite strata. Journal of
Majer, E., Nelson, J., Robertson-Tait, A., Savy, J. and Wong, I. 2012. Petroleum Science and Engineering, 190, p.107056.
Protocol for addressing induced seismicity associated with enhanced Nagaoka, K., Eboshi, T., Takeishi., Y, Tasaki, R., Honda, K. and
geothermal systems. US Department of Energy, 52. Imamura, K. et al. 2017. Carbon-free H2 production from ammonia
Makasis, N., Narsilio, G. A. and Bidarmaghz, A. 2018. A machine -Al2O3 catalyst.
learning approach to energy pile design. Comput. Geotech. 97, 189– Science Advances.3:e1602747.
203. NAGRA. 2019. Implementation of the Full-scale Emplacement
115
115
Experiment at Mont Terri: Design, Construction and Preliminary Sadokierski, S. and Thiffeault, J. L. 2008. Heat Transfer in Underground
Results. NAGRA Technical Report NTB 15-02. Wettingen, 147 pp. Rail Tunnels, 1–31.
Narsilio, G.A. and Aye, L. 2018. Shallow Geothermal Energy: An Sánchez, M., Gens, A. and Olivella, S. 2012. THM analysis of a large-
Emerging Technology, in: Sharma, A., Shukla, A. and Aye, L. (Eds.). scale heating test incorporating material fabric changes. International
Low Carbon Energy Supply. Green Energy Technol., Springer, Journal for Numerical and Analytical Methods in Geomechanics,
Singapore, pp. 387–411. 36(4): 391-421.
NEA 2013. Underground Research Laboratories (URL). Nuclear Energy Sánchez, M., Pomaro, B. and Gens, A. 2022. Coupled THM analysis of
Agency NEA No. 78122. OECD Publishing, Paris. a full-scale test for high-level nuclear waste and spent fuel disposal
Nguyen, V. T., Wu, N., Gan, Y., Pereira, J. M. and Tang, A. M. 2020. under actual repository conditions during 18 years of operation.
Long-Term Thermo-Mechanical Behaviour of Energy Piles in Clay. Géotechnique.
Environmental Geotechnics, 7 (4): 237–48. Santamarina, J.C. and Rached, R. (2021). Energy Geoscience and
Nicholson, D.P., Chen, Q., de Silva, M., Winter, A. and Winterling, R. Engineering, 16th International Conference of the International
2014. The design of thermal tunnel energy segments for Crossrail, UK. Association for Computer Methods and Advances in Geomechanics
Eng. Sustain. 167, 118–134. IACMAG, Eds. M. Barla, A. Di Donna, D. Sterpi. Springer, vol. 1, pp.
Nikolaidis P. and Poullikkas, A. 2017. A comparative overview of 75-97.
hydrogen production processes. Renewable and Sustainable Energy Sanyal, S. K. and Enedy, S. L. 2011. Fifty years of power generation at
Reviews, vol. 67, pp. 597-611. the Geysers geothermal field, California–the lessons learned. In 36th
Notton G., et al. 2018. Intermittent and stochastic character of renewable workshop on geothermal reservoir engineering, Stanford, California.
energy sources: Consequences, cost of intermittence and benefit of Sasaki, T., Shao, B., Elshafie, M., Papadopoulou, M., Yamamoto, K. and
forecasting. Renewable and Sustainable Energy Reviews, vol. 87, pp. Soga, K. 2021. Simulation of axial tensile well deformation during
96-105. reservoir compaction in offshore unconsolidated methane hydrate-
Olivella, S., Gens, A., Carrera, J. and Alonso, E. E. 1996. Numerical bearing formation. Computers and Geotechnics, 129, p.103894.
formulation for a simulator (CODE_BRIGHT) for the coupled Schneider, M. & Froggatt, A. 2021. The world nuclear industry. Status
analysis of saline media. Engineering Computations, 13(7): 87-112. Report 2021. A Mycle Schneider Consulting Project, Paris.
Pan, L., Lei, L. and Seol, Y. 2021. Pore-scale influence of methane (https://www.worldnuclearreport.org/World-Nuclear-Industry-
hydrate on permeability of porous media. Journal of Natural Gas Status-Report-2021-773).
Science and Engineering, 87. Segall, P. and Fitzgerald, S. D. 1998. A note on induced stress changes
Park, S., Vohs, J. M., Gorte, R. J. 2000. Direct oxidation of hydrocarbons in hydrocarbon and geothermal reservoirs. Tectonophysics, 289(1-3),
in a solid-oxide fuel cell. Nature, 404:265-7. 117-128.
Passelègue, F. X., Almakari, M., Dublanchet, P., Barras, F., Fortin, J. and Sellin, P. and Leupin, O. 2013. The use of clays as an engineered barrier
Violay, M. 2020. Initial effective stress controls the nature of in radioactive-waste management – a review. Clays and Clay
earthquakes. Nature Communications, 11 (1), 1-8. Minerals, 61 (6): 477–498.
Pasten, C.R., Santamarina, J.C. (2011). Energy Geo-Storage, KSCE Sellin, P., Westermark, M., Leupin, O., Norris, S., Gens, A., Wieczorek,
Journal of Civil Engineering, vol. 15, no. 4, pp. 655-667. K., Talandier, J. and Swahn, J. 2020. Beacon: bentonite mechanical
Pasten, C. and Santamarina, J. C. 2014. Thermally Induced Long-Term evolution. EPJ Nuclear Sci. Technol, 6, 23.
Displacement of Thermoactive Piles. ASCE Journal of Geotechnical Selvadurai, A. P. S., Couture, C. B. and Rezaei Niya, S. M. 2017.
and Geoenvironmental Engineering, vol. 140 (5). Permeability of Wormholes Created by CO2 Acidized Water Flow
Pastén, C., Castillo, E. and Chong, S.H. 2019. Thermo-Mechanical through Stressed Carbonate Rocks. Physics of Fluids, 29 (9): 096604.
Ratcheting in Soil–Structure Interfaces. Acta Geotechnica, 14 (5): Shafagh, I., Shepley, P., Shepherd, W., Loveridge, F., Schellart, A., Tait,
1561–69. S. and Rees S. J. 2022. Thermal energy transfer around buried pipe
Peysson, Y., André, L. and Azaroual, M. 2014. Well Injectivity during infrastructure, Geomechanics for Energy and the Environment,
CO2 Storage Operations in Deep Saline Aquifers—Part 1: Volume 29.
Experimental Investigation of Drying Effects, Salt Precipitation and Shapiro, S. A. and Dinske, C. 2009.
Capillary Forces. International Journal of Greenhouse Gas Control, diffusion and hydraulic fracturing. Geophysical Prospecting, 57 (2),
22, 291–300. 301-310.
Pimm, A. J., Garvey, S. D. and de Jong, M. 2014. Design and testing of Shin, H. and Santamarina, J. C. 2016. Sediment–well interaction during
energy bags for underwater compressed air energy storage. Energy, depressurization. Acta Geotechnica, 12(4), 883-895.
66, 496-508. SKB. 2010. Buffer, Backfill and Closure Process Report for the Safety
Pollack, H. N., Hurter, S. J. and Johnson, J. R. 1993. Heat flow from the Assessment SR-Site. TR-10-47. Swedish Nuclear Fuel and Waste
Earth's interior: analysis of the global data set. Reviews of Geophysics, Management Co (SKB), Stockholm, 360 pp.
31(3), 267-280. Soga, K. and Rui, Y. 2016. Energy geostructures, in: S. Rees (Ed.), Adv.
Rammal, D., Mroueh, H. and Burlon, V. 2018. Thermal behaviour of Ground-Source Heat Pump Syst., 1st ed., Woodhead Publishing, pp.
geothermal diaphragm walls: Evaluation of exchanged thermal power. 185–221.
Renew. Energy, 1–11. Song, J. and Zhang, D. 2013. Comprehensive review of caprock-sealing
Raupach, M. R., Marland, G., Ciais P., Le Quéré, C., Canadell, J. G. and mechanisms for geologic carbon sequestration. Environmental
Klepper, G. et al. 2007. Global and regional drivers of accelerating Science & Technology, 47(1), 9-22.
CO2 emissions. Proceedings of the National Academy of Sciences, Spangenberg, E., Kulenkampff, J., Naumann, R. and Erzinger, J. 2005.
104:10288-93. Pore space hydrate formation in a glass bead sample from methane
Rossi, E., Jamali, S., Wittig, V., Saar, M. O., and von Rohr, P. R. 2020. dissolved in water. Geophysical Research Letters, 32 (24), L24301.
A combined thermo-mechanical drilling technology for deep Spokas, K., Fang, Y., Fitts, J. P., Peters, C. A. and Elsworth, D. 2019.
geothermal and hard rock reservoirs. Geothermics, 85, 101771. Collapse of reacted fracture surface decreases permeability and
Rui, Y. and Soga, K. 2019. Thermo-hydro-mechanical coupling analysis frictional strength. Journal of Geophysical Research: Solid Earth,
of a thermal pile. Proceedings of the Institution of Civil Engineers- 124(12), 12799-12811.
Geotechnical Engineering, 172 (2), pp.155-173. Sterpi, D., Angelotti, A., Habibzadeh-Bigdarvish, O. and Jalili, D. 2018.
Rutqvist, J. and Stephansson, O. 2003. The role of hydromechanical Assessment of thermal behaviour of thermo-active diaphragm walls
coupling in fractured rock engineering. Hydrogeology Journal, 11 (1), based on monitoring data. J. Rock Mech. Geotech. Eng. 10, 1145–
7-40. 1153.
Rutqvist, J., Barr, D., Datta, R., Gens, A., Millard, M., Olivella, S., Stewart, M. A. and McCartney, J. S. 2013. Centrifuge Modeling of Soil-
Tsang, C. F. and Tsang, Y. 2005. Coupled thermal-hydrological- Structure Interaction in Energy Foundations. Journal of Geotechnical
mechanical analysis of the Yucca Mountain Drift Scale Test – and Geoenvironmental Engineering, 04013044.
comparison of field results to predictions of four different models. Sun, M., C., Xia, C. and Zhang, G. 2013. Heat transfer model and design
International Journal of Rock Mechanics and Mining Science, 42: method for geothermal heat exchange tubes in diaphragm walls.
680-697. Energy Build. 61, 250–259.
Ryu, B. J., Collett, T. S., Riedel, M., Kim, G. Y., Chun, J. H., Bahk, J. J., Tang, A.M., Cui, Y. J., and Le, T. T. (2008) A study on the thermal
Lee, J. Y., Kim, J. H. and Yoo, D.G. 2013. Scientific results of the conductivity of compacted bentonites. 671. Applied Clay Science, 41
second gas hydrate drilling expedition in the Ulleung basin (UBGH2). (3-4): 181-189.
Marine and Petroleum Geology, 47, 1-20. Taniguchi, M., Uemura, T. and Jago-on, K. 2007. Combined Effects of
116
116
Urbanization and Global Warming on Subsurface Temperature in Marine and Petroleum Geology, 120, 104566.
Four Asian Cities. Vadose Zone Journal, 6 (3):591-596. Winter, C. J. 2009. Hydrogen energy—Abundant, efficient, clean: A
Terzariol, M. and Santamarina, J.C. (2020). Multi-well Strategy for Gas debate over the energy-system-of-change. International Journal of
Production by Depressurization from Methane Hydrate-Bearing Hydrogen Energy, 34:S1-S52.
Sediments, Energy J. DOI.org/10.1016/j.energy.2020.119710. Xing, P., McLennan, J. and Moore, J. 2020. In-situ stress measurements
Terzariol, M. Park, JH, Castro GM and Santamarina, J.C. (2020). at the Utah frontier observatory for research in geothermal energy
Methane Hydrate-Bearing Sediments: Pore Habit and Implications, (forge) site. Energies, 13 (21), 5842.
Marine and Petroleum Geology, vol. 116, DOI: Yamamoto K., Terao T., Fujii T. and Ikawa M, et al. (2014). Operational
10.1016/j.marpetgeo.2020.104302. overview of the first offshore production test of methane hydrates in
Terzariol, M., Goldsztein, G. and Santamarina J.C. (2017). Maximum the Eastern Nankai Trough, Offshore Technology Conference, OTC-
Recoverable Gas from Hydrate Bearing Sediments by 25243-MS, Houston, Texas.
Depressurization, Energy, 141, pp. 1622-1628. Yamamoto, K., Kanno, T., Wang, X. X., Tamaki, M., Fujii, T., Chee, S.
Thomas, H. R. and He, Y. 1995. An analysis of coupled heat, moisture S., Wang, X. W., Pimenov, V. and Shako, V. 2017. Thermal responses
and air transfer in a deformable unsaturated soil. Géotechnique, 45: of a gas hydrate-bearing sediment to a depressurization operation. RSC
667-689 Adv. 7 (10), 5554–5577.
Ugueto, G. A., Wojtaszek, M., Huckabee, P. T., Savitski, A. A., Guzik, Yarushina, V. M., and Bercovici, D. 2013. Mineral Carbon Sequestration
A., Jin, G. and Haustveit, K. 2021. An integrated view of hydraulic and Induced Seismicity. Geophysical Research Letters, 40 (5): 814–
induced fracture geometry in Hydraulic Fracture Test Site 2. In 18.
Unconventional Resources Technology Conference, pp. 923-938. Yoon, H. C., Yoon, S., Lee, J. Y. and Kim, J. 2021. Multiple porosity
Valdes, J.R. and Santamarina, J.C. (2007). Particle Transport in a Non- model of a heterogeneous layered gas hydrate deposit in Ulleung
Uniform Flow Field: Retardation and Clogging, Applied Physics Basin, East Sea, Korea: A study on depressurization strategies,
Letters, vol. 90, no. 24. reservoir geomechanical response, and wellbore stability. Journal of
Valente, A., Iribarren, D. and Dufour, J. 2018. Harmonising the Natural Gas Science and Engineering, 96,104321.
cumulative energy demand of renewable hydrogen for robust Yun, T. S., Lee, C., Lee, J. S., Bahk, J. J. and Santamarina, J. C. 2011. A
comparative life-cycle studies. Journal of Cleaner Production, vol. pressure core based characterization of hydrate-bearing sediments in
175, pp. 384-393. the Ulleung Basin, Sea of Japan (East Sea). Journal of Geophysical
Vallin, V., Pereira, J. M., Fabbri, A. and Wong, H. 2013. Numerical Research: Solid Earth, 116 (B2), B02204.
Modelling of the Hydro-Chemo-Mechanical Behaviour of Zemlak, W. M., Lemp, S. P. and Thiessen, R. D. 2002. U.S. Patent No.
Geomaterials in the Context of CO2 Injection. International Journal 6,446,727. Washington, DC: U.S. Patent and Trademark Office.
for Numerical and Analytical Methods in Geomechanics, 37 (17): Zhang, J. 2021. Energy geo-storage: A large- to medium- and small-scale
3052–69. compressed air energy storage and geomechanial analysis. Ph.D.
van Renssen S. 2020. The hydrogen solution? Nature Climate Change, dissertation, University of Nebraska-Lincoln.
10:799-801. Zhang, J., Seo, H., Kim, S., Ko, J. and Kim, S. 2020. Experimental study
Venkataramani, G., Ramalingam, V. and Viswanathan, K. 2018. of pipe-pile-based micro-scale compressed air energy storage (PPMS-
Harnessing Free Energy From Nature for Efficient Operation of CAES) for a building. ICEGT, E3S Web of Conferences, vol. 205, p.
Compressed Air Energy Storage System and Unlocking the Potential 07012, EDP Sciences.
of Renewable Power Generation. Scientific Reports, 8:9981. Zhang, L., Yang, L., Wang, J., Zhao, J., Dong, H., Yang, M., Liu, Y. and
Vilarrasa, V. and Carrera, J. 2015. Geologic carbon storage is unlikely to Song, Y. 2017. Enhanced CH4 Recovery and CO2 Storage via
trigger large earthquakes and reactivate faults through which CO2 Thermal Stimulation in the CH4/CO2 Replacement of Methane
could leak. Proceedings of the National Academy of Sciences, Hydrate. Chemical Engineering Journal, 308: 40–49.
112(19), 5938-5943. Zhang, Y., Choudhary, R. and Soga, K. 2015. Influence of GSHP system
Vilarrasa, V., Olivella, S., Carrera, J. and Rutqvist, J. 2014. Long term design parameters on the geothermal application capacity and
impacts of cold CO2 injection on the caprock integrity. International electricity consumption at city-scale for Westminster, London. Energy
Journal of Greenhouse Gas Control, 24, 1-13. and Buildings, 106, pp.3-12.
Villar, M. V., Iglesias, R. J., García-Siñeriz, J. L., Lloret, A. and Huertas, Zheng, L., Xu, H., Rutqvist, J., Reagan, M., Birkholzer, J., Villar, M. V.
F. 2020. Physical evolution of a bentonite buffer during 18 years of and Fernández, A. M. 2020. The hydration of bentonite buffer material
heating and hydration. Engineering Geology, 264: 105408. revealed by modeling analysis of a long-term in situ test. Applied Clay
Villar, M. V., Sánchez, M. and Gens, A. 2008. Behaviour of a bentonite Science, 185: 105360.
barrier in the laboratory: experimental results up to 8 years and Zhong, Y., Narsilio, G.A., Makasis, N. and Scott, C. 2022. Experimental
numerical simulation. Physics and Chemistry of the Earth, 33: S476– and numerical studies on an energy piled wall: The effect of thermally
S485. activated pile spacing. Geomechanics for Energy and the
Violay, M., Passelègue, F., Spagnuolo, E., Di Toro, G. and Cornelio, C. Environment, Volume 29,100276,
2019. Effect of water and rock composition on re-strengthening of Zhou, M., Soga, K. and Yamamoto, K. 2020. Upscaling techniques for
cohesive faults during the deceleration phase of seismic slip pulses. fully coupled THM simulation and application to hydrate gas
Earth and Planetary Science Letters, 522, 55-64. production tests. Computers and Geotechnics, 124, p.103596.
Waite, W. F., Santamarina, J. C., Cortes, D. D., Dugan, B., Espinoza, N. Zhou, M., Soga, K., Yamamoto, K. and Huang, H. 2020. Geomechanical
D., Germaine, J., Jang, J., Jung, J. W., Kneafsey, T. J., Shin, H., Soga, responses during depressurization of hydrate-bearing sediment for-
K., Winters, W. J. and Yun, T. S. 2009. Physical properties of hydrate- mation over a long methane gas production period. Geomechanics for
bearing sediments. Reviews of Geophysics, 47 (4), RG4003. Energy and the Environment, 23, p.100111.
Wang, B. and Bauer, S. 2017. Compressed air energy storage in porous Zoback, M. D. and Gorelick, S. M. 2012. Earthquake triggering and
formations: a feasibility and deliverability study. Petroleum large-scale geologic storage of carbon dioxide. Proceedings of the
Geoscience, 23 (3), 306-314. National Academy of Sciences, 109(26), 10164-10168.
Wang, Z., Pereira, J. M. and Gan, Y. 2020. Effect of Wetting Transition Zoback, M. D., Townend, J. and Grollimund, B. 2002. Steady-state
during Multiphase Displacement in Porous Media. Langmuir, 36(9): failure equilibrium and deformation of intraplate lithosphere.
2449-2458. International Geology Review, 44 (5), 383-401.
Wentorf Jr. R. and Hanneman, R. 1975. Response: Thermochemical Zoback, M. L. and Zoback, M. 1980. State of stress in the conterminous
Hydrogen Generation: Heat Requirements and Costs. Science, United States. Journal of Geophysical Research: Solid Earth, 85
188:1037-8. (B11), 6113-6156.
White, S. K., Spane, F. A., Schaef, H. T., Miller, Q. R. S., White, M. D.,
Horner, J. A. and McGrail, B. P. 2020. Quantification of CO 2
Mineralization at the Wallula Basalt Pilot Project. Environmental
Science and Technology, 54(22), 14609–14616.
White. M. D., Kneafsey, T. J., Seol, Y., Waite, W. F., Uchida, S., Lin, J.
S., Myshakin, E. M., Gai, X., Gupta, S., Reagan, M. T., Queiruga, A.
F. and Kimoto, S and IGHCCS2 Participants. 2020. An international
code comparison study on coupled thermal, hydrologic and
geomechanical processes of natural gas hydrate-bearing sediments.
117
117
118
118
State of the Art 3: Tailing and Mine Wastes
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Energy geo-engineering
Geotechnics of mine tailings: a 2022 State of the Art
Geo-ingenierie pour l’energie
Géotechnique des résidus miniers: État de L’Art 2022
J. Carlos Santamarina
Energy Resources and Petroleum Engineering, KAUST, Saudi Arabia, [email protected]
Andy Fourie
School of
Adnan Engineering, University of Western Australia, Australia, [email protected]
Aftab
WASM: Minerals, Energy, and Chemical Engineering, Curtin University, Australia
Ramon Verdugo
D. Nicolas
CMGI, Espinoza
Santiago, Chile, [email protected]
Petroleum and Geosystems Engineering, U. Texas at Austin, USA
Annika Bjelkevik
Maurice Dusseault
Earth and
Tailings Environmental
Consultants Sciences,
Scandinavia U. Sweden
AB, Waterloo, Canada
Antonio Gens
Luis and
Civil Alberto Torres-Cruz
Environmental Engineering, U. Politecnica de Catalunya - CIMNE, Spain
School of Civil and Environmental Engineering, University of the Witwatersrand, South Africa
Hussein Hoteit
Energy Resources
Dobroslav and Petroleum Engineering, KAUST, Saudi Arabia
Znidarcic
Department of Civil, Environmental and Architectural Engineering, University of Colorado Boulder, USA
Seunghee Kim
Civil and Environmental Engineering, U. Nebraska-Lincoln, USA
Joo Yong Lee
ABSTRACT: Global interest in the geotechnical stability of tailings storage facilities (TSFs) has burgeoned over the past decade.
In particular,
Climate the widely
Change accessible
Response videoKorea
Division, of the Institute
catastrophic failure of the and
of Geoscience Feijão TSF inResources,
Mineral Brazil in 2019 was a definite
Republic hinge point.
of Korea
Enormous efforts are now focussed on the assessment and improvement of existing facilities as well as the design of new ones. A
Liang
focus ofLeithis paper has been on problems associated with contractive and brittle tailings, how to determine liquefaction
susceptibility,
School and whetherWestlake
of Engineering, currently U.,
available
Chinamonitoring technologies are able to provide meaningful and timely information on a
potential, developing failure. We also believe there are other, important aspects of tailings behaviour that must be understood, and
thus the paper
Guillermo also deals with the consolidation, desiccation and creep behaviour of highly compressible tailings, the importance of
Narsilio
correctly accounting
Department for seepage,
Infrastructure Eng.,andU.the analysis and
Melbourne, design for seismic loading conditions. Although new management
Australia
technologies such as filtered tailings are seeing increasing use, we contend that comprehensive understanding of tailings
geotechnics remains
Jean-Michel paramount to ensuring the stability of all tailings facilities.
Pereira
Navier, Ecole des Ponts, Univ Gustave Eiffel, CNRS, France
RÉSUMÉ : L'intérêt mondial sur la stabilité géotechnique des parcs de stockage de résidus miniers (PSRM) a eclaté au cours de la
dernière décennie.
Marcelo Sanchez En particulier, la vidéo largement diffusée de l´effondrement dévastateur du PSRM de Feijão au Brésil en 2019 a
constitué
Civil andun point de convergence
Environmental certain. Texas
Engineering, D'énormes
A&M,efforts
USA sont désormais concentrés sur l'évaluation et l'amélioration des
installations existantes ainsi que sur la conception de nouvelles installations. Cet article s'intéresse aux problèmes liés aux résidus
miniers compressibles
Kenichi Soga et fragiles, à la déterminaison de la susceptibilité à la liquéfaction et la possibilité des actuelles technologies de
surveillance
Civil disponibles à fournir
and Environmental des informations
Engineering, appropriées
UC Berkeley, USAet opportunes sur une potentielle rupture en cours. Il y a, également,
d'autres aspects importants du comportement des résidus qui doivent être appréhendés, et donc le document traite en outre la
consolidation,
María la dessiccation
Victoria Villar et le fluage de résidus hautement compressibles, l'importance de prendre en compte correctement les
infiltrations, ainsi que les analyses et la conception pour des conditions sismiques. Bien que de nouvelles technologies de gestion,
CIEMAT, Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas, Spain
telles que les résidus filtrés, soient de plus en plus utilisées, une compréhension complète de la géotechnique des résidus demeure
primordiale pour assurer la stabilité de tous les parcs à résidus.
Marie Violay
KEYWORDS:
Lab tailings,
Experimental Rock liquefaction,
Mechanics, brittle,
EPFL, consolidation,
Switzerlandseismic.
ABSTRACT: The geo-science and engineering fields have critical roles to play towards a sustainable energy future. This state-of-
1 the-art review focusses on five areas where the geotechnical community
INTRODUCTION has been
Despite involved
the focus theunacceptable
on the most: the oilrate andofgas sectorofwith
failures TSFs
emphasis on methane hydrates, carbon geological storage, geothermal, energy
provided by geo-storage,
ICOLD Bulletin and 121,
nuclearandwaste
manystorage.
subsequentExtensive
studies
tables
1.1 in unacceptable
The the appendix identify potential
rate of failures of geotechnical
Tailings Storagecontributions to(e.g.,
all energy resources
Azam and (mining
Li, 2010), is critically
failures continue needed to supply
to occur. Indeed,rawnot
materials for the energy sector, yet is not addressed in this review).only
Facilities Energy-related
have failuresapplications
continued to involve
occur at a wide range of high
unacceptably time rates,
and
length scales, coupled thermo-hydro-chemo-mechanical processes, but multi-phase fluids, highofpressure-temperature-stress
the consequences many of the more recent failures conditions,have
The ICOLD
fines (International
migration, Commission
reactive fluid transportonand
Large Dams)
phase Bulletin
transformations, and multi-physics
been devastating, repetitive loads. Analysis
both to human life and and design
to the require
downstream
#121 documented
careful some 221
experimentation case records
(including of either
field-scale incidents
studies) at a
and advanced numerical simulations.
environment. Educational
Two particularly programs
catastrophic musthave
failures address the
occurred
tailings storage facility (TSF) or indeed
knowledge needs in energy geo-engineering. outright failure of a TSF. in the past few years in Brazil. In 2015, the Fundão TSF collapsed,
Published in 2001, this bulletin concluded that, “effective releasing some 43 million m3 of tailings, which flowed into the
reduction of the cost of risk and failure can only be achieved by a
Les domaines des géosciences et de l’ingénierie ont unDoce River, eventually
rôle essentiel reaching
à jouer pour the Atlantic
un avenir Ocean
énergétique some 660
durable. Cettekm
commitment from Owners to the adequate and enforced away (Fonseca do Carmo et al, 2017). This failure resulted in 19
revue de l’état de l’art se concentre sur cinq domaines
application of available engineering technology to the design, dans lesquels la communauté géotechnique a été le plus impliquée : le secteur
pétrolier avec fatalities and was followed by detailed forensic investigations that
construction andun focus of
closure surtailings
les hydrates de méthane,
dams and impoundmentsle stockage
over géologique
identified
de dioxyde
theenprimary
de carbone,
causes
la géothermie,
of the failure,
le géostockage
detailscontributions
of which are
d’énergie et le stockage des déchets
the entire period of their operating life”. radioactifs. Des tableaux détaillés donnés annexe identifient les potentielles
discussed later in this paper.
121
95
The Fundao failure resulted in a study commissioned by the stability of the facility. On the other hand, tailings facilities
ICMM (International Council on Mining and Metals) that set out constructed using the downstream method are totally supported
ten principles to be followed to maintain integrity of TSFs and on the foundation soil and tailings only act as a load against the
minimise the risk of catastrophic failure (ICMM, 2020). Despite retaining embankment. The centerline option is an intermediate
this document being adopted by the ICMM (an organisation that structure, with the downstream shoulder resting on the foundation
represents some 27 large multinational mining companies), and soil and supporting the lateral forces generated by the retained
other initiatives that followed the Fundão failure, in January 2019 tailings and water.
the collapse of the Feijão TSF in Brazil resulted in over 260 Although many upstream tailings facilities have performed
fatalities and widespread downstream environmental damage. The satisfactorily for many years and continue to do so, we note that
disturbingly graphic video footage that was published on the analyses of TSF performance records generally identify the
internet has probably done more for the improvement in tailings upstream construction method as the least reliable one (Islam and
management practice than any single previous initiative. It clearly Murakami 2021, Franks et al. 2021). We believe that this result,
demonstrated the extreme brittleness of the tailings retained in this coupled with the previous observation that upstream dams rely on
TSF and the speed with which the resulting flowslide occurred. adequate strength development in the impounded tailings,
The Feijão failure led to the establishment of an independent, warrants the claim that upstream construction carries implicitly
multi-stakeholder review, and, finally in August 2020, to the higher risks than other methods. Indeed, upstream TSFs do require
publication of the GISTM (Global Industry Standard on Tailings much greater oversight and attention to measurable performance
Management). The GISTM provides a blueprint for a step change than downstream facilities.
in the way TSFs are designed, constructed and operated and is
essential reading for anyone working in the field.
Aside from the two catastrophic TSF failures discussed above,
that occurred in Brazil, the past decade has seen more than thirty
TSF accidents or failures (https://www.wise-
uranium.org/mdaf.html). Notable among these are the failures that
occurred in Canada in 2014 at the Mount Polley TSF, and the
Cadia TSF failure that occurred in Australia in 2018, although
neither of these failures resulted in fatalities. In the case of Cadia,
the tailings release was very confined and limited to the mining
lease, whereas the Mount Polley failure resulted in off-lease
release, with flow occurring into the nearby Hazeltine Creek and
Quesnel Lake. The reason these two failures are specifically
flagged here is that they both occurred in well-regulated mining
environments. Mount Polley is in British Columbia, which has
one of the best regulated tailings management regimes in the
world, and Cadia is in New South Wales, also a well-regulated
and mature mining state. These two failures, while not resulting
in the same quantum of devastation as the Brazilian TSF failures,
certainly required a major re-think of the level of understanding
of key geotechnical principles within the tailings industry. How
could such significant failures occur in jurisdictions with a mature
and technically enlightened regulator, with TSFs designed by
competent and skilled geotechnical engineers? Reasons for this
situation are many and varied, and many of these are discussed in
this paper. It is recognised at the outset that we will not be able to
cover all pertinent tailings related geotechnical topics in this
single paper but believe a state-of-the-art review such as this Figure 1. Conventional TSFs. Common types of dam construction.
provides a much-needed opportunity for reflection and
recalibration of the adequacy of our current understanding of the 1.2 Increasing tailings volumes
engineering behaviour of mine tailings.
Despite various recycling initiatives around the world, demand for
Depending on how tailings are processed, transported,
discharged, and stored, different types of TSFs are generated, most minerals remains almost insatiable. Clearly, mining as a
which have significant variations in terms of cost, volume and human activity is not going away. The volume of tailings
type of materials required for construction, rate of construction, produced annually is increasing significantly, as the grade of
and inherent stability. The current practice of tailings management available ore bodies declines. Robertson (2018) suggested that
is mostly associated with tailings deposits of what may be termed every 30 years the volume of tailings worldwide increases about
the ‘conventional’ type, due to their lower cost and simpler ten-fold, the area covered by TSFs increases five-fold and average
implementation. The conventional system of tailings disposal heights of TSFs doubles. He pointed out that the maximum height
consists of one or several dams that confine and retain the of TSFs in the decade 2001 to 2010 was about 240m, with even
saturated loose, hydraulically deposited tailings (slimes) and higher TSFs planned. Improved understanding of the geotechnical
process water in an impoundment. The retaining component of behaviour of mine tailings will thus continue to be imperative, and
tailings facilities (e.g., the perimeter embankment) are usually to evolve to include the effect of increasingly high confining
constructed using the sandy fraction of the tailings, which may be stresses on tailings characteristics. Furthermore, based on data
accomplished using cyclones or, more often, conventional disclosed by mining houses upon the request of institutional
earthmoving equipment. According to the construction procedure investors, it is estimated that 50 billion m3 of tailings will require
used, it is possible to distinguish between upstream, downstream storage during the five-year period extending from 2019 to 2023
and centerline methods of construction (see Figure 1). The (Franks et al. 2021). Tailings storage will remain a necessity for
upstream method of construction requires a minimum volume of the foreseeable future.
coarser tailings (or borrow material) for embankment
construction, but the geotechnical properties of the tailings,
sometimes including the finer fractions, are involved in the overall
122
1.3 Changes in response to catastrophic failures
As already inferred above, we believe the Feijão failure that
occurred in Brazil in 2019 will be seen as an inflection point in
the attention paid to improved tailings management worldwide.
The publication of the GISTM will ensure the focus of owners and
operators of TSFs, government regulators, financiers and
insurance companies are retained on trying to ensure no future
catastrophic failures occur ever again. Having said that, there are
many existing TSFs that have been (and unfortunately some still
are) poorly managed historically and the potential for failure of
such deposits cannot be ruled out, unless placed under appropriate
operational management.
The GISTM covers many topics aimed at vastly improving
tailings management, many of them non-technical. However,
embodied in the guiding ‘Requirements’ is one that we believe
will have (and indeed already is having) a profound influence on
the way many TSFs are operated. Associated increases in cost,
sometimes substantial, are going to be unavoidable. The
Requirement we refer to is Requirement 4.6 which states,
“Identify and address brittle failure modes with conservative
design criteria, independent of trigger mechanisms, to minimise
their impact on the performance of the tailings facility” (the
italicised words are according to the GISTM). Unfortunately,
many TSFs contain extensive zones of potentially brittle tailings,
a topic we discuss in more detail in section 8.1.1. Ideally, going
forward TSFs would be operated to prevent the accumulation of
brittle tailings, but sometimes this is extremely difficult to Figure 2 Relationship between solids content and yield stress for tailings,
achieve. Operational practices that can be adopted to produce showing indicative ranges of properties and appropriate dewatering
dilative rather than contractive (and brittle) tailings are touched technologies (after Jewell and Fourie, 2015; courtesy Australian Centre
upon throughout this paper. for Geomechanics).
1.4 Alternative and emerging technologies. 2 THE CONTRAST BETWEEN WATER DAMS AND
TAILINGS STORAGE FACILITIES.
As discussed in ICOLD Bulletin 121, many, if not all these
failures could be largely attributed to the effects of too much Water dams are usually built to store water for power generation,
water. There are additionally many other drivers to reduce water irrigation, water supply etc. Historically, water dams have been
consumption in tailings management, e.g., in Chile costs of up to built since time immemorial. The history of tailings dams is, in
$US4/m3 have been reported for water. contrast, much more recent (of the order of 100 years) and they
Over the past two to three decades, significant advances have are built to store tailings, which contain different proportions of
been made in technologies that utilise less water to transport sand, silt and clay-sized particles, and related process water.
tailings to the TSF. Figure 2, taken from Jewell and Fourie (2015), Tailings dams form an impoundment, and the entire facility is
demonstrates the qualitative effects of significantly enhanced commonly referred to as tailings storage facility (TSF), tailings
dewatering of tailings. Solid’s concentration is plotted against a
management facility (TMF), or more recently (the termed used in
measure of strength, in this case shear yield stress, as measured
the GISTM), a tailings facility. For purposes of continuity with
typically using a vane rheometer, usually reported in Pascals.
Most operational TSFs today would classify as ‘conventional current literature, we use the former term in this paper (TSF). A
tailings’ (indicated by ‘segregating’ in Figure 2), with no TSF may also include structures such as spillways and
measurable yield stress of the as-deposited tailings. Additional clarification ponds. This section provides a perspective of the
thickening, using much deeper thickeners than previously used, differences between water dams and tailings dams.
results in higher solids concentrations and thus higher yield
stresses, requiring much more attention to detail of tailings 2.1 Ongoing construction and increases in height
transportation, as the higher solids concentration may sometimes Water dams are normally constructed as a conventional civil
require the use of positive displacement pumps rather than engineering project, i.e. with a defined start and end. Normally the
conventional centrifugal pumps. There are also implications for construction phase lasts from a year to a few years, depending on
deposition practices when using high-density thickened tailings. the height, length and type of dam etc. Once construction is
Added impetus to the adoption of increased dewatering has been completed, the dam is handed over to the owner to operate.
provided by recommendations arising from some of the forensic Tailings dams are designed for life of mine (LOM) (in addition to
investigations of major TSF failures to minimise storage of water a period of post-closure, which may be extremely long, as
on TSFs. Referring again to Figure 2, further thickening requires discussed in section 11) and commonly built in stages over that
use of filters, producing a cake-like material when successful. The time. Firstly, a starter dam(s) is constructed, generating an initial
production of filtered tailings usually comes at significantly capacity for typically one to three years storage of tailings and
increased cost, particularly operational costs due to increased process water. The dam is then raised in discrete increments at
energy use. regular time intervals, or continuously as during cycloning.
Regardless of whether the dam is raised discretely or
continuously, the rationale is to progressively increase the storage
capacity as the demand requires. Construction is therefore an
ongoing process throughout the whole operating phase. As dam
raises often include tailings as a construction material, operations
123
such as deposition of tailings is often part of construction on a stationary cyclone, several small cyclones on the crest,
continual basis. mobile cyclone or a set of cyclones and they can be
LOM may vary from around approximately 10 years up to 50+ horizontal or vertical.
years and, more importantly, LOM will likely change over time Subaqueous deposition means deposition under water.
as a) exploration continues and more ore is found and b) This method is mainly used when the tailings contain
commodity prices change leading to more tailings to store if prices sulphur and are prone to oxidation. The TSF is then a
rise or less to store, or even an early close of operations, if prices water impoundment where tailings is deposited under
decrease. Due to changes in LOM the final height for a tailings water. As the natural slope of tailings deposited under
dam will change with time, which must be considered in the initial water is much steeper compared to subaerial deposition,
design, i.e., that the final crest elevation may increase. the discharge point must be moved regularly to utilize the
capacity of the impoundment effectively. The density of
2.2 Importance of operational procedures the tailings is also looser than subaerial (above water)
deposition.
Tailings dams require ongoing construction during operation, Thickened and paste deposition means deposition of
when the design engineer and the owner/operator should work in tailings slurry at a higher solids content than conventional
parallel to ensure that design, construction and operation become tailings, where the latter often have only some 10-50%
an integrated ongoing process. Close collaboration, solids by weight. To reach a higher solids content a deep
communication, knowledge and experience sharing become very thickener is used to further remove water. If enough water
important. Operational activities of a tailings dam include is removed, it goes from a slurry to paste (with a
planning and management of deposition of tailings in the consistency like toothpaste). The higher the solids
impoundment as well as management of water (process water and content, the steeper the beach slope and the lower the
precipitation). Depending on the design chosen for a particular propensity for particle segregation. Thickened tailings or
tailings dam, the influence of tailings deposition may become paste is used when water is scarce and/or when a steeper
critical, e.g., upstream construction totally relies on the deposited beach slope can increase the capacity of the facility to
tailings beach immediately upstream of the crest of the dam being more effectively use the land footprint.
a part of the structural zone of the dam. Conversely, a dam using Filtered tailings deposition is taking thickened tailings one
downstream construction is more analogous to a conventional step further as even more water is removed, filtering the
earth- or rockfill dam and is thus not reliant on the strength of the tailings after it has been thickened. Filtered tailings are too
deposited tailings. For the upstream construction method, the dense to be pumped and must be transported to the TSF
density and the state of the deposited tailings in the structural zone using, for example, conveyor belts or trucks. In the TSF
are critical for stability. The tailings in the structural zone should the filtered tailings usually need to be compacted to create
be non-liquefiable (see also Section 5.1.1). a stable structure. Dust problems are potentially an issue
Deposition of tailings also influences how efficiently the and in wet climates, logistics and compaction must be
storage capacity is utilized as deposition affects the in-situ dry planned to not saturate the filtered tailings after
density of the stored tailings. Economically, dry density is placement. The resulting deposit is commonly referred to
critically important, as it directly affects the utilisation of storage as a “dry stack” and since it may not include free water, a
volume (air space) and thus the footprint of a tailings dam. TSF failure is likely to result in more limited damage
Different methods of deposition of tailings are used depending on compared to a conventional TSF containing tailings and
the desired outcome. The following are the most common. See process water. However, a saturated “dry stack” is still
also Figure 3. prone to liquefaction.
Co-deposition is here used as an umbrella term for all
Single discharge deposition, i.e., a single slurry pipeline manner of integrated mine waste management, e.g.
from the process plant to the TSF. The outlet may methods where tailings and other mine waste materials,
occasionally be moved to even out the build-up of tailings. like waste rock, are deposited in the same facility. Other
Spigot deposition is a method where the main slurry terms used include co-placement (placed, but not
pipeline is divided into several outlets (like a sprinkler). transported together) and co-mingling (placed and
Normally the main slurry pipeline has several discharge transported together). Renaming of existing technologies
points of which some are open simultaneously. To get an such as these includes, for example, the terms eco-tails
even beach, uniform thickness of deposited layer and and paste rock.
improve consolidation, spigots are rotated frequently,
which also prevents dusting if the tailings beach above
water is kept wet.
Drop-pipe is an extension of a deposition point, a spigot,
where there is a large height difference between the
discharge point and the beach. Too high a difference in
height usually results in erosion on the beach and possibly
on the upstream face of the dam. Using a drop pipe helps
dissipate the energy in the flow.
Spray bar deposition is a refinement of spigot deposition
as each spigot is in turn divided into several outlets. This
reduces flow and velocity even more, which increases the
amount of tailings particles settling early on the beach,
almost invariably producing a steeper beach.
Cyclone deposition is when coarse and fine particles are
separated using a cyclone. In the cyclone, the combined
action of gravity and centrifugal forces separates the
coarse fraction from the fines and most of the water. The
coarse fraction is then used for dam embankment
construction and the fines deposited further out on the
Figure 3. Examples of different tailings deposition methods. Photographs
beach. Cyclone schemes may consist of one large
courtesy of Boliden, LKAB and FL Schmidt.
124
For tailings dams relying on deposited tailings for stability, i.e., virtually independent; thus, when changes to the design are
where deposited tailings are part of the structural zone of the dam, introduced, a clear understanding of the resulting implications and
a suitable deposition method is important, as is cycle time and operational requirements are crucial.
adequate drainage. Cycle time refers to how often a specific
deposition point is moved and/or how long it takes to do a full 3. APPLICATION OF GEOTECHNICAL ENGINEERING
cycle around the perimeter of the TSF. Deposition for a short time PRINCIPLES – DO TAILINGS DIFFER FROM NATURAL
results in a thin layer of tailings being deposited, which allows for SOILS?
rapid drying and consolidation. For a specific facility, deposition
for a short time means moving on to the next deposition point Tailings, from an engineering point of view, can be considered as
faster and therefore coming back to the same/first point quickly, man-made soils. Tailings derived from different parent ores, or
i.e., a short cycle time for the full facility, which means less time from different mineral extraction processes may be extremely
to consolidate before deposition takes place again. Therefore, to different and may vary from clayey to sandy soils, sometimes
ensure adequate consolidation an ideal would be a short exhibiting properties outside the framework of understanding
deposition time, but a long cycle time, which are contradictory developed in conventional geotechnical engineering (e.g. laterite
requirements. An optimization is necessary. Additionally, there is nickel tailings). However, tailings can usually be best
the challenge of managing the rate of rise (RoR), which refers to characterised and analysed as geotechnical materials. In this
the rate of raising of the beach of deposited tailings. If the RoR is context, it is necessary to point out that the main difference
too high, consolidation will be negatively affected, and stability between soils and other engineering materials (for example,
may be jeopardized. The only opportunity to affect these concrete, steel and wood), is associated with the development of
conditions (deposition time, cycle time and RoR), is the area, or volumetric strains when subjected to shear stresses.
capacity, of the impoundment. Namely, the area needs to be big Tailings, like any un-cemented soil, is essentially a particulate
enough to facilitate a long enough cycle time and low enough medium made up of mineral particles with interstices (or voids)
RoR. The current tendency to minimize the footprint of a TSF can, and therefore undergo volume changes when subjected to shear
in this perspective, have a negative impact on stability and safety. stresses due to the movements between particles (rotation,
To maximize consolidation and retain a low phreatic surface sliding), which interact with the voids, inducing in them either a
within the structural zone of the dam, drains are normally used. reduction or an increase of their volume.
Upstream dams are sometimes referred to as “drained dams”, The volume change of sandy soils was first observed by
emphasising that drainage is important. It is important for stability Reynolds (1885), who showed that dense sands tend to expand,
directly, but also indirectly, as good consolidation improves shear increasing their total volume when subjected to shear stress.
strength and drainage speeds up the consolidation process. Reynolds called this phenomenon "dilatancy". However, it was
Drainage can be designed in many ways; as horizontal layers or only in the 1930s that Casagrande realized the real importance of
finger drains at the bottom of the dam or under the entire structural volumetric deformation in the mechanical response of soils,
zone for the upstream construction method. Alternatively, developing the concept of "critical density or critical void ratio"
horizontal layers or finger drains can be built at intervals at (Casagrande, 1936). Later, Roscoe and his co-workers presented
increasing elevations as the dam is raised. In some cases, vertical a conclusive study proving the concept of critical void ratio and
drains can be used to connect horizontal drains at different extended it to clayey soils (Roscoe et al, 1958). Typical test results
elevations, although this option is not common. The drains can be obtained on 1 mm steel balls using simple shear are shown in
constructed using natural gravel, crushed rock, perforated or Figure 4 in terms of void ratio and horizontal displacement for a
slotted pipes or different bespoke products using various constant vertical stress of 1.41 kg/cm2 (138 kPa). As can be
geosynthetic options. Collected water can be drained passively or observed, the volumetric strain can be either positive or negative,
actively pumped from a collection sump. depending on the initial void ratio and the magnitude of
deformation, but when the ultimate state is achieved, the volume
2.3 Changes to design intent change tends to cease and the soil deforms under a constant-
volume condition reaching the critical void ratio associated with
As the conditions at a particular tailings dam may change over the normal stress under which the test is performed.
time, the design details may likewise need to change. Conditions
that typically may lead to the need for a design change include
changes in production rate (higher rates of tailings deposition),
changes in grind size (almost invariably meaning production of
tailings with a finer particle size distribution than at the start of the
project), and availability of construction material, where lack of
adequate waste rock, for example, may precipitate a switch from
downstream to upstream construction. Changes in mine staff,
operators and designers may also result in changes to the design.
Such changes in themselves are not a problem. However, it is
crucially important that everyone with operational responsibilities
understand the geotechnical principles of the structure and how
changes to design may affect the overall functioning and stability
of the structure. The mantra is: ‘operate according to the design
intent’. As already mentioned, an upstream constructed tailings
dam relies on provision of suitable underdrainage to ensure a low
phreatic surface which in turn assures a well consolidated
structural zone of the deposited tailings beach. Conversely,
downstream construction relies on the strength of the deposited
tailings to a much lesser extent, if at all. A particular tailings dam Figure 4. Results of simple shear tests on 1 mm steel balls with normal
will be somewhere between 100% dependent on management of stress 140 kPa. Republished with permission of ICE Publishing, from
Roscoe et al. (1958); permission conveyed through Copyright Clearance
tailings deposition providing an adequate structural zone to
Center, Inc.
125
initial settlements can be missed unless displacement data
With some limited exceptions, tailings are amenable to the acquisition is automated.
application of theories and frameworks of understanding
developed for natural soils. These exceptions are discussed 3.2 Quantifying particle shape
subsequently. Being essentially a two phase (sometimes three
phase) material, i.e. solid particles in a liquid matrix, the key Approaches to measuring particle size distribution are well
building block of geotechnical engineering, namely the effective developed and widely used, e.g. sieve and hydrometer methods.
stress principle, is usually applicable and relevant. Additionally, Not so for measuring particle shape. There is no universally
conventional geotechnical engineering laboratory and in-situ tests accepted procedure, although one of the most widely used
that aid in the mechanical characterisation of natural soils are techniques is that used by Cho et al., 2006. This method is based
generally useful to characterise tailings. on the chart shown in Figure 5 and is based on two parameters,
the sphericity, S, and the roundness, R.
3.1 So what are the differences?
When differences occur, they are usually attributable to the way
tailings originate. Unlike natural soils, which essentially evolve
through a combined process of weathering, transportation, and
deposition (sometimes the former totally dominating of course, as
in the case of residual soils), tailings are invariably derived from
some form of industrial process. The most common of these
processes include some (or all) of the following: crushing,
grinding, chemical treatment, and hydraulic transport in a pipeline
(occasionally in a channel). The resulting particles are therefore
often significantly more angular than most natural soils.
In addition, tailings are often deposited at void ratios that are
much higher than the void ratios encountered in natural soils
Figure 5. Method of characterising particle shape showing: a) sphericity,
familiar to most geotechnical engineers. Indeed, when deposited
S and roundness, R, and b) measurements used to determine S and R.
as a slurry, tailings undergo sedimentation before a true soil
Republished with permission of ASCE, from Cho et al. (2006); permission
deposit is formed for which the effective stress principle is valid.
conveyed through Copyright Clearance Center, Inc.
Even after such a tailings layer is formed, it continues to undergo
significant volume changes under self-weight stresses and Sphericity, S, characterises the gross shape of a particle and
exhibits highly nonlinear compressibility and large strains. Most reflects the similarity between a particle’s length, height and
of the volume change takes place at low effective stresses and is width. It is generally defined as the ratio of the radius of the largest
associated with a significant reduction in the hydraulic circle inscribed in the particle, ris, to the radius of the smallest
conductivity of the material. It is not unusual that, after circle circumscribed around the particle, rcs.
sedimentation, a tailings layer is formed at a void ratio as high as Roundness, R, which is the conceptual inverse of angularity,
5 (for relatively low plasticity materials) and up to 20 or even 30
represents the scale of a particle’s major surface features (such as
for materials with high plasticity (e.g. phosphatic clays, Abu-
Hejleh et al 1996). Furthermore, volumetric strains that exceed
corners and edges). R is quantified using the equation, =
50%, as the applied effective stress increases from 0.1 kPa to 1
kPa are not rare. A reduction of the hydraulic conductivity by five
or more orders of magnitude is common for the effective stress where ri is the radius of each major surface feature and Nsf is the
range expected in the field (e.g. Abu-Hejleh et al 1996). number of surface features (Figure 5). A limitation of the chart
Considering these significant changes, traditional linear, small depicted in Figure 5 is that it is based on a 2D projection of the
strain consolidation theory is not suitable for such materials and particle and as such is unable to capture 3D aspects of particle
the assumption of a constant coefficient of consolidation and a shape.
constant compression index is not always justified. A nonlinear, Altuhafi et al (2013) provide three shape measures, namely
finite strain consolidation theory that accounts for the effect of convexity, sphericity and aspect ratio, for characterising particles
self-weight of the material is thus essential for a rational approach using digital image analysis (DIA). The method appears more
to analyze the behavior of tailings as they transition from a high robust than the Krumbein and Sloss approach, as the parameter R
water content slurry upon deposition to a soil deposit in which the is somewhat scale dependent, an undesirable feature when
principle of effective stress holds true (e.g. Gibson et al 1967). characterising shape. Recently Yang et al (2019) have also used
Consequently, conventional consolidation tests are not suitable DIA techniques to investigate shapes of tailings particles, with
for determining the consolidation parameters of such tailings and promising results. However, much of the available literature
testing procedures and test analyses consistent with the nonlinear correlating particle shape to engineering characteristics is based
consolidation theory must be used for characterization. Such on the parameters R and S, and this paper is thus restricted to the
experimental procedures have been developed and are available existing literature bank.
for routine use as discussed in section 4.2.2.
We emphasize that many tailings streams, particularly those 3.3 Correlation of sphericity and roundness to geotechnical
derived from hard-rock mining, settle out (sediment) very rapidly properties.
and thus do not necessarily require testing according to the
procedures discussed later in this paper and conventional 3.3.1 Maximum and minimum void ratio values
consolidation testing methods are appropriate. Although tailings There is now overwhelming and consistent evidence that S and R
derived from many hard rock ore deposits may contain correlate strongly with the maximum and minimum void ratios,
appreciable fines contents (more than 20% by mass), they still emax and emin respectively. Both these extreme void ratio values
have high hydraulic conductivity (and thus coefficient of increase with decreasing S and R, i.e. as the particle becomes more
consolidation) values because the fines are non-plastic, being rock angular and irregular (or less rounded). As noted by Santamarina
flour. In fact, some (coarser) tailings have such high coefficients and Cho (2004), for mono-sized spheres the loosest stable packing
of consolidation that using a conventional oedometer may be (emax) that can be achieved is 0.91 and the densest (emin) is 0.34. If
problematic because consolidation takes place so rapidly that a range of different sized spherical particles were present, emin
126
could be even lower. Cho et al (2006) presented data from 16 Where m is the slope of the correlation and is necessarily >1 since
natural sands from around the world, 17 crushed sands from by definition emax > emin. The range of void ratios can thus be
Georgia, USA (being crushed, these sands are likely similar to expressed as:
tailings sands and silts), as well as a few other materials such as
glass beads, granite powder and tailings from the Syncrude
emax – emin = m* emin – emin = emin (m – 1) (2)
oilsands operation in Alberta, Canada. Figure 6 clearly illustrates
the inverse relationship between the extreme void ratios and both
sphericity and roundness. A third parameter, , which was
suggested by Cho et al (2006) who called it the ‘regularity’; and
defined it as = (R+S)/2, is also shown in Figure 6. The regularity
is intended to capture both sphericity and roundness in a single
parameter. As can be seen in Figure 6, it has the same inverse
correlation with the extreme void ratios as S and R (which is not
unexpected).
Figure 7. Approximately proportional correlation between emax and emin for
a wide variety of soil types. (after Cubrinovski and Ishihara (2002)).
Since m > 1, it follows that (m – 1) > 0. So (m – 1) is a direct
proportionality constant between the void ratio range (emax – emin)
Figure 6. The effect of particle shape on extreme void ratios. Republished and the minimum void ratio emin. A similar analysis shows that
with permission of ASCE, from Cho et al. (2006); permission conveyed (emax – emin) is also directly proportional to the maximum void
through Copyright Clearance Center, Inc. ratio emax. We can then conclude that an increase in emin (or emax,
as they are proportional) also increases the void ratio range (emax
Doubt is sometimes expressed about the value and relevance – emin). This in turn increases the potential for volume change
of extreme void ratios emax and emin for soils with a high upon shearing of a very loose packing. If shearing occurs under
percentage of silt sized particles (or finer), given that the standard undrained conditions, the volume reduction will generate large
techniques used to determine these parameters were originally positive excess pore pressure, a factor closely linked to the
intended for clean sands. However, the data presented by Cho et mechanism of static liquefaction, discussed in Section 4.3.
The results from Cho et al. (2006) highlight that high particle
al (2006) were all for material with d50 values more than 0.12mm angularity, such as that of tailings, promotes high values of emin,
(most being in the range of 0.3 to 0.5mm), with Cu (coefficient of emax, and therefore of (emax – emin). However, it is worth
uniformity) values less than or equal to 2.5, thus obviating this remembering that particle size distribution also has a significant
concern. Furthermore, the technical literature now contains effect on the extreme void ratios. Namely, broad gradations
multiple examples of experimental programmes that have adopted promote tight packings while narrow gradations promote loose
emin and emax to aid in the characterisation of non-plastic sand-silt ones (Biarez and Hicher 1994). So, the final value of extreme void
mixtures (e.g. Zlatovic and Ishihara 1995, Cubrinovski and ratios, and associated volume compressibility, of tailings will be
Ishihara 2002, Thevanayagam et al. 2002, Papadopoulou and Tika both a function of particle size distribution and particle shape.
2008, Yang et al. 2006, Carrera et al. 2011, Torres-Cruz and
3.3.2 Critical state (constant volume) friction angle
Santamarina 2020). The consistency of the results yielded by
The ratio of deviator stress (q) to mean effective stress (p') that
these programmes strongly suggests that the extreme void ratios
soils reach at critical state is termed the critical stress friction ratio
are useful indices to understand the mechanical behaviour of silt-
(M) (also discussed in Section 4.3.3). M values are affected by
sized non-plastic tailings.Figure 6 shows that particles with lower
particle shape, with particle angularity promoting higher values of
regularity have a larger difference between the extreme void
M. Tailings that result from rock crushing processes have angular
ratio values, e.g. for , the difference is about 0.2, whereas for
shapes and can thus be expected to exhibit relatively high M
= 0.3, this difference is about 0.45. This observation is also
values. Figure 8 presents the M values for a variety of tailings and
supported and expanded by the extreme void ratio data for over
compares them to values of other soil types that include natural
300 soils reported by Cubrinovski and Ishihara (2002). Their data
soils and sands from rock crushers. The median particle size D50
covers a wide variety of soil types, including gravels, clean sands
is a convenient parameter to plot the data although no correlation
and silts. Although Cubrinovski and Ishihara (2002) calculated
is observed between M and D50. There is significant overlap
separate emax vs emin regression lines for different soil types within
between the M values of tailings and that of other types of soils.
their database, the database as a whole reveals a roughly
This is somewhat expected given that angular particles that
proportional correlation (linear and passing through the origin)
produce high M values are also found in soils that are not tailings.
between emin and emax (Figure 7).
Furthermore, the M values of tailings span a relatively wide range
emax = m*emin (1) from 1.12 to 1.78. The data from glass beads serves to highlight
the role of particle shape, with their spherical geometries leading
to the lowest M values of approximately 0.85.
127
1.8 when depositing the tailings hydraulically, as very significant
particle segregation occurs along the tailings beach. This results
1.5 in a steep beach of sandy material near the deposition point, and a
relatively flat (<1% slope) beach of fine tailings, with these fine
Friction ratio M
1.2 tailings showing extremely low rates of consolidation.
0.9 Differences in Gs for different commodities mean that
Other BAU DCU different deposition strategies may need to be adopted. For
HMS HVC MER
0.6 OIL PFE SAU Glass beads example, tailings with a very high Gs will settle very quickly after
PLA STA discharge from the delivery pipe, forming a relatively steep beach;
0.3 this in turn means the height of the discharge point requires
1 10 100 1000 3000 frequent adjustment, adding to costs of tailings management.
Median particle size D50 [ m] Another implication is that the use of ‘general rules’ are not
always applicable. For example, for gold tailings an in-situ dry
density of 1.7t/m3 should provide tailings that are dilative upon
Figure 8. Critical friction ratio (M) against median particle size (D50) for shearing, a highly desirable outcome as discussed later in this
tailings and other soil types. Symbols for tailings are defined as: BAU = paper. Adopting this value of dry density for an iron ore tailings
Brazilian gold tailings, DCU = Deixing copper tailings, HMS = Hilton
with Gs of 3.8, however, results in a void ratio of 1.24. This may
Mines, HVC = Highland Valley Copper tailings, MER = Merriespruit gold
not be sufficiently dense to ensure dilative behaviour.
tailings, OIL = Syncrude oil sand tailings, PFE = Panzhihua Iron tailings,
SAU = South African gold tailings, PLA = platinum tailings, STA = Stava
3.4.2 Water chemistry
fluorite tailings. Data as compiled by Torres-Cruz and Santamarina
The mineral processing options used for extraction of the
(2020).
commodity of interest vary widely across industries. For example,
extraction of heavy minerals from mineral sands ores is usually
3.4 Some other notable differences between tailings and done by gravity separation, which is made possible by the large
natural soils difference in Gs between mineral and non-mineral (i.e. the sand
and clay fractions), whereas commodities such as gold and copper
3.4.1 Specific gravity require utilisation of a chemical processing agent, such as cyanide
Specific gravity, Gs, is the particle density normalised by the that is often used for gold extraction. Perhaps even more extreme
density of water which, although dependant on temperature, is are the processing of nickel laterite deposits, and of bauxite to
commonly approximated to 1 g/cm3. For natural soils, Gs is often produce alumina.
between 2.65 and 2.7, the former value being that of quartz sand. When extracted from a sulphide deposit (in general this
Occasionally a natural soil may be encountered with Gs values indicates a hard-rock deposit), the process for nickel typically
well outside this range, but it is unusual. Tailings, on the other utilises some form of chemical processing agent. However, nickel
hand, routinely exhibit values well outside this range. Table 1 contained within laterite deposits (which are generally near
below provides some examples of the variability encountered. surface and soft rock to soil consistency) is much more difficult
to extract. The most widely used process utilises high pressure,
Table 1. Range of specific gravity values of mine tailings acid leach. The resulting waste stream is more a chemical process
residue than a conventional tailing. For example, Azam et al
Commodity Specific gravity, Gs Reference
(2005) report results from tests on pressure acid leach (PAL)
Coal (fine refuse) 1.5 to 1.8 Soderberg and Busch, 1977 nickel laterite slurries which all had pH values less than unity and
Lead-Zinc 2.8 to 3.4 Soderberg and Busch, 1977 electrical conductivity values more than the upper limit of the
Phosphate 2.5 to 2.8 Bromwell and Oxford, 1977 measuring device (10 mS/cm). The clay minerals originally
Gypsum 2.3 to 2.4 Vick, 1977 present in the ore were completely decayed by this process, while
Gold 2.8 to 3.1 Fourie et al, 2021; Li et al, magnesium sulphates were formed by the reaction between
2018 sulphuric acid and the magnesium released from the ores. The
Iron ore 3.1 to 3.8 Fourie et al, 2021; Li and resulting slurry was thus radically transformed by the extraction
Coop 2019 process. Sedimentation tests were carried out on PAL slurries
Bauxite 2.9 Fourie et al, 2021 from four different countries; the final settled void ratios varied
Copper 2.7 to 3.75 Fourie et al, 2021; Li 2017 between 4 and 13.
Platinum 3.0 to 3.6 Fourie et al, 2021; Torres- Similarly, the extraction of alumina from bauxite ore produces
Cruz and Santamarina, 2020 a process residue, rather than a tailing. Here the process (most
typically the patented Bayer process) comprises heating the ore to
Table 1 shows the high variability across commodities and about 150 to 200°C in a pressure vessel to which sodium
highlights the importance of experimentally measuring Gs as hydroxide is added, resulting in a process residue that has a pH of
opposed to relying on assumed values. The value of Gs is affected around 12, with high levels of calcium and sodium hydroxide. The
process residue consists of two distinct components, ‘red sand’
by the particulars of the mining process. For example, the Gs value
and ‘red mud’. The red mud residue may contain haematite,
of gold and copper tailings will be closer to the upper end of the goethite, gibbsite, rutile, calcite, sodalite and other complex
range provided when significant quantities of pyrite are silicates, but there is usually a complete absence of either quartz
discharged in the tailings stream. In addition to the across- or clay minerals (Newson et al, 2006). Particles are generally silt
commodity variation, there is often also distinct variability for a sized, with some specific residues having appreciable quantities
particular commodity. For example, many mineral sands deposits of clay-sized particles. Red muds typically have low plasticity and
(from which heavy minerals such as zircon (Gs = 3.9 to 4.7) are high specific gravity values (up to 3.3). Despite the lack of clay
recovered) often contain two distinctly different tailings streams, minerals, and low plasticity, red muds demonstrate some
the sand fraction and the very fine fraction (commonly referred to geotechnical properties that are characteristic of clay-rich tailings,
as slimes). Not only are the particle size distributions (PSDs) very such as relatively high compressibility. In contrast, the angle of
internal friction for red muds tends to be similar to sand-like
different, but the Gs values are often also different. This
geomaterials, with reported values being typically 38 to 42°. Red
combination (different PSD and different Gs) presents challenges muds thus exhibit both clay-like and sand-like characteristics,
128
illustrating the dangers of blindly applying empirical correlations
that have been developed over many years for natural soils
(examples being correlations between undrained shear strength
and Atterberg limits). Furthermore, there are strong indications
that many red muds are cemented or aggregated, with in-situ
undrained shear strengths being very high compared to
uncemented, natural clayey soils at equivalent liquidity indices
(Newson et al, 2006). It is also not unusual for red muds to exhibit
high sensitivities (values from 5 to 15 are common), that is, high
ratios of peak to residual undrained strength. Based on chemical
analyses, Newson et al (2006) attribute the cementation to the
action of hydroxysodalite, which is only slightly water soluble but
becomes more soluble in alkaline conditions.
Figure 9 shows results from a CPTu (cone penetration test
with pore pressure measurement) performed at a red mud site in
Australia. Red mud has been deposited for many years at the site,
and until recently was operated solely through cycles of
deposition in relatively thick layers, with minimal time allowed Figure 10. Illustration of increased yield stress due to bonding; above a
for desiccation drying. The few spikes of higher tip resistance (e.g. critical value of vertical effective stress, the bonding is destroyed,
at about 5.8 to 7m and again at 7.8m are likely periods when some resulting in very large decreases in void ratio under small increments of
drying occurred, a practice that is now more common, as stress (after Leroueil and Vaughan, 1990).
evidenced by the higher strengths in the top 1m or so. What is
particularly important in this result (which is not atypical in red Bitumen extraction from the oilsands deposits in northern
muds) is the very gradual, if any, increase in tip resistance with Alberta utilises a process that combines sodium hydroxide and
depth. It would be expected that, as for natural soils, the tip heat. The clay fraction of the resulting tailings tends to segregate
resistance would increase with depth, due to the effect of from the dominant sand fraction (a function of the bimodal
overburden. Equilibrium pore pressures at this site were particle size distribution) and because the bitumen extraction
hydrostatic, so the observed effect (i.e. no increase in tip process causes dispersion of the clay particles, the resulting
resistance with depth) cannot be attributed to excess pore water deposit experiences extremely slow consolidation rates. As an
pressures. Rather, it is suggested here that this is consistent with example, Jeeravipoolvarn et al (2008) discuss results from a 10m
the observation by Nagaraj et al (1998) that the stress-carrying highly experimental column containing fine oilsands tailings,
capacity of bonds in cemented soils does not change during where after 20 years consolidation settlements were still
compressive loading, manifesting as very little, if any, void ratio occurring.
change with increases in effective stresses. This is illustrated in In terms of impact on the engineering characteristics of
Figure 10, which is a result from two natural, lightly cemented tailings, the biggest impact of the above-mentioned chemical
clays reported by Leroueil and Vaughan (1990) which contrasts processes (that produce residue with chemically altered pore
the behaviour in compression of the intact clays compared with fluids) is probably most pronounced in the initial rates of
the behaviour when prepared as a slurry to the same void ratio as sedimentation and consolidation. This is due to the altered states
the intact samples. of components of the residue, primarily clay minerals, such as the
An explanation for the CPTu result shown in Figure 9 dispersion that occurs with oilsands tailings and some coal
therefore is that the void ratio is almost constant with depth, as is tailings, where dispersion occurs because inter-particle repulsive
the undrained shear strength. The cone detects this constant shear forces dominate. The resulting suspensions are very slow settling,
strength, but the unknown corollary is what happens when the sometimes even remaining indefinitely as a suspension (a state
cemented bonds are destroyed, as they almost certainly are during termed ‘stable’ in some industries, whereas in a geotechnical
advancement of the cone. engineering context the term ‘stable’ is of course the exact
opposite). A good example is provided by Blight et al (1995),
where diamond tailings retained such a high suspended solids
content that the water could not be re-used in the processing plant.
If such repulsive interparticle forces are not overcome, the process
of self-weight consolidation is severely inhibited.
In cases where self-weight consolidation, and other processes
that accelerate an increase in density (such as evaporation) do
occur, the effect of pore fluid chemistry begins to diminish. The
shear strength of the residue is governed by factors such as tailings
density, and in particular density relative to the critical density (or
critical state). Having said that, the lingering impact of inhibited
consolidation may result in a material state above the critical state,
so behaviour upon shearing remains impacted by the pore fluid
chemistry.
An effect that is often overlooked in laboratory testing of
tailings is the salinity of the pore fluid. Small salt concentrations
have little to no effect, but at highly elevated salt concentrations
the effects cannot be ignored. Many of the gold mining areas of
Western Australia (WA) utilise hypersaline groundwater for ore
processing. The resulting tailings have extremely high salinity.
Values of total dissolved solids (TDS) of borewater used in some
WA mines exceed 200 g/l. Aside from engineering characteristics,
such high salinity affects even simple characterisations such as
Figure 9. Results from cone penetration test in red mud. Note in particular
measured water content. This occurs because oven drying causes
the relatively constant tip resistance with depth, e.g., below 15m depth. salts to precipitate as solids, and the conventional definition of
129
gravimetric water content w = Mw/Ms, where Mw and Ms are the deposition of thickened, high-density tailings, the yield stress
masses of water and solids content respectively, incorrectly enables the resulting profile of deposited tailings, this profile
includes the salt mass in the value of Ms, whereas it is actually in referred to as the beach, to be steeper than is usually achieved with
solution in the tailings pore fluid. The correct water content (given conventional tailings, potentially increasing the storage capacity
by Mundle et al, 2012) is: of a given footprint of a TSF. The higher as-placed density of
thickened tailings compared with conventional tailings further
= (3) improves the in-situ density of the resulting TSF.
( ) A drawback with many flocculant-treated tailings is that the
resulting material (the underflow) has a loose, fluffy texture, the
where f is the density of the pore fluid. structure of which can be very sensitive to shearing. The
underflow tailings are usually pumped to the TSF, using either
To illustrate the importance of this correction, for a tailings centrifugal or positive displacement pumps, with the transport
with hypersaline pore water (a TDS of 210 g/litre), at a particular distances often being many kilometres. During such transport the
void ratio the correct water content would be 79%, whereas the thickened tailings are thus subjected to relatively significant
conventional (incorrectly determined) water content would have shearing, potentially resulting in a decrease in the yield stress of
been 67%. The discrepancy flows through to calculations of void the tailings deposited at the TSF. Given the importance of yield
ratio and thus dry density. stress on beach slope development, this alteration of yield stress
is critically important to TSF operators. Houman et al (2007)
3.4.3 Effect of addition of flocculants sampled and tested tailings from ten points along the distribution
As noted in 3.4.2, some tailings may exhibit extremely slow system, from the thickener to the TSF, more than 5km away. Their
settling rates, an effect compounded by the dispersion of clay results are shown Figure 11, which illustrates the change in the
yield stress due to different shear inducing actions. The greatest
particles due to the process used to extract the mineral of interest.
effect is the centrifugal pumping from the thickener to the
Ways to deal with this problem have followed two distinct paths. distribution sump with an approximate four-fold drop in yield
The first (and most widely adopted) is to introduce additives to stress. It should be noted that this large drop was bound to occur
the tailings water in large thickening tanks, the second being to early in the distribution process, when the aggregated structure
introduce (additional) additives at the point of deposition, often was at its most fragile. Of importance is the final yield stress of
referred to as in-line thickening. the tailings as deposited, which is about 25Pa, compared with the
underflow value of 250Pa.
3.4.3.1 In-tank thickening
To overcome the inter-particle repulsive forces that occur in some
dispersed clay-rich tailings in particular, the tailings stream from
the process plant is often first diluted (to solids contents of the
order of 8%) before being fed into the feedwell component of a
tailings thickener. Additives, usually flocculants but with pre-
treatment using coagulants not uncommon, are also introduced
into the feedwell, providing the opportunity for thorough contact
with the dilute tailings, enhancing the subsequent interaction
between tailings and additives.
A comprehensive discussion of tailings clay mineralogy and
treatment options is provided by Vietti and Coghill (2015). For
present purposes it is sufficient to note that the addition of high
molecular weight, synthetic polymeric flocculants (commonly
used products include polyacrylamides and its derivatives) adhere
to clay particles, providing a bridging between and agglomeration
of particles, increasing the size of the resulting aggregate, also
termed a ‘floc’, thus increasing the rate of settlement of the
suspension.
Thickeners for providing more rapid settlement of treated
Figure 11. Measured change in partially sheared yield stress throughout
tailings have evolved rapidly over the past two decades or so, and
the tailings distribution system, where SP1 indicates immediately after the
a comprehensive discussion is provided by Bedell et al (2015). thickener and SP10 at the TSF (from Houman et al, 2007; courtesy
These thickeners, sometimes referred to as deep cone thickeners, Australian Centre for Geomechanics).
or paste thickeners, can be very large diameter, e.g. >50m, and
more than 20m high. The height enables a reasonably substantial Accounting for the shear-induced drop in yield stress is trivial,
overburden stress to develop within the thickener, generating if samples are taken at an appropriate location, e.g. the point of
some self-weight consolidation of the newly aggregated tailings.
deposition. During the design phase for a new TSF, the obvious
The aggregation process is further enhanced by the provision of
vertical rakes (pickets) that rotate slowly through the settling precaution is to test fully sheared samples of thickened tailings
slurry, enhancing the drainage and consolidation of the slurry. The prior to undertaking geotechnical tests such as settling and
thickened, or high-density tailings, is extracted from the base of consolidation tests. The effect of partial or no shearing of
the thickener, this being referred to as the underflow. At this stage flocculated suspensions is more acute in applications of the
of the process, the tailings have a measurable, but rather low yield second generic technique of tailings thickening, i.e. in-line
stress, unlike conventional tailings, which usually undergo flocculation.
minimal thickening and certainly insufficient thickening to impart
any measurable yield stress. 3.4.3.2 In-line flocculation
Yield stress is essentially just the undrained shear strength of This technique is now known by various names (e.g. secondary
a geomaterial, but measurements are in the pascal range rather flocculation or polymer treatment); we retain the term in-line
than kilopascal range. Indeed, the preferred method of measuring flocculation. In this technique, additional flocculant is introduced
yield stress has emerged as the vane test, as borrowed from into the tailings stream as it travels through the pipeline, with the
geotechnical engineering practice (Gawu and Fourie, 2004). Upon point(s) of flocculant addition being close to the pipeline end,
130
mainly to minimise shearing of the flocculated material. In-line
flocculation has many challenges, such as differences in the
viscosity of the flocculant and the tailings hindering mixing and
achieving the optimum mixing energy to maximise the
flocculation process but minimise post-flocculation shearing.
Reported applications of in-line flocculation almost invariably
report extremely high flocculant dosages, with values of 250g/t
reported by Wells et al (2015).
The primary reason for using in-line flocculation is rapid
water release and more rapid increase in the shear strength of near-
surface tailings. One of the pioneering applications of in-line
flocculation was to treat the problematic ‘mature fine tailings’, or
MFT that occur within oilsands tailings due to segregation of the
tailings stream upon deposition. Successful reported applications
of the technology are discussed by Wells et al (2015). After an
initial period of enthusiasm for in-line flocculation about ten years Figure 12. Difference between flocculated (PT) and untreated (UT)
ago, this enthusiasm has waned slightly, as operational challenges material (after Reid and Fourie, 2017).
became more apparent, examples being difficulties controlling
beach profile, in addition to the high costs, given the required 4 LABORATORY CHARACTERISATION OF TAILINGS
flocculant dosage concentrations. Nevertheless, the technology BEHAVIOUR.
continues to find some applications, with designers now aware of
potential pitfalls. 4.1 Index tests
Assuming successful implementation of an in-line As already noted, the specific gravity (SG) of mine tailings varies
flocculation deposition system, interest turns to potential changes widely. When preparing samples for SG testing, precautions may
in the geotechnical properties of the resulting tailings, particularly be necessary, for example tailings containing salts or other
as the structure (or fabric) of the material is so different from concentrated chemicals. This becomes even more important when
conventional tailings. working with process residues, such as bauxite red muds.
Most researchers who have investigated changes in hydraulic Similarly, pre-testing treatment may be required when carrying
conductivity report either an increase or no change, for tests out particle size analyses. Some industries have bespoke
undertaken at the same density for both thickened and
techniques that find relatively little application outside of those
conventional tailings (Jeeravipoolvarn et al, 2009, Beveridge et al,
specific industries. One example is the testing of oilsands tailings,
2015, Znidarcic et al, 2015 and Reid and Fourie, 2017). There are
thus two factors potentially benefitting water management and where the methylene blue index (MBI) test is used as an indicator
recovery with an in-line operation, i.e. the initial rapid water of clay content; this is used principally with the tailings fraction
release on deposition, and the increased rate of self-weight referred to as mature fine tailings. We limit further discussion on
consolidation provided by the higher hydraulic conductivity. the topic of industry-specific index tests and rather focus on
What is not so clear are potential changes in long-term strength. geotechnical characterisation.
The polymers used to date in in-line flocculation applications are,
to the authors knowledge, all biodegradables. Over time, the 4.2 Sedimentation, consolidation, desiccation and creep.
fabric induced by the high flocculant dosages used will likely be Laboratory testing techniques to evaluate volume change behavior
retained, as the flocculants degrade. The question is whether this and determine the consolidation parameters to be used in
induced fabric has the potential to be more sensitive (where modeling settlement and void ratio distributions of the disposed
sensitivity reflects the change from peak to post-peak strength, tailings in a TSF continue to evolve. For the design phase of a
usually measured in terms of undrained shear strength) than project, samples for testing should be obtained from a pilot mill
conventional tailings. processing plant which mimics the operational conditions in the
To determine possible long-term shear strength of highly future plant. However, upon starting the mine operation,
flocculated tailings, it is of course first necessary to determine additional samples should be collected and tested to confirm that
whether there are any short-term changes. Reid and Fourie (2017) the consolidation parameters in the design phase were realistic. If
clearly show that the critical state line (CSL) for heavily not, the consolidation analyses from the design phase should be
flocculated material lies significantly above the CSL of repeated and subsequently used to monitor the performance of the
equivalent, untreated material, with the difference in void ratio at disposed tailings. It is noted that some tailings exhibit various
100kPa mean effective stress, p', for example, being levels of particle segregation upon deposition, depending on their
approximately 0.1 (see Figure 12). This difference in void ratio is gradation and water content upon deposition. Particle segregation
significant, particularly when defining the boundary between creates a heterogeneous spatial distribution of the consolidation
potentially contractive and potentially dilative behaviour. For properties upon deposition and the laboratory testing program and
example, material deposited using in-line flocculation that is at a subsequent analyses should recognize this. It is noted again that
p' value of 100kPa and a void ratio of 0.5 would be well below its consolidation parameters of highly compressible tailings are
corresponding CSL; however, it is also well above the CSL of the highly nonlinear, with significant changes in compressibility and
untreated tailings. The question that arises is, if degradation of the hydraulic conductivity values over the range of stresses to which
flocculant results in an unaltered fabric and void ratio, is the the tailings will be subjected, so any analyses using “average”
tailings then potentially highly collapsible (contractive), a natural values of these parameters may often be meaningless. To address
analogy perhaps being the Rissa landslide in Norway in 1978, or this problem rationally, the sample collection and testing
is the process of degradation of flocculant accompanied by a strategies should include samples of various gradations from the
gradual decrease in void ratio, perhaps moving the state condition finest to the coarsest that are collected from the pool and beach
below the CSL of untreated material? This question appears to areas of the impoundment. Thus, a range of consolidation
remain unresolved. parameters is obtained enabling the settlement and void ratio
distribution analyses to “bracket” the expected behavior in the
impoundment or provide a rational method of assessing the
settlement rates and magnitudes in different parts of the
impoundment.
131
The selected samples are then subjected to testing in the The slurry behaves like a soil if the initial void ratio is below
laboratory to get the model parameters necessary for predicting 8. For a higher initial void ratio the material will sediment until
tailings behavior in the impoundment upon deposition. Depending the zero effective stress void ratio is reached. The zero effective
on the consistency or water content of the deposited material and stress void ratio for the kaolin clay is between 8 and 20, depending
the tailings management strategies adopted, upon deposition the on the water content at which the slurry was mixed. This is not a
processes controlling tailings behavior include: sedimentation, surprising result since the particle interaction at the very low
consolidation, desiccation and possibly creep. The goal of the effective stress will be dependent primarily on the characteristics
laboratory testing protocol should be to obtain the relevant of the diffuse double layer attached to the clay surface and the
modeling parameters for all these processes. Subsequently we chemical characteristics of the pore liquid, so appropriate process
discuss the testing protocol for each of these processes. water should be used in settling tests. When more free liquid is
available for each particle in more dilute suspensions, the double
4.2.1 Sedimentation layer will be thicker and the clay particles will reach equilibrium
In typical hard-rock tailings and other low compressibility
at a larger distance apart. This difference in the initial void ratio
tailings, particles tend to settle out along the beach soon after
creates different compressibility curves at low effective stresses.
deposition and thereafter the process of consolidation proceeds.
If a more detailed analysis of the sedimentation part of the process
However, in highly compressible tailings, consolidation is
preceded by sedimentation. is desired, a more elaborate testing protocol needs to be employed
Sedimentation is the process in which the soil particles settle following a procedure similar to the one described by Pane (1985)
in water until at the bottom of the settling column they create a in which he proposes a fluidization test in addition to the
soil layer with developing effective stresses and corresponding consolidation test to characterize the material behavior from
void ratio reduction. Unlike the settling process described by sedimentation to consolidation. In many practical applications the
Stokes’ law, in which the settling velocity of the particles is detailed sedimentation modeling is not critical and can be ignored
proportional to their size, during sedimentation all the particles in the analyses if the volume change associated with the transition
settle at the same velocity proportional to the local concentration from the initial void ratio at which tailings are deposited to the soil
of suspension. The process is properly described by the classical formation void ratio is properly accounted for. This is also valid
Kynch’s (1952) theory of sedimentation which was later for rapid-settling tailings as often very substantial bleed water is
augmented by work of Pane (Pane 1985, Pane and Schiffman released upon deposition on the beach, and the initial void ratio of
1997) who analyzed the combined sedimentation-consolidation the tailings layer is significantly lower than it was upon
process and developed experimental techniques for the deposition. Although not discussed in detail in this paper, properly
determination of soil sedimentation characteristics. accounting for the sedimentation process becomes extremely
The theory provides a convenient tool for the analysis of the important when sizing deep thickeners for clayey (i.e. highly
sedimentation process. However, from the practical point of view compressible) tailings used to produce high density tailings.
in tailings disposal problems the sedimentation process does not
usually play a major role and could be neglected in the analysis. 4.2.2 Consolidation
Namely, compared to the consolidation process that takes place Once a tailings layer is formed in an impoundment, some pore
over a period of several years, sedimentation is usually completed water will be expelled from the material during the consolidation
within several days. However, with the evolving interest in the use process. As stated earlier, the process is described by a nonlinear
of in-pit deposition, recognizing the different approach required finite strain consolidation theory (Gibson et al 1967) and the
for tailings where sedimentation is important is essential. material parameters must be determined by laboratory testing and
Nevertheless, the consolidation analysis will often be the only test analyses procedures that are consistent with the theory to be
analysis needed in the design of highly compressible used for both test interpretation and consolidation modeling. The
impoundments as regards volume changes. It is important to know consolidation process depends on both compressibility and
at which soil density or void ratio the sedimenting slurry becomes permeability characteristics of the soil. Power functions, fitted to
a soil and the effective stress principle applies. Unfortunately, this the experimental data, best represent constitutive properties of soft
value, which was called the fluid limit by Monte and Krizek
soils under one-dimensional compression (Somogyi et al 1981).
(1976), is not a material constant but depends on the initial water
Compressibility:
content of the slurry (Liu, 1990; You, 1993). This void ratio can
be easily determined in a simple experiment in which the slurry is e = A ´B (4)
allowed to sediment in a laboratory beaker and the surface soil is
sampled for the void ratio determination. This void ratio Hydraulic conductivity:
corresponds to zero effective stress and gives the upper bound to
the void ratio at which the mixture exists as a soil. Figure 13 shows k = C eD (5)
the results for a laboratory mixture of kaolin clay with water.
in which e is the void ratio, k is hydraulic conductivity, ´ is the
effective stress and A, B, C and D are material parameters that
need to be determined in laboratory testing.
In addition to these relations the zero effective stress void ratio
must be defined, bringing the number of parameters for
compressibility to three. Liu and Znidarcic (1991) expanded the
compressibility relationship to the form:
e = A( ´+Z)B (6)
so that the zero effective stress void ratio becomes e0 = AZB. There
is general agreement that the void ratio-hydraulic conductivity
relationship for highly compressible materials is well represented
by the stated power function, although several alternative forms
Figure 13. Relationship between initial void ratio and void ratio at onset for the compressibility relationship have been proposed over the
of development of effective stresses (i.e. zero effective stress void ratio) years. Any form of these functions is acceptable if they properly
for a kaolin clay (data from Liu, 1990).
132
model compressibility and hydraulic conductivity relationships
observed in the experiments.
The objective of the experimental work is to determine the
five parameters A, B, Z, C and D. Conceptually, the simplest
approach to determining these parameters is to take a tailings
sample and load it incrementally with several load increments
covering the stress range of interest and then directly measure
hydraulic conductivity at each load increment. This procedure
results in several void ratio-effective stress and void ratio-
hydraulic conductivity data points to which the analytical
expressions could be fitted and the corresponding parameters
determined. This simple procedure is often referred to as the
Slurry Consolidation Test and many variants of the procedure
have been implemented in various laboratories. There are two
challenges with this procedure. First, the test must provide reliable
data in the low effective stress range because for many slurries the
dominant volume change is observed in the low effective stress
range, (less than 10 kPa) and often even less than 1 kPa. Second,
to measure the hydraulic conductivity, a hydraulic gradient must
be created across the sample, which results in a variable effective
stress distribution, and a corresponding variation in void ratio and
variable hydraulic conductivity within the sample. This is
particularly critical at low effective stresses where the hydraulic
conductivity could easily change by an order of magnitude due to
the imposed seepage forces and self-weight of the material. These Figure 14 Seepage induced consolidation test equipment.
challenges often require imposition of low hydraulic gradients and
long testing times in highly compressible, low permeability
tailings to minimize these undesirable effects.
To address these challenges Imai (1979) suggested that
seepage forces could be used to trigger consolidation of the slurry.
By adding pore water pressure measurement needles in the
sample, he could account for the hydraulic conductivity variation
in the test analysis. The procedure required a specialized
apparatus and a very skilled operator to obtain reliable results.
Following the idea of using seepage forces to trigger consolidation
in slurry samples, Liu (1990), Abu-Hejleh and Znidarcic (1994)
and Abu-Hejleh et al (1996) developed a testing procedure and the
corresponding analysis to determine the consolidation
characteristics of slurries from a two-step test and an inverse
process solution approach to determine all five parameters for the
compressibility and hydraulic conductivity relationships. It
requires an initial permeation of the sample under a relatively high
gradient and a final step load test and direct hydraulic conductivity
measurement at the maximum stress level expected in the field.
The testing procedure significantly shortened the testing time and
the inverse problem analysis procedure determined the
Figure 15 Equipment used for the SICTA procedure.
consolidation parameters using the finite strain consolidation
theory. Thus, the experimental efforts and the data analysis Figure 16 presents the compressibility and hydraulic
method are completely consistent with the numerical models to be
conductivity test results for several tailings samples ranging from
subsequently used to analyze behavior of tailings upon deposition.
fine to coarse materials. Figure 16b presents both the step load test
The schematic of the laboratory equipment for the SICTA
results with the direct measurement of hydraulic conductivity at
procedure is presented in Figure 14. It consists of three major
components: the testing cell containing a slurry sample with a each load increment and the seepage induced consolidation testing
piston and load frame, a flow pump for controlling the imposed and analysis (SICTA) results (Abu-Hejleh and Znidarcic 1994).
flow rates and a differential pressure transducer connected to a Both procedures yield very similar test results, but the advantages
data acquisition system. Figure 15 shows an example of the of the SICTA procedure is evident in the results for fine-grained
system components in a laboratory. tailings. For the step load test results an assumption is made that
the effective stress, void ratio and hydraulic conductivity within
the sample are uniform and the data points represent the average
values for these quantities without any regard for the self-weight
stresses or induced seepage forces within the sample. The seepage
induced consolidation test analysis on the other hand correctly
accounts for the variation of these quantities within the sample
throughout testing. This is best illustrated by the data points in the
compressibility plot for fine tailings at the effective stress of 0.1
kPa. The upper data point represents the average void ratio at
steady state under the applied stress of 0.1 kPa, self-weight of the
sample and a very small seepage force during hydraulic
conductivity measurements under a low flow rate. The lower data
point represents the average void ratio at steady state under the
133
applied stress of 0.1 kPa, self-weight of the sample and a seepage shrinkage where the soil remains saturated from the initial void
force due to hydraulic conductivity measurements involving a ratio to the shrinkage limit void ratio. The second stage involves
flow rate four times higher than the rate used for the upper data zero shrinkage, where the soil does not show any volumetric
point. The higher flow rate increased the effective stress at the change from the shrinkage limit void ratio to the end of the
bottom of the sample to about 4 kPa. The SICTA procedure used desiccation process. To account for these two shrinkage stages, an
the data at steady state under the higher flow rate and the data at extension of the nonlinear finite strain consolidation theory as
the maximum effective stress to calculate all five model developed by Gibson et al (1967) was established by Abu-Hejleh
parameters A, B, Z, C and D and generate the continuous curves and Znidarcic (1995), Oliveira Filho (1998) and Yao et al (2002).
covering the whole range of effective stresses and void ratios The overall consolidation and desiccation process of soft soils is
observed in the test. modeled in this theory with consecutive phases which correspond
chronologically to the phases that a soft soil layer undergoes in
the field after deposition. These phases are consolidation under
one-dimensional compression, one-dimensional shrinkage,
propagation of desiccation cracks with tensile stress release, and
three-dimensional shrinkage. Several material functions are
required in the solution process for a given field situation. In
addition to the compressibility and hydraulic conductivity
relations for consolidation analysis, the shrinkage relations under
one- and three-dimensional conditions are needed. Additional
laboratory tests are required to determine these relations, but for
slurries they are often found to be very similar to the
compressibility relations used in the consolidation analysis and
additional tests might not be necessary (Oliveira Filho 1998). If a
detailed analysis of desiccation crack propagation is desired,
parameters for a cracking function must be determined. The
function that relates the void ratio at which the crack will open for
a given depth (total stress) and its parameters can be obtained from
a desiccation test conducted in a geotechnical centrifuge under
increased self-weight stresses. However, given the relative
scarcity of geotechnical centrifuge testing facilities, this step is not
commonly conducted.
Figure 16 Compressibility (a) and hydraulic conductivity (b) relationships
for several tailings samples.
The effect of seepage forces during permeation of the sample
to get the hydraulic conductivity characteristics is significant for
fine grained tailings under low effective stresses (< 1 kPa).
Neglecting these forces in the test analysis could lead to incorrect
representation of the consolidation characteristics of tailings. In
addition, the SICTA procedure requires a much shorter testing
time than the step load test with the direct hydraulic conductivity
measurement at each load increment. Depending on the hydraulic
conductivity of the material, a step load type of test may take
several months to complete while the seepage induced Figure 17 Shrinkage data for china clay (after Abu-Hejleh and Znidarcic,
consolidation test is typically completed with a couple of weeks 1995).
for fine grained materials (Abu-Hejleh et al 1996).
4.2.4 Creep
4.2.3 Desiccation In some instances, tailings processing includes treatments that
When the upward flow of water, due to the consolidation process create flocculated structures with significant viscous deformation
is completed, the desiccation process starts. Although we usually of the flocs. Examples include treated fine-grained oilsands
associate desiccation with evaporation, it is important to realize tailings (often referred to as mature fine tailings, or MFT), and
that the lowering of the groundwater table could cause desiccation fine-grained mineral sands tailings (‘slimes’). The nonlinear finite
as well. In general, the desiccation process may be defined as the strain consolidation theory does not properly capture this behavior
process in which soil compression takes place under increasing and an extended consolidation model that includes creep is needed
suction or negative pore water pressure. Thus, any time the pore (Murphey et al, 2021). It is also possible that some tailings,
water pressure drops below the atmospheric pressure the soil especially of high plasticity, exhibit time dependent compression
undergoes desiccation. As suction increases, the soil shrinks but even without any additional chemical treatment that would create
remains saturated until the air entry value is reached, after which a flocculated structure. In such cases the laboratory testing must
desaturation begins. This simplifies the analysis significantly. be expanded to include time dependent compressibility
Experimental shrinkage data, i.e. void ratio versus water content, characterization for modeling purposes. By necessity, such tests
are shown in Figure 17. They suggest that soil shrinkage can be require prolonged testing times, and a proper balance should be
modeled with two stages. The first stage involves normal found between test duration and the trustworthiness of the data. In
any case, any observed and documented creep effects in the
134
laboratory must be extrapolated by the creep model to longer provide some redundant data as well as creep data collection at all
times to accommodate time scale in the field. four load levels. The other extreme would be to test only one
A creep model recently developed for tailings accounts for the sample and collect creep data at each load level, implying a four-
time dependent compressibility by changing the three fold increase in testing time. Assuming that a minimum of three
compressibility parameters A, B and Z into time dependent weeks of creep data collection is desired, the four parallel tests
equivalents defined by the following relations (Gjerapic et al would require about six weeks of testing time, while for a single
2021): sample, the test would take six months to complete. It is also noted
that any creep model will rely on extrapolation beyond the testing
time in the laboratory up to the times of interest in the field. The
Af, Bf and Zf parameters in the proposed model (Eqs. 7 to 9) should
be considered as parameters to be used in sensitivity analyses
rather than the “true” parameter values at infinite time, as clearly
they can never be truly verified.
In a soil exhibiting creep, the total void ratio change is an
additive decomposition of the void ratio change due to the
effective stress change and the void ratio change due to creep per
se. The continuity condition requires that this total void ratio
change be consistent with the imposed hydraulic gradient and the
hydraulic conductivity of the material. The creep-induced void
ratio change is not associated with changes in the hydraulic
gradient and does not speed up the pore pressure dissipation
process. Creep only causes void ratio changes to be larger, but it
Equations (7), (8) and (9) allow for the time-dependency of does not significantly affect the void ratio change rate until the
constitutive properties A, B and Z by introducing additional consolidation process ceases. This does not imply that there are
(creep) parameters. At time equal to zero, parameters A, B and Z no creep induced void ratio changes during the consolidation
are defined by their initial values A0, B0 and Z0. Experimental process, as will be demonstrated later. To better elucidate this
evidence suggests that there is a limit in the change of parameters peculiarity a thought experiment is helpful. Let us consider a real
A, B and Z. Therefore, parameters Af, Bf and Zf are introduced to soil with infinite hydraulic conductivity. Such a soil would, under
define the final (limiting values) of parameters A, B and Z at the imposed stress change, such as self-weight or externally
infinite time. Parameters a1, a2, b1, b2, z1, and z2 are found by applied load, undergo instantaneous consolidation followed by
fitting laboratory data to Equations (7), (8), and (9) creep induced void ratio change directly correlated to elapsed
A set of laboratory tests on a sample, or multiple samples, of time. In that case the two separate void ratio changes would be
a slurry are required to collect data for the estimation of these completely uncoupled and easily distinguishable from each other.
constitutive parameters. In order to determine all parameters in the In a real soil however, the void ratio change will progress
proposed model including A0, B0, Z0, C, D, Af, Bf, Zf, a1, a2, b1, b2, gradually whereby the rate of change is controlled by Darcy’s law,
z1, and z2, a sample must be compressed under several load and any creep induced void ratio change will be undistinguishable
increments. The hydraulic conductivity should be measured at from the effective stress induced change. The creep induced void
each load increment and the loads should be sustained for a long ratio change will only very slightly affect the initial consolidation
enough period to collect sufficient data to model the long-term
compressibility, i.e. creep. From our experience with testing soft theoretically never ends, the effective stress induced void ratio
soils, we determined that for routine testing a set of four load changes become negligible on a relatively small laboratory
increments is sufficient to properly characterize the specimen after a period controlled by the hydraulic conductivity
compressibility behavior of these materials. At the low end of the of the material. In the past, 24 hours or the time required for
stress range, a seating load of 0.1 kPa is considered the lowest dissipation of the measurable excess pore water pressure was
value for which reliable data could be collected. The highest stress often considered to be long enough to assume the consolidation
of interest should correspond to the highest effective stress process to be completed. However, for slurries, taller samples, and
expected in the project. In between these two extreme values, two very low hydraulic conductivity materials that time can be
additional load increments are desirable to better characterize the substantially longer. Nevertheless, at some point the void ratio
material in the mid-stress range and to provide some data change with time can be considered to be caused by creep only
redundancy. Likewise, measuring the hydraulic conductivity at and the experimental data collected after that time can be used to
each of these loads provides sufficient data to correctly fit the creep model and determine the model parameters. The time
characterize the void ratio – hydraulic conductivity relationship. after which the creep model is fitted to the experimental data
When measuring the hydraulic conductivity the imposed should be determined by performing consolidation analysis for
hydraulic gradient should be as low as practical to minimize the each load increment to find at what approximate time the
void ratio variation within the sample during testing. This is consolidation process is essentially completed. Note, only a rough
particularly critical at low effective stress range at 0.1 and 1 kPa estimate of this time is sufficient.
load increments. As pointed out above, the SICTA procedure Once the creep model parameters are determined from the
(Abu-Hejleh et al. 1996) accounts for variations in effective stress data in the creep observation period, the model is used to
due to seepage. However, this procedure is only applicable for extrapolate the creep void ratio change to infinite time (AF, BF and
materials that do not exhibit creep and cannot be applied for this ZF) and to zero time (A0, B0 and Z0). While the parameters at
model without further development to account for creep in the infinite time should be used as variables for parametric studies,
analysis. the parameters at zero time have a role of separating the creep
The long duration tests required to capture creep behavior induced void ratio changes from those caused by the effective
come with a set of challenges that may compromise the quality of stress change during the consolidation process when experimental
the data. This includes potential gas generation, bacterial growth separation of the two effects is not possible. It is noted that while
or other long-term changes that might take place in the laboratory. this might appear to be a somewhat undiscriminating process, its
Thus, in each case a compromise must be found between the effect on any subsequent consolidation/creep analyses will be
quantity and quality of the data. In ideal circumstances, four minimal as only the total void ratio change is relevant for realistic
simultaneous tests would be performed, and each test would be predictions.
targeted for a prolonged testing time at one load level. This would
135
An example of the creep effect and its modeling for a water pressure. Loose soils and tailings will, when sheared
laboratory prepared bentonite slurry is shown in Figure 18. undrained, produce large positive increments in excess pore water
pressure, this often being referred to as shear-induced pore
pressures. More discussion on what constitutes ‘loose’ follows in
the section immediately after this.
Undrained triaxial compression tests on loose samples will
generally mobilise a peak deviator stress well before the effective
stress strength envelope is reached. This is best explained by
reference to a stress path plot, as shown in Figure 19, which plots
mean effective stress p´ against deviator stress q. The sample
shown has been isotropically consolidated to a mean effective
stress p0´ and then loaded in undrained compression. Being loose,
the undrained effective stress path follows the path shown,
reaching a peak deviator stress at a ratio that is much less than
would be achieved in an undrained compression test on a denser
specimen (indicated by the dashed line in the figure). A straight
line through the origin and the peak deviator stress value (with
multiple triaxial tests done at different effective confining stresses
to produce more points, preferably) defines the instability line,
often defined by:
= (10)
Undrained effective stress stability analyses are often carried
out using the vertical effective stress-dependent undrained shear
strength ratio,
Figure 18. Example of creep tests and associated modeling results
(parameters A, B and Z given in equations 7 to 9).
4.3 Shear strength under monotonic loading
Testing of tailings to determine appropriate strength parameters
has evolved rapidly over the past decade or so. Before discussing,
in some detail, the application of the critical state soil mechanics
(CSSM) framework to tailings testing, we discuss other aspects of
monotonically loaded tailings specimens.
4.3.1 Effective stress (drained) strength parameters
Stability evaluations of TSFs were for many years confined to
effective stress slope stability analyses, where key inputs were the
strength parameters c´ (apparent cohesion) and the angle of
internal friction, and the elevation of the phreatic surface.
Effective stresses were evaluated based on the phreatic surface
level. No account was paid to pore water pressures generated due
to shearing, which we later discuss in some detail, which are
crucial when considering the stability of loose, contractive and
particularly brittle tailings.
Unfortunately, it is still not uncommon to see stability Figure 19. Illustration of stress path of an undrained triaxial compression
evaluations based solely on effective stress (sometimes termed test on specimen that was initially looser than the critical void ratio.
drained) considerations, ignoring undrained stability evaluations,
despite the dangers of this approach having been emphasised Many national guidelines for TSF design and analysis now
many times over. Fortunately, this is changing. specifically call for both drained (effective stress) and undrained
The drained shear strength parameters are almost invariably stability analyses (ANCOLD, 2012, CDA, 2020), with the same
derived from results of triaxial compression tests, which may be Factor of Safety (1.5) being required for both conditions. In the
drained or undrained. An issue of concern, even with these case of ANCOLD it is specifically applicable for TSFs where loss
relatively simple tests, is the use of multi-stage loading tests, of containment is possible should a slope instability occur.
where a single sample is tested at three values of confining stress, The first experimental results showing the exceptional stress-
and only the peak deviator stress is determined. The justification strain response of a loose sandy soil subjected to an undrained
often made for multi-stage loading tests is limitations on the loading is shown in Figure 20 (Castro 1969). This sample of sandy
number of samples available for testing. However, it must be soil was first isotropically consolidated under 400 kPa, reaching a
recognised that any post-peak strength loss and the strain over void ratio after consolidation equal to 0.714, which is associated
which it occurs is thus lost, rendering the data useless for with a relative density of 37%. Over 14 minutes the sample was
implementation in any form of numerical modelling. axially loaded by means of small dead-load increments that
yielded a typical stress-strain curve with a peak vertical stress of
4.3.2 Undrained strength parameters around 230 kPa at approximately 1% of axial strain. Thereafter,
When the intrinsic tendency of volume change is suppressed in a the next small load increment triggered a sudden failure of the
saturated soil mass, for example due to fast loading due to seismic specimen and within a fraction of a second (0.18 seconds) it
disturbances, the soil response is undrained, and therefore, the reached about 20% axial deformation and the strength dropped to
volume change tendency is transformed into changes in pore a value of 80 kPa.
136
= log (12)
There is no fundamental physical basis for Equation 11. It is
merely intended to provide an empirical fit to the underlying data
points. Indeed, Equation 11 predicts physically inadmissible
values at very low stresses (e e < 0).
Empirical alternatives to Equation 11 are available to model CSLs
that exhibit curvature in e-log(p') space (e.g. Li and Wang 1998,
Bauer 1996). Naturally, the greater versatility of these alternative
equations comes at the expense of introducing a third fitting
parameter. Choosing between Equation 11 or more elaborate
alternatives largely depends on visual inspection of the data and
the envisaged application of the CSL. Ultimately, it is a matter of
Figure 20. Load-controlled CIU (isotropically consolidated, undrained) how closely the idealised CSL needs to follow the underlying data
triaxial test on loose sand (modified after Castro 1969). points, and whether the scatter and curvature of such points
warrants adopting a more complex alternative to Equation 11. It
As can be observed, the pore water pressure increases is interesting to note that even when characterising the exact same
significantly during the test, reaching a constant value close to tailings, some laboratories adopt Equation 11, while others adopt
90% of the confining pressure. From this type of test result, the more elaborate alternatives, such as a power law alternative (Reid
true liquefaction or flow failure was established, as well as the et al. 2021).
concept of critical state that provides the framework to describe
the behaviour of sandy soils and low-plasticity silts. Experimental 4.3.3.1 The Steady State of Deformation
results have shown that it is possible to establish an instability line The terms ‘steady state’ and ‘critical state’ are both used to
(Figure 19), which, if reached under an in-situ drained loading describe behaviour of soils at large strains. It is often suggested
regime (stress-controlled), will trigger a sudden increase in that they are one and the same, but in this paper we use the
strains, and generate positive excess pore pressures. Because the opportunity to take a short diversion to point out that there are
rate of pore pressure generation is greater than the time required differences. Whether these differences are of consequence when
to dissipate these shear-induced pore pressures, the drained characterising the engineering behaviour of tailings was
condition is almost instantaneously modified into an undrained extensively debated by the co-authors of this paper.
response of the soil, reducing the stresses to the undrained residual The term “steady state of deformation” was introduced by
strength. Poulos (1971) and it was defined as the state at which a particulate
A distinction has been made between ‘loose’ and ‘dense’ material deforms continuously under a constant state of effective
samples. Since these are relative terms, we must clarify what stress at constant velocity and at constant void ratio. The steady
constitutes loose and dense. Invoking the concept of CSSM state of deformation is achieved only after all particle orientation
provides a convenient way to make this distinction and to utilise has reached a statistically steady-state condition and after all
the resulting differences in behaviour in routine design and particle breakage, if any, is complete, so that the shear stress
analysis and interpretation of laboratory and field data. needed to continue deformation and the velocity of deformation
remains constant (Poulos 1981).
4.3.3 Application of critical state soil mechanics principles Roscoe and collaborators (Roscoe et al, 1958) introduced the
Critical state soil mechanics (CSSM) is commonly adopted as a critical void ratio line to refer to the locus in e-p'-q space
framework within which to understand the strength and volume associated with the ultimate state reached by a soil that is sheared
change characteristics of tailings (Jefferies and Been, 2015, and continues to deform at constant stress and constant void ratio,
Robertson et al. 2019, Bedin et al. 2012). The key tenet of CSSM as already discussed in this paper.
is that when soils are subjected to large shearing deformations, Later, Schofield and Wroth (1968) coined the term "critical
they eventually reach their critical state line (CSL) which is state soil mechanics", stating that the kernel of this concept is that
typically defined in a three-dimensional space: void ratio (e), soil and other granular materials, if continuously distorted until
mean effective stress space (p'), and deviator stress (q). It is they flow as a frictional fluid, will reach a well-defined critical
common to model the e-p' projection of the CSL by adopting a state that can be represented by two straight lines: one in the q-p´
semi-logarithmic linearization: plane and the other in the e-log p´ plane.
According to these definitions, both conditions (steady state
and critical state) seem to be identical in that both emphasize the
condition of an ultimate state at large strains of continuous
= log (11) deformation at constant void ratio and constant stresses. However,
the definition given by Poulos (1981) for the steady state has more
conditions to be satisfied, being:
where 10 is the slope and thus a measure of compressibility, p* is
a reference stress that ensures a dimensionless argument for the
- Constant velocity
logarithm, and p* is the void ratio when p' = p*. Equation 11 is
often used without explicitly including p*, which effectively - Particle orientation to reach a specific fabric
amounts to adopting p* as one unit of stress. The value of p* is - All particle breakage, if any, is completed
immaterial to the resulting CSL, although stronger correlations
between p* and soil properties emerge when p* takes on p' values According to these definitions, the critical state would be a
at which the underlying triaxial tests reached critical state (Cho et state that does not include the rate of deformation and orientation
al. 2006, Torres-Cruz 2019). As for the base of the logarithm, of particles, which are important factors that affect the mechanical
besides the base 10 used in Equation 11, it is also common to use behavior of clayey materials. It is worth mentioning that because
a natural logarithm. Changing the base b of the logarithm the critical state concept was applied successfully to develop the
well-known Cam-Clay model, its applicability for clayey soils is
produces a different slope b
often considered completely valid, regardless of the effects of
137
deformation rate and particle re-orientation. This situation needs Additionally, an example presented by Schofield and Wroth
some clarification, which likely is associated with the complex (1968) is reproduced in Figure 23, where it is clear that the
behavior of clayey materials, part of which is summarized below. residual strength developed by clayey soils with platy particles
The Japanese Kawasaki clay (Gs = 2.69, LL = 55.3, PI = 29.4, differs from the strength established at the critical state.
FC = 83.9%) has been tested at different rates of deformation
(Nakase and Kamei, 1986). Figure 21 shows the results of a series
of undrained triaxial tests started from a K0-consolidated
condition (zero lateral strain) and loaded in either compression or
extension. Three different rates of deformations were applied
during loading: 7x10-1, 7x10-2 and 7x10-3 %/minute.
Figure 23. Critical state strength and residual strength. Republished with
permission of ICE Publishing, from Skempton (1964); permission
conveyed through Copyright Clearance Center, Inc.
Regarding the requirement associated with the completion of
the particle breakage, if it exists, we deem it to not be relevant in
the case of fine-grained materials.
According to the above discussion of the behavior of clayey
materials, it is possible to assert that the critical state is applied to
these soils without actually accomplishing steady state behavior.
Therefore, the critical state is a less rigorous framework, but is
sufficiently satisfactory even for practical applications in fine-
grained materials.
Figure 21. Undrained response of Kawasaki clay at different rates of
Conceptually, the steady state definition is a better
deformation (after Nakase and Kamei, 1986).
representation of the fundamental behavior of soils. However, if
These results clearly show that the rate of deformation has an applied rigorously, a non-unique (rate dependent) steady state will
important effect on the stress-strain curves and on the undrained be obtained in the case of fine-grained soils, although in the case
strength; the higher the rate of deformation the higher the of sandy soils its application is more direct, because rate effect
undrained strength. However, the angle of friction mobilized at and re-orientation of platy particles are not particularly relevant to
the ultimate state or the critical state is not influenced by the rate these soils. Particle breakage may occur in sandy soils, and if it
of deformation. This means that the critical state line in the q-p´ occurs this effect is reflected in the resulting steady state line.
plane is unaffected by the rate of deformation, but the critical state Therefore, it may be concluded that the critical state only has
line in the e-log p´ plane is definitely influenced by the rate of two requirements: the ultimate state of a soil can be represented
deformation. For clayey soils, the systematic dependence of the by one straight line in the q-p´ plane and another in the e-log p´
undrained strength on the rate of strain has been reported by plane. This simplicity is probably the key to the practical
different researchers (e.g. Bjerrum 1969; Sheahan et al, 1996; applicability of the CSSM, even in fine-grained soils.
Díaz-Rodríguez et al, 2009; Chow et al, 2012). In the case of tailings, the applicability of either of these
On the other hand, particle orientation in platy particles of frameworks depends on whether behavior is closer to a fine-
clays has been identified as the cause of the development of a grained soil or to a granular soil. Considering that the most
“residual strength” that is reached at significantly large catastrophic failures that have affected tailings dams are
deformations and under very low rates of deformation (Bishop et associated with liquefaction, the tailings that behave as sandy
al, 1971; Lupini et al, 1981; Skempton, 1985). As an example, in materials are those that are of greatest interest in terms of their
Figure 22 the stress ratio-displacement curve obtained from a ring mechanical response when assessing physical stability of tailings
shear test is presented where, after about 30 days of shearing, a deposits.
residual strength associated with a friction angle, r = 8.6o , was A comprehensive study showing the robust framework
mobilized in Kalabagh clay. provided by the steady state of deformation was reported by
Ishihara (1993) using the Japanese standard Toyoura sand. In
Figure 24 a series of CIU triaxial test results are presented in terms
of stress-strain and effective stress-path curves for samples of
different void ratios consolidated at the same isotropic pressure of
1 MPa. The tremendous effect of the density is apparent in the
mobilized undrained residual shear strength, which in this case
varies from zero to 2 MPa, associated with soil responses from
contractive to dilative.
Figure 22. Development of residual strength in Kalabagh clay.
Republished with permission of ICE Publishing, from Skempton (1985);
permission conveyed through Copyright Clearance Center, Inc.
138
plotting the experimental data with a logarithmic horizontal axis.
For instance, Ishihara (1993) reports the e-p' projection of the SSL
(essentially the same as the CSL, as discussed previously) of
Toyoura Sand for p' values that exceed 3 MPa (Figure 26). When
p' is plotted on an arithmetic scale, the SSL is almost linear with
a slightly concave up shape. There is certainly no evidence of an
inflection point. Conversely, if p' is plotted on a logarithmic scale,
the data in Figure 26 plot as a distinctly curved SSL. In this case
Figure 24. CIU triaxial test results at a confining pressure of 1 MPa and at least (and likely in many more), the SSL (or CSL) curvature is
different void ratios (after Verdugo, 1992). not necessarily related to particle breakage. This assertion is
backed up by results of a numerical study using the Discrete
Another series of CIU triaxial test results are presented in Element Method (DEM) reported by Nguyen et al (2018). They
Figure 25, where the samples were tested with the same void ratio, simulated drained and undrained triaxial compression tests to
e = 0.833, (after consolidation) and at consolidation pressures of determine the CSL of a collection of varying sized ellipsoid
0.1, 1, 2 and 3 MPa. These experimental results show that for a particles. The resulting particle size distribution of the particles
given void ratio, the undrained strength mobilized at large strains had a d50 = 1mm and a coefficient of curvature of 1.5, producing
is unique regardless of the initial confining pressure. Additionally, a simulated material not unlike a uniformly sized tailings material
the effective stress-paths vary from dilative to contractive (aside from the particle shape). The key aspect of this work is that
depending on the initial confining pressure with respect to the particle breakage was not allowed. Nevertheless, the distinctly
ultimate state, or steady state of deformation. curved CSL shown in Figure 27 was produced; another indicator
that curvature of the CSL may have nothing to do with particle
breakage.
Figure 25. CIU triaxial test results at different confining pressure and same
Figure 26. Steady state projections of Toyoura sand (after Verdugo, 1992).
void ratio after consolidation. Republished with permission of ICE
Publishing, from Ishihara (1993); permission conveyed through Copyright
Clearance Center, Inc.
4.3.3.2 Implications of a curved CSL
Although it is often reasonable to use the semi-logarithmic
representation of the CSL (equation 11), accounting for curvature
of the CSL, when it clearly occurs, does have important
implications. As hypothesised by Robertson (2017), tailings
exhibiting a curved CSL are less brittle at higher effective stresses
than at low stresses. When sheared undrained, an element of loose
tailings at a high mean effective stress, p´, will decrease to a lower,
but still substantial value of p´, whereas a loose element at an
initially low value of p´ reduces to a very small, perhaps even zero
value, i.e. true liquefaction.
Curvature of the CSL has often been attributed to particle
breakage (Bedin et al, 2012 and Schnaid et al, 2013, among
others). Valenzuela (2015) suggests that at pressures (presumably
vertical stresses) greater than 1MPa, breakage of the more angular Figure 27. Curved CSL derived from triaxial simulations (using DEM) of
a collection of ellipsoidal particles (after Nguyen et al, 2018).
edges of tailings particles occurs. However, the curvature of the
CSL may, in many cases, be nothing more than an artefact of
139
4.3.3.3 The q-p' projection of the CSL. CSL variability. Admittedly, such under estimation is unlikely to
Selection of an equation to model the q-p' projection of the CSL be significant if at least one intermediate gradation is also tested.
is straightforward because unequivocally the data define a In contrast to adopting particle size distribution to guide the
proportional trend: sample selection process, the analysis of a database of over 150
soils suggests that, regardless of silt content, the minimum void
ratio (emin) is a suitable index property to assess the variability of
CSLs in non-plastic soil deposits including tailings (Torres-Cruz
= (13) 2019, Torres-Cruz and Santamarina 2020). Namely, emin
correlates linearly to p* and to the upper bound of 10. The
determination of emin in tailings may seem an unusual proposition
where the slope M is termed the critical state friction ratio. As M because standardised emin protocols are generally designed for
depends on particle shape (Cho et al. 2006, Sadrekarimi and Olson sands, as opposed to the often silt-dominated gradations of
2011), it is conceivable that the CSL may deviate from the simple tailings. However, there are numerous precedents in the technical
linear trend proposed by Equation 13 when the stresses are high literature that support the reliability of emin determinations in silts
enough to induce particle breakage. However, we are not aware (e.g. Zlatovic and Ishihara 1995, Cubrinovski and Ishihara 2002,
of any practical cases in which this has been observed. Thevanayagam et al. 2002, Papadopoulou and Tika 2008, Yang et
M depends on the magnitude of the intermediate stress al. 2006, Carrera et al. 2011, Torres-Cruz and Santamarina 2020).
relative to the other two principal stresses. Typically, the triaxial
2 3) condition is taken as a reference and the 4.3.3.4 CSL Experimental procedures.
resulting slope denoted Mtc. This reference condition can then be Considerable effort has been devoted to refine the procedures used
scaled to other stress conditions (Jefferies and Been 2015). to measure the CSL. The following are some of the most salient
aspects of the available guidance.
4.3.3.4 Sample selection
CSLs depend on intrinsic particle characteristics such as a. In principle, both dilatant and contractive specimens
mineralogy, particle shape, and particle size distribution. approach the same critical state line (Been et al. 1991,
Specifically, the e-p' projection is sensitive to variations in any of Verdugo and Ishihara 1996). However, dilatant specimens
these characteristics while the q-p' projection is only sensitive to are prone to strain localisations that imply the routinely
variations in mineralogy and particle shape (Sadrekarimi and measured global void ratio of the specimen may not be
Olson 2011, Li et al. 2015, Torres-Cruz 2019). Investigations into representative of the localised zone of the specimen that
the variability of CSLs within a single tailings deposit generally reached critical state (Desrues et al. 1996). Accordingly,
show that the different gradations exhibit different e-p' CSL determinations should rely on contractive specimens
projections, but essentially the same q-p' projection (Fourie and (Been and Olivera 2015).
Papageorgiou 2001, Carrera et al. 2011, Li and Coop 2019, b. The large shear strains required to reach the critical state
Torres-Cruz and Santamarina 2020). The positive practical destroys the effect of the initial fabric of the specimen (Been
implication of this is that any potential variations in mineralogy et al. 1991, Verdugo and Ishihara 1996). Indeed, numerical
and particle shape within the deposit are sufficiently constrained simulations indicate that, for a constant intermediate
to prevent a significant effect on the q-p' projection of the CSL.
principal stress ratio, soils reach a constant fabric at critical
This seems reasonable in TSFs that receive tailings from a single
state (Zhao and Guo 2013). Consequently, the CSL is
plant or at least from a single ore body. On the negative side, there
considered independent of the initial fabric (i.e. preparation
is the implication that thoroughly characterising the e-p'
projections of the CSLs of any given tailings dam requires testing method and consolidation stress path) of the specimen. The
multiple gradations. This immediately leads to the question: what preferred preparation method becomes the one that provides
is a robust approach to select the samples whose CSLs will be greatest control over the initial void ratio, as opposed to the
experimentally determined? This requires careful consideration one that best mimics the in-situ depositional environment.
because CSL determination is a fairly resource intensive The moist tamping method has proven very effective for this
enterprise, with each point of the CSL requiring one triaxial test purpose (Been and Olivera 2016). It is worth noting, though,
with careful void ratio measurements (Been et al. 1991, Verdugo that the thin layering that often occurs during hydraulic
and Ishihara 1996, Reid et al, 2021). The investigations into recent deposition may lead to differences between the CSLs of
tailings dam failures suggest that, for practical purposes, CSLs are undisturbed and moist-tamped specimens (Verdugo, 2009,
defined with anywhere between 4 and 8 points (Morgenstern et al. Reid and Fanni, 2022). We also note that although the
2016, Jefferies et al. 2019, Robertson et al. 2019). Occasionally, undercompaction method (Ladd 1978) is often used to
even more tests may be required to achieve a well-defined CSL. promote greater density uniformity across the specimen, the
Clearly, samples must be chosen carefully before committing to method does not appear to have a systematic effect on the
such comprehensive testing. resulting CSL (Reid et al. 2021).
Ideally, the gradations selected for CSL testing should enable c. Both drained and undrained specimens are useful in CSL
assessment of the variability of CSLs within the tailings deposit. determinations (Been et al. 1991, Verdugo and Ishihara
As the focus is on the variability of the e-p' projection, and 1996, Been and Olivera 2016). In general, the low stress
assuming Equation 11 is valid, the selected gradations should portion of the CSL can be more easily determined with
disclose the variability of p* and 10 within the tailings deposit. undrained specimens, while the CSL at higher stresses can
A common approach is to select samples that span the be readily determined using drained tests. Once M is known
predominant range of gradations in the deposit, from the finest to from initial tests, drained tests have the advantage of
the coarsest material (Jefferies and Been 2016, Jefferies et al. allowing estimates of the initial consolidation stress p'i
2019, Robertson et al. 2019). In plastic tailings, this approach required to reach a target p' value at critical state. Indeed,
seems reasonable because the finer material will typically be more for a specimen that is initially consolidated anisotropically
10
to a stress state (p'i, qi), tested along a straight effective
(Schofield and Wroth 1968). However, the silt content of non- stress path with a slope sl, and which aims at having a mean
plastic soils does not have a systema
effective stress p'cs at critical state; the required value of p'i
exhibit a monotonic correlation with p* (Thevanayagam et al.,
is given by:
2002). Accordingly, testing only the coarsest and finest gradation
of a tailings deposit may lead to a systematic under estimation of
140
+ g. The large strains required to reach a critical state (i.e.
= (14) constant e, p', and q) imply that triaxial tests often do not
reach true critical state conditions, or perhaps steady state
conditions, as discussed in section 4.3.3.1 (Verdugo and
For the common case of isotropic consolidation (qi = 0) and Ishihara 1996, Murthy et al. 2007). Accordingly, assessment
shearing under constant cell pressure (sl = 3), equation 14 of critical state conditions may be slightly improved by
becomes: extrapolating triaxial test results to conditions of zero
volume change or zero pore water pressure change, for
drained and undrained tests, respectively (Carrera et al.
(3 ) 2011). While the extrapolations are logically sound, there
= (15)
3 appear to be no systematic investigations into the change of
CSL parameters that result from performing these
extrapolations.
Equations 14 and 15 are particularly useful when running
tests that aim at populating specific portions of the p' 4.3.3.5 Reproducibility of CSL measurements
domain of the CSL. As the engineering of tailings dams becomes increasingly reliant
d. Reliable void ratio measurements are essential for CSL on critical state soil mechanics concepts, it becomes critically
determination. Given the difficulties in measuring the void important to assess the reliability with which the CSL can be
ratio changes that occur during specimen saturation, the measured. For the practicing engineer, the question becomes one
recommended approach is to determine the void ratio by of assessing the level of agreement that can be expected amongst
measuring the water content at the end of the test. This can the CSLs determined by different laboratories. That is, the
be accomplished by freezing the specimen at the end of the reproducibility of CSL determinations. The issue of CSL
test (Sladen and Hanford 1987, Been et al. 1991), or by reproducibility was recently investigated by means of a round
increasing the cell pressure at the end of the test to squeeze robin testing scheme in which 15 soil laboratories, both
out as much water as possible, thus making the specimen commercial and academic, from around the world tested the same
easier to handle upon disassembly of the triaxial cell tailings (Reid et al. 2021). The laboratories were instructed to
(Verdugo and Ishihara 1996). determine the CSL of a sample of sandy silt gold tailings using
e. The use of enlarged and lubricated end platens has been between 4 and 7 triaxial tests and focusing on the p' stress range
recommended to promote constant radial strains along the between 50 and 800 kPa. No additional instructions on how to
height of sand specimens (Been and Olivera 2015). measure the CSL were given to the laboratories. As such, the
round robin programme provided unique insights into how the
However, these modified platens are not entirely effective
variability in CSL testing protocols affect reproducibility. Herein
(Rees 2010) and we are not aware of any investigation
we merely comment on the differences in reproducibility of the e-
showing that they influence the resulting CSL. Quite the
p' and q-p' projections.
opposite, an investigation into CSL reproducibility showed
that enlarged and lubricated end platens do not significantly
affect the CSL (Reid et al. 2021).
f. Detailed descriptions of CSL experimental procedures
generally do not address the connection between the loading
ram and the top cap (Been et al. 1991, Verdugo and Ishihara
1996, Been and Olivera 2015). However, a recent study has
specifically recommended the adoption of an embedded
connection between the top cap and loading ram as
illustrated in Figure 28 (da Fonseca et al. 2021). The
rationale for recommending this type of connection is that it
eliminates tilting of the top cap and misalignment between
the loading ram and the specimen. This type of connection
is different from the more widely adopted connection in
which the top cap is allowed to rotate (Reid et al. 2021).
Importantly, the embedded connection appears to yield
slightly, but consistently, higher values than those yielded
by rotating connections (da Fonseca et al. 2021, Reid et al.
2021). This issue thus appears to be worthy of further
investigation.
Figure 28. Loading ram embedded into top platen: a) Sketch and b)
Figure 29. e-p' data points from a CSL testing round robin exercise. a) All
photograph. Source: da Fonseca et al. (2021).
13 laboratories with available data and b) 11 screened laboratories. Data
source: Reid et al. (2021).
141
Figure 29 shows the critical state e-p' data points measured by CSLs, the effect of the scatter can be formalised by computing the
the 13 laboratories whose detailed results were included in the standard error of p* and of 10 which provide a statistical
supplementary files of Reid et al. (2021). As might be expected, assessment of the precision of p* and of 10 (Torres-Cruz and
the scatter of the dataset as a whole is significantly larger than the Santamarina 2020, Reid et al. 2021). Since the data points are
scatter of the data of any individual laboratory. Indeed, most of centred around p' = 100 kPa, we adopt p* = 100 kPa, and therefore
the data points fall within a band that has a height of = 0.1. we work with 100. As shown in Figure 31, the two data points of
Reid et al. (2021) noted that three of the laboratories adopted the Fugro dataset with the largest values of p' are excluded from
procedures that are known to significantly compromise the quality the regression as they may suggest a curvature that cannot be
of CSL determinations, and one laboratory generated a clearly captured with Equation 11. The resulting standard errors of 100
outlying dataset. On this basis, Reid et al. (2021) excluded the data (SE- 100) and of 10 (SE- 10) are shown in Figure 31. As expected
from these four laboratories from further analysis and focused on from the scatter, the values of SE- 100 and of SE- 10 of the BGC
the 11 datasets shown in Figure 29b. The screening reduced the dataset are lower than the corresponding values of the Fugro
scatter in the data, with most data points now falling within a band dataset. That is, the BGC dataset enables more precise estimates
that has a height of = 0.07. However, the data suggests that of 100 and 10.
there is still room for improvement, as single laboratories are
generally able to measure CSLs with a scatter of less than =
0.02. See for example the data from UWA, BGC, Golder Perth, 0.8
KCB, UPorto, and Golder Vancouver in Figure 29. Another
important lesson to be taken from this data is the need for BGC:
consultants to query the experimental procedures used by 100 = 0.63, SE- 100 = 0.002
laboratories since deviation from good practice can result, as 0.7 10 = 0.082, SE- 10 = 0.004
detected by Reid et al. (2021), in spurious CSLs.
Figure 30 shows all the available critical state points in q-p'
space. Even without excluding data from any of the laboratories,
there is much better convergence in q-p' space than in e-p' space 0.6
Void ratio e
(compare Figures 29a and 30). Since there are no void ratio Fugro:
calculations involved in the q-p' projection of the CSL, it appears 100 = 0.60, SE- 100 = 0.008
that the lower reproducibility of the e-p' CSL projection is due to 10 = 0.044, SE- 10 = 0.012
0.5
difficulties in reliably measuring the relevant volumes, and the
associated changes, of the specimens.
Excluded from
0.4 regression
3000 BGC
UWA BGC
UC Lab A Fugro All
Golder Perth KCB 0.3
2500
UPorto UofT 1 10 100 1000 10000
PUC UC Berkeley
Fugro Golder Vanc. Mean effective stress p' [kPa]
Deviator stress q [kPa]
2000 UTFSM
Figure 31. CSL data points that exhibit different degrees of scatter. Data
source: Reid et al. (2021).
1500
Comparisons between SE- 100 and SE- 10 require
normalisation of the values. A possible normalisation is to divide
1000 SE- 100 and SE- 10 100 10,
respectively, that can be observed in soils in general. This yields
the normalised values NSE- 100 and NSE- 10 (Torres-Cruz and
Santamarina 2020). In the absence of a database that covers the
500
full spectrum of soil types, we rely herein on a CSL database of
160 non-plastic soils (Torres-Cruz 2019, Torres-Cruz and
Santamarina 2020). This database suggests that, for non-plastic
0
100 10 varies from
0 500 1000 1500 2000
virtually 0 to 0.3 (range of 0.3). Previous analyses had put the
Mean effective stress p' [kPa] 100 at 0.2 (Torres-Cruz and Santamarina 2020,
100
Figure 30. All available q-p' data points from a CSL testing round robin
numerical models, not in actual soils. So the slightly revised lower
exercise. Data source: Reid et al. (2021).
bound of 0.3 is adopted herein.
The NSE- 100 and NSE- 10 of the BGC and Fugro datasets are
4.3.3.6 Confidence in estimated CSL parameters. plotted in Figure 32 along with the values that correspond to 91
An issue that is often overlooked in the determination of CSL CSLs analysed by Torres-Cruz and Santamarina (2020). The data
parameters is the confidence with which they can be estimated show that the NSE- 10 is always larger than NSE- 100. That is, the
considering the scatter of the underlying points. This is 100 estimates is better than the relative
particularly relevant to the e-p' projection of the CSL since this is 10 estimates. Indeed, most of the data points indicate
the projection prone to greater scatter (compare Figures 29a and that NSE- 10 is between three to twelve times NSE- 100.
30). To illustrate this point, consider two of the datasets from the
CSL round robin programme mentioned above (Reid et al. 2021):
BGC and Fugro. Figure 31 shows the critical state data points of
these two datasets in e-p' space. Less scatter is apparent in the
BGC dataset, suggesting that these data points lead to a more
precise definition of the CSL. Adopting Equation 11 to model the
142
Figure 32 100 10. Data
sources: Torres-Cruz and Santamarina (2020) and Reid et al. (2021).
Figure 32 provides a useful benchmark against which to
assess the effect of scatter on the fitting parameters of Equation
11 100 10. In Figure 32, the closer to the lower left corner a Figure 33. Illustration of the effect of assumed residual strength ratio on
100 10. the resulting cost of a stabilising buttress (after Sarantonis et al, 2020).
However, it is important to keep in mind that precision and
accuracy do not always go hand in hand. That is, low scatter in the There are two generic approaches to determining the residual
e-p' points does not necessarily mean that the CSL has been shear strength of a material. One is the use of empirical
measured correctly. This point was also illustrated in the Reid et correlations based on back-analyses of documented failures, and
al. (2021) round robin exercise in which several CSL entries were the other is direct measurements using either in-situ tests or
correctly disregarded due to lack of adherence to proper testing laboratory tests (sometimes both). There are strong advocates of
protocols, regardless of scatter. Clearly, accuracy should take both generic approaches, e.g. advocates of the former insist that
precedence over precision. As the advice of the statisticians goes: void ratio localisation in most laboratory element tests render
it is better to be approximately right, than exactly wrong. interpretation of such results potentially misleading (Weber,
2015). Laboratory testing advocates point to the enormous
4.3.4. Residual strength of tailings. uncertainties in mobilised shear strength in a failing mass, e.g. due
As already mentioned, there is a growing call to evaluate the to varying rates of shearing, together with many of these analyses
stability of existing TSFs on the basis of residual (large strain) not accounting for effects of momentum when evaluating the
strengths, when potentially liquefiable tailings exist in the zone of shear strength that is required to produce the observed failed
the TSF that may be incorporated in a potential failure. If geometry. Additionally, Robertson (2010) pointed out that many
liquefiable tailings only exist in the centre of a TSF, distant from failure case histories involved retrogressive sliding, the
the retaining perimeter embankment(s), then this issue is Merriespruit TSF failure in South Africa in 1994 being a good
example (Fourie et al., 2001), which is rarely accounted for in the
irrelevant. What is of concern is the presence of potentially
backanalyses. Evaluation of the liquefied shear strength based on
liquefiable tailings near the perimeter. The emerging approach is
such examples is thus questionable. From results presented by
thus, ‘if the tailings can liquefy, assume they will’. For example,
Been and Jefferies (2006), estimating residual strength ratios
some time ago, Silvis and de Groot (1995) suggested that based on backanalyses of failures provides a (sometimes very
triggering should always be assumed if the soils (tailings) are significantly) conservative estimate when compared with
susceptible to strength loss, and Morgenstern (2018) strongly predictions using laboratory test data and a critical state
counselled along similar lines, suggesting, “that liquefiable constitutive model. Since the work of Been and Jefferies (2006)
deposits that can liquefy be assumed to do so and that containment there have been further attempts to reconcile backanalysed
be provided by a buttress of non-liquefiable unsaturated tailings strengths with laboratory test data, but a unified approach appears
and/or compacted dilatant material”. to remain elusive. Perhaps the complexity of the stress-strain
Emphasis is moving away from trying to engineer out all behaviour of very brittle, loose soils and tailings and the attendant
plausible mechanisms that may trigger liquefaction and rather to implications of realistically analysing CPTu data, as discussed
assume that some unknown (even potentially unknowable trigger) later, will make such a rapprochement impossible, at least in the
will occur at some time. In many cases this might simply be the short to medium term.
MDE (maximum design earthquake) or even lower magnitude Laboratory data, primarily from triaxial compression tests and
earthquakes. However, the cost implications of such an approach direct simple shear (DSS) tests are currently widely used in
can be significant, with the authors aware of many recently practice for residual strength assessments. Some examples, drawn
completed and currently under construction buttresses that cost in from experience with copper tailings in Chile are provided here to
the order of tens-of-millions of USA dollars. Furthermore, as illustrate the range of factors to consider when testing.
demonstrated so well by Sarantonis et al (2020), the importance Figure 34 presents values of the residual undrained strength
of the assumed post-liquefaction strength ratio on the resulting as a function of the void ratio for copper tailings, considering
buttress cost is very significant. They provide a summary of their different fines contents (where ‘fines’ refer to particles finer than
analyses of a hypothetical TSF, buttressed using rockfill, and plot highlight the effect of particle size
the cost (in millions of dollars) per 100m length of buttress against distribution on mechanical response.
the assumed liquefied (residual) strength ratio (Figure 33). As
noted by Sarantonis et al (2020), as the residual strength ratio is
varied between 0.05 and 0.2 (a plausible range), there is a three-
fold increase in cost. Clearly, accurate and reliable determination
of residual strength of tailings is a key challenge to the industry.
143
particularly by Dyvik et al (1987), is that changes in vertical stress
required to ensure constant volume conditions (i.e. no change in
height) are equal to the measured pore water pressure in truly
undrained simple shear tests. This means that tests in the DSS can
even be carried out on dry specimens and if constant volume
conditions are maintained (by adjusting the vertical total stress as
shearing progresses) the pore pressures that would be developed
during undrained tests can be determined.
The DSS device does, however, have drawbacks. In
particular, effective stress strength parameters cannot be correctly
determined as the principal stresses are unknown (and
unknowable explicitly). In fact, principal stresses rotate during
shearing, with the extent of rotation a function of the friction angle
(Budhu, 1984). Atkinson et al (1991) showed that the friction
Figure 34. Residual undrained strength of Chilean copper tailings. angle obtained in a DSS is smaller than that obtained from a
triaxial test, results confirmed by Dounias and Potts (1993).
These results are particularly relevant in applications that use Despite these limitations the DSS test has, for many years
cyclones to reduce the fines content in the sandy tailings, as used been widely used in Scandinavia in the design of TSFs, with the
in the construction of tailings sand dams, which is a widely view apparently being that it provides a conservative estimate of
practiced in Chile and Peru. shear strength (Knutson, 2018). However, this extra degree of
Although widely used in practice, the strain limitations of ‘safety’ cannot be quantified, and use of the triaxial test for
typical triaxial and direct simple shear devices are such that it is determining effective stress strength parameters for TSF design is
not always possible to reach steady state (or indeed critical state) now becoming more widespread in Sweden, particularly after the
conditions in these tests. There is also emerging evidence that comprehensive work of Bhanbro et al (2017) on sandy tailings,
boundary conditions, particularly in the DSS may be such that which clearly showed the extent of these differences.
post-peak behaviour measured in these devices is not Notwithstanding its limitations, the DSS remains a tool
representative. Although a ring shear device, particularly such as worthy of consideration particularly for determining undrained
described by Suits et al (2009), holds promise in determining strength ratios for the horizontal and near-horizontal sections of
residual strengths at very large strains, the authors were not aware potential failure planes in slope stability analyses. It is worth
of any measured undrained residual (or large strain) shear noting that DSS tests have been extensively used in many of the
strengths of tailings using such equipment. This we regard as an recent forensic evaluations of TSF failures (Morgenstern et al,
important research need. 2016, Jefferies et al, 2019).
4.3.5 Simple shear testing 4.3.5 Loading conditions other than triaxial compression and
Considering that a significant portion (often the majority) of simple shear
potential failure surfaces may be sheared under conditions much Although not common in tailings engineering, other loading
closer to simple shear, there is an overreliance on triaxial conditions that have been used include triaxial extension, plane
compression testing. Ladd (1991) showed how soil tested under strain and hollow cylinder torsional shear tests.
different loading conditions (e.g. triaxial compression, triaxial Triaxial extension tests provide strength data that are
extension and simple shear) resulted in different values of considered relevant to a relatively small length of a typical failure
undrained strength ratio. He presented results as a function of plane in a TSF slope stability analysis, typically the downstream
plasticity index (PI) and showed the divergence increased as PI tail of potential failure surface, see Sadrekarimi (2016). As shown
decreased. In other words, the difference will be greatest for by Ladd (1991), the undrained shear strength ratio in triaxial
non-plastic tailings, typically silty and sandy tailings, exactly the
extension is lower than both triaxial compression (TC) and SS,
type of material particularly prone to brittle behaviour if loose of
but difficulties in carrying out triaxial extension tests (typically
critical.
stress concentrations causing sample necking, which renders area
An alternative to triaxial compression tests, particularly to
characterise undrained strengths along horizontal or near- calculations nigh impossible, and some difficulties ensuring
horizontal sections of a potential failure surface was suggested as sample contact with the loading top cap) have usually meant that
being simple shear (SS) testing. In a true SS device, the sample is strength data from either TC or SS tests are used for this section
sheared under conditions of pure shear, with no complementary of the failure surface.
shear stresses on any of the boundaries. The Cambridge SS device Intuitively, a plane strain loading apparatus is highly suited
as described by Roscoe (1953) comes close to imposing true for determining shear strengths relevant to TSF slope stability
simple shear loading conditions. However, practicality has analyses, as these are usually well represented using a two-
resulted in other SS devices being developed, with pioneering dimensional simulation. However, plane strain devices are not
work completed at both the Norwegian Geotechnical Institute common, and usually only available in research institutions. Much
(NGI) and Swedish Geotechnical Institute (SGI). In these devices, valuable work on plane strain testing of static liquefaction of loose
a cylindrical specimen is confined in the horizontal direction by sands (not specifically tailings) has been carried out by
either a reinforced rubber membrane or a stack of thin, teflon Wanatowski and his co-workers (e.g. Wanatowski and Chu, 2007,
rings. The intention is to provide restraint while assembling a 2011).
specimen, but minimise confinement during application of shear. The most versatile loading apparatus is the torsional shear
These devices, and various modified versions, are commonly hollow cylinder (TSHC) (often referred to simply as a hollow
referred to as direct simple shear (DSS) devices, a term that is used cylinder device), although its use for testing tailings has been
in this paper. limited. The TSHC can reproduce a range of stress paths,
In a DSS test the known stresses are the (imposed) vertical including TC, true SS and plane strain. A drawback is the cost of
stress and the shear stress developed on a horizontal plane during the device and the complexity of operation (particularly compared
the shearing process. The shear strength along a horizontal plane with SS and triaxial devices). However, some suppliers are now
through the specimen is thus determined (i.e. the undrained shear producing compact and fully automated devices. Given the
strength) and the vertical total stress is known. An important importance of evaluating a range of possible stress paths for
assumption, which has been verified by several researchers, but
144
stability of TSF slopes, we foresee much more widespread uptake Tailings Dams failed due to the Valparaiso Earthquake which
of the TSHC in the next decade, if not sooner. occurred on March 3, 1985, of Mw = 8.0 (Castro and Troncoso,
Section 4.3 has focused on various approaches to determining 1989), the Antiguo Tailings Dam failed due to the Punitaqui
the shear strength of mine tailings under static loading conditions. Earthquake on October, 1997, Ms = 7.0, generating a flow of
More detailed discussion on the importance of various stress paths nearly 60 000 m3 of tailings (Villavicencio et al, 2014), the Las
followed during testing is provided in section 6, where we briefly Palmas Tailings Dam collapsed causing 4 casualties during the
discuss some of the results from well-documented forensic Maule Earthquake, Mw = 8.8, which occurred on February 27,
investigations. 2010. It is important to mention that all these tailings dams were
constructed using the upstream method.
4.4 Shear strength under dynamic loading The tailings dams that have experienced seismic failures have
been grouped considering the earthquake magnitude in
The experience gained in regions of high seismicity clearly shows
combination with the distance to the epicenter (Conlin 1987). For
that tailings dams are sensitive to earthquakes; in fact, several
seismic failures have been reported, as discussed below. However, earthquakes of moment magnitude greater than 8.0, the epicenter
almost all of these seismic failures occurred in tailings dams can be a large area (rupture zone), with a more representative
constructed with the upstream method (Lyu et al., 2019). In this measure being the minimum distance to the rupture surface.
regard, it is important to note that properly designed tailings dams Figure 36 summarises data from failures due to earthquakes,
constructed with the downstream or centerline methods have providing guidance on when potential seismic failure may occur
experienced stable behavior under strong seismic loading (Valenzuela et al, 2020) and, accordingly, special attention must
conditions (Valenzuela et, 2020). be considered in the design of TSFs.
Figure 35 shows the location of the world's mining areas and
of recorded earthquakes. It is evident that there are many countries
that have an important mining industry and at the same time have
active seismicity. Among these countries are: New Zealand,
Indonesia, Japan, China, Canada (west coast), United States (west
coast), Philippines, Mexico, Colombia, Peru and Chile).
Figure 36. Seismic failures of tailings sand dams (after Valenzuela et al.
2020).
On the other hand, there are a significant number of countries
with an important mining industry and with a relatively negligible
seismicity, for example Brazil, Australia, many African countries
and several European countries. In these countries, with low or
practically zero seismicity, it is common to find TSFs constructed
with tailings that are in a contractive state, and that can be
seriously affected by disturbances that mobilize the undrained
response. There are many examples, such as the catastrophic
failures of Los Frailes in Spain, Samarco and Brumadinho in
Brazil, Merriespruit in South Africa and Cadia in Australia,
among many others. The Los Frailes and Cadia failures have
strong similarities; first, the foundation soil failed and generated
significant displacement that triggered liquefaction of the stored
Figure 35. Location of a) mining areas and b) major earthquakes (where tailings, which contributed significantly to the dam failure
black dots are magnitude >9, red dots magnitude between 8 and 8.9, mechanism (Gens et al, 2006; Jefferies et al, 2019). Thus, when
orange between 7 and 7.9 and yellow between 6 and 6.9). contractive materials are present, there is an elevated risk of
failure if the residual undrained resistance is triggered, and the
In Japan, the failure of Dikes 1 and 2 of Mochikoshi TSF were stability analyses have not considered the possibility of the
associated with the 1978 Izuoshima-Kinkai Earthquake (Ishihara, mobilisation of this resistance.
1984). More recently, also in Japan, the seismic failures of the
Kayakari, Zenigame and Gengorou Tailings Dams were due to the 4.4.1 Liquefaction in terms of Cyclic Mobility.
2011 Great East Japan Earthquake of Mw = 9.0 (Ishihara et al, The term liquefaction was coined by Hazen (1920) to describe the
2014). In Chile, several tailings dams have experienced collapse failure of the hydraulic fill sand of the Calaveras Dam on 24
due to seismic disturbances: the Barahona Tailings Dam suffered March 1918. In this failure, the upstream toe of the Calaveras
a breach due to liquefaction on December 1, 1928, induced by the Dam, which was under construction near San Francisco,
Mw = 7.6 Talca Earthquake (Aguero 1929), the El Cobre Tailings California, failed suddenly, and approximately 700,000 m3 of
Dam failed catastrophically due to the La Ligua Earthquake, Mw material flowed for around 90m. Apparently no special
= 7.4, on March 28, 1965, causing more than 200 casualties disturbance was noticed at the time of the failure, indicating that
(Dobry and Alvarez, 1967), the Cerro Negro and Veta del Agua this phenomenon can occur in the absence of earthquakes. Since
145
that failure, the term liquefaction has been used in a broad sense amplitude of cyclic stress. Figures 38 and 39 present typical
to describe two different phenomena that may occur in saturated results of CRR curves obtained in tailings samples, considering a
cohesionless soils. These phenomena have in common a strain level criterion. Figure 40 presents a CRR curve considering
significant pore-pressure buildup and large ground deformations. the 100% pore water pressure build-up criterion.
Nevertheless, to understand the actual soil behavior, it is
important to distinguish between the so-called flow failure, or
static liquefaction, where a sudden loss of strength takes place,
and the term cyclic mobility, which is associated with a
progressive strain softening with cycles of loading, without any
loss of strength.
Figure 38. Cyclic stress ratio for ±2.5% axial strain vs number of cycles.
Tailings retrieved from Kayakari dam after the 2011 earthquake.
Reprinted from Ishihara et al. (2014), with permission from Elsevier.
Figure 37. Undrained cyclic response of loose (upper plots) and dense
(lower plots) sandy soils (Ishihara 1985).
Typical results of undrained cyclic torsional shear tests are
presented in terms of both stress-strain curves and effective stress-
paths in Figure 37 (after Ishihara 1985). In the case of the loose
specimen (Dr = 47%), there is a gradual migration of the effective
stress-path toward the origin, which is more pronounced in the
first cycle and after the phase transformation line is met. When the
cyclic effective stress path touches the phase transformation line,
a significant change in the cyclic response takes place. During
loading, the effective stress path turns upwards and to the right,
indicating a dilative response, while during unloading it turns left,
toward the origin, which indicates a strong contractive behavior Figure 39. Cyclic resistance ratio at = 3.75% obtained in cyclic simple
associated with a large increase in pore water pressure. Once the shear for gold tailings (James et al, 2011).
phase transformation line is crossed, this phenomenon is repeated
and at each cycle of loading and unloading the effective stress path
moves upward and downward closely along the failure line until
eventually it starts to pass though the origin, indicating a condition
of zero effective stress at the instant of zero acting shear stress.
That is, a pore water pressure build-up equal to the confinement
stress only occurs when the shear stress is zero. Accordingly, this
condition is momentary and does not induce failure of the soil.
Although this condition is associated with zero effective stress, it
should not be considered as a condition of zero shear resistance.
As soon as there is an acting shear stress, the pore water pressure
decreases and the shear resistance of the soil is again mobilised.
The same general pattern is observed for the sample with Dr
= 75%, except that the phase transformation line is crossed during
the first cycle and thereby the large change in the effective mean
stress occurs much earlier than in the specimen of loose sand.
The cyclic responses outlined above can be classified as
"cyclic mobility" according to Casagrande (1970; 1975). This is Figure 40. Cyclic resistance ratio for 100% pore water pressure build-up,
because they only involve degradation of stiffness, not strength, in simple shear for tailings samples retrieved with Shelby tube (Been
which can be significant or moderate depending upon the density 2016).
of the sandy soil and the level of the applied cyclic loading, among
other factors. 4.4.2 Effect of fines content on cyclic resistance ratio.
If the specimen is strained sufficiently, there is a tendency for Experimental evidence indicates that the mechanical properties of
the soil to dilate and regain stiffness and strength. In this context, silty sands are largely controlled by the fines content and by the
when cyclic mobility is the phenomenon under evaluation, the plasticity of these fines (Troncoso et al., 1985; Polito et al., 2001;
cyclic strength is normally defined as the Cyclic Resistance Ratio Verdugo et al., 2004; Ni et al, 2004, among others). The results of
(CRR), which is evaluated as the number of cycles required to a series of cyclic triaxial tests performed on a copper tailings sand
reach a certain amount of shear deformation, or 100% of pore using samples prepared at the same void ratio, with different low
water pressure build-up, when a sample is subjected to a certain
146
plasticity fines content, are shown in Figure 41 (after Troncoso et Plasticity Index equal or greater than 7. However, in tailings
al., 1985). materials further research is needed to confirm this criterion.
Figure 43. Strain softening in clayey material. Republished with
permission of ASCE, from Boulanger and Idriss. (2006); permission
conveyed through Copyright Clearance Center, Inc.
Figure 41. Effect of low-plasticity fines on the cyclic strength of copper For practical purposes, fine-grained soils with a sand-like
tailings (after Troncoso et al, 1985). behavior can be treated as sands to assess their liquefaction
potential, so the empirical charts based on SPT and CPT are
A more comprehensive study was carried out by Verdugo and
applicable. On the other hand, fine-grained soils with a clay-like
Viertel (2004) using a copper tailings sand. The tailings sand was behavior could undergo strain-softening during the application of
mixed with various contents of low-plasticity tailings fines: 2; 10; cyclic loads. It is recommended to evaluate the potential
18 and 28%. For each mixture, different densities were selected. occurrence of strain softening by means of cyclic tests in the
The results are presented in Figure 42. laboratory. For fine-grained tailings this is considered the most
reliable procedure.
4.4.3 Accounting for earthquake magnitude.
Since the cyclic resistance to liquefaction is a function of the
number of loading cycles, it is necessary to establish the number
of cycles that corresponds to the design earthquake. Research
suggests that the effective number of equivalent uniform cycles is
predominantly related to the earthquake magnitude, as shown in
Figure 44. This issue is discussed in section 8.3.1.
Figure 42. Cyclic strength as a function of void ratio for different fines
contents (after Verdugo and Viertel (2004).
For the range of fines content tested, the cyclic strength
decreases consistently as the fines content increases, regardless of
the void ratio of the samples.
Extremely large volumes of tailings are produced from the
extraction processes of copper, gold and nickel. These tailings
vary mainly from sandy silt to silty clay, with the fine tailings
usually being non-plastic or of low plasticity. There are other mine
tailings that include significant fine-grained particles of medium
to high plasticity, such as tailings from the extraction of alumina
from bauxite ore (red mud), many iron ore tailings, and the oil Figure 44. Number of equivalent stress cycles vs earthquake Magnitude.
sand fine tailings resulting from the extraction of crude bitumen Reprinted from Idriss et al. (2004), with permission from Elsevier.
from oil sands, among others.
5 FIELD TESTING.
It has been recognized that fine-grained soils can undergo
liquefaction or significant deformation during earthquakes (Bray Concerns about the potential instability of upstream TSFs is
et al., 2004, 2006; Boulanger and Idriss, 2006). Depending on the rightly currently focussed on existing TSFs. Sampling of tailings,
type of fines, the cyclic behavior of these soils can be separated particularly cohesionless silty or sandy tailings (which constitute
into two groups: the "sand-like" and the "clay-like" fine-grained a very significant proportion of existing TSFs) without disturbing
soils. The first exhibits liquefaction susceptibility characteristics the very characteristic that is of interest, i.e. void ratio, is
similar to those observed in sandy soils. Although clay-like fine- extremely difficult. Although interesting work is currently
grained soils, when subjected to cyclic loading, produce some underway using, for example, a modified Sherbrooke-type
increase in pore pressure and can deform significantly, they do not sampling device, such techniques are far from commonplace or
liquefy like sandy soils, rather they undergo a cyclic softening, proven effective. Furthermore, even if truly undisturbed sampling
"strain-softening" phenomenon, as observed in Figure 43 were feasible, there would still be considerable uncertainty
(Zergoun et al., 1994; Boulanger and Idriss, 2006). A simple regarding the in-situ stress levels which also affect mechanical
criterion proposed by Boulanger et al., (2006) to identify fine- response. Such uncertainty particularly affects estimates of lateral
grained soils with a clay-like behavior is that they present a stresses (see Section 5.4). Recourse is therefore made to the use
of in-situ testing.
147
Until relatively recently, the SPT (standard penetration test) of soils. Furthermore, Ottawa Sand is a clean and uniformly
was used extensively for estimating the in-situ strength and even graded sand with D50 = 0.35 mm, sub-rounded particles, and
susceptibility to liquefaction of tailings. As with many procedures composed primarily of quartz (Cunning 1994). These
used in tailings, many of the SPT correlations in use were derived characteristics are significantly different from those of many
from work on essentially clean (low fines content) sands. With tailings which tend to be siltier, more broadly graded, with angular
recent rapid and significant advances in CPT (cone penetration particles, and possess a more complex mineralogy (Torres-Cruz
test) technology, the SPT is not recommended for tailings and Santamarina 2020). Figure 45 illustrates the effect that K0 and
characterisation. The primary methods currently in use are the particle grain characteristics can have on the contractive-dilative
CPT and field vane, with lesser but increasing use of techniques boundary as inferred from Fear and Robertson (1995). To this end,
such as full-flow penetrometers, dilatometers and very Figure 45 considers K0 = 0.5 and K0 = 0.8, and two distinctly
occasionally, pressuremeters. The discussion that follows is different sands: Ottawa Sand and Alaska Sand. The latter is an
skewed toward the two former techniques, with less coverage of angular and highly compressible carbonate marine tailings sand
the other methods, given their relative sparsity of use.
(Fear and Robertson 1995). It is worth noting that in Figure 45,
Equation 16 does not match the curve of Ottawa Sand at K0 = 0.5.
5.1 Cone Penetration Testing of Tailings.
This is due to differences in converting from shear wave velocity
The cone penetration test (CPT) is the most used in-situ test to to cone tip resistance between Fear and Robertson (1995) and
characterise tailings. Early versions of the CPT were limited to Olson and Stark (2003). The variety of contractive-dilative
measurements of tip resistance (qc) and sleeve friction (fs). boundaries in Figure 45 highlights the potential shortcomings of
Modern CPTs almost invariably include pore water pressure using Equation 16 as a generally applicable contractive-dilative
measurements (known as CPTu) and routinely include seismic boundary.
wave velocity measurements (SCPTu). More recently,
qc1 (MPa)
incorporation of P-wave (i.e. compressional wave) measurements
0 2 4 6 8 10 12 14
have become more common (thankfully with no further change to 0
the acronym). The most common location for the pore water
pressure sensor is on the shoulder of the cone (u2), although
alternative locations have also been used mostly for research
100
purposes (Lunne et al. 1997).
Given that the CPT does not perform direct measurements of
the mechanical behaviour of the soil, its use relies on
'v0 (kPa)
correlations between CPT results and the soil behavioural 200
aspects of interest. In tailings engineering, particularly in
assessments of liquefaction potential, the focus is generally on
the state, the peak (or yield) undrained strength, and the residual 300
(or liquefied) undrained strength. In this context, state refers to
whether the material has a contractive or dilative tendency when
subjected to shearing. Contractive materials are prone to 400
significant strength loss under undrained conditions whereas
dilative materials are not. Different methodologies are available Figure 45. Non-uniqueness of contractive- v0-qc1
to conduct these three stages of a liquefaction potential space. Notes: Boundaries of Alaska and Ottawa Sands as per Fear and
assessment. Herein we critically review some of the most Robertson (1995). Dilative soils are expected to plot to the right of the
common approaches. boundaries and contractive soils to the left.
5.1.1 Assessment of state Robertson (2010) proposed the criterion Qtn,cs = 70 as a
Olson and Stark (2003) proposed a generally applicable conservative contractive-dilative boundary (Figure 46), where
contractive-dilative boundary defined as Qtn,cs is the normalised clean sand equivalent tip resistance defined
.
below. This criterion was considered in the post-failure
= 1.10 × 10 ( ) (16)
investigation of the Fundao Dam (Morgenstern et al. 2016) and,
although not the preferred approach, in the post-failure
where v and qc1 have units of kPa and MPa respectively and v investigation of the Brumadinho tailings dam (Robertson et al.
has an upper bound of about 350 kPa. qc1 is the overburden- 2019).
normalised tip resistance defined in Olson and Stark (2002) as
1000
1.8
Normalised cone resistance Qtn
q = (17)
0.8 +
(Qtn - 11)(1 + 0.06F)17 = 70
100
where pa
have noted the limitations of using Equation 16 as a generally
applicable contractive-dilative boundary (Olson 2009, Robertson 10
2010). Notwithstanding, Equation 16 was considered in the post-
failure investigation of the Fundão Dam (Morgenstern et al. 2016)
Qtn,cs = 70
and it is thus deemed pertinent to address this equation herein. The 1
limitations of Equation 16 stem from its origins (Torres Cruz 0.1 1 10
2016). This boundary was initially proposed, in terms of SPT Normalised friction ratio F
results, by Fear and Robertson (1995) specifically for Ottawa
Sand under K0 = 0.5 stress conditions. That is, Equation 16 was Figure 46. Comparison of the contractive-dilative boundaries proposed by
not originally intended to be generally applicable to a wide variety Robertson (2010) and Robertson (2016), shown by the solid and dashed
lines respectively.
148
Development of the Qtn,cs = 70 criterion relied on a database This implies that, at the same Qtn, sands with fines are less likely
of 36 flow liquefaction case histories that were grouped into five to classify as contractive (Qtn,cs
different classes of CPT data quality. In descending order of rationale for attributing a greater liquefaction resistance to silty
reliability, the five classes were defined as follows. Class A sands than to clean sands was "the greater compressibility and
corresponded to cases with electric CPT results that included decreased permeability of silty sands" (Robertson and Wride
sleeve friction readings. Class B included cases with mechanical 1998). However, non-plastic soils, which include many types of
or electric CPT results but no sleeve friction readings. Classes C, tailings, do not exhibit a direct correlation between fines content
D, and E included cases for which CPT results were not available and compressibility as indexed by the slope of the critical state
but rather inferred from SPT, relative density, or otherwise line (Reid 2015, Torres-Cruz and Vermeulen 2018). Furthermore,
estimated, respectively. Robertson (2010) relied mostly on the even when containing significant amounts of fines, the
Class A case histories to justify the Qtn,cs = 70 criterion. permeability of non-plastic soils may remain high enough to
enable virtually drained (Bq
These observations suggest that the use of Kc in Equation 18 to
The normalised clean sand equivalent tip resistance Qtn,cs is
compensate for the assumed higher compressibility and lower
defined as
permeability of non-plastic silty tailings can lead to potentially
unconservative assessments of state.
, = (18) Robertson (2010) indicated that the case history database
agreed with the Qtn,cs = 70 boundary because most flow
where Qtn is the stress-normalised tip resistance liquefaction case histories, including all case histories with high
quality CPT data, exhibited Qtn,cs
however, is that since flow liquefaction case histories do not
provide information on the performance of dilative soils, the data
= (19)
supports the adoption of Qtn,cs = 70 as an upper bound for
contractive soils, without necessarily implying that it is also a
lower bound for dilative soils. That is, the case history data
qt is the cone tip resistance corrected for unequal area effects and
presented by Robertson (2010) is not sufficient to establish that
and n is a stress exponent capped at 1 calculated with the
Qtn,cs = 70 is indeed a boundary. We acknowledge, however, that
following equations (Robertson 2009):
without additional information, the cautious approach is to treat
Qtn,cs = 70 as a boundary. That is, to assume that all soils with Qtn,cs
= 0.381 ( ) + 0.05 ( ) 0.15 (20) indeed contractive.
Robertson (2016) updated the contractive-dilative boundary
to Equation 9 and noted that it is mainly applicable to soils with
little or no microstructure (e.g. cementation).
= (3.47 log ) + (log + 1.22) (21)
70 = ( 11)(1 + 0.06 ) (24)
Figure 46 compares Equation 24 to the Qtn,cs = 70 criterion.
= 100% (22) The two criteria bear strong resemblance for Fr
corresponds to values often measured in tailings deposits (Reid
2015). Accordingly, the objections raised herein to the Qtn,cs
criterion, are likely to remain valid to Equation 24 when applied
Ic is a soil behaviour type (SBT) index and F is the sleeve friction
to tailings. However, we note that Equation 24 is written in terms
ratio. Given the interdependence between Qtn, n, and Ic an iteration
of Qtn, as opposed to Qtn,cs. This reflects the fact that Robertson
process is required to find their final values. While Robertson
(2016) did not adopt Equations 18 and 23. We believe this is a
(2009) explained the rationale for Equation 20, we note the
step in the right direction given the difficulties noted above to
absence of supporting experimental data points that would enable
correlate fines content to compressibility and permeability.
an assessment of its reliability.
Robertson (2016) does retain Equation 20, however, for which we
noted the lack of validation data. Despite this limitation, we
The second factor in Equation 18, Kc, is defined as
believe that its agreement with the case history database presented
1.64, = 1.0 by Robertson (2016), makes Equation 24 a useful empirical
> 1.64, = 0.403 + 5.581 21.63 + 33.75 17.88 (23) criterion.
State is also commonly characterised by the state parameter
which is defined as the void ratio difference between the current
It is worth pointing out that although Equation 23 includes a state of the soil and the critical state line (CSL) at the same mean
fourth degree polynomial whose coefficients involve 19 effective stress p' (Been and Jefferies 1985). Positive values of
significant figures, the equation is approximate, was proposed indicate states that plot above the CSL (contractive), whereas
based on qualitative arguments and lacks any supporting data negative values of indicate states that plot below the CSL
points (Robertson and Wride 1998). Similar to Equation 20, the (dilative). It has been suggested that soils with -0.05 should
uncertainty of Equation 23 is thus not quantifiable. be deemed contractive (Shuttle and Cunning 2008). While
Equation 23 defines a direct correlation between Kc and Ic. In calculating for a laboratory specimen is relatively
general, coarse-grained soils take on low values of Ic whereas fine straightforward, inferring from CPT readings is anything but.
grained soils take on high values, with Ic = 2.6 considered an Been et al. (1987) first proposed a systematic framework to
approximate boundary between sand-like and clay-like soils
(Robertson and Wride 1998). Accepting this premise, soils with may be of limited use in tailings wherein drained penetration
high fines contents are expected to yield high values of Ic, which cannot be generally guaranteed. Plewes et al. (1992) suggested a
lead to high values of Kc, and consequently high values of Qtn,cs. simplified screening procedure to assess from CPT readings
149
while accounting for excess pore water pressures. The method, based estimates of are "broadly aligned." Improved agreement
which was slightly updated by Jefferies and Been (2016), relies is observed in the data from Jefferies et al. (2019), although
on a correlation between the normalised sleeve friction ratio F and unfortunately the data clusters in a small portion of the graph,
the slope of the critical state line CSL. However, recent additions limiting the potential of this data to validate the approach.
to the -F dataset indicate that the uncertainty of the correlation is Shortcomings in the widget method to characterise loose sands (
higher than reflected by the original dataset (Reid 2015). > 0) may stem from the fact that it is anchored on calibration
Furthermore, error propagation analyses show that even when the chamber testing programmes that involved, almost exclusively,
Plewes method is implemented with values measured in the dense sands ( < 0) (Shuttle and Jefferies, 2016) and sands that
laboratory, the variability of within tailings deposits impedes were both clean (low to zero fines) and nearly all quartz sands
estimates of that are accurate enough for detailed (Been, 2016).
characterisation (Torres-Cruz 2021). These observations highlight Appraisal of Figure 47 requires acknowledgement that the
the screening nature of the procedure. More rigorous methods to measured from a sample is not a perfect validation benchmark.
estimate the state parameter from CPT results can be broadly Shuttle and Cunning (2007) mention sample disturbance as a
divided into those that assume drained penetration and those that potential explanation for the mismatch reflected in Figure 47. This
assume undrained penetration. It is the latter that is often used in would have required the disturbance to consistently induce
tailings because, as noted above, drained CPT results cannot be dilation of the samples. However, it appears unlikely that samples
guaranteed at the standard penetration rate of 20 mm/s. Within this that were deemed contractive by both the widget method and by
undrained CPT category, the method that appears to have gained water content logging would have been prone to dilation due to
the widest acceptability is the one proposed by Shuttle and disturbance. This is borne out by the results presented by Mohajeri
Cunning (2007), further developed by Shuttle and Jefferies and Ghafghazi (2012) for a natural silt from Vancouver Island,
(2016), and also described in Jefferies and Been (2015). The Canada, that was sampled and handled with extreme care. Despite
procedure relies mostly on CPT results, in-situ measurements of
shear wave velocity (Vs), knowledge (or estimates) of lateral stress which corresponded to about one third of the void ratio difference
conditions, laboratory testing of disturbed samples, and numerical between the loosest and densest possible void ratios of this soil.
simulations of triaxial tests and of spherical cavity expansion. Shuttle and Jefferies (2016) estimated the of the high-quality
Because the spherical cavity expansion simulations can be samples by simulating relevant triaxial tests and determining the
performed with a publicly accessible widget, we will refer herein best fit . They noted that uncertainty was introduced because the
to this approach as the "widget method." in-situ void ratio of the triaxial specimen was taken as equal to the
The widget method was used in the investigation of the Cadia average void ratio over the 0.8 m length of the sampler. It is
NTSF embankment failure (Jefferies et al. 2019) and was unlikely, though, that this difference between average and local
regarded as the "most reliable" approach to assess the in-situ state void ratio can account for the systematic underestimation of
of the Brumadinho Dam (Robertson et al. 2019). As such, the observed in Figure 47. Jefferies et al. (2019) computed the of
method is likely to be viewed as the best alternative by many high-quality samples by comparing the sample void ratio to that
practicing geotechnical engineers. To assess the capabilities of the of the critical state line at the same mean effective stress. Two
method, we collected data from three different sources that report CSLs were considered because layering in the samples introduced
values inferred using the widget and values estimated from uncertainty regarding which was the applicable one. The
high quality samples. The comparison is shown in Figure 47, in horizontal bars in Figure 47 reflect this. Estimates of lateral
which the vertical bars are centred at the mean estimated by the stresses (see Section 5.4) are another source of uncertainty when
widget method over the length of the sampler. The upper and inferring from high-quality samples. Despite limitations in
lower ends of the vertical bars are placed at one standard deviation 47 suggests that there may be
from the mean. Of the three references considered in Figure 47, an unconservative bias in the widget method and is a reminder that
the length of the sampler was only reported in Shuttle and Jefferies the validation of this method in real silt deposits is a pending task
(2016) as 0.8 m. The same length was assumed for the other two for the geotechnical community.
references. It is worth noting that the three references considered in
Figure 47 used different versions of the widget method to estimate
0.4 . For example, Shuttle and Jefferies (2016) updated the mapping
Shuttle and Cunning (2007) of the limiting cavity expansion pressure to CPT resistance, and
0.3
Shuttle and Jefferies (2016) Jefferies et al. (2019) reported that the pore water pressure
Jefferies et al. (2019) measurement location considered by the widget changed from the
face of the cone to the shoulder of the cone. The widget version
from widget
0.2 used by Jefferies et al. (2019) made the predictions that plot
closest to the identity line in Figure 47. However, validation data
0.1 from additional case histories is clearly required to reliably assess
the predictive capability of the widget.
0 A recent analysis of the 1974 Tar Island slump in Canada aims
at validating the computation of liquefaction using finite elements
coupled with critical state soil mechanics (Shuttle et al. 2022).
-0.1
Although this analysis uses the widget method to infer from
-0.1 0 0.1 0.2 0.3 0.4
CPT measurements, it does not constitute a validation of the
from sample
widget because no comparisons are presented between as
measured in high-quality samples and predicted by the widget.
Figure 47.
and based on high quality samples. 5.1.2 Assessment of undrained yield strength
When sheared under undrained monotonic conditions, contractive
Figure 47 shows that for two of the datasets (Shuttle and soils mobilise a yield (or peak) strength Su that, upon further
Cunning 2007, Shuttle and Jefferies 2016) the widget method shearing, drops to a residual value Sr (Verdugo and Ishihara 1996,
significantly and systematically underestimates the Olson and Stark 2003). It is common to normalise the yield
contractiveness of the samples. This is at odds with the undrained strength Su by the pre-failure vertical effective stress
observation by Shuttle and Jefferies (2016) that CPT and sample- v, thus leading to the yield strength ratio (YSR). When a
150
contractive soil reaches its YSR, liquefaction is triggered and the 5.1.3 Assessment of residual undrained strength
strength of the soil drops to its residual value Sr. Static loads, Olson and Stark (2002) presented an analysis of residual (or
deformations, or dynamic loads can all move a soil towards its liquefied) strength Sr based on the same database of 33 flow
YSR and, importantly, reaching the YSR does not necessarily failure case histories considered by Olson and Stark (2003) for
require the loading conditions to reach Su (Olson and Stark 2003). their yield strength analysis. They identified 21 case histories with
Yield strength assessments are often disregarded in liquefaction enough information to assess the mobilised residual shear strength
potential assessments because designers and owners adopt the by means of a rigorous back-analysis which on occasions included
cautious view, warranted in our opinion, that if liquefaction can kinetic considerations. Based on this data they proposed a
be triggered (soil is saturated and contractive) then it will occur liquefied strength ratio (LSR)-qc1 correlation which they
(e.g. Robertson 2010). This appears to have led to much more complemented with upper and lower limits to account for
emphasis in the literature on the liquefaction susceptibility and variability. Figure 49 presents these 21 case histories in LSR-qc1
residual strength analyses than on the yield strength analysis. One space and shows that there is only a weak (R2 = 0.21) linear
of the few methods available to assess the yield strength ratio of correlation. That is, 79% of the variation in LSR is unaccounted
contractive soils was proposed by Olson and Stark (2003). This for by qc1. Furthermore, of the 21 case histories, CPT readings
approach was considered in the post-failure investigation of the were only available for seven of them. For the remaining 14 case
Brumadinho tailings dam (Robertson et al. 2019). histories, qc1 was assessed from correlations to SPT, relative
Olson and Stark (2003) back analysed 33 flow failure case density, or otherwise estimated. As noted above, this inevitably
histories involving mostly non-plastic and low plasticity soils. introduces uncertainty. It is also apparent from Figure 49 that the
They identified eight case histories whose pre-failure conditions LSR-qc1 correlations proposed by Olson and Stark (2002) are
corresponded to yield strength conditions and on this basis steeper than the linear regression computed herein. That is, the
proposed a YSR-qc1 correlation which they complemented with regressed line suggests a weaker dependence of LSR on qc1 than
upper and lower limits to account for variability. Figure 48 indicated by the correlation proposed by Olson and Stark (2002).
summarises these eight case histories in YSR-qc1 space and shows Again, the implication for LSR assessments is that qc1 can be
that there is only a weak (R2 = 0.15) linear correlation. That is, bypassed as an explanatory variable and should instead take note
85% of the variation in YSR is unaccounted for by qc1. that Olson and Stark (2002) reported LSR values that ranged from
Furthermore, of the eight case histories shown in Figure 48, CPT 0.027 to 0.12 with a mean value of 0.083. Wider upper and lower
readings were only available for two of them. For the remaining bounds result from considering the uncertainties of each case
six case histories, qc1 was assessed from correlations to SPT, history as reported by Olson and Stark (2002). The applicable
relative density, or otherwise estimated. Such correlations and LSR value must then be chosen within these bounds, or beyond if
estimations further reduce the reliability of the correlation. It is warranted, while keeping in mind that qc1 does not provide robust
also apparent from Figure 48 that the YSR-qc1 correlations guidance. As for the YSR, we note that while our comments do
proposed by Olson and Stark (2003) are steeper than the linear not provide additional guidance in the selection of LSR, they will
regression computed herein. That is, the regressed line suggests a hopefully prevent undue overreliance on the LSR-qc1 correlation
weaker dependence of YSR on qc1 than indicated by the proposed by Olson and Stark (2002).
correlation proposed by Olson and Stark (2003).
The implication for YSR assessments is that instead of 0.25
attempting to correlate YSR to qc1, it should be recognised that the Data point labels indicate case history ID as per Olson & Stark (2002).
flow failure case history database analysed by Olson and Stark
(2003) yielded YSR values that ranged from 0.24 to 0.28 with a 0.2 Average and limit trends (Olson & Stark 2002):
Liquefied strength ratio LSR
YSR = 0.0143qc1 + 0.03 ± 0.03
mean value of 0.26. Wider upper and lower bounds result from
considering the uncertainties of each case history as reported by
Olson and Stark (2003). The applicable YSR value must then be 0.15 Best fit line: YSR = 0.0093qc1 + 0.053
(R2 = 0.21)
chosen within these bounds, or beyond if warranted, while
3
keeping in mind that qc1 does not provide robust guidance. 26 8 2 15
Doubtless a significant challenge. While these comments do not 0.1 31 33 25 30
29 32 19
provide additional guidance to estimate the actual YSR, we hope 28 6 20
that they will prevent undue overreliance on the YSR-qc1 12 22
9
correlation proposed by Olson and Stark (2003). 0.05 24
10 21 Lacks CPT data
Has CPT data
0
0 1 2 3 4 5 6 7
Corrected tip resistance qc1 (MPa)
Figure 49. Flow failure case histories analysed by Olson and Stark (2002)
with adequate to back-calculate the liquefied strength ratio.
Based on 13 flow failure case histories with CPT records (data
classes A and B), Robertson (2010) proposed the following
correlation for the lower bound of liquefied strength ratio LSR
which is valid for Qtn,cs :
0.0003124 , + 0.02199
= = (25)
0.0001783 , 0.02676 , + 1
Figure 48. Flow failure case histories analysed by Olson and Stark (2003)
Equation 25 has a lower bound of 0.03 and an upper bound of
and deemed to be close to their yield strength envelope.
tan . Given the uncertainties involved in back-analysing case
151
histories, it is a positive feature of Equation 25 that it aims at correlations that fit soils with different compressibilities: stiff,
providing a lower bound of Sr, rather than a best estimate. intermediate, and compressible (Figure 51a).
However, Equation 25 relies on Qtn,cs and thus inherits the There are at least three points of concern in the approach
objections that were previously made regarding this parameter. It suggested by Jefferies and Been (2016). Firstly, like Olson and
is also worth noting that Equation 25 was considered during the Stark (2002, 2003), most of the case histories (8 out of 12) used
post-failure investigation of the Brumadinho TSF but ultimately to develop the correlations do not have electric CPT readings. This
not used due to a lack of match with laboratory results. The leads to reliance on results from mechanical CPT, SPT, and, in
discrepancy was attributed to the mineralogy and bonding of the one case, assessment of tip resistance from relative density. That
Brumadinho tailings (Appendix E of Robertson et al. 2019) and is, there is a strong reliance on the data classes B, C, and D as
may point to limitations in the applicability of Equation 25. defined by Robertson (2010) (Figure 51b). Secondly, the
Robertson (2010) noted that Equation 25 can be overly approach relies on in-situ determinations of , the feasibility of
conservative in sensitive clays and for such soils proposed: which remains unproven even when high quality CPT data are
available (Figure 47) and seems even less likely when back
= = (26) analysing case histories that fall into data classes B, C, and D.
Lastly, because the CSL of 10 of the 12 case histories is unknown,
Jefferies and Been (2016) relied on the assumption that finer
10. However, this
Robertson (2010) justified Equation 26 with the observation assumption is at odds with experimental data (Papageorgiou and
that for most clay soils fs Sr. The supporting reference cited in
Fourie 2001, Reid 2015, Torres-Cruz and Vermeulen 2018,
Robertson (2010) leads to two datasets: one from the Santa
Torres-Cruz 2019).
Barbara Channel site originally reported by Quiros and Young
(1988) and later reproduced by Lunne et al. (1997), and another
a) 0.3
from the Scoggins Dam in Oregon, USA originally reported by
Compressible ( = 0.17) Compressible
Farrar et al. (2008) and later reproduced by Robertson (2009). At 10
Intermediate
the Santa Barbara Channel site, the undrained residual strength
v
Stiff
was evaluated with triaxial tests, whereas at Scoggins Dam it was
Liquefied strength ratio Sr/ Intermediate
evaluated using field vane testing. Figure 50 summarises these 0.2 ( 10 = 0.12)
datasets and provides an indication of the degree of scatter around
the fs = Sr line. We note the absence of supporting data from clayey
tailings deposits and suggest that such datasets would be a useful
addition to the literature, especially considering that the fs Sr 0.1
assumption has been challenged on the grounds that there are
factors independent from Sr that significantly influence fs. Such
Stiff ( = 0.05)
factors include the pore water pressure on the end areas of the 10
sleeve, the roughness of the sleeve, and the extent of soil 0
remoulding (Lunne, 2007). -0.1 0 0.1 0.2 0.3
Characteristic state parameter k
160
Residual undrained strength Sr (kPa)
Santa Barbara Channel
Scoggins Dam b) 0.3
120 Compressible ( 10 = 0.17) Class A
Class B
v
Class C
Liquefied strength ratio Sr/
Intermediate Class D
80 0.2 ( 10 = 0.12)
40
0.1
0
0 40 80 120 160 Stiff ( 10 = 0.05)
Sleeve friction fs (kPa)
0
-0.1 0 0.1 0.2 0.3
Figure 50. Comparison between the sleeve friction fs and the residual
shear strength Sr from two non-tailings sites. Sources: Lunne et al. Characteristic state parameter k
(1997) and Robertson (2009).
Figure 51. Best practice LSR- k correlations proposed by Jefferies and
Acknowledging that perfectly undrained conditions cannot be
Been (2016) with data discriminated by a) assumed soil compressibility,
guaranteed in the field, Jefferies and Been (2016) resorted to a and by b) data class.
database of 12 flow failure case histories to develop correlations
between the characteristic state parameter k and the liquefied 5.1.4 Partial drainage during CPTu
strength ratio LSR. In their framework, k represents the value of Analyses of CPTu test data are usually carried out assuming either
that effectively controls the mechanical behaviour of a soil fully drained penetration (no generation of excess pore water
deposit. Jefferies and Been (2016) suggest a value between the pressure) or fully undrained penetration (zero volume change
80th and 90th percentile of for liquefaction analyses. That is, they during penetration). Perhaps this is a hangover from the desire to
suggest that the weakest 20% or 10% of a soil deposit determines think of soils as either sands (drained) or clays (undrained).
its stability. Their approach accounts for the effect of soil However, as emphasised many times in this paper, many mine
properties, especially compressibility, on the LSR- k correlation. tailings are either silts or silty sands, with permeability often in
Accordingly, they use the case history database to develop three the range from 10-5 to 10-8 m/s (Schnaid, 2021). The traditional
152
cone penetration rate of 20mm/s may not be appropriate for such Distinction between fully drained, partially drained, and fully
tailings, potentially leading to partial drainage during the undrained conditions can also be aided by the CPTu-based soil
penetration process. This is especially important when testing classification chart proposed by Schneider et al. (2008, 2012). The
loose, contractive materials. When compared to undrained chart is defined in Q-U2 space and its regions are given by the
conditions, partial drainage increases penetration resistance and dotted lines in Figure 53. Schneider et al. (2008) explain that the
affects the excess pore water pressure generated during Bq = 1 contour line is not a boundary, but rather is intended as a
penetration and the form of any subsequent dissipation test carried link to the Q-Bq chart proposed by Robertson (1991). Q, U2, and
out. A key implication is that undrained strengths may be Bq are normalised measures of CPTu results and are defined as:
overestimated.
The issue of partially drained CPTu penetration has been = (28)
investigated by several researchers, including Finnie and
Randolph (1994), who suggested a normalised velocity, V, to
account for variables that affect penetration bearing resistances. 2= (29)
Their equation was
= (30)
= . (27)
where v is the penetration velocity (usually 20mm/s for field
CPTu tests), d is the diameter of the cone and cv is the vertical
coefficient of consolidation measured in a 1-D consolidation test.
Results are often presented schematically in a form such as that
shown in Figure 52, where a logarithmic scale is usually used for
the horizontal axis. The tip resistance is usually normalised by
either a reference tip resistance value (in Figure 52 the drained tip
resistance is used) or by vertical effective stress. As indicated,
below a certain threshold normalised velocity the CPTu test is
fully drained; above another threshold velocity, the test is fully
undrained. Between these thresholds, penetration generates partial
drainage. From our experience, insufficient acknowledgement of
the potential for a CPTu test in tailings (particularly silts or silty
sand tailings) to be in the region between drained and undrained,
i.e. partially drained, has been given in many testing programmes.
Figure 53. Q-U2 soil classification chart. Note: Dotted lines correspond to
original boundaries (Schneider et al. 2008), and solid lines correspond to
Bq contours that approximate or replace the original boundaries. Plotted
data correspond to sounding CPTU-05-05N_Updated reported in
Robertson et al. (2019).
Figure 52. Illustration of effect of normalised penetration velocity on The Q-U2 chart has five regions: one drained region, two
normalised cone tip resistance. undrained regions (clays and sensitive clays), one region for
transitional soils, and one region shared by low rigidity clays and
Finnie and Randolph (1994) suggested the transition from transitional soils. The term "transitional soils" refers to the wide
drained to partially drained penetration occurred at V 0.01 and variety of soil types in which the CPTu is partially drained and
from partially drained to undrained at V 30. includes clayey sands and silts, silts, and residual soils (Schneider
During a CPTu it is common to occasionally stop the test and et al. 2008).
allow excess pore water pressures to equalise, referred to as a Contours of constant Bq provide reasonable approximations to
dissipation test. From the rate of dissipation, the coefficient of the curved boundaries of the Q-U2 chart (Robertson 2016).
consolidation due to horizontal drainage, ch, may be obtained, Furthermore, results from non-plastic tailings and coal ash
using e.g. the procedure suggested by Teh and Houlsby (1991). indicate that the boundaries of the drained region can be replaced
The parameters cv and ch are usually not equivalent, due to by Bq contours (Torres-Cruz and Vermeulen, 2018). Combined,
anisotropy, although this effect is often considered insubstantial these two observations suggest that the Q-U2 chart as proposed
enough to ignore. An alternative approach to determine whether by Schneider et al. (2008, 2012) can be simplified if all the
cone penetration is undrained is to use the time to 50% dissipation boundaries are replaced by Bq contours as shown in Figure 53. It
of excess pore pressure, i.e. t50, as an indicator. DeJong and follows that Bq can be used to readily verify whether, according to
Randolph (2012) and Robertson (2012) have suggested that when the Q-U2 chart, CPTu soundings are being conducted under
t50 is greater than 50 seconds the penetration is essentially drained, undrained or partially drained conditions. Table 2
undrained.
153
summarises the expected drainage conditions for different values 5.2 Field vane testing
of Bq.
As mentioned earlier, field vane testing is used extensively for
fine-grained soils and is appropriate for clayey tailings. However,
Table 2. Approximate correspondence between Bq and drainage conditions use of the vane for silty and sandy tailings is problematic if the
based on the Q-U2 chart proposed by Schneider et al. (2008, 2012). standard testing rate of typically 0.1° per second is used. Blight
(1962) realised this problem when working with silty gold tailings
Bq range Drainage conditions and derived a relationship between degree of consolidation U
<-0.01 Partial drainage
during shearing and a dimensionless time factor T, given by:
-0.01 to 0.01 Fully drained
0.01 to 0.05 Partial drainage = (31)
0.05 to 0.2 Partial drainage or undrained
>0.2 Undrained where tf is the time to failure, i.e. time to reach peak strength and
D is the vane diameter. Blight (1962) derived this relationship
both empirically and theoretically and the validity of Blight’s
Figure 53 includes data from a CPTu sounding performed at approach has been confirmed by Chandler (1988). Using his
Dam I at the Córrego do Feijão iron mine near Brumadinho, Brazil approach, Blight (1962) suggests a T value smaller than about
(Robertson et al. 2019). Regardless of whether one considers the 0.05 is required to ensure undrained shearing occurs. This is easily
original Q-U2 boundaries or the Bq contours, the chart indicates achieved for clayey tailings, but for a typical silty tailings, which
that partial drainage conditions prevailed during some portions of may have a cv value of about 100 m2/year, the required time to
the sounding. Naturally, plots of Bq versus depth can aid in failure to ensure undrained conditions using a 5 cm diameter vane
locating the layers in which partial drainage is occurring. is about 40 seconds, i.e. much faster than achieved using the
Based on the discussion in this section, there appears to be a standard rate of 0.1° per second. Using a larger diameter vane
need for greater awareness amongst tailings engineers that it is increases the time to failure, but for the silty tailings used in the
often warranted and necessary to modify the standard cone example, even this would not ensure undrained shearing for any
penetration rate of 20 mm/s to produce results that are compatible reasonably sized vane. Clearly, increased rates of shearing are
with interpretation procedures that assume either fully drained or required to ensure truly undrained shear strengths are obtained. If
fully undrained conditions. CPTu data are available, it is possible to substitute ch for cv in the
above equation, with little loss of relevance.
5.1.5 CPT testing in partially saturated tailings Given the enormous number of cone penetration tests and
It is crucially important to remember that CPTu testing, and field vane tests that have been carried out on silty and silty sand
indeed other field tests such as vane or full-flow penetrometer tailings worldwide, the majority of which were likely carried out
tests, have all been developed for application to saturated at inappropriate testing rates, one can only wonder about the
conditions, which for all practical purposes can be assumed to validity of some of the stability analyses carried out using such
begin below the phreatic surface. Above the phreatic surface, data and the millions of dollars that have perhaps been wasted as
where negative pore water pressures relative to atmospheric exist, a result.
there can be a substantial contribution to strength provided by A related issue that the authors have often encountered in
suction effects. Simplistically, suction increases the effective practice is the use of field vane data to ‘calibrate’ cone
stress (being a negative pore water pressure) and thus shear penetrometer data. If a CPTu test is carried out at a rate such that
strength. Interpretations of material parameters from tests such as the test is truly undrained, the peak undrained shear strength, Cu,
CPTu tests done above the phreatic surface can thus be misleading. is often estimated from the equation
Unfortunately, most currently available interpretation software
does not adequately account for this issue. = (32)
The suggestion by Robertson (2010) that ‘soils with a constant
value of Qtn,cs have essentially a similar state parameter and hence where qc is the cone tip resistance, v0 is the total vertical stress
a similar response to loading’, should not be extended to and Nkt is a bearing capacity factor, usually taken to be between
conditions above the phreatic surface. Doubtless Robertson 11 and 20.
(2010) never intended for such extrapolations to be made, but However, it is not uncommon to see the undrained shear
unfortunately they sometimes are. Aside from work such as that strength obtained from field vane tests, which are implicitly
by Pournaghiazar et al (2011) who describe a calibration chamber assumed to be truly undrained, to back-calculate Nkt values using
suitable for testing tailings, there has, until recently, been limited the known v0 and measured qc values. Values of Nkt higher than
research on cone penetration testing of partially saturated silts or 40 are not uncommon when this approach is adopted and the value
silty sand tailings. Miller et al (2018) provide a useful summary derived is nothing more than an artefact of either the penetrometer
of tests done on natural soils, varying from sands to clays. They or the vane, or both, not being truly undrained tests. We strongly
conclude that cone tip resistance generally increases with discourage this practice unless it can be confirmed that both the
increasing matric suction (as would be expected), so ignoring this CPTu and vane tests were truly undrained.
effect would result in an overestimate of soil strength and Field vane tests provide another option for determining
underestimate the state parameter. They further conclude that the residual strengths. Although traditionally used primarily for
tip resistance dependency on suction is greater for soils with measuring peak shear strength, increasing interest in residual
significant plasticity compared to non-plastic soils. This is strength determination has meant the vane is now widely used in
consistent with more plastic soils being capable of sustaining this regard as well. Improvements such as high precision control
higher suctions than sandy and silty soils. This observation is of rate of rotation and down-hole torque measurement (thus
perhaps something of a saving grace when dealing with common eliminating rod friction effects) have improved confidence in vane
silty and sandy tailings – the suction hardening effect may be test results. However, given the above discussion about the rate of
relatively muted and the potential for incorrectly evaluating shearing, particularly when dealing with silty materials, it is
strength parameters not as great as with clayey tailings. conceivable that even if the vane were rotated sufficiently fast to
ensure measured peak strength values were truly undrained
154
values, the same may not automatically be true for residual As noted above, measurement of lateral stresses is one of the
strength values. The much greater time (related to much greater applications of the PMT. Such measurements are necessary for a
degree of rotation) required to mobilise the large strain strength thorough interpretation of the CPT (Jefferies and Been, 2015).
may also be enough time to enable some degree of consolidation Furthermore, the relative magnitudes of the vertical and horizontal
to occur, thus providing an overestimate of strength. effective stresses determine the levels of deviatoric stress which,
in turn, affects stability. Despite the importance of measuring
5.3 Full-flow penetrometers lateral stresses, a review of the literature and consultations with
geotechnical practitioners suggest that such measurements are
As discussed by Schnaid (2021), an alternative for estimating
seldom made. Indeed, Graham (1992) loosely cited Jamiolkowski
residual shear strength is full-flow penetrometers, the two most
as saying that horizontal stress measurements are the 'black hole'
common being ball penetrometers and T-bar penetrometers (see
of geotechnical engineering. Fast forward to 2022 and horizontal
Stewart and Randolph, 1994). These were initially developed
stresses remain as elusive as ever.
primarily for the offshore oil industry, to measure strengths of soft
Figure 54 presents plots of v vs h corresponding to an oil
seabed soils. Recommended procedures and interpretation
sand tailings deposit (Shuttle et al. 2022). Given their similarities
methods for determining peak and remoulded strengths (and thus
with tailings deposits, measurements in hydraulic sand fills are
sensitivity, alternatively expressed as Bishop’s Brittleness Index
also included in Figure 54 (Graham and Jefferies 1986). All h
IB, discussed in Section 8.1.1) are discussed by Randolph and
measurements in the figure were performed with a SBPMT. The
Andersen (2006) and Zhou and Randolph (2009) for tests in clays.
figure shows that although it is standard practice to assume a
With suitable adjustments, such as rate of cycling, these full-flow
constant value of the h/ v ratio throughout an entire tailings
penetrometers are likely to see increased use for testing of some
deposit (e.g. Jamiolkowski and Masella 2015, Lipinski and
tailings. However, the authors’ experience with a ball
Wdowska 2018, Robertson et al. 2019), significant deviations can
penetrometer has been that in slightly stiff tailings, the borehole
occur from the mean value.
behind the penetrometer does not always close (violating the
concept of ‘full-flow’) and upon cycling of the device, very low Horizontal effective stress 'h [kPa]
or even zero readings are detected (because there is no resistance 0 200 400 600 800
to extraction). Although potentially very useful, indiscriminate 0
use of full-flow penetrometers is inadvisable. Vertical effective stress 'v [kPa]
Hydraulic fill (Alerk)
Mahmoodzadeh and Randolph (2014) suggest a similar 50 Hydraulic fill (Amauligak)
relationship between normalised resistance and normalised
Oil sand tailings (dyke crest)
penetration for the ball penetrometer as for a cone penetrometer. 100 Oil sand tailings (beach)
Relevant values of ch may be determined directly from dissipation
tests carried out using the ball penetrometer as most ball 150
penetrometers include a pore pressure sensor. However, the
location of this sensor is key, as sometimes it is located at mid- 200
height (referred to as the equator location), sometimes on the face
of the ball and sometimes on the shaft, the latter being most 250
similar to the location on a typical cone penetrometer. Using data
from Mahmoodzadeh and Randolph (2014), a suitable penetration
300
rate can be determined to ensure fully undrained penetration
occurs with the ball penetrometer.
Figure 54. h v in hydraulically placed oilsand tailings and
5.4 Pressuremeter testing in tailings. other hydraulic fills. Data sources: Shuttle et al. (2022) and Graham and
Jefferies (1986).
Pressuremeter tests (PMT) are conducted by expanding a
membrane against the walls of a borehole. There are three types Further to the scarcity of lateral stress measurements in
of PMTs: 1) prebored (PBPMT), in which the borehole is first tailings reported in the literature, part of the problem appears to
made and the PMT device subsequently introduced; 2) self-boring lie with the inherent difficulty of measuring lateral stress in soils
(SBPMT), in which boring and device insertion occur in general. There are significant uncertainties in assessing the in-
simultaneously; and 3) full-displacement (FDPMT), in which the situ value of h from PMT results as well as from other in-situ
PMT device is fitted with a cone that allows it to be pushed into tests such as the Iowa stepped blade, and the flat dilatometer
the ground (Mair and Wood, 1987). (Marchetti 2015, Mair and Wood 1987). The uncertainties stem
PMTs have been used in a variety of ways to characterise both from the fact that lateral stresses are very sensitive to the
deposited tailings and the underlying foundation soils. Its disturbance that necessarily results from the insertion of any kind
applications include assessment of the friction angle (Gomes et al. of measurement probe into a formed soil. That is, the act of
2003, Chakraborty et al. 2020), flow slide potential (Debasis et al. measuring alters the parameter being measured. Engineered fills
2002), state parameter (Wride et al. 2000, Ghafghazi and Shuttle such as tailings deposits allow the positioning of a measurement
2008), elastic moduli (Morgenstern et al. 2015, Chakraborty et al. probe, such as a total stress cell, before depositing the tailings,
2020), and lateral stress (Graham and Jefferies 1986, Wride et al. thus eliminating the need for disturbance to make the
2000, Chakraborty et al. 2020, Shuttle et al. 2022). Clearly some measurement (e.g. Chen et al. 2021). However, even in this case,
advantage has been taken of the potential of pressuremeter tests to the effect of the relative stiffness of the stress cell and the soil
better characterise TSFs. However, our consultations with results in the measurement of lateral stresses not being
industry practitioners show that, overall, the use of PMTs in TSFs straightforward (Clayton and Bica, 1993).
remains limited. This is also reflected by the fact that the post-
failure geotechnical investigations of the embankment at CADIA
Valley Operations (Jefferies et al. 2019) and of the Brumadinho 5.5 Relative density as a screening tool.
dam (Robertson et al. 2019) did not consider PMT results. The Relative density, RD, is defined as
question as to whether PMTs can contribute more to the
geotechnical characterisation of TSFs thus remains open.
= (33)
155
where emax and emin are the maximum and minimum void ratios caution against unquestioning adherence to the use of empirical
determined using standard laboratory testing techniques and e is correlations when determining tailings design parameters, such as
the void ratio of interest (e.g. the in-situ void ratio Therefore, a state and strength. Empirical correlations remain useful as
min screening tools and as a reality check, but our strong preference is
emax). Relative density is used universally as a state for a mechanics-based approach rather than slavish use of
index parameter for non-plastic soils, despite its inability to empirical correlations. The techniques for deriving correlations
normalise mechanical response. However, it is not a good between CPTu results and soil state, as advocated by Shuttle and
indicator of behaviour for silts. As discussed by Shuttle and Cunning (2007) and Shuttle and Jefferies (2016) amongst others,
Jefferies (1998), one sand (Hilton Mines sand) produced the same implicitly account for influences of aspects such as particle shape.
CPT resistance at 60% relative density as another sand (Monterey For example, the Norsand parameters that are used to calculate
sand) at 40% relative density, both sands having a few percent limiting pressures from spherical cavity expansion solutions
fines. (which are subsequently used to estimate the relationship between
The limitations of relative density as an indicator of state are normalised tip resistance and the state parameter) will be
further illustrated by the results presented by Liu and Lehane influenced by particle shape. Predictions of limiting cavity
(2012). They report results from centrifuge tests on four uniformly expansion pressure will therefore also change with changes in
graded silica materials, each with its own characteristic particle constitutive model input data.
shape (one being spherical glass beads, SGB). A result is shown
in Figure 55 for three of the four materials tested, which is for a 6 MODES OF FAILURE.
centrifuge test carried out at 50g with samples prepared at a
relative density of 20% (i.e. expected to be contractive). As can There are a variety of potential failure modes to consider when
be seen, there are significant differences, with the cone resistance evaluating the stability of both proposed and existing TSFs. To
increasing with the degree of angularity of the particles. Using the systemize risk analysis and operation of TSFs, failure modes are
terminology discussed in Section 3.2, the roundness, R, and generally divided into categories. These categories can be
sphericity, S, are summarised in Table 3, together with index different from site to site. However, the following generalized
properties that illustrate the otherwise similar nature of the three eight failure modes are commonly used:
materials tested.
Overtopping
Slope instability
Internal erosion
External erosion
Foundation
Environmental
Human
Other
A risk analysis considering relevant failure modes can, and
should, be used as a basis for the design of surveillance and
monitoring programs. The aim is to be able to detect a potential
failure scenario as well as initiate and execute measures to prevent
failure. Below follows a description of what each failure mode
suggested here may comprise.
6.1 Overtopping
Figure 55. CPT end resistances (qc) observed in centrifuge tests run at 50g
on samples prepared at 20% relative density. Republished with permission Water management, e.g. water balance and management of
of ICE Publishing, from Liu and Lehane (2012); permission conveyed extreme flows, is crucial for all dam structures. Tailings dams
through Copyright Clearance Center, Inc. Table 3. Properties of three typically have a smaller catchment area compared to water dams,
materials used in centrifuge tests (after Liu and Lehane, 2012). but water discharge may be restricted due to water quality, for
example, leading to positive water balances. There are, however,
Table 3. Properties of three materials used in centrifuge tests (after Liu instances, where a TSF has a significant upstream catchment, such
and Lehane, 2012). as valley TSFs, or facilities storing high or extreme rainfall events
Material d50 (mm) Cu R S
(e.g. facilities without spillway) and these facilities may be more
susceptible to overtopping. Freeboard needs to be designed to
SGB 0.12 1.05 1 1 retain extreme design floods and water management under
UWA sand 0.15 1.4 0.7 0.62 extreme conditions should preferably be automatic, i.e. not require
GB1 0.45 1.13 0.62 0.54 human intervention.
Examples of overtopping failures include Baia Mare,
Rumania in 2000 and Merriespruit, South Africa in 1994.
Considering Figure 55, below about 50mm (equivalent to
2.5m depth at full scale), there is more than 100% difference 6.2 Slope instability
between the SGB and GB1 materials, with the natural UWA sand Material properties in the structural zone of the dam, pore
material also being substantially lower than the GB1 results. pressures, deformations and static and seismic loads affect TSF
Similar differences were seen in centrifuge tests run at 100g and slope stability. Deposition practices, consolidation, compaction
200g. and drainage affect the strength of the tailings and must be
Cone tip resistance can therefore be a poor indicator of in-situ understood to correctly estimate material properties. The
state if a crude framework such as relative density is used. It is elevation of the phreatic surface is equally important, and a
too dependent on unquantified influences, such as particle shape realistic assessment is crucial. As many tailings dams (typically
and compressibility. As argued in several places in this paper, we upstream dams) are drained structures (or at least designed to be
156
so) the pore pressure is not necessarily hydrostatic. In such cases, Examples of foundation failures are Aznalcollar, Spain in
the use of piezometers to adequately assess the porewater pressure 1997, Mount Polley, Canada in 2016 and Cadia, Australia in 2018.
regime is an absolute necessity. Seismic hazard/response analyses
also need to be done correctly, as dealt with in detail in section 6.6 Environmental
8.2.
The thinner the structural zone (i.e. the narrower the beach of Environmental failure modes include situations where the facility
unsaturated tailings) the more sensitive the dam becomes to exceeds the expected impact on the environment. For example,
potential instability. This is particularly relevant in upstream dust generation, uncontrolled seepage, emergency release of water
dams, where the operation of the dam has a significant impact on during extreme events, limited settling time before discharge
the width of the structural zone. causing increased levels of contaminants in the discharge water
Examples of slope instability failures include Brumadinho, Brazil etc. This failure mode, typically, does not lead to catastrophic
in 2019 and Fundao, also Brazil, in 2015. failures such as a dam breach, but is nevertheless important to
consider.
6.3 Internal erosion (piping)
6.7 Human
Seepage occurs through virtually all TSFs and if the hydraulic
gradient is high enough, the water may erode fine particles if Intervention, or lack of intervention, by mine personnel may cause
adequate filters are not in place. Ongoing erosion can lead to a critical situation, worsen the situation or in worst case a dam
piping, and an open channel may develop from the water pond to failure. Education, training and emergency preparedness planning
the downstream side through the dam and/or the foundation and is therefore crucial.
ultimately cause failure.
An example of internal erosion failure is Omai, Guyana in 6.8 Other
1992. This includes failure modes not categorized above. For example,
natural hazards including rock, soil, or snow avalanches into the
6.4 External erosion
TSF impoundment leading to overtopping, or instability of the
Extreme floods in adjacent streams may erode the toe of the dam. dam.
High intensity rainfall may also cause erosion on the downstream Actual failures show that a combination of factors is normally
slope. Both external erosion events can potentially lead to dam the reason for failure, and in hindsight the cause of the failure
instability. If a large free water pond exists, or excessive wave could, in many cases, have been predicted. (ICOLD, 2001).
action occurs, it could possibly cause erosion and instability of the
upstream slope leading to overtopping. 6.9 Trigger mechanisms.
6.5 Foundation In order to initiate a failure mode a trigger is required. Figure 56
shows a number of different triggers, such as: 1) high pond water
Instability of the foundation can be due to undetected weak layers, levels, 2) inhomogeneous tailings mass, 3) low density in the
incorrect strength assumptions, pore pressure generation, artesian tailings mass, 4) heavy rain, 5) low permeability layers in the
water pressures and/or incorrect seismic hazard/response analysis. structural zone, 6) weak zones in the foundation and 7) material
Complex geologic formations (including clay layers, weak transport in the dam.
bedding planes etc.) increase the risk as well as soils that may be Figure 56. Mechanisms and processes that contribute to the failure of
sensitive to static and/or cyclic liquefaction, such as loose tailings dams. From Santamarina et al. (2019). Reprinted with permission
granular soils. from AAAS.
157
6.10 Lessons from forensic investigations of TSF failures downstream of the TSF received inadequate warning of the
developing failure.
As discussed in section 9 (monitoring), various combinations of Morgenstern et al. (2016) present an in-depth forensic
instrumentation and visual observation may be utilised to gain evaluation of the Fundão failure, working through several
advance warning of one or more of the above generic failure hypothetical failure modes and steadily eliminating all but one
modes developing. However, failures of actual TSFs provide based on the weight of evidence gathered. The process followed
additional lessons on failure modes that are sometimes overlooked is a lesson in itself; some forensic investigations appear to have
or not anticipated. It is therefore useful to evaluate the likelihood rushed into deciding on a particular failure mode explanation
of a particular failure mode occurring, but it is not a substitute for without working through a process of elimination as Morgenstern
a good understanding of geotechnical engineering principles. A and his co-investigators did. The failure was attributed to lateral
few of the recent TSF failures and associated forensic extrusion of slimes at depth causing horizontal extension (and thus
investigations illustrate this point. lateral unloading) of overlying, loosely placed sand tailings.
Given these sandy tailings were extremely brittle, the amount of
6.10.1 The Fundão TSF failure in Brazil in 2015 strain required to initiate collapse due to static liquefaction was
On the afternoon of 5th November 2015, the Fundão TSF in small. It appears these movements certainly did not manifest as
Brazil collapsed. The resulting flowslide of 32 million m3 of cracks on the surface of the TSF or noticeable bulging at the toe
tailings flowed for some 660 km downstream (after entering a of the TSF.
river), engulfing the town of Bento Rodrigues and resulting in 19
known fatalities. It was probably the worst industrial accident in 6.10.1.2 Lateral extrusion as a failure mechanism.
Brazil until that time. Figure 57 shows a lateral extrusion type stress path, such as may
The Fundão TSF was a cross-valley embankment that abutted have been experienced in the brittle sandy tailings overlying the
natural slopes on either end. The original plan was to deposit ductile slimes. Strain compatibility between the slimes and the
sandy tailings behind a compacted earthfill starter dam and then sandy tailings as the slimes deforms causes a reduction in
raise the TSF height through a sequence of upstream raises. horizontal stress but the vertical total stress remains unaltered; the
Deposition of sandy tailings from the embankment crest was to resulting stress changes produce a decrease in mean effective
continue while deposition of a separate tailings stream, consisting stress and an increase in deviator stress. Failure is triggered at
of finer, clay-like tailings, often referred to as ‘slimes’, took place. point A, well below the effective stress failure envelope.
The intention was to use the deposited sand tailings to retain the
slimes, maintaining a wide zone of free-draining sands adjacent to
the embankment. This goal is consistent with good practice when
operating upstream TSFs, i.e. maintaining a large, well-drained
and unsaturated tailings beach adjacent to the perimeter
embankments. Unfortunately, this design intent was not adhered
to.
6.10.1.1 Findings from the forensic investigation
At the time of the failure there were several workers either on the
slope of the embankment or on the embankment abutments. From
the various eyewitness accounts, Morgenstern et al. (2016)
reported the following sequence of events:
Figure 57. Illustration of lateral extrusion stress path showing triggering
of failure when instability line is reached.
On the afternoon of the failure, November 5th,
2015,
several workers were on the facility and in a position to see
Prior to the failure developing, pore pressure measurements
along the length of the embankment crest. The first thing
were stable; loading appeared to be proceeding in a drained
noticed by many workers was a cloud of dust rising from
fashion, i.e. no generation of excess pore water pressure.
the left (looking downstream) abutment.
A worker reported seeing waves developing in the central However, when failure was triggered, the failure was undrained.
portion of the reservoir, accompanied by cracks forming Shear-induced excess pore pressures in the contractive sandy
on the left side and blocks of sand moving up and down on tailings could not be sufficiently quickly dissipated, despite the
the left abutment setback. tailings having a relatively high permeability. It illustrates the
Another worker saw a crack open along the crest of a step- dangers of adopting an attitude that was still widespread in the
in embankment that had been constructed adjacent to the tailings community at the time, that ‘sands cannot fail undrained’.
left abutment, propagating in both directions. With the enormous increase in research activity over the five years
At another location, workers experienced an avalanche of since Fundão, hopefully attitudes such as this no longer exist and
mud-like tailings cascading down from the left abutment. the risks posed by brittle tailings are now acknowledged.
Other observers reported seeing a sudden jet of dirty water
“explode” out of an underdrain. 6.10.2 The Feijão TSF failure in Brazil in 2019.
A worker standing on the tailings beach reported the The Feijão failure, which occurred on 25th January 2019, has
ground begin to move beneath him and crack around him, already been mentioned in the introduction to this paper. It caused
detaching from the step-in slope and moving downstream. over 260 fatalities, many of them members of downstream
communities. The graphic video that emerged of the failure and
The report by Morgenstern et al. (2016) contains additional its immediate aftermath has doubtless been viewed many millions
eyewitness accounts, providing an excellent timeline of events as of times. It clearly demonstrates the suddenness with which the
they unfolded. Only the above observations, which were the first failure occurred. A vast volume of the tailings was clearly brittle.
in a sequence of many, are included here to illustrate a specific Failure of the TSF occurred at about 12:30 pm local time.
point. At Fundão it is clear the failure evolved over a few minutes. Unfortunately, this meant there were several people in a meal
Initiating an Emergency Response Plan (ERP) was not possible. room located a short distance downstream of the TSF when it
Some workers on the TSF managed to scramble to safety, but failed. The video footage shows the failure occurred over about
unfortunately, others did not. Additionally, communities 80% of the face of the TSF in only five seconds, with the release
158
of tailings occurring through a sequence of retrogressive failures 6.10.2.3 Reconciling the EPR and CIMNE findings
in only about five minutes. Once again, timeous implementation The findings from the two forensic reports are very different and
of an ERP was not possible. sometimes contradictory. For example, CIMNE conclude, “The
Unusually, the Feijão failure has seen two detailed forensic set of numerical analyses performed lead to the conclusion that
investigations carried out by different investigators. The first of the drilling of borehole B1-SM-13 is a potential trigger of the
these was commissioned by the TSF owners Vale and released in liquefaction that caused the collapse of the dam” and furthermore,
December 2019 (Robertson et al, 2019). This report is referred to “The analyses carried out have not been able to identify other
as the Expert Panel Review (EPR) in following discussions. The liquefaction triggers. In particular, the calculations performed
second report resulted from a Term of Cooperation signed by Vale incorporating only the effects of enhanced rainfall and creep,
with the Federal Public Prosecutor’s Office (MPF) to contribute either alone or in combination, have not resulted in a general dam
to and collaborate with the authorities on a second, independent failure”.
forensic investigation. This second investigation was carried out As noted earlier, EPR on the other hand attribute the failure to
by the Centro Internacional de Métodos Numéricos (CIMNE) and a combination of creep and loss of suction due to rainfall.
their report was released in August 2021 (CIMNE, 2021). This Furthermore, they specifically discard borehole over-
will be referred to as the CIMNE report. pressurisation as a trigger, concluding, “The localized drilling of
The Feijão TSF was built as an upstream valley impoundment the borehole on the day of the failure would not have triggered the
and raised in 15 stages over 37 years, starting in 1976. No wall observed global failure of the dam, as confirmed by computer
raises occurred after 2013 and tailings deposition ceased in 2016; simulations”.
the facility was thus inactive at the time of the failure. The fourth Design engineers would, quite reasonably, be concerned
wall raise include a step-in, much like the portion of the Fundão about this disparity in conclusions from the two forensic
TSF referred to above. investigations. Although the authors of the current paper have
personal opinions on which of the trigger mechanisms is most
6.10.2.1 Findings from the EPR forensic investigation plausible, we cannot comment in detail, not being privy to all the
The EPR concluded that the sudden loss of strength and the available data and investigations. A rational approach to the
resulting failure w ere a result of a critical combination of dilemma resulting from the different conclusions is perhaps to
ongoing internal strains, i.e. creep, as well as a strength reduction adopt the approach suggested some time ago by Silvis and de
due to loss of suction in the (surface and near-surface) unsaturated Groot (1995) and more recently by Morgenstern (2018), amongst
zone due to intense rainfall towards the end of 2018. others, which is – if there are tailings that are liquefiable, assume
The EPR notes that operations allowed ponded water to get they will liquefy. Rather than trying to predict plausible trigger
close to the crest of the TSF at times, resulting in the deposition mechanisms, simply assume liquefaction will be triggered.
of weak tailings near the crest, the above-mentioned step-in Although this is certainly a seemingly logical suggestion, it comes
resulting in upper portions of the slope being over weaker fine at a sometimes-substantial cost, for example when stabilising
tailings (slimes) and a lack of significant internal drainage buttresses are required to address stability under conditions of
resulting in a persistently high phreatic surface. They also flag the assumed mobilised large-strain strengths (see Figure 33).
likelihood of bonding between tailings particles occurring due to
the iron content, and this bonding producing a stiff, but collapsible 7 FIELD BEHAVIOUR AND ASSOCIATED MODELING.
structure which could manifest as a very brittle material if sheared
undrained. 7.1 Consolidation
6.10.2.2 Findings from the CIMNE forensic investigation Modeling of the field consolidation behavior of tailings serves
The CIMNE investigators commissioned additional laboratory several purposes. It provides information on the storage capacity
and field tests and undertook extensive, detailed numerical of a tailings impoundment as well as the rate of rise of the
modeling of the TSF. The CIMNE findings are very different from resulting tailings dam. Analyses also provide information on
those of the EPR and specifically discard both creep and bonding spatial and temporal distributions of the void ratio and densities
as contributory causes of the failure. within the impoundment. The information is then used to evaluate
Their report unequivocally identifies over-pressurisation of a both static and seismic stability of the resulting dam at various
borehole on the day of the failure as the trigger mechanism that stages of its construction. The void ratio distribution is also used
resulted in the catastrophic failure that resulted. Problems to establish the hydraulic conductivity values within the
associated with attempted installation of horizontal drains in 2018 impoundment to perform a realistic seepage analysis. Finally,
had been reported, with the compressor pressurising fluid during knowing the pre-closure void ratio distribution is essential to
drilling causing pressures of as much as 2MPa. In the CIMNE evaluate long-term post closure settlements in magnitude and
study they triggered (numerically) a very localised pore water rates. Detailed consolidation process modeling provides all these
over pressurisation at the tip of the borehole that was being drilled distributions through a rational approach, rather than through
on the day of the failure (borehole B1-SM-13). This overpressure empirical correlations.
propagated throughout the TSF perimeter zone, resulting in a very To gain an insight into the field behavior of tailings
rapid progressive failure. The predicted locations of rapid and consolidating under self-weight, several simple analyses are
significant movement match closely the observations provided by presented. The first example is a 5 meter tall column filled with
the video footage referred to previously. Interestingly, they tried tailings instantaneously. The material characteristics for the
triggering a failure by over-pressurising other boreholes that had analysis are presented in Figure 16 for both fine and coarse
been installed, but in all these cases the zone of overpressure grained tailings and the model parameters are listed in Table 4.
remained relatively localised. In their concluding discussions they The material parameters of the fine tailings are representative of
suggest, “It is likely the occurrence of liquefaction and its highly compressible and plastic tailings such as some oil sand
subsequent uncontrolled propagation requires a particularly tailings or kimberlite tailings. Conversely, the material parameters
unfavourable combination of circumstances”. Perhaps we can of the coarse tailings are representative of low compressibility and
take some solace from this suggestion (but still not ignore risks low plasticity tailings such as those characteristic of hard rock
associated with brittle tailings). mining. The analyses assume an impervious bottom boundary in
which settlement is caused by self-weight only.
159
Table 4. Parameters for finite strain modeling of consolidation example. change from a concave upward shape to a convex upward shape for both
fine and coarse tailings.
Parameter A B Z C (m/day) D Gs
(kPa)
7.2 Desiccation and creep
Fine tailings 8.00 -0.254 0.164 1.89x10-6 4.00 2.60
Coarse 1.48 -0.188 0.296 5.48x10-3 2.69 2.60 7.2.1 Desiccation
tailings Desiccation of tailings is widely acknowledged to improve the
mechanical characteristics of soils and tailings through suction-
Figure 58 presents the results of two analyses in terms of induced overconsolidation (Daliri et al. 2014). However,
column heights versus time for both fine and coarse grained experimental data that quantifies the magnitude of desiccation-
materials, for instantaneous filling. In both cases, an impervious induced suction as well as its effect on the mechanical response of
bottom boundary is assumed and the settlement is caused by the the tailings is relatively scarce. Efforts to conduct real-time
self-weight only. The settlement rate and the time to consolidated monitoring of the development of suction pressures at a South
state is controlled by the hydraulic conductivity values for the two African platinum tailings dam have been recently reported
tailings, while the ultimate height is controlled by their (Basson et al. 2021). The approach partly relied on the use of low-
compressibility characteristics. The physical processes cost tensiometers that enable the measurement of negative pore
controlling the consolidation behavior upon deposition are better water pressures as low as -1.7 MPa (Jacobsz 2019). Regarding
understood by looking at the void ratio distributions at selected laboratory assessments of the beneficial effects of desiccation on
times presented in Figure 59. strength, Daliri et al. (2014) reported improvements under both
monotonic and cyclic loading. Daliri et al. (2014) also noted that
mechanical overconsolidation yielded a mechanical response that
was qualitatively different from that of desiccation-induced
overconsolidation. Importantly, they cautioned that simulating
desiccation through mechanical overconsolidation can result in
unconservative designs. More recently, results from monotonic
direct simple shear tests on undisturbed samples of desiccated
tailings have quantified the increase in the peak undrained
strength ratio (Su v) induced by desiccation (Reid et al. 2018a,
b). The results from Reid et al. allow an estimation of the
preconsolidation vertical effective stress beyond which the
benefits of desiccation-induced overconsolidation become
negligible. Their results also highlight that the effective
preconsolidation pressure is lower in silt-sized materials (e.g. gold
tailings) than in more clayey materials (e.g. iron ore and bauxite
tailings) (Reid et al. 2018b).
To demonstrate the desiccation process we analyze the same
5 m columns (i.e. initially saturated and with an impervious
bottom boundary) that were previously considered when
Figure 58. Height vs time for 5m tall columns of fine and coarse tailings. discussing the consolidation procedure. Again, both fine and
coarse tailings are considered (Table 4) and an evaporation rate of
0.3 mm/day is assumed that eventually leads to the development
of capillary forces. Figure 60 presents the height vs time
relationships for both the fine and coarse tailings, while Figure 61
presents the void ratio distributions at various times during the
consolidation and desiccation process.
Figure 59. Void ratio distributions at selected times for 5 m tall columns
of fine and coarse tailings.
The consolidation process starts at the bottom boundary, despite it being
impervious. Initially an upward hydraulic gradient equal to the critical
gradient for the material is created in the slurry. The upward flow
maintains the condition of zero effective stress and a uniform void ratio Figure 60. Height vs time for 5 m tall columns of fine and coarse tailings
throughout the column, except at the impervious bottom where the flow subjected to surface evaporation at 3 mm/day.
rate is zero. This condition persists in the consolidating column until the
consolidation front reaches the top of the column. During that time, the
surface settlement rate is also constant and equal to the critical gradient
multiplied by the hydraulic conductivity of the slurry at the initial void
ratio. In a settling column, this initial velocity can be used to calculate the
hydraulic conductivity at the beginning of the consolidation process. Once
the consolidating front reaches the top boundary, the void ratio profiles
160
profile develops above that point. Figures 62 and 63 present the
results of such an analysis for both fine and coarse tailings.
Figure 61. Void ratio distributions at selected times for 5 m tall columns Figure 62. Height vs time for 5m tall columns of fine and coarse tailings
of fine and coarse tailings subjected to surface evaporation at 3 mm/day. subjected to drainage to atmospheric pressure at the bottom.
Comparing Figures 58 with 60 and 59 with 61, several
observations can be made. Firstly, in the initial stages of 5
Fine tailings - 20 days Fine tailings - 100 days
consolidation, there is no difference in settlement rates and void 4.5
Fine tailings -150 days Fine tailings - 350 days
Fine tailings 500 days Fine tailings - 700 days
ratio distributions between the self-weight consolidation and self- Fine tailings 1000 days Fine tailings 4344 days
weight consolidation with surface evaporation of 3 mm/day. 4 Coarse tailings 20 days
Coarse tailings - 60 days
Coarse tailings - 40 days
Coarse tailings - 80 days
During that time, the self-weight consolidation expels the pore 3.5 Coarse tailings 100 days Coarse tailings - 559 days
water, for both materials, at a rate that is higher than the
evaporation rate. Thus, evaporation has no effect on the tailings’ 3
Height (m)
behavior. Once the settlement rate drops below the evaporation 2.5
rate the desiccation process starts, and evaporation “pulls” water
out at the constant rate of 3 mm/day and consequently the 2
settlement rate equals the evaporation rate. The time at which the 1.5
evaporation rate becomes larger than the settlement rate due to
1
self-weight consolidation can be discerned from the change in
slope in the curves of Figure 60. For the fine tailings, this change 0.5
in slope occurs at approximately 230 days, whereas for the coarse
0
tailings it occurs at approximately 90 days. The settlement process 0 2 4 6 8 10 12 14
continues until the shrinkage limit is reached, at which point the Void ratio
material starts desaturating and the evaporation efficiency
significantly decreases or completely stops. After the desiccation Figure 63. Void ratio distributions at selected times for 5 m tall columns
process starts, the void ratios at the tailings surface reduce due to of fine and coarse tailings subjected to drainage to atmospheric pressure
suction development and a crust is formed. The crust thickness at the bottom.
depends on the relative values of hydraulic conductivity at the
tailings surface and the evaporation rate. The higher the The results show larger settlements due to larger increases in the
evaporation rate relative to the hydraulic conductivity, the thinner effective stresses when compared to self-weight consolidation
the crust. This is seen in Figure 61 where a distinct desiccated only. The void ratio profiles do not show the development of a
crust is formed for fine-grained material with lower hydraulic distinct crust at the surface as the suction development is more
conductivity, while for coarser material with higher hydraulic gradual than in the case of high evaporation rate at the surface.
conductivity the same evaporation rate produces more uniform
drying with depth. 7.2.2 Creep
The desiccation process also leads to tension crack To demonstrate the creep effect on the settlement and void ratio
development at the tailings surface due to the release of tensile distributions of a bentonite slurry (a highly plastic material) a five-
stresses created in the desiccation process. The thinner the crust, meter high column, impervious at the bottom, was analyzed using
the shallower and more closely spaced the cracks, while deeper the CONCREEP model (Gjerapic et al 2021). The results are
and broader spaced cracks are observed when the evaporation rate presented in Figures 64 to 66. Figure 64 compares the height-time
is lower relative to the hydraulic conductivity of tailings. Detailed results for both consolidation with creep and consolidation-only
description of the cracking process analysis is beyond the scope predictions for the material. The time-dependent compressibility
of the present paper, but a rational experimental approach and relations are presented in Figure 18 and the corresponding model
numerical model development has been presented elsewhere parameters are listed in Table 5.
(Abu-Hejleh and Znidarcic 1995, Oliveira-Filho, 1998 and Yao et
al 2002). Table 5. CONCREEP model parameters for a bentonite slurry.
Tailings can be subjected to suction stress development upon Creep and compressibility Hydraulic conductivity Specific Initial
deposition even in the absence of significant evaporation at the (m/day) gravity void ratio
(kPa)
surface. If the surface water is readily removed from the tailings
and the bottom boundary is drained to the atmosphere (i.e. the A0
Af
17.47
10.53
B0
Bf
-0.44
-0.35
Z0
Zf
0.29
0.16
C 2.62x10-7
2.40 30
bottom boundary constitutes a water table), suction develops near a1 3.00 b1 1.50 z1 7.00
D 2.43
a2 -0.50 b2 -0.30 z2 -0.60
the tailings surface as the groundwater level is gradually lowered
below the surface. Ultimately, the groundwater level reaches the
water table at the bottom of the deposit and a hydrostatic suction
161
7.3 Tailings storage facility filling analyses.
The nonlinear finite strain consolidation theory developed by
Gibson et al (1967) is a one-dimensional theory with spatial
variability in the vertical direction. Although it rigorously treats
the consolidation process in slurries, it cannot account for the
three-dimensional nature of tailings storage facilities. However, it
should be recognized that any water drainage in horizontal
directions in such facilities will be minimal as the depth to the area
ratio is very small in all cases. Even when the depth of such
facilities exceeds a few hundred meters, the horizontal lengths are
several times larger and any horizontal flow due to consolidation
can be neglected. Still, the three-dimensional geometry must
sometimes be considered, as the depth of tailings will vary across
the facility. Gjerapic et al (2008) developed an analysis algorithm
Figure 64. Height vs time for a 5m tall column of a bentonite slurry in which a tailings facility was discretized into several columns
analyzed with and without creep effects. and filling of each column analyzed using the one-dimensional
nonlinear finite strain consolidation theory. The algorithm
As expected, the creep effect causes larger settlement than
accounts for the interaction of the adjacent columns during the
consolidation only. However, the initial rate of settlement,
filling process in the sense that while they are filled
controlled by hydraulic conductivity, is the same in both analyses
simultaneously, the taller column will undergo larger settlement
and the creep effect cannot be separated from consolidation only,
for the first sixty years. Figures 65 and 66 present the and will require more tailings to be placed in it to maintain a
corresponding void ratio distributions at selected times for both leveled tailings surface. The algorithm is implemented into the
consolidation and creep, and consolidation only analyses. Again, numerical model FILCON that provides the filling curves for
there is very little difference in the void ratio distributions up to several scenarios of the filling process. It first provides the tailings
sixty-one years, but the creep effects are significant beyond that elevations if the tailings are deposited at an initial solids content
point. The creep effects are especially noticeable at the low and no water is “bled” from the slurry in the sedimentation
effective stress range (shallower depth) as expected from the time process. Thus, this is the filling rate for the case when an
dependent compressibility curves in Figure 18 for which the creep incompressible liquid is filling the facility. The second filling
effect diminishes with the increase in the effective stress. For this curve presents the case when the solids content in the slurry is
example, the maximum effective stress at the bottom is 2.26 kPa, increased instantaneously due to sedimentation and the freed
creating a creep induced reduction in void ratio from the effective water is bled out and removed from the facility. For that case, any
stress induced ultimate value of 11.6 to the value of 7.78 after consolidation beyond the zero effective stress void ratio is
2740 years. neglected. The third curve in the output presents the actual
projected tailings elevations under the given production rate that
accounts for simultaneous filling and consolidation. Finally, the
filling curve for a 100% consolidated tailings is also given. That
would be a filling scenario when the production rate is so slow
that any deposited tailings instantaneously consolidates under its
self weight.
Figure 67 presents an example of elevation-area curve for a
tailings storage facility that is discretized into four columns to
perform the filling analysis. The total volumes for the actual
facility and the discretized model are identical at each discretized
elevation.
Figure 65. Void ratio distributions for a 5m tall column of a bentonite
slurry with creep.
Figure 67. Elevation-area data for a tailings storage facility discretized into
four columns with equivalent volume for filling analyses.
Figure 68 presents the results of the filling analysis assuming
an ore production rate of 2000 tonnes per day and the tailings
disposal at an initial solids content of 10% (initial void ratio of
Figure 66. Void ratio distributions for a 5m tall column of a bentonite 23.4). The material characteristics are given in Table 4 for Fine
slurry without creep. Tailings. The results show that the 120m deep facility would be
162
filled in about 1800 days if no water is bled out during filling, differences were particularly acute near the toe of the slope,
while it would take about 3200 days to reach maximum elevation exactly where failures often initiate.
if all the water released in the sedimentation process is removed The analyses presented by Gjerapic and Znidarcic (2020)
simultaneously. When accounting for self-weight consolidation assumed isotropic hydraulic conductivity values for the various
during filling, the top elevation will be reached in about 4350 layers in the analysis. Beckett et al (2016) compared phreatic
days. If the production rate were much lower so that the tailings surface profiles measured in geotechnical centrifuge tests with
could completely consolidate under its own weight during filling, those predicted using standard software (SEEP/W). The soil
the facility could accommodate 2.8 times larger tailings volume chosen for testing was a relatively uniform silt (coefficient of
(12200 days/4350 days). This is clearly seen from the post-filling uniformity of five) and thus almost incompressible above about
settlement curve for the given production rate where the ultimate 25kPa vertical effective stress. As shown in Figure 69, the
tailings height due to the self-weight only is significantly reduced. agreement was generally extremely good for all the equilibrium
For this example, the ultimate height was 41.4 meters that would water levels that were set in the constant head reservoir on the left
of the centrifuge.
be reached after two thousand years.
Figure 68. Filling analyses results for the tailings storage facility
discretized into four columns.
Figure 69. Predicted and measured results for steady state seepage, after
7.4. Seepage Beckett et al (2016).
The prime focus of seepage analyses of TSFs is usually the The good correlation observed in Figure 69 gives confidence
potential environmental impacts of contaminated water on water in using standard software for seepage analyses. However, the
resources. This paper does not discuss these issues as it focuses conditions analysed in this comparison were idealised, in
on geotechnical aspects of tailings behaviour. However, correct particular the isotropic nature of hydraulic conductivity. However,
seepage analyses are also crucial for rational evaluation of TSFs, as many natural soils, may have anisotropic and
geotechnical stability. Shear strength of tailings (whether heterogeneous hydraulic conductivity characteristics. This is
characterised in terms of effective stresses or through undrained partly due to effects of particle shape (e.g. more clayey tailings
stability analyses) is dictated by effective stresses; correct will have plate-like particles rather than being bulky, resulting in
evaluation of pore water pressures is thus critical. As these are particle fabric governed by settling conditions) and partly due to
usually obtained by estimates of the phreatic surface evaluated flow conditions on a tailings beach, where meandering flow
through interpretation of spot measurements of pore pressure (e.g. produces variable rates of particle deposition. True hydraulic
using piezometers), assumptions inherent to such interpretations conditions may thus sometimes best be characterised as
should be challenged and verified. anisotropic (higher horizontal hydraulic conductivity).
As discussed in Section 9, pore water pressures are usually Furthermore, where tailings compressibility is significant, it is
measured at specific locations (and depths) using electronic necessary to characterise a reduction in hydraulic conductivity
piezometers, which are increasingly being automated rather than with depth, as illustrated previously in Figure 16. Another often-
requiring manual readings. The pressures are converted to head, overlooked factor is the effect on hydraulic conductivity of
i.e. water level, to define the location of the phreatic surface. particle segregation during flow down a beach. Blight (1997)
Somewhat inexplicably, at many TSFs the authors are aware of, a illustrates this effect, where hydraulic conductivity decreases
single piezometer is installed along a given cross-section. This is exponentially down the beach according to the observed decrease
of no use when trying to compare predicted versus measured pore in particle size. The resulting phreatic surface, when compared
pressure profiles. There are some TSFs where two piezometers are with a phreatic surface derived from assumptions of homogeneous
installed on a given cross-section and, albeit rarely, sometimes hydraulic conductivity was starkly different, as illustrated in
three or more. When two piezometers are available for a given Figure 70.
cross-section, it is common for the phreatic surface to be derived
by simply joining the inferred water levels. As noted by Gjerapic
and Znidarcic (2020), this is a fundamentally incorrect approach
as it violates continuity of mass. They illustrate the implications
further by taking two approaches to the interpretation of the
phreatic surface shape; these were the commonly-used linear
interpolation and an interpretation based on seepage analyses that
predicted pore pressures that matched the measured values. The
two approaches provide distinctly different outcomes, with the
linear approach underestimating the height of the phreatic surface
and thus overestimating effective stresses and strength. The
163
saline tailings. They found that with inhibited evaporation,
tailings consolidation was due to self-weight consolidation alone.
This invariably results in only low densities being achieved and
consolidation may take extremely long. Initial evaporation from
newly deposited tailings occurs through loss of pure water, often
resulting in precipitation of a distinct salt crust on the tailings
surface responsible for inhibiting further evaporation. The top
layer of the deposit does not achieve any significant strength
under such circumstances, with tailings below the crust remaining
at an extremely high water content.
With relatively fresh water, diligent management of
deposition layer thickness and cycle time can produce dense
tailings. However, even relatively moderate levels of salinity in
Figure 70. Comparison of phreatic surface resulting from assumption of the tailings water can cause actual evaporation rates to fall
homogeneous, isotropic hydraulic conductivity with assumption that significantly below the potential evaporation rate, Ep. This
hydraulic conductivity decreases exponentially down the beach. reduced evaporation rate occurs even with an open body of saline
Republished with permission of ICE Publishing, from Blight (1994); water and is exacerbated by crust formation. Firstly, the crust is
permission conveyed through Copyright Clearance Center, Inc. almost always white and the surface reflectivity (albedo) values
as high as 0.83 (Fujiyasu, 1997), meaning most of the incoming
In Figure 70, the assumption of isotropic permeability solar radiation is reflected. Secondly, once the salt crust dries, it
produces a clear overestimate of the phreatic surface height and acts somewhat like a surface mulch, meaning moisture flow must
thus underestimate of effective stress and strength. Although this occur in vapour form, further reducing the rate of evaporation.
may be a conservative outcome, it is surely no way to design The vastly constrained reduction in evaporation that occurs
major structures, i.e. blind faith in implied conservatism. It also due to crust formation means the tailings consolidate extremely
illustrates the folly of trying to fit a seepage model that is based slowly and in-situ void ratios remain high for some time, a highly
on assumptions of homogeneous, isotropic hydraulic conductivity undesirable scenario. Some mining operations have therefore
to measured pore water pressures of a system that is more akin to instituted intervention techniques to mechanically break up the
the exponential relationship evaluated in Figure 70. salt crust, often referred to as ‘mud farming’. Equipment has been
Another critical aspect of seepage analyses is the presence of specifically developed for this role, such as amphiroles. Newson
unsaturated zones within TSFs and the associated reduction in and Fahey (1998) demonstrated the benefit of crust removal by
hydraulic conductivity. While the presence of unsaturated zones measuring actual evaporation rates both before and after salt crust
might in some instances improve the stability of tailings (as removal at a Western Australian gold mine TSF. Figure 71 shows
suction increases effective stress and thus shear strength), seepage the results, and compares the measured evaporation rates with
of the pore-water towards downstream slopes will cause the measured pan evaporation rates (using tailings water). As can be
initially unsaturated zones to re-saturate and create zones with seen, the salt crust causes a decrease in evaporation of about 5 or
entrapped air. The effects of air entrapment on hydraulic 6 times, thus significantly reducing loss of water and strength
conductivity in the zones of positive pore-water pressures have gain.
been studied and reported by Hwang (1999, 2002), Bicalho
(1999), Bicalho et al (2011) and Lee (2011). The air entrapment
creates zones of reduced hydraulic conductivity, especially in the
areas where pore-water pressure is lower near the phreatic line and
towards a downstream slope. This lower hydraulic conductivity in
turn restricts the pore-water flow and increases the phreatic line
especially near downstream slopes of TSFs. Collins (1997) and
Collins and Znidarcic (2004) demonstrated the importance of
accounting for unsaturated seepage conditions when analysing
rain infiltration and its effects on triggering landslides. Similar
mechanisms may affect the stability of TSFs.
We suggest that a great deal more attention needs to be placed
on rational seepage analyses, taking account of measurable
hydraulic conductivity variations, including anisotropy and
unsaturated conditions, where possible. This is not a trivial
expectation, but enormous attention to aspects such as in-situ Figure 71. Measured evaporation at a gold mine TSF in Western Australia
strength and state characterisation have greatly improved current (after Newson and Fahey, 1998).
practice; perhaps it is time similar attention was paid to improving
seepage predictions and interpretations of in-situ measurements. 7.6 Aging
7.5 Crust formation. The increase over time of the resistance and/or stiffness of soils
has been identified as aging (Mitchell et al., 1984; Skempton
The salinity of tailings water in most TSFs is generally relatively 1986; Schmertmann, 1991). Furthermore, it has also been
low. However, there are many operations where water salinity is recognized that the resistance to liquefaction increases with the
extremely high, an excellent example being many mining
age of the deposit (Kramer et., 1998; Leon et., 2006). In particular,
operations in the Western Australian goldfields, where salinity
may be as high as 300 g/l total dissolved solids. Other examples Troncoso and co-workers studied the effect of aging on the
include many process residues, such as bauxite red mud, which liquefaction resistance of a copper tailings sand using cyclic
also have high process water salinities. triaxial tests (Troncoso et al., 1988). This study was carried out
The key concern with high salinity tailings water is the based on a series of cyclic triaxial tests performed on
inhibition of evaporation. Newson and Fahey (1998) reported on ‘‘undisturbed’’ samples retrieved from an old tailing dam at
an extensive study of various mining operations in Western different depths, which basically means different age of
Australia, with a prime focus being on effects of evaporation from deposition. Additionally, tests on fresh samples reconstituted in
164
the laboratory were conducted. The test results are summarized in 7.7.2 Tailings Deposition.
Figure 72, where the cyclic stress ratio required to produce a state Tailings deposition in cold temperatures may:
of softening with 5% double amplitude strain has been used as
criterion of liquefaction. Restrict the length of the slurry pipelines to prevent freezing
of the pipeline, if measures are not taken (e.g. isolating the
pipeline).
Limit the possibility of regularly changing the deposition
point, resulting in deposition occurring for extended periods
from the same deposition point.
Cause freezing of the beach, potentially resulting in embedded
permafrost layers if overly thick layers are deposited before
thawing occurs.
The first two issues are more of an operational nature, which
Figure 72. Curves of cyclic resistance ratio (a), and increase in cyclic can be addressed through appropriate procedures, whereas the
resistance with time (b), (after Troncoso et al., 1988). third may have a significant impact on stability of the tailings dam.
Embedded frozen layers will restrict permeability, or even result
in effectively impermeable layers, which will affect the rate of
It can be observed that cyclic resistance ratio tends to increase consolidation and result in layers of loose, saturated, contractive
by a factor of 3.5, 2.4 and 2 for the samples of 30, 5 and 1 years materials in the deposited tailings. If these layers are in the
of deposition, respectively. These experimental results indicate structural zone of the tailings dam, stability will be compromised.
that for stability analyses during the period of post closure, the If the layer stays frozen, the strength of the layer will be high, but
aging effect should be considered. This type of study can be done possible future thawing will cause an increase in pore water
when the tailings dam has been in operation for several years, pressure, reduced effective stress, and increased water content
when it is possible to retrieve samples at different depths, which within the material. For these reasons, the risk associated with
are associated with different years of deposition. Testing this impermeable layers such as strength loss during thawing, frozen
batch of ‘‘undisturbed’’ samples, it should be possible to establish layers, or permafrost, should be avoided in deposited tailings near
or within the structural zone of the tailings dam.
the variation of the cyclic strength with the age of deposition,
Knutsson (2018) examined tailings deposition in cold regions
which allows an estimation of the improvement of the cyclic
and did a numerical study of the potential formation of permafrost.
resistance with time.
The risk of generating permafrost due to tailings deposition exists
even in regions with no natural permafrost, as material being
7.7 Freezing
frozen during winter might not fully thaw during the following
Cold weather is a condition affecting tailings management in summer, due to added height of the tailings on the surface (e.g.
several ways with regards to both dam construction and deposition continued tailings deposition). The study simulated conditions in
of tailings. northern Sweden above the arctic circle with an average yearly
temperature of 0.1 C. Tailings deposition was studied for an
7.7.1 Dam Construction annual raise of 1, 2 and 3m, as well as for the yearly raise being
Freeze and thaw processes may reduce the density of dam deposited during 4, 8 and 12 months. For the first two scenarios,
construction material to a depth related to the material properties, deposition starting in January (winter) and July, respectively,
the temperature over time as well as the depth of snow cover. were studied. The following tailings characteristics were
Effects to depths of up to 3-4 m, if not more, at locations with low considered:
temperatures (below 0 °C) over long periods of time (6 months or
more) are not unusual. This means that the zone of material Specific gravity (Gs) – 2.8
affected by freeze and thaw must be re-compacted before the next Quartz content (q) – 40%
raise can be done. If the freeze depth is more than the rate of rise, Degree of saturation (Sr) – 33 and 100%
this may mean re-compaction is required for material at a depth Dry density ( d) – 1.3 and 1.8 t/m3
below the new tailings surface (Figure 73). A solution may be to
place a layer of “frost protection material” on the dam crest, which for four different scenarios, “Wet & Loose”, “Wet & Dense”,
is then removed before the next raise. With frost protection, only “Dry & Loose” and “Dry & Dense” which together cover a wide
the top surface of the material (dam crest) needs re-compaction range of in-place tailings possibilities. The results from the study
before the next raise is constructed. demonstrated:
Decreased water content results in less permafrost generation,
which has a larger impact than density.
Thickness of permafrost layers is nearly proportional to the
deposition rate (RoR).
A shorter deposition period generates more permafrost when
starting in winter.
Deposition starting (for 4 or 8 months deposition periods) in
January (northern hemisphere winter) results in more
permafrost compared to starting in July.
Figure 73. Illustration of a possible situation where freeze depth (here
When considering a 12 month deposition period, “Wet &
illustrated as 3m) is more than a yearly rise (here illustrated as 2m) results
Loose” material always generated permafrost for the studied
in a zone affected by freeze and thaw (and therefore reduced compaction)
deposition rates (1,2 and 3 m/a). Conversely, “Dry & Dense”
corresponding to 1.5 dam raises, e.g. 1.5 dam raises need to be re- material never generated permafrost (see Table 6).
compacted (zone to be re-compacted) before next raise can be done.
165
Table 6. Permafrost thickness [m] for 12 month deposition (e.g. constant considered as a first approximation in the verification of the TSF
RoR of tailings surface with time). stability and should be carried out with full knowledge of the
ultimate strengths of the materials involved, including those of the
Material Deposition rate (RoR) foundation soils.
1 m/a 2 m/a 3m /a Recently, Sadrekarimi (2016) proposed a modification to the
Wet & Loose 0.35 0.8 1.2 conventional, widely used limit equilibrium method of analysis to
Wet & Dense 0 0.425 0.875 account for various factors, including modes of shear (plane strain
Dry & Loose 0 0.25 0.65
compression, simple shear and plane strain extension). His
method allows for re-assigning shear strengths of regions (‘slices’
Dry & Dense 0 0 0
in the vernacular of the limit equilibrium method) where the peak
shear strength has been mobilised, to values equal to the large
strain shear strength. The focus was particularly on undrained
Freezing and thawing may, however, be used to improve shear strengths. The possibility offered by the modification
consolidation and shear strength of wet, soft tailings. Cyclic proposed by Sadrekarimi was to incorporate some degree of
freeze-thaw has been shown to significantly dewater the tailings. progressive failure into the simple limit equilibrium method.
During freezing, ice lenses develop and when they thaw, they Although the method does assist in providing insight into
leave “channels” in the material through which water can be vulnerabilities of some TSF slopes, it is overly brutal in the
drained. These “channels” then collapse and a volume decrease is instantaneous switch from peak strengths to large strain (residual)
achieved. For maximum effect it is crucial to make sure the strengths. No allowance for strain compatibility is made. The
released water is drained away from the tailings. Practically, this issue of brittleness of some tailings materials is over-simplified
is achieved by letting the top surface of wet, soft tailings freeze, using this approach.
thaw and drain a number of cycles (years) until maximum
consolidation is achieved. This method has been studied and used 8.1.1 Brittleness
in northern Canada to facilitate and reduce costs of TSF closure As noted in Section 1.3, the GISTM requires that brittle potential
capping (Beier and Sego, 2008, Proskin et.al., 2010). failure modes be identified and addressed. No clarity is provided
on what constitutes brittle behaviour however. There are two
8 METHODS OF ANALYSIS. aspects that we believe must be considered to address this
This section discusses methods of analysis considering static question, the first being the amount (percentage) of loss of
loading and dynamic loading conditions, with analyses strength post-peak, and the strain over which this strength loss
approaches that vary from relatively straightforward (e.g. limit occurs. Unfortunately, it seems that too often only the former
equilibrium) through empirical approaches, to advanced aspect is considered.
numerical modelling. However, each of these topics, particularly One outcome of the forensic investigations into the Aberfan
the latter, warrant a state-of-the-art paper in their own right, so this disaster in Wales in 1966, in which a flowslide of colliery residue
discussion is necessarily kept as brief as possible. destroyed large parts of a nearby village and resulted in 144
fatalities, was the definition by Bishop (1967) of what he termed
8.1 Static loading conditions the Brittleness Index, IB:
The limit equilibrium method assumes a potential slip surface and
the mobilised shear strength along the surface is determined and = (34)
compared with the available shear strength. The soil (tailings) is
assumed to be a perfect rigid-plastic material, and therefore, when
materials of quite different stress-strain behavior are involved, an where is the resistance to shear on the potential failure surface
engineering judgment is required to establish the most probable and the subscripts f and r relate to peak (failure) strength and
strength that is mobilised simultaneously by these materials along residual strength respectively. Thus, a material that exhibits no
the considered slip surface. The most common method is to post-peak loss in strength has a brittleness index of zero.
discretize the soil mass into a series of slices that are analyzed as
free bodies. To satisfy both the equilibrium of forces and moments
in the slices, it is necessary to transform an inherently
indeterminate problem into a determinate one. Depending on the
simplifications and assumptions made to satisfy the equilibrium
of forces and moments in the slices, different procedures have
been proposed, among which the one developed by Morgenstern
and Price (1965) is probably the most widely used in tailings
engineering.
Along the slip surface under analysis, both the available
mobilized shear strength and the shear stress required for
equilibrium (driving stresses) are evaluated. Then, the evaluation
of a Factor of Safety (FoS) is straightforward as the ratio between
these quantities. The FoS is a parameter, or an index, that provides
a simple indication of the state (in terms of its stability) of the
earth structure. Figure 74. Illustration of brittleness, which is a function of both the amount
The reports on the failures of Aznalcóllar (Los Frailes), Mount of post-peak strength loss, and the strain over which this strength loss
Polley, Samarco, Brumadinho and Cadia tailings dams (Gens, occurs.
2019; Hoffman, 2015; Morgenstern et al, 2016; Robertson et al,
2019; Jefferies et al, 2019) allow us to point out that when the Figure 74 illustrates this concept, which can be applied to
ultimate (large strain) strengths of the materials are properly either drained or undrained representations of soil strength. In the
established, the FoS obtained from the limit equilibrium analyses latter case the undrained strength is often normalised with respect
were systematically lower than one, consistent with the failures to the relevant vertical effective stress. Material A in this figure
that have occurred. In this context, the authors' opinion is that the shows a significant loss of strength (with a brittleness index close
use of the limit equilibrium analysis is useful, but it should be to unity) over a very small increment of strain. Such behaviour
166
was demonstrated for Merriespruit tailings (where failure of a TSF
in South Africa in 1994 caused 17 fatalities) by Fourie and
Tshabalala (2005). Tests were carried out on K0 consolidated
specimens subjected to undrained triaxial compression, with
results shown in Figure 75. As can be seen, the peak strength was
mobilised at axial strains less than ½%, followed by very rapid
strength loss, with IB values ranging from about 0.4 to 0.8. These
tailings exhibit both a high brittleness (high IB) and a rapid loss of
strength.
Figure 75. Deviator stress vs axial strain for undrained triaxial Figure 76. Variation of deviator stress with axial strain for undrained
compression tests on K0 consolidated specimens of Merriespruit tailings triaxial compression test on anisotropically consolidated bauxite residue.
(after Fourie and Tshabalala, 2005). Numbers in the legend refer to the
void ratio prior to shearing. Both the Merriespruit (gold) tailings and the bauxite residue
referenced above were anisotropically (K0) consolidated prior to
Material B in Figure 74 shows approximately the same values shearing. One exhibited extreme brittleness, with very rapid loss
of peak and residual strength as material A, but the strain over of strength, whereas the other was more ductile. However, it is not
which the strength loss occurs is many times greater than that of only material type that influences brittleness. The same material,
material A. Is material B therefore brittle? In terms of the consolidated along different stress paths prior to being subjected
percentage strength loss, yes it is. However, the loss occurs to some form of undrained loading, may exhibit different degrees
gradually (in terms of strain, not necessarily time). Perhaps a key of brittleness. This is illustrated in Figure 77, after Been (2016).
distinction between material A and material B is that post-peak Figure 77 shows the same material, prepared loose, subjected to
behaviour of material B is likely to provide some indicators – three different consolidation stress paths prior to being sheared.
visual and otherwise – of the loss of available strength, because Path A was isotropic consolidation. Paths B and C both involved
significant movements will occur prior to the terminal (residual) drained, constant mean effective stress consolidation, but sample
strength being mobilised. Measurements using inclinometers, B was not loaded into the ‘instability zone’ whereas sample C was.
shape accelerator arrays (SAA’s), or InSAR techniques are likely The instability zone is the region between the line defining the
to detect movements prior to residual strength being mobilised critical friction ratio (i.e. ‘M’) and the instability locus. As can be
over the entire potential slip surface. More about monitoring seen from Figure 77b, the slightest perturbation applied to sample
techniques in Section 9. C results in immediate loss of strength, whereas in sample B a
Behaviour similar to that of material B is shown in Figure 76,
small amount of strain is required to fully mobilise peak available
which is for an undrained triaxial compression test on a sample of
strength; for sample A, even more strain is required. Similarly,
bauxite residue (red mud) anisotropically consolidated to a
experimental data suggests that the brittleness index of nonplastic
vertical effective stress of 600kPa. Once again, the peak strength
is mobilised at a very small axial strain (less than 1%), but even soils is a function of their state parameter normalised by the slope
after 25% strain, the residual strength had not yet been fully of their CSL (Jefferies and Been 2015).
mobilised.
There is thus a great deal of difference between material A
and material B in Figure 74, in terms of the potential risk of a
catastrophic failure. Material A is likely representative of the
Feijão tailings, which exhibited both high brittleness and rapid
strength loss. A challenge facing TSF designers is to determine
what the threshold of truly brittle behaviour is. Assuming there is
the potential for significant loss of strength (high IB), over what
strain does this become a significant concern? Is it 1% strain
(probably expressed as shear strain in order to enable comparison
with numerical models – see section 8.2), is it 5%; 15%? This is a
major challenge for the industry. Simply designing a stabilising
buttress for any material that exhibits some strength loss (i.e.
strain softening), irrespective of the amount of strain required to
fully mobilise residual strength does not seem logical. For less Figure 77. Illustration of importance of consolidation stress path on
brittle materials exhibiting a degree of strength loss, it may be subsequent stress-strain behaviour, particularly the strain at which stress
more prudent to complement limit equilibrium analyses with loss is triggered (after Been, 2016; courtesy Australian Geomechanics
appropriately sophisticated numerical modelling (where Society).
‘appropriate’ refers to the ability to capture key features of
According to Robertson (2010), cohesive soils (tailings) with
behaviour, in this case being strength loss with strain).
a shear strain to reach peak undrained strength greater than about
5% and a gradual drop-off in resistance after reaching peak
167
strength are less likely to experience flow liquefaction, which
applies, for example, to many high-PI clays. As with any guidance
rule, the 5% shear strain indicator mentioned here may be
misleading. For example, when consolidated isotropically, rather
than anisotropically, the bauxite residue shown in Figure 76
reached peak strength at a much higher axial strain than the K0
consolidated specimen shown.
There is also the issue of structure (or fabric) and its impact
on brittleness. According to Leroueil and Hight (2003), a feature
of sensitive fine-grained soils involved in flow failures is their
relatively low plastic limit and small strain required to reach peak
undrained shear strength. Although some guidance does exist, it
may be expecting too much of the CPTu test to provide
meaningful information on brittleness of a tailings material.
Shuttle and Cunning (2007) and Jefferies and Been (2006) suggest
that when a soil has a state parameter greater than -0.05, strain
softening and strength loss in undrained shear can be expected.
They do not speculate on rate of post-peak strength loss however.
In a similar vein, Robertson (2010) suggests that soils with a
Qtn,cs>70 are likely dilative and strain hardening in undrained
shear is likely. Gens (2019) provides some extremely valuable
insights to this problem. Using a finite element approach with the
constitutive model CASM, as described by Yu (1998), he reports
the response to CPTu of soils having the same properties, except
for the strain required to reach peak strength and the strain over
which strength loss occurs. The various idealisations are
illustrated in Figure 78, where sample G has an IB value close to
zero and sample A an IB of about 0.66. Significant detail is
provided on the resulting profiles of tip resistance and excess pore
pressure generated during the simulations, a summary of which
Gens (2019) provides in Figure 79, which shows an increase in Qt
with decreasing IB, i.e. the more brittle the soil, the smaller the tip Figure 79. Variation of normalised tip resistance (Qt) and pore pressure
resistance generated, even for the same value of state parameter. parameter Bq with brittleness index (after Gens, 2019).
Also note, the peak undrained shear strength is the same for all the
simulations, whereas the resulting Qt value is not. Gens (2019) As before, the best way to address this issue is using a
emphasises that these results are preliminary, but the implications mechanics-based approach to analysing CPTu results. The
are potentially extremely important, as it throws into question our ‘widget’ approach of Shuttle and Jefferies (2016), as discussed
understanding of the limitations of the CPTu test. We expect this before, uses analyses of the spherical cavity expansion problem
topic is likely to be a focus of active research in coming years. with realistic soil models to derive a surrogate of CPTu behaviour,
with a scaling factor being required to make the transition from
spherical cavity to CPTu. Although the ability of the widget to
predict in situ remains a pending task (Figure 47), perhaps a
focus on an approach such as this provides a way to accommodate
materials with various rates of strength loss.
The role of bonding in determining brittleness
No discussion of brittleness would be complete without some
discussion of potential bonding (or cementation) in tailings
producing a brittle response to undrained loading, particularly in
light of the suggestions made by the ERP (Robertson et al, 2019)
relating to the Feijão TSF failure that the iron ore tailings
developed a bonded structure after deposition. As summarised by
Schnaid (2021), the bonding may have rendered the tailings stiff
and potentially brittle, which he suggests is consistent with the
lack of observable deformations prior to failure and the sudden
and rapid behaviour once failure was triggered. However, the
CIMNE report (2021) comes to the opposite view, as mentioned
before, and concludes there was no evidence of bonding.
To provide guidance on whether a soil is structured (whether
due to bonding, cementation or even aging – see for example
Figure 78. Stress-strain curves of soils used in CPTu simulations, as Troncoso et al, 1998), Schneider and Moss (2011) suggested a
reported by Gens (2019). rigidity index KG, which was slightly modified by Robertson
(2016) to , where
=( )( ) . (35)
and G0 = (Vs)2, Vs is the shear wave velocity and is soil (tailings)
density.
Using this approach, it is suggested that if KG > 330, the soil
can be classified as structured, and generalised CPTu based
168
empirical correlations may have less reliability (Robertson, 2016). example, is often defined for a 2% probability of exceedance
However, the situation is further complicated if the material is within a lifespan of 50 years, which corresponds to a return
partially saturated. As discussed by Robertson et al (2017), period of 2475 years.
suction hardening increases the small strain stiffness (and thus Maximum Credible Earthquake (MCE): This is the largest
Vs), an effect that could erroneously be interpreted as possible earthquake that could occur at the project site. It is
cementation. recommended to establish the MCE using the deterministic
approach. This earthquake is associated with the post-closure
8.2 Seismic stability analyses condition of the TSF and represents the ultimate possible
seismic loading under which the tailings facility must remain
8.2.1 Defining loading conditions in place without collapsing. Nevertheless, in countries with
To carry out seismic stability analyses of a tailings dam, it should very high seismicity such as Perú, Chile, Japan and Indonesia,
be recognised that there are two clearly different stages: operation the MCE must be considered during the operation of the TSF,
and post-closure. The operational phase, which often lasts a few but accepting important damages that will not compromise the
decades (20 to 40 years in many cases, with some mining confinement of the stored tailings and water.
operations being over one hundred years old), is practically
always linked to the construction of the facility and the global The Global Industry Standard on Tailings Management
growth of the TSF. During this stage it is possible to foresee the (GISTM) recommends the seismic design criteria summarised in
existence of a routine situation (normal operation) and special Table 1. It is worth noting that the application of an earthquake
situations, e.g. associated with some critical but brief construction with a return period of 10,000 years is associated with a 1%
issue (such as excavation at the toe of the existing dam to remove likelihood of probability of exceedance in a lifespan of 100 years,
weak foundation soils for the new stages). To face these practical which can be considered unnecessarily conservative. A return
situations two seismic events are normally considered: period of 5,000 years is associated with a 2% likelihood of
Operational Basis Earthquake (OBE) and the Maximum Design probability of exceedance in a lifespan of 100 years, equivalent to
Earthquake (MDE). The MDE is more demanding than the OBE, a 1% exceedance in 50 years, which is commonly referred to as
and it must be ensured that for this seismic event, considering the the Safe Shutdown Earthquake that is applied in seismic design of
special conditions of the TSF, the tailings and water remain critical structures such as chemical industries, liquefied natural
perfectly stored. In Australia and New Zealand, the MDE is gas industries and power plants, among others. Even though it is
termed the SEE (Safety Evaluation Earthquake); the associated an extreme seismic event, its use in highly sensitive facilities
descriptions are that the OBE only results in minor damage and makes sense, and therefore, it can also be considered applicable to
the TSF and appurtenant structures and equipment remain TSFs.
functional, whereas the SEE causes damage, which is acceptable,
but there is no uncontrolled release of water (or tailings) from the Table 7. Seismic design criteria.
facility.
Consequence classification Seismic criteria – annual exceedance probability
Additionally, the long-term stability of decommissioned (also Operations and closure Passive – closure (passive
termed ‘closed’) TSFs as well as those that have been truly (active care) care)
‘abandoned’, i.e. there is no identifiable owner who can safely Low 1/200 1/10 000
close the facility – hopefully fast becoming a problem consigned Significant 1/1 000 1/10 000
to history, constitutes another design consideration. Seismic High 1/2 475 1/10 000
stability must be guaranteed in the long term, and therefore, long Very high 1/5 000 1/10 000
Extreme 1/10 000 1/10 000
return periods in the selection of the earthquakes must be
considered. Accordingly, the Maximum Credible Earthquake
(MCE) must be estimated and used in the stability analysis. The seismic hazard analysis has to consider the different
The probabilistic seismic hazard analysis (PSHA) and potential seismic sources in the area of the project. For example,
deterministic (DSHA) methods for the assessment of seismic in subductive seismic environments, there are three potential
hazard analysis are usually presented as antagonistic approaches. seismic sources: interplate (thrust), intraplate (intermediate depth
However, they may in fact complement one another when in-slab earthquakes caused by down-dip tension in the subducting
estimating the ground motions for design. In regions of high plate) and cortical (shallow crustal earthquake). Because
seismicity, where continental active faults and/or tectonic intraplate and cortical earthquakes attenuate much faster than
boundaries generate large seismic events, which have been interplate earthquakes, the appropriate selection of the attenuation
systematically reported in the last 200 - 300 years, the DSHA is law becomes relevant. Additionally, when considering intraplate
more suitable. This is because in regions of high seismicity, the earthquakes, the resulting seismic disturbance at the site of the
empirical evidence of extreme seismic events is reliable and dam is highly sensitive to the definition of the hypocenter
provides valuable real information about the ground motion locations. This is an important issue where no consensus has been
effects (Wang, 2010). However, in areas where seismicity is low achieved yet, and in the meantime the worst case is usually
or diffuse, the PSHA is more appropriate for establishing the adopted resulting in the selection of an intraplate earthquake just
potential earthquakes that can occur at the project site. In general, below the site of the project.
the different seismic disturbances that must be considered are: It is common to select a time history that provides the largest
Operational Basis Earthquake (OBE): This is a probable peak ground acceleration. However, the frequency content of the
earthquake to which a tailings facility can be expected to be selected earthquake could be a more relevant criterion, especially
exposed to during its design life; therefore, the tailings when analyzing high dams that tend to have a low predominant
disposal facility is expected to remain fully operational when vibration frequency (or large period). In this case, it is
subjected to such an earthquake. This earthquake could be recommended to first estimate the natural period of the tailings
defined, for example, for a 10% probability of exceedance dam, and then the seismic hazard analysis should identify those
within a 50 year period, which corresponds to a return period potential earthquakes that contain intensive energy at around the
of 475 years. natural frequency of the dam. These can be earthquakes of distant
Maximum Design Earthquake (MDE), also termed SEE, as epicenters that generate and impose surfaces waves, which result
discussed previously: The main purpose of designing against in seismic events containing low frequencies.
this earthquake is to safeguard loss of life, uncontrolled A balance must be struck between an earthquake with the
failure, collapse and loss of confinement, which may lead to highest possible acceleration and with a frequency content as
catastrophic consequences. This strong ground motion, for close as possible to the predominant frequency of the TSF. In
169
summary, two different design earthquakes must be defined: one There are several commercial software packages that can be
with maximum PGA and another with a frequency content that used to perform seismic analyses, but it is important to point out
matches, or it is close enough to, the predominant frequency of that the constitutive laws that are adopted to model the stress-
the dam. Therefore, the seismic analysis should consider at least strain behavior of the materials have a significant effect on the
six records (OBE x 2; MDE (SEE) x 2; MCE x 2). results. The discussion of the different constitutive laws that have
been developed is beyond the scope of this paper. We stress
8.2.2 Pseudo-Static Analysis – Limit Equilibrium however, that the selected constitutive law must show a
This procedure was very popular before computers were compromise between the predictive capacity of the model and the
developed to the level that made possible a dynamic stress-strain practicality of evaluating the input parameters required by the
analysis in only a few hours. In this context, many engineers see model. Interesting reviews on constitutive models have been
no reason to continue using the pseudo-static approach to verify reported by Lade (2005) and Prevost and Popescu (1996).
the stability of TSFs. Although perhaps true, it is necessary to Under static loading conditions, the mechanical response
retain “the big picture” before entirely eliminating the pseudo- depends mainly on the geometry, the magnitude of the acting
static analysis approach as one of the alternatives for assessing the forces and the stress-strain behavior of the materials involved,
stability of TSFs. with the physical stability ultimately controlled by the distribution
Seismic stability evaluated by pseudo-static analysis requires of stresses and their magnitudes relative to the mobilized
the use of seismic coefficients, which represent inertial forces strengths. On the other hand, in the case of seismic disturbances
through the application of permanent forces using a limit of earth structures, the shaking induces inertial forces that are
equilibrium approach as discussed in section 8.1. Strictly largely dependent on the amplitude and frequency content of the
speaking, the replacement of a seismic disturbance by a static seismic load and on the predominant natural frequency of the earth
force is theoretically incorrect. Nevertheless, from an engineering structure. The well-known phenomenon of resonance occurs
design point of view, it has been used successfully in countries when there is a perfect match between the frequency of the load
with high seismicity, in both building design and in the stability and the natural frequency of the structure. This issue is important
assessment of earth structures (Sano, 1906; Housner, 1984; in the selection of the design earthquakes as discussed above.
Towhata 2008). The seismic coefficients (horizontal and vertical) In addition, every dynamic response is affected by the loss of
were traditionally estimated from the expected peak horizontal energy that normally takes place during vibrations. This
acceleration of the site, but more recently the allowable level of phenomenon is called damping, which for simplicity is usually
seismic displacements of the slopes has also been included (Bray characterized as viscous type behaviour, and therefore increases
and Travasarou, 2009; Papadimitriou et al, 2014). Considering the as the strain rate increases. It is necessary to point out that soils
high seismicity of Chile, Saragoni (1993) proposed the following develop a loss of energy that is of the hysteretic type (it does not
relationships for the horizontal seismic coefficient, kh, for depend on the speed of deformation). In soils, the damping ratio
interplate (shallow thrust) earthquakes: is defined as the area of one stress-
normalized by the area of the equivalent input linear elastic energy
(Figure 80a, after Kokusho 1980, Ishihara 1996). This hysteretic
damping ratio is strongly dependent on the level of deformation
as shown in Figure 80b, (after Kokusho 1980, Ishihara 1996). At
(36) large deformations close to failure, the damping ratio reaches
values as high as 25%, which means a significant loss of energy,
which in general, is favorable for the attenuation of the seismic
response.
The vertical seismic coefficient is commonly assumed as 1/2
or 2/3 the horizontal seismic coefficient. In general, the
incorporation of this vertical seismic coefficient does not have a
real effect of the resulting factor of safety.
The pseudo-static analysis procedure is certainly attractive
because of its simplicity and because empirically it has been
shown to provide a reasonable estimation of the level of stability
of an earth structure subjected to earthquakes. Nevertheless, due
to the rather crude assumptions inherent to this method, its main
disadvantage is associated with the need for engineering
judgment, which is not assured if applied by inexperienced
professionals. Additionally, for tailings dams where liquefaction Figure 80. Damping ratio definition (a), and damping ratio of Toyoura
can be a major problem, or where foundation soils are complex, sand (b), (after Kokusho 1980).
pseudo-static analysis must be used with careful attention to the
strength of weak materials. Similarly, however, in these The output of a dynamic analysis can be greatly affected by
situations, strain-strain numerical models have also to be applied the mathematical formulation adopted to model the damping. In
with the same care, as the variety and complexity of input the dynamic analysis of earth dams (including tailings dams), it is
parameters and the range of available modeling options reduces common to use the so-called Rayleigh Damping (RD), which is
the likelihood of similarity of outcomes when undertaken by mathematically very attractive, but is difficult to estimate the
different engineers (including those with extensive experience). values that represent the real problem under analysis. The RD is
essentially a viscous damping that is established as a linear
8.2.3 Numerical modelling combination of mass and stiffness. Thus, the damping matrix C is
With the development of computers with large memory and given by:
computing capacity, the numerical methods of finite elements and
finite differences in 2D or 3D can feasibly be used to analyse and
predict the behavior of TSFs under both static and seismic stress [ ]= [ ]+ [ ] (37)
loading conditions, including different scenarios of seepage and
stages of construction.
170
where, and are constants of proportionality and M and K are It is of great importance to note that in the data used in this
the mass and stiffness matrices respectively. chart, which comes from different earthquakes that occurred in
different parts of the world, there are no cases of liquefaction for
varies with the response frequency. Therefore, it is important to values of qc1-Ncs greater than 170. This allows the conclusion to be
adjust this damping ratio to the fundamental frequency of the drawn that there is a threshold value of the tip resistance (qc1-Ncs =
analyzed problem, otherwise the results could be significantly 170) above which liquefaction would not occur, regardless of how
overdamped in terms of lower accelerations and smaller strong the earthquake is. A corollary for areas of high seismicity
deformations. Importantly, the RD is different from the hysteric is that compaction of non-plastic materials (e.g., tailings) to qc1-
damping ratio indicated in Figure 80a (after Kokusho 1980, Ncs > 170, prevents the occurrence of liquefaction.
Ishihara 1996), and, therefore, the number that is adopted for the For earthquakes of magnitude other than 7.5, overburden
RD requires experience and a deep understanding of the variables pressures different than 1 atm and nonzero driving static shear, the
of the problem under analysis. It is not an exaggeration to say that CRR obtained from this chart must be modified using the
Magnitude Scaling Factor, MSF, the effective overburden stress
in a dynamic problem, damping is a fundamental parameter,
factor, K , and the shear stress factor, K , respectively (National
which is usually not well characterized. Nevertheless, when
Academies US, 2016). Accordingly, once the representative value
constitutive models are used that reproduce the stress-strain curve
of penetration resistance qc1-Ncs has been established, the cyclic
during loading and unloading, damping results automatically from resistance ratio, CRRM=7.5, is obtained from the corresponding
the load-unload loops that the model generates. In the absence of chart. This value is then corrected as follows to obtain the cyclic
liquefaction, the analysis of these loops allows verification of the resistance ratio, CRR, representative of the actual conditions:
capacity of the model to reproduce the behavior of the soil or
tailings. = × × × (38)
.
The use of stress-strain analyses is a clear improvement on the
limit equilibrium analysis method, but it is extremely important to As can be observed from the above expression, the effects of
keep in mind that stress-strain analyses must always be verified earthquake magnitude, overburden pressure and static shear
by alternative means. A first verification of the capability of the modify the CRRMw=7.5, which means that the curve of Figure 81
model is to check how well it reproduces the results of the moves vertically when these effects are considered. Therefore, the
performed laboratory tests. Also, it is recommended to verify the threshold value of the tip resistance (qc1-Ncs = 170) discussed
model in simple conditions where the soil response is known. In above holds valid for higher confining pressure, static shear and
addition, it should be reviewed by independent experts to verify earthquake magnitude.
the quality of the analysis.
8.3.1 Magnitude Scaling Factor, MSF.
8.3 Empirical approaches To account for increased likelihood of liquefaction with an
Given the difficulty in recovering high quality undisturbed increase in the magnitude of an earthquake, Seed and Idriss (1982)
introduced the Magnitude Scaling Factor, MSF, which is based on
samples from the field for laboratory testing, for the estimation of
the concept that the greater the magnitude of the earthquake the
the CRR the use of in-situ tests is usually recommended, for
greater its duration, and therefore the larger the number of cycles
example, SPT, CPTu, or the vane shear test. In this context, a
of the seismic disturbance. It has been recognized that the MSF
significant amount of research has been carried out using the depends on several factors such as type of earthquake, distance to
CPTu to evaluate the onset of liquefaction through the CRR. the epicenter, soil type and characteristics of the site, among
Figure 81 presents a chart for estimating the CRR to magnitude others. However, the actual effect of these factors is still under
7.5 earthquakes based on the cone tip resistance normalized by investigation. Boulanger and Idriss (2014) incorporated the soil
overburden pressure and fines content, qc1-Ncs. characteristics by considering the values of the tip resistance qc1Ncs
and/or (N1)60cs in the MSF as shown in Figure 82.
Figure 81. CRR for Mw = 7.5, 'v = 1 atm and no static shear stress, as a
function of the equivalent clean sand qc1Ncs, (after Boulanger and Idriss Figure 82. MSF as a function of the penetration resistance. Reprinted from
2014). Boulanger and Idriss (2014), with permission from Elsevier.
This chart has been obtained from sites than are not associated 8.3.2 Overburden stress factor, K .
with tailings materials. Therefore, its use for tailings should be Results of laboratory cyclic tests have systematically shown that
carried out with due caution. The authors´ opinion is that for liquefaction resistance decreases with increasing confining stress,
tailings that classify as sandy soils or silty soils of low plasticity, as shown in the test results on tailings sand presented in Figure
the chart provides a good estimate of the CRR. However, in 83. Because the empirical field-performance database utilised
tailings containing plastic fines, the chart is likely to give cases associated with overburden pressures of around 1 atm (100
misleading results.
171
kPa), Seed and co-workers introduced the overburden correction However, as shown in Figure 85, it has been recognized that
factor, K , defined as the ratio between the CRR at a given K is a function of the relative density (Vaid et al., 2001). Different
effective vertical stress to the CRR at an effective vertical stress studies have provided more data that show that other factors, such
of 1 kg/cm2 (1 atm), considering the same conditions of density as the stress history, also influence the value of K (Boulanger
and other variables than may affect the CRR (Seed 1983). 2002, Hynes et al., 1998, 1999; Montgomery et., 2012; Vaid et al.,
2001; Seed et al., 1990, among others).
It is important to mention that there is experimental evidence
indicating that tailings materials have a less pronounced decrease
in cyclic stress ratio with confining pressure (Riemer et al, 2008;
Verdugo and Santos 2009; Verdugo 2011). This condition can be
explained by the greater compressibility shown by tailings
materials, that results in a fairly pronounced increase in density
with the confining pressure. Figure 86 presents the isotropic
consolidation curves of three natural sands and one copper tailings
sand. For a better comparison between these sands, the void ratios
have been normalized by the initial preparation void ratio of each
sand. From these results, the higher compressibility of copper
tailings sands is clear, at least in the range of pressures used up to
5 MPa.
Figure 83. Effect of confining pressure on the cyclic strength of a Peruvian
tailings sand. Republished with permission of ASCE, slightly modified
from Riemer et al. (2008); permission conveyed through Copyright
Clearance Center, Inc.
For natural sands, reported values of K are presented in
Figure 84, where significant scatter can be observed (Montgomery
et al., 2012). Figure 86. Compressibility of natural and tailings sands (after Maureira et
al., 2012).
A particular characteristic of tailings particles is their
angularity and sharp edges, which may crush during shear as
shown in Figure 87. Therefore, it is possible to postulate that the
angularity of the tailings sands results in a greater compressibility
generated by the breakage of the sharp edges, which increases the
density and packing, favouring a better arrangement of the
particles and their contacts.
0.1 mm
0.1 mm
Figure 84. Reported values of K .(after Montgomery et al., 2012).
Figure 87. Illustration of angularity of particles of copper tailings, before
triaxial test (left), and after test (right).
Accordingly, for tailings materials it is highly recommended to
assess the effect of the overburden pressure by means of suitable
laboratory tests.
8.3.3 Shear stress factor, K .
The chart developed by Seed and co-workers to estimate the
potential of liquefaction based on SPT measurements was
established for level ground conditions in which there is no static
driving shear stress. However, it is well accepted that the static
shear stress has a significant effect on the generation of pore
pressure, as well as on the development of cyclic strains (Vaid et
al., 1983, Pillai, 2001, Boulanger 2002, Idriss et al., 2003, Hosono
et al., 2008, Verdugo 2011, Park, 2020). Given this, a significant
amount of research has been done on this topic. Nevertheless, the
Figure 85. Effect of Dr on K values (after Vaid et al., 2001) reported results are not definitive, and it has even been considered
to jointly include the effects of confining pressure and static shear,
for which the coupled factor K has been proposed. Considering
172
that the cyclic resistance is sensitive to the static shear stress, and
it varies from sand to sand, in the case of tailings materials is
recommended to perform suitable laboratory tests.
9 MONITORING.
The Global Industry Standard on Tailings Management (GISTM)
devotes a substantial section of the document to issues of
monitoring (Principle 7). Of relevance to the discussions
presented in this paper, Requirement 7.2 requires, “Design,
implement and operate a comprehensive and integrated
engineering monitoring system that is appropriate for verifying
design assumptions and for monitoring potential failure modes.
Full implementation of the Observational Method shall be
adopted for non-brittle failure modes. Brittle failure modes are
addressed by conservative design criteria”.
9.1 Displacement monitoring.
9.1.1 Remote monitoring.
The GISTM provides no illumination on what constitutes ‘brittle’
behaviour, probably reflecting the lack of consensus on this issue).
However, the distinction made in the above quote is critically
important. It reflects the understanding provided by the Feijão
forensic reports that failure of extremely brittle tailings is unlikely
Figure 88. Experimental data from modified direct simple shear device
to be preceded by sufficient warning. Although there are a number
showing minimal vertical strain prior to rapid collapse of specimen (after
of publications that claim to have successfully predicted the Feijão
Reid and Fourie, 2019).
failure, mainly using Interferometric Synthetic Aperture Radar
(InSAR) satellite imagery, these studies have one thing in Despite the likely limitations of InSAR discussed above, there
common. Their predictions were done with the benefit of is likely significant value to be gained by using InSAR for
hindsight. None of the authors of this paper are aware of accurate monitoring of more ductile systems, as long as these systems
predictions of failure of a TSF made (and published) before the (where ‘system’ is used to indicate the combination of tailings,
failure. If such papers do exist, we would warmly welcome underlying foundation soils and any other soil or rock zone of
notification of such work. relevance, or even an embedded structure such as a decant facility)
In their forensic investigation of the Feijão failure, CIMNE are correctly characterised.
highlight a number of potential difficulties in using InSAR
imagery to detect an early indication of potential instability of 9.1.2 Local measurements of displacement.
brittle material, including the line-of-sight (N-S) of the satellites Although many mines utilise an array of surface survey
were in the same direction as the movement of the TSF, they do monuments for monitoring displacements, our experience is that
not separate different motion components, and the histories of such systems are rarely useful, particularly if used in isolation.
line-of-sight displacement measurements from the different Great value can be obtained from instruments such as
studies noted above vary quite significantly. It is noted that it is inclinometers, which are installed in vertical casings, or the more
unclear (using InSAR alone) if the area of the TSF of interest recent technology known as shape accelerator arrays (SAAs);
moved by 5, 15 or 30mm in the year prior to the failure. Holden both of these systems enable detection of horizontal movement
et al (2020) point out the challenges of correctly accounting for (primarily) with depth, thereby providing a continuous profile of
signal-to-noise ratios of InSAR measurements in many cases. horizontal movement with depth. These devices are particularly
InSAR has many potentially useful applications when dealing useful when a well calibrated numerical model has been used to
with ductile materials, but the application of the technology to produce predictions of expected movement, ahead of the
more brittle material remains questionable. measurements being made. This constitutes a true application of
the much-advocated Observational Method (Peck, 1969), which
Reid and Fourie (2019) report on a laboratory study using a
is rarely correctly implemented.
purpose-built direct simple shear (DSS) device in which stress
control of both horizontal (shear) and vertical stresses was
9.2 Pore pressure monitoring.
possible. Specimens were prepared at a positive state parameter
and consolidated under an applied shear stress (sometimes Although there are still numerous TSFs where pore pressures are
referred to as a stress bias). Samples were then gradually brought monitored using standpipe piezometers, which need to be read
to failure (drained conditions were maintained during this manually, our observations are that vibrating wire piezometers
process) by gradually reducing the vertical total stress (referred to (VWPs) are increasingly being used, with data acquisition being
in the paper as a CSD stress path test). A typical result is provided automated. This is the preferred option; obtaining continuous
in Figure 88, which shows vertical strain as a function of time. records of pore water pressure, with the ability to check data
Initially, there is a slight expansion (negative vertical strains) due whenever required, is immensely beneficial to a robust tailings
to elastic unloading, but when failure is triggered, contractive management system. However, it is extremely important that the
strains occur extremely rapidly, too rapidly to track accurately, data from VWPs are interrogated by an experienced engineer,
and complete liquefaction occurs. It is highly unlikely that such with issues such as regular checks of calibration values being
small pre-failure deformations could be detected, even by a made, and interpretation of data being made in the light of a
sophisticated system such as InSAR. reasonable seepage model. Given the discussions provided in
section 7.4, it is unlikely that interpretations made on the
assumption of homogeneous and isotropic hydraulic
conductivities will automatically be rational. Emphasis must be
173
placed on developing and maintaining a calibrated seepage model dam-break studies. The results obtained from a dam-break
wherever possible. analysis have been shown to be sensitive to the following
parameters and conditions (this list is not necessarily exhaustive):
9.3 Other instrumentation.
Amount of both supernatant tailings liquor and free water in
Monitoring systems based primarily on a combination of the impoundment.
displacement monitoring and/or pore pressure sensors are Extension, shape, and location of the considered breach.
currently common at many TSFs worldwide. Recent high-profile Residual strength of stored tailings.
failures of TSFs, as discussed at some length in this paper, have Topography of the zone downstream of the dam
galvanised the introduction of many more, largely unproven
technologies. Some of these may well prove highly applicable and The strength (viscosity) of tailings released in a flowslide
economically viable to implement, such as some of the event as well as the downstream topography are variables to be
geophysical based techniques, but until a solid database of proven quantified when undertaking a dam break study. The viscosity is
success becomes available, caution is warranted. The University usually characterised as being deterministic in nature. However,
of Western Australia currently hosts a website focussed on the amount of fluid retained within an impoundment is
‘Meaningful Monitoring’, accessible at: probabilistic as well as being perhaps the most important
www.tailingsmonitoring.com, where advantages and condition that controls the run-out distance of the tailings, should
disadvantages of existing as well as emerging technologies are a flowslide occur (Rourke et al., 2015). The characteristics of the
discussed. breach that forms, enabling tailings release, cannot be measured
but rather requires a prediction that will have to rely to some
10 GOVERNANCE extent on engineering judgement. Therefore, a dam-break analysis
This paper focussed on geotechnical engineering aspects of safely can range from excessively conservative to excessively
unconservative. Given this, we strongly recommended that dam-
managing TSFs. However, we are cognisant that such
break analyses be done that include sensitivity studies allowing
considerations form only a part of the decision-making processes
for the variations noted above.
that most mining companies are currently implementing. There
The consequence category may, however, sometimes be
are increasingly onerous compliance issues, whether these be chosen without a dam break analysis if it is obvious which
regulatory (e.g. imposed by a sovereign government) or sector led, category the facility falls under. A dam break analysis should,
such as the GISTM, which originated from actions initiated by the however, be carried out as a basis for emergency preparedness and
Church of England pension fund. The intention of this short response planning (EPRP).
section is not to go into any detail about such governance issues, Different countries have different systems of consequence
but primarily to highlight the framework that increasingly governs classification, and we do not explore this issue here. It is also a
tailings management. rapidly evolving endeavour, with changes to regulations occurring
on an ongoing basis.
10.1 Consequence classification
Approaches to safely managing TSFs, from both regulators and 11 CLOSURE AND RESIDUAL LIABILITIES.
interested sector groups (such as the GISTM), are increasingly Closure of TSFs is probably the most challenging and most
based on the concept of a Consequence Category difficult aspect of overall mine site closure. It propels tailings
classification. The intention of this approach is to understand the management into a design category similar to that of management
potential damage that a TSF failure would inflict on its of nuclear waste, as the time perspective becomes ‘in perpetuity’.
environment regardless of the likelihood of such failure. This GISTM (2020) reinforces the concept of “designing for closure”,
approach has been used in some countries for many years and is which influences the project from inception (e.g. site and
often used to differentiate between facilities posing high risk from technology selection) as well as influencing design alternatives
those posing less risk to population, environment, society, considered for the operational phase, based on likely impacts on a
infrastructure, economies etc. successful closure strategy.
Recognising the extremely long design life of a closed TSF
A consequence classification is undertaken in order to: (particularly when compared with typical civil engineering
Inform selection of design loading criteria infrastructure) a number of challenges face designers of these
Inform dam safety management operations and programs facilities, as well as owners, including:
Meet regulatory and stakeholder requirements
Loading cases
The classification is based on the consequences of a possible Land use after mining
TSF failure and the associated runout, should a flow-type failure Responsibility for land after mining
occur. The likelihood of a TSF failure is not considered, but the Approval of closure design
failure mode needs to be physically possible, referred to as a
‘credible failure mode’. It is important to understand that the TSF The current strong focus on mine closure is relatively new (the
consequence classification category is not a measure of risk first International Mine Closure Seminar was held in Perth in
(which considers likelihood). On the contrary, the dam 2006). Few TSFs have been closed and relinquished, i.e.
classification category is used to support design and operating including transfer of responsibility from the mining company to
decisions intended to reduce the likelihood of failure and hence the state. Nevertheless, planning for closure is today, in most
reduce risk. A dam owner can, if so inclined, base the design and countries, a regulatory obligation and a closure plan is a
surveillance program on a higher consequence category, which in requirement for getting an operating permit. Knowledge and
some cases may be relevant based on a cost-benefit evaluation. competence, as well as requirements and expectations, are
Not always, but often, the potential runout from a TSF failure constantly increasing, placing more and more attention and focus
is determined using a dam break analysis which in turn aids with on closure as well as changing the perspective from merely
consequence classification. Until recently such analyses were, at closing a facility to designing for closure.
best, rudimentary. However, recent efforts by various research Planning for closure up front includes input from a wide range
groups and consulting engineering companies have seen of specialists in addition to the tailings and/or mine engineers. It
significant improvements in our ability to conduct meaningful should include specialists in surface and groundwater,
174
hydrologists, geochemists, biologists, social planners, community isotropic, homogeneous hydraulic conductivity values. In reality,
relations planners, landscape architects, and others who may hydraulic conductivity may be anisotropic (usually more
provide valuable input to the closure plan. Additional to including permeable horizontally than vertically), may decrease with depth,
a broader range of specialists, the closure process will invariably and may decrease with proximity to the decant pond. All, or some
benefit from being integrated with the design and operation of these factors could render an oversimplified seepage model
processes. useless. Calibration of a seepage model, using monitored and
We flag the issue of closure in this paper, because we strongly correctly interpreted piezometer data should be the starting point
believe that sustainable closure of a TSF will require for correctly determining phreatic conditions. Anything else may
implementation of sound geotechnical engineering principles. provide a false sense of security.
Topics discussed in this paper, such as correct laboratory and field Whilst many countries in South America, as well as some
testing, methods of analysis and appropriate application of other parts of the world have faced the challenges of designing
empirical approaches apply to ensuring the stability of closed and TSFs to withstand very high seismic loading conditions, for other
relinquished TSFs as well as to operating facilities. Given current
parts of the world (e.g. Australia and South Africa) it remains
climate variations, TSFs are likely to be subjected to extreme
largely an unlikely, but possible event. The paper therefore walks
weather events beyond what occurs during the operational life of
a line between a fully mechanistic, laboratory-informed approach
a facility. Geotechnical robustness is essential to ensure future
generations are not faced with problems of repairing and to seismic design and analysis, and one that reflects current
remediating inherently inappropriate geotechnical structures. practice in seismically active countries, which includes the
widespread use of empirical approaches, such as charts relating
12 DISCUSSION penetrometer resistance to liquefaction susceptibility. Another
aspect that is flagged is the continuing use of pseudo-static
At present, the focus of many, perhaps most tailings engineers is coefficients in seismic analyses, at least for screening-level
on how to best deal with upstream TSFs containing brittle tailings, evaluations. Retaining such empirical approaches respects the
and rightly so. The spate of recent failures, particularly in Brazil philosophy espoused by Burland (1989) when discussing his
in 2015 and 2019 has highlighted how sudden these failures can concept of the ‘soil mechanics triangle’, at the centre of which is
occur and how wide-reaching the impact can be. Before delving what he refers to as ‘well winnowed experience’.
into a deeper discussion about this specific topic, we reflect on a Operating a TSF in an extremely cold climate, such as that in
few other issues we believe should not be lost sight of when northern Sweden and northern Canada present special challenges,
designing a new TSF, or indeed evaluating the stability of an including embedment of frozen layers during winter operations.
existing TSF. These topics include volume change of soft tailings, While still frozen, these layers become effectively impermeable,
seepage considerations, stability in highly seismic regions, and potentially resulting in preferential horizontal flow of seepage
impacts of climatic conditions on TSF stability. water, with attendant stability implications. If these frozen layers
Many TSFs are now more than 1km2 in area, sometimes many subsequently thaw, they may result in increased pore water
square kilometres. A large percentage of such TSFs is therefore pressures and thus a reduction in shear strength. At the opposite
essentially subject to one-dimensional consolidation, with end of the climatic spectrum, many TSFs are in arid or semi-arid
principal stress directions remaining constant throughout the locations, where evaporation dominates over precipitation. This
growth of such TSFs. Tailings derived from the processing of water imbalance can contribute to strength gain through the
sulphide ore (i.e. hard rock) and composed of silty and/or sandy process of desiccation noted above. However, crust formation due
particles will usually settle rapidly upon deposition and to desiccation, which is exacerbated by use of highly saline
conventional methods of consolidation analysis are adequate. groundwater, results in restricted drying of tailings below the crust
However, for a large range of mine tailings, including mineral layer. This has negative implications on TSF stability as gains in
sands slimes, fine iron ore tailings, oilsands fine tailings (or shear strength that would otherwise occur are inhibited.
mature fine tailings, MFT), and many phosphate tailings, Let us now return to the key issue addressed in this paper, how
deposition takes place at extremely high void ratios and the to address the design of upstream TSFs and how to evaluate the
tailings are highly compressible. Properties such as hydraulic stability of existing TSFs containing potentially liquefiable
conductivity vary over orders of magnitude with increasing tailings.
effective stress, and conventional approaches to calculating The greatest of these two challenges is evaluating the stability
volume change are deficient. The paper demonstrates laboratory of existing TSFs. There is often a paucity of relevant information
testing techniques to characterise these properties and numerical about the operational history of many TSFs, with issues such as
modelling procedures to predict likely field behaviour. Such poor decant pond control, decant pipe breakages, or slope
evaluations are relevant not only to volume calculations but also instabilities often undocumented. Obtaining a reliable evaluation
to stability assessments. There are instances where the volume of existing conditions is thus difficult. We contend that in-situ
decrease renders the tailings dilative (located below the critical testing techniques are the preferred first option, as extracting
state line) and thus inherently less susceptible to liquefaction. As samples for laboratory testing may destroy the very property of
embodied in critical state soil mechanics, volume change and interest, i.e. brittleness. Of the available in-situ testing methods,
shear strength gain cannot be separated. the preferred option currently is the CPTu. This is probably due to
Another, often overlooked aspect of volume change is that the great deal of both empirical and theoretical information to
induced by desiccation. Solar drying, which induces desiccation, facilitate estimation of properties such as undrained strength and
can produce large volume changes, particularly in fine-grained, state. However, practitioners should be aware of the limitations of
compressible tailings, providing large gains in shear strength. these methods some of which we covered under Section 5.1.
Through manipulation of deposition layer thickness desiccation Furthermore, the vast majority of studies using calibration
can be enhanced, providing both improved shear strength and chamber tests to correlate CPTu readings with known, controlled
increased storage of tailings (as the achieved in-situ dry density is state conditions were carried out on clean, quarzitic sands. Many
higher). tailings of interest when it comes to potential brittleness are
Seepage is an issue we suggest is often inadequately predominantly silts or silty sands. Recently, research has begun to
addressed. Given the critical importance of the shape and location focus on this category of tailings using calibration chamber tests,
of the phreatic surface to the stability of upstream TSFs, it is e.g. Ayala et al (2020). Existing limitations in the interpretation of
surprising that many seepage models are based on assumptions of in-situ tests highlight the need of proactively managing a TSF to
175
prevent it from reaching a state where it is susceptible to failure, where they identify the primary trigger for liquefaction as
catastrophic failure. Measurable indicators that can inform this being over-pressurisation of a vertical borehole during
type of proactive management have been described by Boswell installation. This particular trigger mechanism has not previously
and Sobkowicz (2018). been documented, to our knowledge, but required detailed
Evaluations of the potential for static liquefaction at an laboratory testing and numerical modelling.
existing TSF are currently predominantly dependent on CPTu Quite rightly, the need for better monitoring techniques of
testing, backed up with interpretations utilising fundamental TSFs to provide sufficient early warning of developing, or
mechanics, such as critical state constitutive models in numerical impending failure likelihood is currently attracting a lot of
simulations of the cone penetration process. Evaluations of attention and research effort. Unfortunately, there has also been a
potential seismic instability are different. Although evaluation of great deal of shameless opportunism, with some companies
in-situ state is carried out, as with static liquefaction assessments, making outrageous and unproven claims of their proprietary
there is much greater reliance on empirical correlations, an technology’s ability to provide the required early warning. We
example being Figure 81 in this paper. These empirical counsel scepticism when considering implementation of unproven
correlations are understandable, as there have been many technologies to such important structures as TSFs containing
earthquakes in which liquefaction of sands has occurred, as well brittle tailings. We recognise that some of the emerging
as many where liquefaction has not occurred, with cone technologies may ultimately provide the desired accuracy,
penetration data also being available. Drawing correlations using relevance and reliability, but this is not currently the case for many
such data pairs is thus an obvious strategy. However, once again, technologies. Traditional instrumentation, such as vibrating wire
the majority of these correlations are based on earthquake-induced piezometers for pore pressure measurement, and inclinometers or
liquefaction of relatively clean sands. Extrapolation to silty shape acceleration arrays for displacement measurements, should
materials usually requires accounting for any relevant differences be retained (where appropriate). Coupling these conventional
between silts and sands. monitoring approaches with some of the emerging solutions is a
Laboratory testing is of course the only way to characterise sensible approach.
tailings behaviour during the design of a new facility, but can also Although this State of the Art paper is lengthy, there are many
be useful for evaluations of existing facilities, as long as it is clear topics we could still have covered. These include more detailed
what can reasonably be measured. Disturbed samples are widely discussions on high-density thickened tailings and filtered
used to develop the critical state line, but may not provide tailings, as these tailings management strategies are generating
representative information about brittleness (particularly in terms increased interest as owners tackle issues such as risk of failure
of both strength loss and the strain over which strength loss and competing demands for available water. Other emerging
occurs). Undisturbed samples are necessary to determine these solutions include the use of blended, or co-mingled waste
characteristics, and these are notoriously difficult to obtain in silts (combining waste rock with tailings). We suggest that most, if not
and silty sands. all of the approaches to characterisation of tailings discussed in
In characterising brittle (i.e. strain-softening) soils, focus is on this paper are also applicable to thickened tailings, filtered
both the peak strength and the residual, large strain strength, with tailings, or co-mingled waste. An aspect not discussed in our
many designers now regarding the latter as the governing stability paper which will certainly find increasing application is
assessment criterion. This posed a question for the authors of this characterisation of unsaturated properties, particularly for projects
paper, ‘are we (our industry) prepared to do whatever it takes to using filtered tailings for example.
ensure stability’? We have also not specifically addressed the topic of
Many currently operating TSFs, if evaluated using the transitional materials, where the term is used to define soils
residual (or liquefied) shear strength (Sr), would be found to have (tailings) that are characterised by incomplete convergence to a
a factor of safety against slope instability of less than unity. We unique normal compression line and/or critical state line in simple
say this based on our personal experience with reviews of many laboratory tests (Todisco and Coop, 2019). Although Li and Coop
existing facilities. What course of action should be taken in such (2019) found no evidence of transitional behaviour in their tests
cases? Many of the larger mining companies have chosen as first on iron ore tailings from the Panzhihua TSF in China, it is a
resort the construction of a stabilising buttress on the downstream consideration we believe warrants further investigation for a range
slope of the existing TSF, with some going even further and of tailings. Given the current reliance on evaluations of the critical
constructing large earthen ‘deflector’ walls, to cater for a potential state line for tailings, any dependence of the CSL on initial void
flowslide failure. However, is this necessarily an appropriate ratio must be identified and quantified.
blanket solution (buttress construction), or should we adopt a Finally, we note the extreme importance of disseminating
more nuanced approach for materials that require very large knowledge to those operating TSFs in resource-poor countries,
strains to fully mobilise Sr (e.g. material B in Figure 74)? Clearly where we use ‘resource’ in this context to denote access to
an answer to this question depends on factors that include cost, information, availability of suitable laboratories and skilled,
potential consequences of a failure, the risk appetite of different competent contractors (e.g. for conducing CPTu tests). The
companies and whether a credible failure mode for a particular performance record of TSFs leaves no doubt that significant
TSF truly exists. failures can occur in well-established mining jurisdictions in
Although case studies of the various TSF failures that have which access to competent engineering services is available (e.g.
been well-documented provide invaluable lessons to all tailings Australia, Brazil and Canada). Clearly the occurrence of high-
engineers, relying on a reactive approach such as this is not profile accidents in resource-poor countries is also a distinct
sufficient. There is a clear need for trained geotechnical engineers possibility. Preventing a TSF failure in any country should be a
who understand the subtleties of mine tailings, and have a sound concern of the entire mining industry. As evidenced by the
fundamental understanding of geotechnical engineering international reaction to the Feijão failure, everyone becomes
principles. Ideally, this would equip the next generation of tailings tarred with the same brush when such a failure occurs.
engineers with the ability to identify possible failure mechanisms,
or triggers for a failure, based on an understanding of geotechnical 13 CONCLUDING COMMENTS
engineering rather than just by comparison against a check-list of
previously documented failures. An excellent example of the In the discussion section preceding these concluding comments
former is the report by CIMNE into the Brumadinho (Feijão) TSF we reviewed the main points of our paper and discussed these in
a little more detail. Here we offer three overriding concluding
176
comments. They are based on the numerous discussions held used empirical correlations are based, a cautionary approach is
between the five co-authors over a period of many months. appropriate. We suggest that empirical correlations are a useful
We believe that, in general, the key underlying principles that starting point for tailings characterisation, but should be
govern the behaviour of mine tailings (geotechnically speaking; confirmed with a comprehensive suite of testing, including both
geochemical considerations are beyond our scope for this paper) laboratory and in-situ tests wherever possible.
are well understood. The concerns generated by failures such as
Feijão, where spontaneous, or static liquefaction occurred with 14 ACKNOWLEDGEMENTS
such sudden and devastating consequences, caused consternation
for many engineers, who perhaps wondered if there was an The authors would like to thank Mr Edi Nuryatno for assistance
underlying principle that had until now remained unrecognised. with drafting many of the figures and for assistance with final
The relatively poor appreciation, at least until recently, of static formatting of the paper, Mr Christopher O’Donovan for assistance
liquefaction may be due to the training at most universities of the with obtaining copyright clearances and for authors who kindly
principles of soil mechanics, which often endorse the assumption granted us copyright clearance to reproduce figures of theirs,
that ‘loading is drained in sands and undrained in clays’. With being Antonio Gens, Ross Boulanger, Gonzalo Castro and
many tailings being sand-like in that they consist of nonplastic Michael Etezad. The authors also gratefully acknowledge many
silty mixtures, it may have been tempting to assume that they of our students and colleagues who have contributed through
always exhibit drained behaviour. However, if we pause to numerous discussions over the years to the development of ideas
consider the definition of drained loading, which is loading that is presented in the paper.
applied at a rate slow enough that excess pore pressures dissipate
quicker than they are generated, it becomes clear that sands can 15 REFERENCES
fail in an undrained fashion, which is the essence of static
liquefaction. A key to understanding static liquefaction is that Abu-Hejleh, A.N. and Znidarcic, D. 1994. Estimation of the Consolidation
shearing of contractive soils induces positive excess pore water Constitutive Relations, Computer Methods and Advances in
pressures. Once such shear-induced pore pressures are triggered, Geomechanics, Siriwardane & Zaman (eds) Balkema, Rotterdam,
they may propagate faster than even the most permeable sand is 499-504.
capable of dissipating. Abu-Hejleh, A. N. and Znidarcic, D. 1995. Desiccation Theory for Soft
Our second point relates to the first. We contend that the Cohesive Soils, Journal of Geotechnical Engineering, ASCE, New-
training of many tailings engineers is inadequate. Not recognising York, Vol. 121(6), 493-502.
the coupling between volume change and shear strength is a Abu-Hejleh, A. N., Znidarcic, D., and Barnes, B.L. 1996. Consolidation
deficiency of many undergraduate programmes. Although it is Characteristics of Phosphatic Clays, Journal of Geotechnical
certainly not faultless, critical state soil mechanics provides an Engineering, ASCE, New-York, 122(4), 295-301.
elegant connection between these two aspects of soil behaviour. Agüero, G. 1929. Formation of Tailings Deposits in El Teniente Mine.
Inclusion of critical state soil mechanics, non-linear finite strain Anales del Instituto de Ingenieros de Chile. 164-187. (in Spanish).
consolidation theory, and a better appreciation of the strengths and Altuhafi, F., O’Sullivan, C. and Cavarretta, I. 2013. Analysis of an image-
limitations of widely used laboratory and in-situ tests would based method to quantify the size and shape of sand particles. J.
further aid in improving the quality of tailings engineering Geotech. Geoenviron. Eng., 139(8), 1290-1307.
practice. Likewise, seepage analyses for stability calculations Atkinson, J.H., Lau, W.H.W., and Powell, J.J.M. 1991. Measurement of
require far greater attention to the details of anisotropy, stress and soil strength in simple shear tests. Canadian Geotechnical Journal,
saturation dependent hydraulic conductivity in both formal 28, 255-262.
education and engineering applications. While routine seepage Ayala, J., Fourie, A.B. and Reid, D. 2020. Cone penetration testing on silty
models might be adequate for water balance calculations, often tailings using a new small calibration chamber. Geotechnique Letters,
they fail to identify critical issues that affect stability of TSFs. 10, 492-497.
Needless to say that if it were not for water in tailings, we would Azam, S. and Li, Q. 2010. Tailings dam failures: a review of the last one
not have stability problems. Given the rapid uptake of certain hundred years. Geotechnical News, December 2010, 50-53.
numerical modelling packages, many of which include highly Basson, J.A., Broekman A., and Jacobsz S.W. 2021. TD-DAQ: A low-
complex constitutive laws, we are concerned that a lack of cost data acquisition system monitoring the unsaturated pore pressure
understanding of fundamental soil mechanics principles will regime in tailings dams. HardwareX.
result in misleading, potentially dangerous interpretations of the https://doi.org/10.1016/j.ohx.2021.e00221
results generated by such packages. Bauer, E. 1996. Calibration of a Comprehensive Hypoplastic Model for
Our third, and final point is that there is still a great deal of Granular Materials. Soils and Foundations. 36(1): 13-26.
work to be done, techniques and theories to be developed and Beckett, C., Fourie, A.B. and O’Loughlin, C.D. 2016. Centrifuge
refined. Use of in-situ tests, particularly the cone penetrometer, to modelling of seepage through tailings embankments. International
tailings applications has greatly advanced the understanding of Journal of Physical Modelling in Geotechnics, 16(1), 18-30.
such tests, which has a benefit to all geotechnical engineering Bedell, D., Fawell, P., Slottee, S. and Schoenbrunn, F. 2015. Thickening.
practitioners. Issues of partial drainage during penetration, and the In: Paste and Thickened Tailings – A Guide, Third Edition. Ed. R.J.
challenge of interpreting cone test results when carried out in Jewell and A.B. Fourie, Australian Centre for Geomechanics.
partially saturated media are everyday experiences when working Bedin, J., Schnaid, F., da Fonseca, A.V., and de Costa Filho, L.M. (2012).
on TSFs. Similarly, use of conventional consolidation theory may Gold tailings liquefaction under critical state soil mechanics.
often be inappropriate and misleading, particularly when working Géotechnique, 62(3): 263–267. doi:10.1680/geot.10.P.037.
with soft, highly compressible mine tailings. Volume change Been, K. 2016. Characterising mine tailings for geotechnical design.
calculations, and thus estimates of storage capacity of a TSF, may Proceedings Geotechnical and Geophysical Site Characterisation 5,
be grossly in error if the consolidation behaviour of such tailings Sydney, Australia, Australian Geomechanics Society.
is not correctly characterised. We strongly caution against the Been, K., Jefferies, M.G. and Hachey, J.E. 1991. The critical state of
extrapolation of information on material behaviour from one site sands. Géotechnique, 41(3), 365–381.
to another without an evaluation of the validity of such Been, K. and Jefferies, M.G. 1985. A state parameter for sands.
extrapolation, preferably based on appropriate laboratory and/or Géotechnique, 35(2), 99–112.
in-situ testing. As discussed in our paper, tailings particle shape Been, K., Jefferies, M.G., Crooks, J.H.A. and Rothenberg, L. 1987. The
can be highly angular. Coupled with the fact that many tailings cone penetration test in sands: Part II, general inference of state.
have a wide distribution of particle sizes, unlike the clean, Géotechnique 37(3), 285-299.
relatively well-rounded sands on which many of our currently
177
Been, K., and Olivera, R. 2015. Appendix B: Laboratory testing to Carrera, A., Coop, M., and Lancellotta, R. 2011. Influence of grading on
determine the critical state of sands. In Soil Liquefaction: A critical the mechanical behaviour of Stava tailings. Géotechnique, 61(11):
state approach. CRC Press. 935–946.
Beier, N. and Sego, D. 2008. Cyclic freeze-thaw to enhance the stability Casagrande, A. 1936. Characteristics of Cohesionless Soils Affecting the
of coal tailings. Cold Regions Science and Technology. Elsevier 2008- Stability of Slopes and Earth Fills. Journal of the Boston Society of
08-16. Civil Engineering, January, pp. 13-32.
Beveridge, A., Mutz, P., and Reid, D. 2015. Tailings co-disposal case Casagrande, A. 1970. On Liquefaction Phenomena," Lecture, reported by
study – art or science? In Proceedings of the 18th International Green and Ferguson, Geotechnique, Vol. 21, No.3, (1971), pp. 197-
Seminar on Paste and Thickened Tailings, Cairns, Australia, 5–7 May 202.
2015. Edited by R.J. Jewell and A.B. Fourie. Australian Centre for Casagrande, A. 1975. Liquefaction and Cyclic Deformation of Sands - A
Geomechanics, Perth, Australia, 505–520. critical review," Fifth Panamerican Conference on Soil Mechanics
Bhanbro, R. 2017. Mechanical behaviour of tailings – Laboratory Tests and Foundation Engineering, Buenos Aires, Argentina.
from a Swedish Tailings dam. PhD thesis, Luleå University of Castro, G. 1969. Liquefaction of sands, PhD thesis, Harvard University,
Technology. USA.
Biarez, J., and Hicher, P.Y. 1994. Elementary mechanics of soil Castro, G. and Troncoso, J. 1989. Effects of Chilean Earthquake on three
behaviour: saturated remoulded soils. A. A. Balkema. tailing dams”. International Seminar on Dynamic Behavior of Clays,
Bicalho, K.V. 1999. Modeling Water Flow in an Unsaturated Compacted Sands & Gravel. Kitakyushu, Japan.
Soil, Ph.D. Thesis, University of Colorado Boulder. Chakraborty et al 2020. In Situ Investigation for Geotechnical
Bicalho, K.V., Znidarcic, D. and Ko, H.Y. 2011. One-dimensional flow Characterization of Consolidated Dry Mine Tailing Slurry Backfill.
infiltration through a compacted fine grained soil, Soils and In Geotechnical Characterization and Modelling.
Foundations, 51(2), 287-295. Chandler, R.J. 1988. The in-situ measurement of the undrained shear
Bishop, A.W. 1967. Progressive failure -- with special reference to the strength of clays using the field vane. Vane shear testing in soils: field
mechanism causing it. Proc. Geotech. Con, Oslo, 2: 142-150. and laboratory studies, ASTM STP 1014, A.F. Richards, Ed., ASTM,
Bishop, A. W., Green, G. E., Garga, V. K., Andresen, A. & Brown, J. D. Philadelphia, 1988, 13-44.
1971. A new ring shear apparatus and its application to the Chen, Z., Guo, X., Shao, L., Shunqun, L. and Tian, X. 2021. Design of a
measurement of residual strength. Geotechnique 21(4), 273-328. three-dimensional earth pressure device and its application in a
Bjerrum, L. 1969. Effect of rate of strain on undrained shear strength of tailings dam construction simulation experiment. Acta Geotechnica,
soft clays. Second contribution to panel discussion, Main Session 5, 16, 2203-2216.
7th International Conf. on Soil Mechanics and Found. Eng. Mexico. Cho, G-C, Dodds, J. and Santamarina, J.C. 2006. Particle shape effects on
Blight, G.E. 1962. A note on field testing of silty soils. Canadian packing density, stiffness and strength: natural and crushed sands. J.
Geotechnical Journal, 3, 142-149. of Geotechnical and Geoenv. Eng, 132(5), 591-602.
Blight, G.E., Boswell, J.E.S. and Zenon, A. 1995. Underwater Chow S., Alonso-Marroquin F. and Airey D. 2012. Viscous Rate Effects
construction of an embankment to extend the life of a tailings in Shear Strength of Clay. Proc. 11th Australia New Zealand
impoundment. Proc. Instn. Civ. Engrs. Geotech. Engng. 113, 80-85. Conference on Geomechanics.
Blight, G.E. 1997. The master profile for hydraulic fill tailings beaches. CIMNE 2021. Computational analyses of Dam I failure at the Corrego de
Proc. Instn. Civ. Engrs, Geotechnical Engineering, 107, 27-40. Feijao mine in Brumadinho, Final Report, CIMNE, August 2021.
Boswell J. and Sobkowicz J. 2018. Leading versus lagging indicators of Clayton CRI and Bica AVD 1993. The design of diaphragm-type
tailings dam integrity. In Proceedings of the 22nd International boundary total stress cells. Géotechnique, 43(4), 523-535.
Conference on Tailings and Mine Waste. Tailings and Mine Waste Collins, B.D., 1997. Landslide Triggering Mechanism, M.S. Thesis,
'18. Keystone, Colorado, USA, September 30–October 2, 2018. University of Colorado Boulder
Boulanger, R. W. 2002. Evaluating Liquefaction Resistance at High Collins, B.D., and Znidarcic, D., 2004, Stability Analyses of Rainfall
Overburden Stresses. Proc. 3rd U.S.-Japan Workshop on Advances Induced Landslides, Journal of Geotechnical and Geoenvironmental
Research on Earthquake Engineering for Dams, San Diego, CA., 1- Engineering, ASCE, New-York, 130(4), 362-372.
12. Conlin, B. 1987. A review of the performance of mine tailings
Boulanger, R. and Idriss, I. 2006. Liquefaction Susceptibility Criteria for impoundments under earthquake loading conditions. Proceedings
Silts and Clays. J. of Geotechnical and Geoenvironmental Eng., Vancouver Geotechnical Society Seminar on Earthquake. Geot.
132(11). Canada.
Boulanger, R. and Idriss, I. 2014. CPT and SPT based liquefaction Cubrinovski, M., and Ishihara, K. 2002. Maximum and minimum void
triggering procedures. Report No. UCD/CGM-14/01. University of ratio characteristics of sands. Soils and Foundations 42(6):65-78.
California at Davis. Cunning, J.C. 1994. Shear wave velocity measurements of cohesionless
Bray, J., Sancio, R., Durgunoglu, T., Onalp, A., Youd, L. Stewart, J., Seed, soils for evaluation of in situ state. MSc thesis. University of Alberta.
R., Cetin, O., Bol, E., Batuary, M., Christensen, C. and Karadayilar, Edmonton, Canada.
T. 2004. Subsurface characterization at ground failure sites in da Fonseca A.V., Cordeiro D., and Molina-Gómez F. 2021.
Adapazari, Turkey. Journal of Geotechnical and Geoenvironmental Recommended procedures to assess critical state locus from triaxial
Engineering. 130(7), 673–685. tests in cohesionless remoulded samples. Geotechnics. 1: 95-127.
Bray, J.D., and R.B. Sancio. 2006. Assessment of the liquefaction Daliri, F., Kim, H., Simms, P., and Sivathayalan, S., 2014. Impact of
susceptibility of fine grained soils. Journal of Geotechnical and Desiccation on Monotonic and Cyclic Shear Strength of Thickened
Geoenvironmental Engineering, 132(9), 1165–1177. Gold Tailings. Journal of Geotechnical and Geoenvironmental
Bray J, and Travasarou T. 2009. Pseudostatic coefficient for use in Engineering, 140(9).
simplified seismic slope stability evaluation. Journal of Geotechnical Debasis R., Campanella R.G., Byrne P.M., and Hughes J. 2002. Undrained
and Geoenvironmental Engineering, 135(9). anisotropic monotonic behavior of sand from in situ tests. J. Geotech.
Bromwell, L.G., and Oxford, T.P. 1977. Waste clay dewatering and Geoenviron. Eng. ASCE.
disposal. In: Proc. ASCE Specialty Conf. on Geotechnical Practice DeJong, J.T. and Randolph, M. 2012. Influence of partial consolidation
for Disposal of Solid Waste Materials, Ann Arbor, MI during cone penetration on estimated soil behavior type and pore
Budhu, M. 1984. Nonuniformities imposed by simple-shear apparatus. pressure dissipation measurements. J. Geotech. Geoenviron. Eng.,
Geotechnique, 20, 125-137. 138(7): 777-788.
Burland, J. B. 1989. Nash lecture: The teaching of soil mechanics – A Desrues J, Chambon R, Mokni M, and Mazerolle F. 1996. Void ratio
personal view. Proceedings of the ninth European Conference on Soil evolution inside shear bands in triaxial sand specimens studied by
Mechanics and Foundation Engineering, Dublin, 1987, Vol.3, pp computed tomography. Géotechnique 46(3):529-546.
1427-1447.
178
Díaz-Rodríguez, J., Martínez-Vasquez J. and Santamarina J. 2009. Strain- Methods, Tailings and Mine Waste ‘20, November 15-18, UBC
Rate Effects in Mexico City Soil. Journal of Geotechnical and Studios, University of British Columbia, Vancouver BC, 413-424.
Geoenvironmental Engineering, 135(2). Gjerapic, G., Znidarcic, D., Peric, D., and Murphy, F.D., 2021,
Dobry, R. and Alvarez, L. 1967. Seismic Failures of Chilean Tailings Development of an engineering model for consolidation and creep in
Dams. Journal of Soil Mechanics and Foundation Division, ASCE, tailings, Tailings and Mine Waste ’21, November 7-10, Banff, AB,
93(6), 237-260. Canada, ISBN: 978-1- 55195-466-0, 841-851.
Dounias, G.T. and Potts, D.M. 1993. Numerical analysis of drained direct Gomes, R.C., Ribeiro, L.F.M., Albuquerque Filho, L.H. and Rezende,
and simple shear tests. Journal of Geotechnical Engineering, 119(12), C.H. 2003. Geotechnical Aspects of Eroded Sediments Relevant to
1870-1891. Tailings Disposal Design, Natural Hazards Review, 4(2), 65-70.
Dyvik, R., Berre, T., Lacasse, S. and Raadim, B. 1987. Comparison of Graham, J. 1992. The key material properties in geotechnical engineering
truly undrained and constant volume direct simple shear tests. need to be measured instead of being specified. In Predictive Soil
Geotechnique, 37(1), 3-10. Mechanics.
Farrar, J.A., Torres, R., and Crutchfiedl, L.G. 2008. Cone penetrometer Graham, J. and Jefferies, M. 1986. Some examples of in situ lateral stress
testing, Scoggins Dam, Tualatin Project, Oregon. Engineering determination in hydraulic fills using the self-boring pressuremeter.
Geology Group Bureau of Reclamation, Technical Services Center, Hazen, A. 1920. Hydraulic-fill dams, ASCE Transactions, 83, 1713–1745.
Denver, Colo. Report No. 86–68320. Hoffman, A. 2015. Mount Polley Mine, tailings storage facility breach.
Fear, C.E., and Robertson, P.K. 1995. Estimating the undrained strength Investigation Report of the Chief Inspector of Mines. Ministry of
of sand: a theoretical framework. Canadian Geotechnical Journal, Energy and Mines.
32(5), 859-870. Holden, D., Donegan, S. and Pon, A. 2020. Brumadinho Dam InSAR
Finnie, I., and Randolph, M.F. 1994. Punch-through and liquefaction study: analysis of TerraSAR-X, COSMO-SkyMed and Sentinel-1
induced failure of shallow foundations on calcareous sediments. Proc. images preceding the collapse. Slope Stability 2020, Australian
Int. Conf. on Behavior of Offshore Structures, Boston (USA): 217- Centre for Geomechanics, Perth, Australia.
230. Hosono, Y. and Yoshimine, M. 2008. Effects of anisotropic consolidation
Fonseca do Carmo, Fl., Kamino, L. H. Y., Tobias Junior, R., de Campos, and initial shear load on liquefaction resistance of sand in simple shear
I., Fonseca do Carmo, Fe., Silvino, G., Xavier de Castro, K. J. d. S., condition. Geotechnical Engineering for Disaster Mitigation and
Mauro, M. L., Rodrigues, N. U. A., Miranda, M. P. d. S., Pinto, C. E. Rehabilitation. Liu, Deng and Chu (eds).
F. (2017). Fundão tailings dam failures: the environment tragedy of Houman, J., Paterson, A. and van Sittert, F. 2007. Quantifying the Effect
the largest technological disaster of Brazilian mining in global of Pump Shear on Slurry Rheology — A Case Study at the Combined
context. Perspectives in Ecology and Conservation, 15, 145-151. Treatment Plant. Proc. 10th International Seminar on Paste &
Fourie, A.B., and Papageorgiou, G. 2001. Defining an appropriate steady Thickened Tailings, March 2007, Perth, Australia. Ed. R. Jewell and
state line for Merriespruit gold tailings. Canadian Geotechnical A.B Fourie, Australian Centre for Geomechanics.
Journal, 38(4):695–706. Housner, G.W. 1984. Keynote lecture, 8th World Conf. Earthq. Eng., San
Fourie, A.B., Blight, G.E. and Papageorgiou, G. 2001. Static liquefaction Francisco, Post Conference. Volume 1, pp. 25–39.
as a possible explanation for the Merriespruit tailings dam failure. Hwang, C., 1999, Transient Seepage in Embankment Dams, M.S. Thesis,
Canadian Geotechnical Journal. 38(4), 707-719. University of Colorado Boulder.
Fourie, A.B. and Tshabalala, L. 2005. Initiation of static liquefaction and Hwang, C., 2002, Determination of Material Functions for Unsaturated
the role of K0 consolidation. Canadian Geotechnical Journal, 42(3), Flow, Ph.D. Thesis, University of Colorado Boulder.
pp 892-906. Hynes, M. E., Olsen, R. S. and Yule, D. E. 1998. The Influence of
Fourie, A., Reid, D., Ayala, J., Russell, A., Vo,T., Rahman, M. and Vinod, Confining Stress on Liquefaction Resistance. Proc. of the 30th Join
J. 2021. Improvements in estimating strengths of loose tailings: Meeting of the U.S.-Japan Cooperative Program in Natural
results from the TAILLIQ research project. Proc. Mine Waste and Resources Panel on Wind and Seismic Effects, NIST SP 931, 168-
Tailings, AusIMM, Brisbane, Australia, 207-217. 184.
Franks, D.M., Stringer, M., Torres-Cruz, L.A., Baker, E., Valenta, R., Hynes, M. E. and Olsen, R. S. 1999. Influence of Confining Stress on
Thygesen, K., Matthews, A., Howchin, J. and Barrie, S., 2021. Liquefaction Resistance. Proc. International Workshop on Physics
Tailings facility disclosures reveal stability risks. Scientific reports, and Mechanics of Soil Liquefaction, Rotterdam, The Netherlands,
11(1), 1-7. 145-152.
Fujiyasu, Y.1997). Evaporation behaviour of tailings. PhD thesis, Idriss and Boulanger, 2003. Estimating K for use in evaluating cyclic
University of Western Australia, Perth, Australia. resistance of sloping ground. In: Hamada, O’Rourke, Bardet (Eds.),
Gawu, S.K.Y. and Fourie, A.B. 2004. Assessment of the modified slump Proc. of the 8th US-Japan Workshop on Earthquake Resistant Design
test as a measure of the yield stress of high-density thickened tailings. of Lifeline Facilities and Countermeasures against Liquefaction,
Canadian Geotechnical Journal, 41(1), 39-47. Report MCEER-03-0003, MCEER, SUNY Buffalo, NY, pp. 449–
Gens, A. 2019. Hydraulic fills with special focus on liquefaction. Keynote 468.
Lecture. Proceedings of the XVII ECSMGE, Reykjavik, 1-31. Idriss I. and Boulanger R. 2004. Semi-empirical Procedures for Evaluating
Ghafghazi and Shuttle 2008. Evaluation of soil state from SBP and CPT: Liquefaction Potential During Earthquakes. Proc. 3rd International
A case history. Canadian Geotechnical Journal, 45(6). Conference on Earthquake Geotechnical Engineering. Berkeley,
Gibson, R.E., England, G.L., and Hussey, M.J.L. 1967. The theory of one- USA.
dimensional consolidation of saturated clays. 1. Finite Non-linear Imai, G., 1979, Development of a new consolidation test procedure using
Consolidation of Thin Homogeneous Layers. Géotechnique, 17, 261- seepage force, Soils and Foundations, Tokyo, Japan, 19(3), 45-60.
273. Ishihara, K. 1984. Post-Earthquake Failure of a Tailings Dam Due to
GISTM Global industry standard on tailings management. 2020. ICMM Liquefaction of the Pond Deposit. Int. Conference on Case Histories
(International Council on Mining & Metals), UNEP (United Nations in Geotechnical Engineering. St. Louis. USA.
Environment Programme) and PRI (Principles for Responsible Ishihara, K. 1993. 33rd Rankine Lecture: Liquefaction and flow failure
Investment). during earthquakes. Geotéchnique, Vol. 43, No. 3, pp. 351—415.
Gjerapic, G., Johnson, J., Coffin, J., and Znidarcic, D., 2008, Ishihara, K. 1996. Soil Behaviour in Earthquake Geotechnics. Clarendon
Determination of Tailings Impoundment Capacity via Finite-Strain Press, Oxford. ISBN 0-19-856224-1.
Consolidation Models, Proceedings of the Geo Institute Congress, Ishihara, K. 1985. Stability of natural deposits during earthquakes. Theme
New Orleans, 2008. 9-12 March 2008, New Orleans, Louisiana, CD- Lecture, Proceeding of the XI ICSMFE, Vol.2, 321-376.
ROM Ishihara K., Ueno K., Yamada S., Yasuda S. and Yoneoka, T. 2014.
Gjerapic, G., and Znidarcic, D., 2020, Assessment of Liquefaction Breach of a Tailing Dam in the 2011 Earthquake in Japan. 34th United
Triggering for Upstream Tailings Dams Using Limiting Equilibrium
179
States Society on Dams Annual Meeting & Conference Seismic Li, X. S. and Wang, Y. 1998. Linear representation of steady-state line for
Design of Tailings Dams. San Francisco, USA. sand. J. Geotech. Geoenviron. Engng., 124(12), 1215–1217.
Islam, K. and Murakami, S., 2021. Global-scale impact analysis of mine Lipinski and Wdowska 2018. Hybrid approach for evaluation of tailings
tailings dam failures: 1915–2020. Global Environmental Change, 70, state on the basis of shear wave velocity measurement. In Proceedings
102361. of SGEM Conference 2018.
Jacobsz SW 2019. TUKS tensiometer measures to -1.7 MPa. Civil Liu, J.C., 1990, Determination of Soft Soil Characteristics, Ph.D. Thesis,
Engineering Magazine – SAICE 27(1). University of Colorado Boulder.
James M., Aubertin M., Wijewickreme D. and Wilson G. 2011. A Liu, J.C., and Znidarcic, D., 1991. Modeling One-Dimensional
laboratory investigation of the dynamic properties of tailings. Can. Compression Characteristics of Soils. Journal of Geotechnical
Geotech. J. 48: 1587–1600. Engineering, ASCE, 117, No. 1, pp. 164-171.
Jamiolkowski and Masella 2015. "Geotechnical Characterization of Liu, Q.B. and Lehane, B.M. 2012. The influence of particle shape on the
Copper Tailings at Zelazny Most Site." Third International (centrifuge) cone penetration test (CPT) end resistance in uniformly
Conference on the Flat Dilatometer. graded granular soils. Geotechnique, 62(11), 973-984.
Jeeravipoolvarn, S., Chalaturnyk, R.J. and Scott, J.D. 2008. Consolidation Lunne, T., 2007. Can we measure remoulded strength or sensitivity with
modeling of oil sands fine tailings: history matching. Proc. CPTU sleeve friction and friction ratio? Norwegian Geotechnical
GeoEdmonton ’08, Edmonton, Canada, 190 - 197. Institute. Report No.: 20061023-1.
Jeeravipoolvarn, S., Scott, J.D., and Chalaturnyk, R.J. 2009. Geotechnical Lunne T, Robertson PK, and Powell JJM 1997. Cone penetration testing
characteristics of laboratory in-line thickened oil sands tailings. In in geotechnical practice. CRC Press.
Proceedings of Tailings and Mine Waste 2009, Banff, Alta., 1–4 Lupini, J. F., Skinner, A. E. and Vaughan, P. R. 1981. The drained residual
November 2009. Information Technology, Creative Media: strength of cohesive soils. 31, No. 2, 181-213.
Vancouver, B.C., 813–828. Lyu, Z., Chai, J., Xu, Z., Qin, Y. and Cao, J. 2019. A comprehensive
Jefferies M., and Been K. 2015. Soil Liquefaction: A critical state review on reasons for tailings dam failures based on case history.
approach. Second Edition. CRC Press. Advances in Civil Engineering, 2019, 1-18.
Jefferies, M., Morgenstern, N., Van Zyl, D. and Wates, J. 2019. Report on Mahmoodzadeh, H. and Randolph, M.F. 2014. Penetrometer testing:
NTSF Embankment Failure Cadia Valley Operations for Ashurst effect of partial consolidation on subsequent dissipation response. J.
Australia. Geotech. Geoenviron. Eng., 2014, 140(6): 04014022.
Jewell, R.J. and Fourie, A.B., Paste and Thickened Tailings – A Guide, Mair R.J. and Wood D.M. 1987. Pressuremeter testing: methods and
2nd edition, Australian Centre for Geomechanics, Perth, 2006. interpretation. CIRIA.
Knutsson, R., 2018. On the Behaviour of Tailings Dams – Management in Marchetti S. 2015. Some 2015 updates to the TC16 DMT report 2001.
Cold Regions. Doctoral Thesis, Luleå University of Technology, Third International Conference on the Flat Dilatometer.
Sweden. Maureira, S. and Verdugo, R. 2012. Cyclic response of tailings sands
Kokusho, T. 1980. Cyclic triaxial test of dynamic soil properties for wide tested at high pressure. VII Chilean Geotechnical Congress,
strain range. Soils and Foundations, 20, 45-60. Concepcion (in Spanish).
Kramer, C., and Arango, I. 1998. Aging effects on the liquefaction Miller, G.A., Tan, N.K., Collins, R.W. and Muraleetharan, K.K.2018).
resistance of sand deposits: A review and update. Proc., Eleventh Cone penetration testing in unsaturated soils. Transportation
European Conf. on Earthquake Engineering, Paris. Geotechnics, 17, 85 – 99.
Kynch, E.J., 1952, A Theory of Sedimentation, Transactions of the Mitchell, J. K., and Solymar, Z. V. 1984. Time-dependent strength gain in
Faraday Society, 48, pp. 168-176. freshly deposited or densified sand. Journal of Geotech. Eng., 11011,
Ladd, R. 1978. Preparing test specimens using undercompaction. 1559–1576.
Geotechnical Testing Journal, 1(1), 16–23. Mohajeri, M. and Ghafghazi, M. 2012. Ground sampling and laboratory
Ladd, C.C. 1991. Stability evaluation during staged construction. Jnl. testing on low plasticity clays. Proc. 15th World Conference on
Geotechnical Engineering, 117(4), 540 – 615. Earthquake Engineering, Lisbon, Portugal.
Lade, P. 2005. Overview of constitutive models for soils. ASCE Montgomery, J., Boulanger, R. and Harder, I. 2012. Examination of the
Geotechnical Special Publication No. 128. Center for
Lee, J., 2011, Limits to Continuity of Water Flow in Unsaturated Soils, Geotechnical Modeling, University of California, Davis, California.
Ph.D. Thesis, University of Colorado Boulder. Report No.UCD/CGM-12/02.
Leon, E., Gassman, S. and Talwani, P. 2006. Accounting for Soil Aging Monte, J.L., and Krizek, R.J., 1976, One-dimensional mathematical model
When Assessing Liquefaction Potential. Journal of Geotechnical and for large-strain consolidation, Geotechnique, 26(4), 495-510.
Geoenvironmental Engineering, 132(3). Morgenstern N.R., Vick S.G., Van Zyl D. 2015. Report on Mount Polley
Leroueil, S., and Hight, D.W. 2003. Behaviour and properties of natural Tailings Storage Facility Breach. Province of British Columbia.
soils and soft rocks. In Proc. of the 2nd International Workshop on Morgenstern, N.R. 2018. Geotechnical risk, regulation and public policy.
Characterisation and Engineering Properties of Natural Soils, Soils and Rocks, 41(2), 107 – 129.
Singapore, 2–4 December 2002. Ed. by K.K. Phoon, D.W. Hight, S. Morgenstern, N., Vick, S., Viotti, C. and Watts, B. 2016. Fundão Tailings
Heroueil, and T.S. Tan. A.A. Balkema, Rotterdam, the Netherlands. Dam Review Panel, Report on the Immediate Causes of the Failure of
Vol. 1, 29–254 the Fundão Dam.
Leroueil, S. and Vaughan, P.R. 1990. The general and congruent effects Morgenstern, N. and Price, V. 1965. The Analysis of the Stability of
of structure in natural soils and weak rocks. Geotechnique, 40(3), 467- General Slip Surfaces. Geotechnique, 15(1),
488. Mundle, C., Chapman, P., Williams, D., and Fredlund, D. 2012. The
Li, W. 2017. The mechanical behaviour of tailings [online]. Ph.D. thesis, impact of high density liquors on standard soil mechanics
Department of Architecture and Civil Engineering, City University calculations. Proc. 16th International Conference on Tailings and
of Hong Kong. Available from Mine Waste. Information Technology, Creative Media, Keystone,
https://scholars.cityu.edu.hk/en/theses/the-mechanicalbehaviour-of- Colorado. 281–290.
tailings(d3d2338a-6cf2-4e19-993b-a3f67dad274a).html. Murphy, F.D., Gjerapic, G., Znidarcic, D., and Sorta, A., 2021, Validation
Li, W., and Coop, M.R. 2019. The mechanical behaviour of Panzhihua of an engineering model for consolidation and creep in oil sand
iron tailings. Canadian Geotechnical Journal, 56(3): 420–435. tailings, Tailings and Mine Waste ‘21, November 7-10, Banff, AB,
Li, W., Coop, M.R., Senetakis, K., and Schnaid, F. 2018. The mechanics Canada, 833-840.
of a silt-sized gold tailing. Engineering Geology, 241: 97–108. Murthy, T.G., Loukidis, D., Carraro, J.A.H., Prezzi, M., and Salgado, R.
Li, G., Liu, Y.-J., Dano, C., and Hicher, P.-Y. 2015. Grading-dependent 2007. Undrained monotonic response of clean and silty sands.
behavior of granular materials: from discrete to continuous modeling. Géotechnique, 57(3): 273–288.
Journal of Engineering Mechanics, 141(6): 04014172.
180
Nagaraj, T.S., Pandian, N.S. and Narashima, R. 1998. Compressibility Quiros, G.W. and Young, A.G., 1988. Comparison of field vane, CPT, and
behaviour of soft cemented soils. Geotechnique, 48(2), 281-287. laboratory strength data at Santa Barbara Channel site. ASTM
Nakase, A. and Kamei, T. 1986. Influence of strain rate on undrained shear International.
characteristics of Ko-consolidated cohesive soils. Soils and Randolph, M. F., and Andersen, K. H. 2006. Numerical analysis of T-bar
Foundations, 26, 85-95. penetration in soft clay. Int. J. Geomech., 6(6), 411–420.61
National Academies of Sciences, Engineering, and Medicine. 2016. State Rees SD. 2010. Effects of fines on the undrained behaviour of
of the Art and Practice in the Assessment of Earthquake-Induced Soil Christchurch sandy soils. PhD Thesis. Department of Civil and
Liquefaction and Its Consequences. Washington, DC: The National Natural Resources Engineering. University of Canterbury. New
Academies Press. doi: 1017226/23474. Zealand.
Newson, T.A. and Fahey, M. 1998. Saline tailings disposal and Reid, D., 2015. Estimating slope of critical state line from cone penetration
decommissioning. MERIWA Project No. M241, Australian Centre for test—an update. Canadian Geotechnical Journal, 52(1), 46-57.
Geomechanics. Reid, D. and Fanni, R., 2022. A comparison of intact and reconstituted
Newson, T., Dyer, T., Adam, C. and Sharp, S. 2006. Effect of structure on samples of a silt tailings. Géotechnique, 72(2), 176-188.
the geotechnical properties of bauxite residue. Journal of Reid, D. and Fourie, A.B. 2017. Laboratory assessment of the effects of
Geotechnical and Geoenvironmentral Engineering, 132(2), 143-151. polymer treatment on geotechnical properties of low-plasticity soil
Nguyen, H.B.K., Rahman, M.M. and Fourie, A.B.2018). Characteristic slurry. Canadian Geotechnical Journal, 53: 1718–1730.
behavior of drained and undrained triaxial compression tests: DEM Reid, D. and Fourie, A.B. 2019. A direct simple shear device for static
study. Journal of Geotechnical and Geoenvironmentral Engineering, liquefaction triggering under constant shear drained loading.
144 (9). Geotechnique Letters, 9(2), 142-146.
Ni, Q., Tan, T., Dasari, G. and Hight, D. 2004. Contribution of fines to the Reid, D., Fourie, A., Ayala, J.L., Dickinson, S., Ochoa-Cornejo, F., Fanni,
compressive strength of mixed soils. Géotechnique, 54 (9) (2004), pp. R., Garfias, J., Da Fonseca, A.V., Ghafghazi, M., Ovalle, C., Riemer,
561-569. M., Rismanchian, A., Olivera, R., and Suazo, G. 2021. Results of a
Oliveira Filho, W.L., 1998, Verification of a Desiccation Theory for Soft critical state line testing round robin programme. Géotechnique,
Soils, Ph.D. Thesis, University of Colorado Boulder. 71(7), 616-630.
Olson SM. (2009). Strength ratio approach for liquefaction analysis of Reid, D., Fourie, A., Castro, J. and Lupo, J. 2018a. Undrained shear
tailings dams. Proceedings of the 57th Annual Geotechnical strength evolution with loading on an undisturbed block sample of
Engineering Conference. Minneapolis, Minnesota, USA. desiccated gold tailings. Proceedings of Tailings and Mine Waste,
Olson SM, and Stark TD. 2002. Liquefied strength ratio from liquefaction 2018.
flow failure case histories. Canadian Geotechnical Journal Reid, D., Fourie, A., and Russell, A. (2018b). Effects of desiccation on
39(3):629-647. shear strength of tailings – comparison of clayey and sandy tailings.
Olson SM, and Stark TD. 2003. Yield strength ratio and liquefaction Proceedings of Tailings and Mine Waste, 2018.
analysis of slopes and embankments. Journal of Geotechnical and Riemer M., Moriwaki, Y. and Obermeyer, J. 2008. Effect of High
Geoenvironmental Engineering, 129(8):727-737. Confining Stresses on Static and Cyclic Strengths of Mine Tailing
Pane, V., 1985, Sedimentation and Consolidation of Clays, Ph.D. Thesis, Materials. Proc. Geotechnical Earthquake Engineering and Soil
University of Colorado Boulder. Dynamics IV. GSP 181 ASCE.
Pane, V., and Schiffman, R.L., 1997, The permeability of clay Robertson P.K. 1991. Soil classification using the cone penetration test:
suspensions, Géotechnique, Vol. 47, 273-288. Reply. Canadian Geotechnical Journal 28(1):176-178.
Papadimitriou A., Bouckovalas G., Andrianopoulos K. 2014. Robertson P.K 2009. Interpretation of cone penetration tests – a unified
Methodology for estimating seismic coefficients for performance- approach. Canadian Geotechnical Journal 46:1337-1355.
based design of earth dams and tall embankments. Soil Dynamics and Robertson P.K. 2010. Evaluation of Flow Liquefaction and Liquefied
Earthquake Engineering. 56, 57–73. Strength Using the Cone Penetration Test. Journal of Geotechnical &
Papadopoulou, A., and Tika, T. 2008. The effect of fines on critical state Geoenvironmental Engineering 136(6):842-853.
and liquefaction resistance characteristics of non-plastic silty sands. Robertson, P.K. 2012. Evaluation of Flow Liquefaction using the CPT: an
Soils and Foundations, 48(5): 713–725. update. Keynote presentation at Tailings and Mine Waste Conference.
Park, S, Nong, Z. and Lee, D. 2020. Effect of vertical effective and initial Colorado.
static shear stresses on the liquefaction resistance of sands in cyclic Robertson P.K. 2016. Cone penetration test (CPT)-based soil behaviour
direct simple shear tests. Soils and Foundations. 60(6), 1588-1607. type (SBT) classification system – an update. Canadian Geotechnical
Peck, R.B. 1969. Advantages and limitations of the observational method Journal, 53(12):1910-1927.
in applied soil mechanics. Geotechnique, 19, 171-187. Robertson, P.K. 2017. Evaluation of flow liquefaction: influence of high
Plewes, H.D., Davies, M.P. and Jefferies, M.G. 1992. CPT based stresses. Proc. 3rd International Conference on Performance-based
screening procedure for evaluating liquefaction susceptibility. In: Design in Earthquake Geotechnical Engineering, Vancouver,
Proceedings of the 45th Canadian Geotechnical Conference. Canada, Canada.
Toronto. Robertson, P.K., Viana de Fonseca, A., Ulrich, B. and Coffin, J. (2017).
Polito, C. and Martin, J. 2001. Effects of Nonplastic Fines on the Characterisation of unsaturated mine waste: a case history. Can.
Liquefaction Resistance of Sands. Journal of Geotechnical and Geotech. J., 54, 1752-1761.
Geoenvironmental Engineering. 127(5). Robertson P.K., and Wride C. 1998. Evaluating cyclic liquefaction
Poulos, S.J. 1971. The stress-strain curves of soils. Geotechnical potential using the cone penetration test. Canadian Geotechnical
Engineers Inc., Winchester, Mass. USA. Journal, 35(3):442-459.
Poulos, S.J. 1981. The steady state of deformation. Journal of the Robertson, P.K., De Melo, L., Williams, D. and Wilson, G. 2019. Report
Geotechnical Engineering Division. 107(5). of the Expert Panel on the Technical Causes of the Failure of Feijão
Pournaghiazar, M, Russell, AR & Khalili, N. 2011, Development of a new Dam I (Brumadinho).
calibration chamber for conducting cone penetration tests in Roscoe, K. 1953. An apparatus for the application of simple shear to soil
unsaturated soils, Canadian Geotechnical Journal, 48(2), 314–321. samples. Proc. 3rd Int. Conf. Soil Mech. 1, No. 2, 186-191.
Prevost, J. and Popescu, R. 1996. Constitutive Relations for Soil Roscoe, K. H., Schofield, A N. and Wroth, C. P. (1958). On Yielding of
Materials. The Electronic Journal of Geotechnical Engineering. Soils. Geotéchnique, 8(1), 22-53.
Inaugural Volume. Rourke, H. and Luppnow, D. 2015. The Risks of Excess Water on Tailings
Proskin, S., Sego, D. and Alostaz, M., 2010. Freeze-thaw and Facilities and Its Application to Dam-break Studies. Tailings and
consolidation tests on Suncor mature fine tailings (MTF). Cold Mine Waste Management for the 21st Century. Sydney, Australia.
Regions Science and Technology, 63(3), 110-120.
181
Sadrekarimi, A. 2016. Static Liquefaction Analysis Considering Principal Skempton, A. 1985. Residual strength of clays in landslides, folder strata
Stress Directions and Anisotropy. Geotechnical and Geological and laboratory. Geotechnique 35, No. 1, 3-18.
Engineering. 34(4). Sladen, J.A. and Handford, G., 1987. A potential systematic error in
Sadrekarimi A. and Olson, S. 2011. Yield strength ratios, critical strength laboratory testing of very loose sands. Canadian Geotechnical
ratios, and brittleness of sandy soils from laboratory tests. Canadian Journal, 24(3), pp.462-466.
Geotechnical Journal, 48(3):493-510. Soderberg, R.L. and Busch, R. A. 1977. Design guide for metal and
Sano, T. 1906. Report on earthquake damage in California, USA. Journal nonmetal tailing disposal. Department of the Interior, Bureau of
of Architecture and Building Science. 238, 646–656 (in Japanese). Mines, USA.
Santamarina, J.C., Torres-Cruz, L.A., and Bachus, R.C. 2019. Why coal Somogyi, F., Keshian, B., and Bromwell, L.G., 1981, Consolidation
ash and tailings dam disasters occur. Science, 364(6440): 526–528. Behavior of Impounded Slurries, Presented at ASCE Spring
Santamarina, J. C., and Cho, G. C. 2004. Soil behaviour: The role of Convention, New York, New York.
particle shape. Advances in geotechnical engineering: The Skempton Stewart, D. P., and Randolph, M. F. 1994. T-bar penetration testing in soft
Conference, R. J. Jardine, D. M. Potts, and K. G. Higgins, eds., Vol. clay. J. Geotech. Engrg., 120(12), 2230–2235.
1, Thomas Telford, London, 604–617. Suits, L.D., Sheahan, T.C., Sadrekarimi A. and Olson, S. 2009. A new ring
Saragoni, R. 1993. Análisis del Riesgo Sísmico para la Reconstrucción del shear device to measure the large displacement shearing behavior of
Puerto de Valparaíso. Sextas Jornadas Chilenas de Sismología e sands. Geotechnical Testing Journal, 32(3).
Ingeniería Sísmica, Santiago (In Spanish). Teh, C.I. and Houlsby, G.T. 1991. An analytical study of the cone
Sarantonis, E., Etezad, M. and Ghafghazi, M. 2020. The effect of assumed penetration test in clay. Geotechnique, 41(1), 17-34.
residual strength on remediation cost of a typical tailings dam. Thevanayagam, S., Shenthan, T., Mohan, S., and Liang, J. 2002.
Proceedings Tailings and Mine Waste 2020, 671-678. Undrained fragility of clean sands, silty sands, and sandy silts.
Schmertmann, J. H. 1991. The mechanical aging of soils. Journal of Journal of Geotechnical and Geoenvironmental Engineering,
Geotechnical Engineering., 179, 1288–1330. 128(10): 849–859.
Schnaid, F. 2021. On the geomechanics and geocharacterisation of Todisco, M.C. and Coop, M.R. (2019). Quantifying “transitional” soil
tailings. Proceedings of 6th International Conference on behaviour. Soils and Foundations, 59, 2070-2082.
Geotechnical and Geophysical Site Investigation, 25pp. Torres Cruz LA 2016. Use of the cone penetration test to assess the
Schnaid, F., Bedin, J. Viana da Fonseca, A. J. P. and Costa Filho, L. M. liquefaction potential of tailings storage facilities. PhD Thesis.
2013. Stiffness and Strength Governing the Static Liquefaction of University of the Witwatersrand. Johannesburg, South Africa.
Tailings. J. Geotech. Geoenvironmental. Engng., 139 (12): 1943- Torres-Cruz, L.A. 2019. Limit void ratios and steady-state line of non-
5606. plastic soils. Proceedings of the Institution of Civil Engineers -
Schneider, J.A., Randolph, M.F., Mayne, P.W. & N.R. Ramsey 2008. Geotechnical Engineering, 172(3): 283–295.
Analysis of factors influencing soil classification using normalized Torres-Cruz, L.A., 2021. The Plewes Method: a Word of Caution.
piezocone tip resistance and pore pressure parameters. J. Geotech. Mining, Metallurgy & Exploration, 38(3), pp.1329-1338.
Geoenviron. Eng. 134(11): 1569–1586. Torres-Cruz LA and Santamarina (2020). The critical state line of
Schneider, J.A., Hotstream, J.N., Mayne, P.W. & Randolph, M.F. 2012. nonplastic tailings. Canadian Geotechnical Journal, 57(10).
Comparing CPTU Q-F and Q- 2 v0 soil classification charts. Torres-Cruz, L.A. and Vermeulen, N., 2018. CPTu-based soil behaviour
Géotechnique Letters 2:209–215. type of low plasticity silts. In Cone Penetration Testing 2018, 635-
Schneider, J.A. and Moss, R.E.S. 2011. Linking cyclic stress and cyclic 641. CRC Press.
strain based methods for assessment of cyclic liquefaction triggering Troncoso J.H., Verdugo R. 1985. Silt Content and Dynamic Behavior of
in sands. Geotechnique Letters, 1, 31-36. Tailings Sands. Proc. XI International Conference on Soil Mechanics
Schofield, A.N. and Wroth C.P. 1968. Critical State Soil Mechanics. and Foundation Engineering. Vol. 3. 1311-1314.
McGraw-Hill, London. Troncoso J.H, Ishihara K and Verdugo R. 1988. Aging Effects on Cyclic
Seed, H. 1983. Earthquake-resistant design of earth dams. Proc. Symp. Shear Strength of Tailings Materials. Proc. IX World Conference on
Seismic Des. of Earth Dams and Caverns, ASCE, New York, 41–64. Earthquake Engineering, Vol. 3, 121–126. Tokyo, Japan.
Seed, H. and Idriss, I. 1982. Ground motions and soil liquefaction during Towhata, I. 2008. Geotechnical Earthquake Engineering. Springer-Verlag
earthquakes. EERI. Monograph, University of California, Berkeley. Berlin Heidelberg. ISBN 978-3-540-35782-7
Seed, R. and Harder, L. 1990. SPT-Based Analysis of Cyclic Pore Vaid, Y.P., and Chern, J.C. 1983. Effect of static shear on resistance to
Pressure Generation and Undrained Residual Strength. In: Duncan, liquefaction. Soils and Foundations, 23(1): 47–60.
J.M., Ed., Proc. H.B. Seed Memorial Symposium, Vol. 2, BiTech Vaid, Y., Stedman, J. and Sivathayalan, S. 2001. Confining stress and
Publishers, Richmond, 351-376. static shear effects in cyclic Liquefaction. Can. Geotech. J. 38: 580–
Sheahan, T., Ladd, C. and Germaine, J. 1996. Rate-dependent undrained 591.
shear behavior of saturated clay. Journal of Geotechnical Valenzuela, L. (2015). Tailings Dams and Hydraulic Fills. The 2015
Engineering, 122(2), 99-108. Casagrande Lecture. 16th International Conference on Soil
Shuttle, D.A. and Cunning, J., 2007. Liquefaction potential of silts from Mechanics and Geotechnical Engineering. Buenos Aires, Argentina.
CPTu. Canadian Geotechnical Journal, 44(1), pp.1-19. Valenzuela L., Verdugo R., Campaña J. and Opazo C. 2020. Safety
Shuttle DA, and Cunning J. 2008. Reply to the discussion by Robertson performance of dams in Chile highly seismic environment. Proc. 4th
on "Liquefaction potential of silts by CPTu." Canadian Geotechnical International Dam World Conference, Lisbon, Portugal.
Journal 45:142-145. Verdugo R. 1992. Characterization of sandy soil behavior under large
Shuttle, D. and Jefferies, M. 1998. Dimensionless and unbiased CPT deformation. PhD Thesis. University of Tokyo, Japan.
interpretation in sand. Int. Jnl. Num. and Anal. Methods in Geomech., Verdugo, R. 2009. Seismic performance based-design of large earth and
22, 351-391. tailings dams. Proc. International Conference on Performance-Based
Shuttle, D. and Jefferies, M., 2016. Determining silt state from CPTu. Design in Earthquake Geotechnical Eng. – from case history to
Geotechnical Research, 3(3), 90-118. practice. Tsukuba, Japan.
Shuttle D., Martens S., and Jefferies M. (2022). Geostatic stress in oil-sand Verdugo, R. 2011. Seismic stability analysis of large tailings dams. Proc.
tailings. Proceedings of the Institution of Civil Engineers – Theme Lectures of 5th International Conference on Earthquake
Geotechnical Engineering. https://doi.org/10.1680/jgeen.21.00114. Geotechnical Engineering. Santiago., Chile.
Silvis, F. and de Groot, M.B. 1995. Flow slides in the Netherlands: Verdugo R, and Ishihara K. 1996. The steady state of sandy soils. Soils
experience and engineering practice. Can. Geotech. J., 32, 1086- and Foundations 36(2):81-91.
1092. Verdugo, R. and Santos, E. 2009. Liquefaction resistance of thickened
Skempton AW 1964. Long-term stability of clay slopes. Geotechnique, tailings of copper mines. Proc. 17th International Conference on Soil
14(2), pp.77-102. Mechanics and Geotechnical Engineering. Alexandria, Egypt.
182
Verdugo, R. and Viertel, P. 2004. Effect of density and fines content on
the cyclic strength of copper tailings. V Chilean Geotechnical
Congress.
Vietti, A. and Coghill, M. 2015. Slurry Chemistry. In: Paste and
Thickened Tailings – A Guide, Third Edition. Ed. R.J. Jewell and A.B.
Fourie, Australian Centre for Geomechanics.
Villavicencio G., Espinace R., Palma P., Fourie A. Valenzuela P. 2014.
Failures of sand tailings dams in a highly seismic country. Canadian
Geotechnical Journal. 51: 449–464.
Wanatowski, D. and Chu, J. 2007. Static liquefaction of sand in plane
strain. Can. Geotech. J., 44, 299-313.
Wanatowski, D. and Chu, J. 2011. Pre-failure instability behaviour of sand
in strain path testing under plane-strain conditions. Soils and
Foundations, 51(3), 423-435.
Wang, Z. 2010. Seismic Hazard Assessment: Issues and Alternatives.
Pure and Applied Geophysics. Published on line. DOI
10.1007/s00024-010-0148-3
Weber, J.P., 2015. Engineering evaluation of post-liquefaction strength.
PhD thesis, University of California, Berkeley.
Wells, P.S., Charlebois, L., Diep, J., Moyls, B., Omotoso, O., Revington,
A. and Weiss, M. 2015. Inline Flocculation. In: Paste and Thickened
Tailings – A Guide, Third Edition. Ed. R.J. Jewell and A.B. Fourie,
Australian Centre for Geomechanics.
Wride C.E., Robertson P.K., Biggar K.W., Campanella R.G., Hofmann
B.A., Hughes J.M.O., Kupper A., and Woeller D.J. 200. Interpretation
of in situ test results from the CANLEX sites. Canadian Geotechnical
Journal, 37(3).
Yang, S.L., Sandven, R., and Grande, L. 2006. Steady-state lines of sand–
silt mixtures. Canadian Geotechnical Journal, 43(11): 1213–1219.
Yang, Y., Wei, Z., Fourie, A.B., Chen, Y., Zheng, B., Wang, W. and
Zhuang, S. 2019. Particle shape analysis of tailings using digital
image processing. Environmental Science and Pollution Research,
26, 26397-26403.
Yao, D.T.C., 1998, Desiccation Process for Soft Soils, Ph.D. Thesis,
University of Colorado Boulder.
Yao, D.T.C , Oliveira-Filho, W.L., Cai, X.C. and Znidarcic, D., 2002.
Numerical solution for consolidation and desiccation of soft soils,
International Journal for Numerical and Analytical Methods in
Geomechanics, 26, (2), 139-161.
You, Z., 1993, Flow Channeling in Soft Clays and its Influence on
Consolidation, Ph.D. Thesis, University of Colorado Boulder.
Yu, H.S. 1998. A unified state parameter model for clay and sand.
International Journal for Numerical and Analytical Methods in
Geomechanics, 22, 621-653.
Zergoun, M., and Vaid, Y. 1994. Effective stress response of clay to
undrained cyclic loading. Can. Geotech. J., 31, 714–727.
Zhao J. and Guo N. 2013. Unique critical state characteristics in granular
media considering fabric anisotropy. Géotechnique, 63(8), 695-704.
Zhou, H., and Randolph, M.F. 2009. Resistance of full-flow penetrometers
in rate-dependent and strain-softening clay. Geotechnique, 59(2), 79-
86.
Zlatovic´, S., and Ishihara, K. 1995. On the influence of nonplastic fines
on residual strength. In Proceedings of the 1st International
Conference on Earthquake Geotechnical Engineering, Tokyo, Japan.
AA Balkema, Rotterdam, the Netherlands. 239–244.
Znidarcic, D., 2015, If it creeps, does it matter? Tailings and Mine Waste
‘15 conference proceedings, October 25-28, Vancouver, BC, Canada,
pp.229-235.
Znidarcic, D., Adkins, D., Utting, L. and Catling, M., 2015, Rheomax®
ETD technology, a laboratory study of application performance and
associated geotechnical characteristics for polymer assisted tailings
deposition of oil sands MFT, Tailings and Mine Waste ‘15, October
25-28, Vancouver, BC, Canada, 508-520
183
184
State of the Art 4: Environmental Geotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Protecting the environment from contamination with barrier systems: advances and
challenges
Protéger l'environnement de la contamination avec des systèmes de barrière : avancées et défis
Geoengineering Centre and Queen's-RMC, Dept. of Civil Engineering, Queen's University, Kingston, Ontario, Canada.
Stephan Jefferis
Environmental Geotechnics, Banbury, UK and Dept. of Engineering Science, University Oxford, UK.
With sections contributed by:
Fady B. Abdelaal
Geoengineering Centre and Queen's-RMC, Dept. of Civil Engineering, Queen's University, Kingston, Ontario, Canada
Andrea Dominijanni
Department of Structural, Geotechnical and Building Engineering, Politecnico di Torino, Torino, Italy
Kelvin Legge
Department of Water and Sanitation, Pretoria, South Africa
Ennio M. Palmeira
University of Brasília, Brasília, Brazil
John Scheirs
ExcelPlas, Highett, Victoria, Australia
Charles D. Shackelford & Joseph Scalia IV
Department of Civil and Environmental Engineering, Colorado State University, Fort Collins, Colorado, USA
Tony L.T. Zhan
Department of Civil Engineering, Zhejiang University, Hangzhou, China
Craig H. Benson
School of Engineering, University of Virginia, Charlottesville, Virginia, USA
Abdelmalek Bouazza
Monash University, Clayton, Victoria, Australia
Gemmina Di Emidio
Ghent University and Pollux Consulting, Ghent, Belgium
Jeffrey C. Evans
Bucknell University, Lewisburg, Pennsylvania, United States
Evelina Fratalocchi
Facoltà di Ingegneria, Università Politecnica delle Marche via Brecce Bianche Ancona
Maria G. Gardoni,
Federal University of Minas Gerais, Belo Horizonte, Brazil
Takeshi Katsumi
Graduate School of Global Environmental Studies, Kyoto University, Kyoto, Japan
Mario Manassero
Department of Structural, Geotechnical and Building Engineering, Politecnico di Torino, Torino, Italy
Denys Parra
Anddes Perú and National University of Engineering, Lima, Peru
ABSTRACT: The state-of-the-art for barriers systems control of pollution migration and hydraulic structures including liner systems
for landfills, mine waste, and dams/lagoons are discussed along with some comments on current significant environmental issues.
Recent research on geomembranes, geosynthetic clay liners, and composite liners and the practical implications are discussed. The
need to design taking account of the interactions between the liner system components is highlighted. Regional differences relating
to environmental geotechnics of municipal solid waste and mining waste and the implications of differences in regulations and
climate are discussed. Vertical barrier systems may be regarded as mature technologies and research funding is much more limited.
A bibliography of key papers on cut-off walls has been compiled. It is recognized that major changes to cement-bentonite
formulations will be necessitated by the drive to reduce carbon footprints together with the dwindling availability of pulverised fuel
187
ash and blastfurnace slag. The implications for environmental geotechnics associated with recycled materials, waste minimization,
and some overarching issues are discussed. The paper finishes with some thoughts about the future of environmental geotechnics
and some concluding comments.
RÉSUMÉ : L'état de l'art des barrières pour le contrôle de la migration de la pollution et des structures hydrauliques, y compris les
barrières pour les installations de stockage de déchets, les déchets miniers et les barrages / lagunes, est discuté avec quelques
commentaires sur les problèmes environnementaux actuels importants. Des recherches récentes sur les géomembranes, les
géosynthétiques bentonitiques et les étanchéités composites ainsi que les implications pratiques sont discutées. La nécessité de
concevoir en tenant compte des interactions entre les composants de la barrière est soulignée. Les différences régionales relatives à
la géotechnique environnementale des déchets solides municipaux et des déchets miniers sont discutées. Les systèmes de barrières
verticales peuvent être considérés comme des technologies matures. Une bibliographie des principaux articles sur les murs
d’étanchéité a été synthétisée. Il est reconnu que des changements majeurs dans les formulations ciment-bentonite seront rendus
nécessaires par la volonté de réduire l'empreinte carbone. Les implications pour la géotechnique environnementale associées aux
matériaux recyclés, à la minimisation des déchets et à certaines questions globales sont discutées. L'article se termine par quelques
réflexions sur l'avenir de la géotechnique environnementale et quelques commentaires conclusifs.
KEYWORDS: GCLs, geomembranes, barrier systems, cut-off walls, slurry systems, cement- and soil-bentonite, recycled materials
1 INTRODUCTION research groups across the World and their input has been
integrated into this paper.
The last quarter of the 20th century saw a growing concern in Preventative environmental geotechnics for waste
many countries with the previous thoughtless "out of sight, out management emerged from the need to develop better landfills
of mind" approach to waste disposal. This approach frequently for the disposal of hazardous and municipal solid waste. Since
involved the co-disposal of liquid hazardous waste, heavy the early 1980s landfills have been at the forefront of the
metals, and municipal solid waste. To make matters worse, waste development of what we now know as geoenvironmental
had often been disposed in worked-out sand and gravel pits engineering and the more geotechnically focused area of
without any liner with the mantra "dilution is the solution to environmental geotechnics. Regulations that were developed
pollution". After many decades of inappropriate waste disposal, generally banned the disposal of liquid hazardous waste and
the consequent impacts on the environment and human health required its solidification and stabilization in some manner
became apparent. Awareness eventually triggered highly before disposal in an engineered landfill. Early liner systems
publicized removal of people from their homes and the closure focused initially on the use of natural or compacted low
of schools such as occurred in response to President Carter's permeability clay liners but soon also incorporated
declaration of two states of emergency at the infamous Love geomembranes to form composite liners. The development of
Canal site located in Niagara Falls, New York State, USA geosynthetic clay liners (GCLs) and the paucity of suitable
(Cohen et al. 1987; Colten and Skinner 1996; Phillips et al. 2007) natural clays for compacted clay liners in many areas saw the
and at Lekkerkerk in the Netherlands (Brinkmann 1981). adoption of GCLs as an alternative to compacted clay in
Situations such as these and the identification of many composite liners.
"Superfund sites" in the USA prompted governments to develop With recognition of the risks of relying on only a single, albeit
regulations for waste management (e.g., The Resource composite liner, for the disposal of hazardous wastes, the use of
Conservation and Recovery Act, RCRA, passed in October 1976 double composite liners with a leak detection/secondary leachate
to address the increasing problems faced by the USA from the collection system became enshrined in many regulations.
growing volume of municipal and industrial waste). Four Experience in the operation of these facilities and their
primary consequences of the recognition of the potentially monitoring, academic research, and innovations by
harmful effects of waste disposal practices were the recognition manufacturers have substantially advanced the development of
of the need to: geomembranes and GCLs in the 21st century. Today's high
(i) classify and separate different types of waste (e.g., density polyethylene (HDPE) geomembranes (GMBs) are much
construction and demolition, municipal solid waste, more sophisticated than those of the 1970s to late 1990s. The use
hazardous waste, radioactive waste etc.); of geomembranes has expanded to a range of products broadly
(ii) design waste disposal facilities to minimize the risk of defined as HDPE but with different resins and additive packages
contaminant escape to the environment and remove past suitable for different applications. A similar situation applies
nuisances (e.g., fires, vector rodents, birds) associated with with respect to linear low density polyethylene (LLDPE) and
facility operations; and polyvinyl chloride (PVC). Thus, for important projects, the
(iii) contain and remediate past problems, geomembrane needs to be selected for the specific end-use. Key
(iv) proactively consider and address potential future issues and advancements concerning HDPE geomembranes are
problems. reviewed in §2.
The last three of these four consequences fall within the realm In parallel with the growth in HDPE use in landfill
of environmental geotechnics. Thus, this paper is divided into applications, there was a growth in the use of bituminous
three parts. Part A deals with preventative environmental geomembranes (BGMs) for lining canals. Success in this
geotechnics (i.e., actions to prevent future pollution and application has prompted consideration of their use in a range of
addresses the second consequence above). Part B deals with other applications such as landfill covers, pond liners, and covers
remedial environmental geotechnics (actions to minimize the for mine waste. §3 explores BGMs 20 years into the 21st century.
impact of past contamination) and addresses the third Like HDPE geomembranes and BGMs, GCLs have advanced
consequence above. Finally, Part C deals with some over-arching significantly, as has the range of applications for GCLs. Of the
issues relating to future developments in environmental many types of GCL available, needle-punched GCLs with a layer
geotechnics and begins to address the fourth consequence above. of bentonite sandwiched between two geotextiles and held
Contributions on these three topics were sought from leading together by needle-punching have become dominant. However,
they have also evolved with an extensive range of different
188
products (one manufacturer alone produces 50 different GCLs). these regions, in large part in this century, they are still struggling
Differences include standard sodium bentonite and polymer with issues that more developed countries addressed last century.
amended bentonite, mass of bentonite, geotextile characteristics §7 explores these issues from the perspective of those living in
and mass, amount of needle-punching, and the presence or Asia, Africa, and South America.
absence of a polyolefin layer (anything from 200 g/m² to 2000 While landfills have been the focus of the developments and
g/m²) either applied in a molten state to form a coating or glued applications for barrier systems, the volume of waste generated
to the GCL to enhance the performance of the GCL. §4 discusses by mining is enormous compared to the waste contained in
these advances with a particular focus on the use of enhanced landfills. Although there are notable exceptions, the mining
bentonites. industry has generally been far less concerned with
Initially, leachate collection systems involved relatively environmental impact or even the risks to human life and
primitive toe drains and French drains to allow leachate property from mine waste than has been the case for municipal
collection and minimize the escape of contaminated fluids to the and hazardous waste facilities. This is particularly true in the
environment. However, it was soon found that these systems less-developed parts of the world. For example, South America
experienced significant clogging and build-up in the leachate is a significant source of minerals. However, several recent tragic
head (Rowe and Yu 2012). This prompted a movement to have failures have forced insurance companies to pressure mining
blanket drains initially of pea gravel but clogging of these companies to pay more attention to the risks to human life and
leachate collection systems still occurred (Fleming et al. 1999; property and there is growing concern regarding their
Rowe and Yu 2010). The solution to this problem was the use of environmental impact.
relatively coarse uniform gravel as a drainage layer. However, The use of geosynthetics generally, and especially in barrier
initial applications of the solution began to compromise the systems, has grown substantially this century even with the
service life of geomembranes by increasing punctures and, absence of appropriate regulations and/or the enforcement of
arguably of more importance, inducing tensile strains that can regulations in many countries. §8 examines geosynthetic
become stress cracks in the future. Issues such as this began to applications in mining. It notes the lack of regulations and
raise awareness of the interactions between the barrier system standards relevant to the many facilities at high altitudes with
components themselves, and between the components and the steep side slopes and tailings storage facilities that pose a risk
environment. §5 seeks to raise awareness of the importance of from liquefaction.
designing barriers as a system, carefully considering the many Remedial environmental geotechnics emerged from the need
interactions and how they affect the system's overall to clean up the problems such as those triggered by investigations
performance. of the landfilling of chemical waste at Love Canal (see the
Key contaminants of concern in developing most landfill analysis of the social and technical issues set out in Mazur
regulations last century were chlorinated solvents, heavy metals, (1998)). Remediation of contaminated land has been and remains
and metalloids. Regulations and modern waste management a preeminent issue in some areas/countries on which massive
practices have substantially reduced the concentrations of these sums have been and are yet to be expended. However, although
contaminants in the waste stream. Consequently, the inputs for this state-of-the-art review were sought from
predominant contaminants in modern municipal waste leachate researchers and practitioners from around the world, a notable
are volatile fatty acids (from the biodegradation of organic gap in the responses received was on the clean-up of
waste) and salts (such as sodium chloride) with relatively trace contaminated land. As a result, the lead authors are left with the
amounts of the contaminants of primary concern in the last impression that from an academic research perspective,
century. contaminated land clean-up is either (a) regarded as a mature
Generally, people did not concern themselves with salt as a discipline where innovation is primarily by consultants and
contaminant of concern. Very few regulations explicitly list clean-up contractors and managed by regulators who, in many
chloride as a critical contaminant to be controlled by the landfill countries, now have substantial experience and powers under
design. A notable exception is Regulation 232/98 (MOE 1998) environmental regulations (though not always the necessary
in Ontario, Canada, which prescribes chloride as a significant funding), or (b) research funding is largely limited to areas such
critical contaminant. The technical experts formulating MOE as biomediated and bioinspired geotechnics.
regulations at the time were aware of the potential for both Part B then considers the present state of research and practice
known unknowns and unknown unknowns. They recognized that for:
chloride is one of the most mobile contaminants in landfill vapour barriers used to minimize the impact of residual
leachate and present at concentrations far above drinking water volatile organic compounds when redeveloping
objectives. Thus, it was an excellent tracer and by keeping brownfield sites (§9).
chloride release very low, we were also keeping other releases vertical barrier walls (cut-off walls) formed by slurry
low. In 2022 we now recognize that landfill leachate contains trench techniques. These are mature technologies with
many contaminants of emerging concern. These are very few reported problems in service. The approach,
predominantly very useful and widely used manufactured therefore, has been to seek input from experts to
chemicals that were subsequently found to have potential health identify significant issues on which a fuller listing of
implications. These contaminants include per- and papers complements a brief discussion (§10).
polyfluoroalkyl substances (PFAS), bisphenol-A (BPA), It is unlikely that 40 years ago anyone would have predicted
polybrominated diphenyl ether (PBDE), polychlorinated the path that has been taken by what is now the subdiscipline of
biphenyls (PCBs), and carbon nanoparticles. §6 discusses the preventative environmental geotechnics. Thus, it would seem
implications of these contaminants for existing landfills and foolish to try and predict where we will be in another 40 years.
future landfills. Even with the banning or restriction on the However, we can look forward a decade and Part C considers:
production of many of these chemicals, they can be expected to increased pressures to recycle materials and some
be entering landfills and in landfill leachate for many decades to unintended consequences (§11),
come. waste minimization (§12),
While highly developed countries are struggling to deal with other emerging issues (§13), and
contaminants of emerging concern, a substantial proportion of finally, some thoughts on the future of preventative
the human population lives in less-developed countries, environmental geotechnics as a discipline (§14).
particularly in Asia, Africa, and South America. While there are Just as Covid-19 may change hard land use, global warming,
some examples of well-developed regulations and practices in and the current drive to eat less meat and become more
189
vegetarian may significantly impact the use of greenspace. For PART A: PREVENTATIVE ENVIRONMENTAL
example, about 70% of the land is managed by farming in the GEOTECHNICS
UK, and less than 10% is urban development. Reduction in the
numbers of grazing animals may allow tracts of land to be 2 EVOLVING HDPE FORMULATIONS AND
returned to ‘nature’ (re-wilding) if not used for crops or unless MULTILAYERED GEOMEMBRANES
someone is paid to manage them. Therefore, we must expect the
growth of scrubland with a return to climax vegetation over a Over the past decade, there has been a substantial increase in
timescale measured in centuries. We must plan for this, and in the range of what are generically called 'HDPE
particular, we need to be undertaking back-casting studies. If the geomembranes'. This includes new tailored resins developed to
future impacts of current plans seem unacceptable, then change address issues and shortcomings identified with standard
is needed now and should not be deferred to when problems HDPE and multilayered geomembranes with white and
arise. conductive layers. This section seeks to make the reader aware
of some of the materials presently available on the market as
1.1 Contaminating lifespan and service life well as some of the challenges and insights that have been
developed this century, notwithstanding the fact that most
An important consideration in this paper is the service life of regulations or guidelines relating to HDPE geomembranes are
the components of barrier systems and the overall service life still heavily rooted in what was known and available in the last
of the system. Barrier systems should be designed to control century.
contaminant escape to a negligible level or the ingress and egress This sub-section will discuss the various polymers and
of water or gases to an acceptable level. The contaminating additives used in the manufacture of HDPE geomembranes and
lifespan of a waste disposal facility is, by definition, the period the functional performance properties they confer.
during which the barrier system must control contaminant escape
to a negligible level (MOE 1998; Rowe et al. 2004). 2.1 Polymers and additives used for manufacturing state-of-
The service life of a component of the barrier system depends the-art HDPE geomembranes
on the material properties of that component and the
environmental conditions that will cause it to fail. For example, for High density polyethylene (HDPE) geomembranes are the
an HDPE geomembrane the critical property for containment is the most widely manufactured and the most widely used polymeric
stress crack resistance (SCR). When the geomembrane cracks liners in the geosynthetics industry. To ensure their long-term
extensively, it no longer serves its containment function and can stability and durability (service life), HDPE geomembranes
be said to have reached its service life. However, whether the stress contain proprietary blends of specialized additives:
cracks occur depends both on the stress crack resistance that is antioxidants, stabilizers, carbon black, and chemicals to
remaining (a material property) and the tensile stresses applied to improve the base polymer's processing stability, welding
the geomembrane. Consequently, a geomembrane with a given stability.
stress crack resistance may have reached its service life in one The long-term service lifetime of HDPE liners is mainly
application and not in another depending on the design and the dependent on the level and effectiveness of the additive package.
consequent tensile stresses and strains within the geomembrane. The manufacturers do not disclose the additives they employ in
Thus, when talking about a material characteristic in isolation it is their different HDPE geomembranes to the end-user. However,
better to use the term time to nominal failure to define when the there is a body of technical knowledge regarding what
material has reached a critical time in its life. constitutes good practice and best practice in the formulation of
Often the time to nominal failure is taken to be the time when HDPE geomembranes.
some engineering property of significance has reduced to 50% of
the original or specified value. As the critical property (e.g., stress 2.1.1 The Base Resin
crack resistance for a geomembrane or hydraulic resistance for a There is a large range of potential candidate base resins for
GCL) reduces below its design value the likelihood of it reaching manufacturing HDPE geomembranes. Over 100 different
its service life increases. Thus, the service life of a component in polyethylene resins are purported to be suitable for HDPE
a barrier system is related to the time to nominal failure but, in geomembrane production. However, prudent selection of the
general, is not equal to the time to nominal failure. base polymer is required to achieve the correct balance of
For a good design, the service life will generally be longer processability, mechanical properties, and perhaps most
than the time to nominal failure. For poor design and/or importantly the outstanding stress crack resistance (SCR)
construction and/or operations the service life could be less than required for HDPE geomembrane long-term durability.
the time to nominal failure. It is important to keep this distinction HDPE resin gain favour with the early use of geomembranes
in mind in discussions of material behaviour and service life. in landfill liners because its high crystallinity gave it very good
Barrier systems are often designed recognizing that chemical resistance. However, the Achilles heel of the high
components may reach their service life before the contaminating crystallinity of HDPE resin was its propensity to stress crack,
lifespan of the facility is reached. A good design is one in which which is cracking failure at a level of stress lower than the yield
the other components of the system will then control the release stress. For this reason, there has been a move away from the use
of contaminants to an acceptable level. Thus, the service life of of HDPE resin and the base resins used today for HDPE
the system is when the barrier system can no longer control the geomembranes are generally medium-density polyethylene
release of contaminants to a negligible level. (MDPE) with a fractional melt index (i.e., an MFI of less than 1
g/10 mins under 2.16 kg mass). The term HDPE is retained not
1.2 Notation because of the resin but because the density of the geomembrane
once the carbon black is added to the resin falls into the category
This paper contains many symbols and terms that may be new to defined by ASTM as HDPE.
the reader. Key symbols are defined in the appendix “Notation”. The designer needs to be judicious in the selection of the
References in one part of this paper to a section another part of geomembrane resin to ensure long-term resistance to stress
the paper are indicated by §. cracking which is an inherent property of the base resin and not
the additive package. Table 1 shows some typical PE base resins
used for HDPE geomembranes as well as some newer higher
temperature capable polyethylene for raised temperature (PERT)
resin grades.
190
In recent developments, there are some new additives such as a carbon black pigment of the correct particle size
diene terpolymer (DTP) from Exxon Mobil where just 3% by (typically 22-25 nm);
weight of DTP added to a geomembrane increases the NCTL- a titanium dioxide pigment (for white geomembranes);
SCR from 1500 hours for additive-free resin to greater than to a processing stabilizer such as a hindered phosphite
3000 hours for the DTP blend. a thermal stabilizer which comprises a primary antioxidant
such as a hindered phenolic antioxidant;
2.1.1.1 PERT Resins
a ultra-violet radiation, UV, stabilizer such as a hindered
For higher temperature applications such as brine ponds or solar
amine stabilizer (HAS, formerly HALS) which also serves
heating ponds, the base polymer chosen for the HDPE
as a long-term thermal stabilizer;
geomembrane should be a Polyethylene of Raised Temperature
(PERT) resin that optimizes tie-chain concentration through the an acid neutralizer (to scavenge acidic residues from the
introduction of co-monomers. Example PERTs include Dowlex, PE catalyst); and
Dow’s Intrepid 2498 or Dowlex 2344 or LyondellBasell’s a polymer processing aid (PPA) such as a fluorinated
Hostalen 4731B (a new-generation PERT resin). elastomer.
2.1.2.1 Carbon Black
2.1.1.2 Recycled Polyethylene Carbon black is added into geomembrane formulation for
The virgin geomembrane base resin is often supplied to the general stabilization purposes, particularly for ultraviolet-
geomembrane manufacturer in a pellet form which is then mixed radiation screening, but it also has some free radical trapping
with carbon black, antioxidants, and stabilizers in a carrier resin capabilities. Therefore, it is generally added as a preformulated
and "let down" into the formulation. concentrate in pellet form. The "masterbatch" comprises a carrier
Regrind or rework flake/pellets (previously processed by the resin with carbon black and possibly the antioxidants/stabilizers.
same manufacturer in-house but never used in the field) may be The carrier resin should be the same as the host resin insofar as
added to the extruder during melt processing without adverse density and melt flow are concerned. However, sometimes
effects. However, this contrasts with reclaimed, recycled, or higher flow carrier resins are used to ensure good mixing with
post-consumer recycled material which should not, under any
the geomembrane base resin.
circumstances, be allowed in the geomembrane formulation.
The acceptable level of carbon black content in an HDPE
Table 1. Typical PE base resins used for "HDPE" geomembranes geomembrane is 2-3% of the total weight of the formulation.
Name of It is critical that, in addition to the amount of carbon black
Manufacturer Comments used, the particle size is also carefully chosen and specified.
Resin
Carbon blacks are classified by the primary size as classified in
Linear MDPE; Unimodal and ASTM D1765. For instance, the commonly used N330, N550,
Marlex Chevron broad MWD; the most widely and N660 grades have particle size ranges of 26-30 nm, 40-48
K306/K307 Phillips used geomembrane resin nm, and 49-60 nm, respectively.
worldwide The choice of carbon black matters! For example, the 60 nm
Borstar
Borealis
Linear MDPE; bimodal and broad N660 carbon black used in some geomembranes made in Asia is
FB1350 MWD too coarse for optimum UV screening. N660 carbon black has
Metallocene linear MDPE; been associated with failures from UV-related degradation.
Enable 3505 Exxon Mobil However, as carbon black particle size decreases, the specific
Unimodal and narrow MWD
Daelim Daelim,
surface area increases substantially. The smaller carbon blacks
Metallocene bimodal MDPE require more antioxidants/stabilizers to be added to the resin to
XP9000 South Korea
compensate for the proportion of additives adsorbed by carbon
Intrepid black particles' high surface area. Since the carbon black for
2498 Dow Bimodal HDPE PERT resin
geomembranes needs to be of the correct particle size to impart
NT/2499 NT
optimum UV protection, the N330 grade is the most preferred for
New Generation PERT HDPE geomembranes. However, given the very small primary particle
resin Ethylene-octene copolymer, size of this carbon black, it is prone to agglomeration (Fig. 2.1),
produced in the proprietary and thus, the extruder must provide good dispersive and
Dowlex solution process. It has a unique distributive mixing capabilities to break down carbon black
Dow
2344 molecular structure with a agglomerates, which can act as stress concentration points and
controlled side-chain distribution, initiate cracking (Fig. 2.2).
which provides excellent stress
crack resistance properties
Total Hexene-based PERT copolymer
Total, France
XRT70 resin
Eltex Ineos, Hexene-based PERT copolymer
TUB220-RT Switzerland resin with narrow MWD
Hostalen Lyondell New Generation PERT HDPE
4731B Basell resin
2.1.2 The Additive Package
The additive package that is mixed with the geomembrane base
resin is usually delivered to the processing extruder via a
concentrate called a masterbatch. The masterbatch for an HDPE
geomembrane application generally comprises the following Figure 2.1 Photomicrographs showing poor carbon black dispersion on
essential ingredients: the left and good carbon black dispersion in HDPE on the right.
a resin carrier (usually a matched flow index MDPE or a
higher flow linear low density polyethylene, LLDPE);
191
Figure 2.2 Stress crack initiated at a carbon black agglomerate in an
HDPE geomembrane (inset with scale bar showing the size of the carbon
black agglomerate).
Figure 2.4 OIT curves obtained by DSC show that the OIT increases
2.1.2.2 Titanium Dioxide proportionally with Irganox 1010 added to the HDPE geomembrane.
White-surfaced HDPE geomembranes contain a concentrated
titanium dioxide white pigment known as coated rutile. Since A more preferred antioxidant is the trifunctional Irganox 1330,
rutile has semi-conductor properties that can produce free which does not contain ester groups in its structure (Figure 2.6)
radicals that can degrade the geomembrane when exposed to UV and hence is resistant to hydrolysis. A recent study by Wang et al.
light, titanium dioxide particles are coated with a passivating (2020) has shown that Irganox 1330 has excellent long-term
barrier layer of alumina, silica, and/or zirconium. The better the ageing resistance compared to other non-ester-based alternatives
coating, the lower the propensity to produce free radicals, like Irganox 3114 and Irganox 1024. Since Irganox 1330 does
reducing photoinduced degradation in the geomembrane. Thus, not contain ester groups in its structure, it is ideally suited to act
the titanium dioxide pigments of choice, known as 'durable' as a long-term heat stabilizer in HDPE liners in applications with
grade, have coherent passivating layers. The importance of using long-term water contact such as full or partial immersion and
a durable grade of titanium dioxide in HDPE geomembranes is moderate to high temperatures such as exposed applications on
crests and batters. Furthermore, since Irganox 1330 has no polar
highlighted in Figure 2.3.
ester linkage, it is significantly more hydrophobic than Irganox
Since titanium dioxide is not as effective a UV screen as
1010 and hence less easily extracted by water.
carbon black and not all the photocatalytic activity of titanium
dioxide can be suppressed by the passivating coatings, white
surface geomembranes generally contain very high levels of
HAS up to twice that required in black HDPE.
Figure 2.5 Tetrafunctional chemical structure of the Irganox 1010
molecule with four active hindered phenolic sites (numbered) and the
Figure 2.3 Photomicrograph showing the surface of a white HDPE four ester groups outlined by dotted ovals.
geomembrane on the left with the incorrect type of white pigment (not a
properly coated rutile) and on the right HDPE with a well-selected
durable coated rutile after exposed to 1600 hours UV.
2.1.2.3 Primary Antioxidant
A primary antioxidant protects the HDPE from high-temperature
thermal oxidation during extrusion and welding and contributes
some long-term thermal stabilization.
A primary antioxidant commonly selected for geomembranes
is Irganox 1010 (BASF), a tetrafunctional phenolic antioxidant
(i.e., four active sites per molecule). The higher the loading of
Irganox 1010, the longer the Oxidative Induction Times (OIT),
as shown in Figure 2.4. However, the ester bridges in Irganox
1010 (Figure 2.5) renders it susceptible to hydrolysis (i.e.,
hydrolytic breakdown), particularly at higher temperatures in the
presence of water and high pH. The products of hydrolysis of Figure 2.6 Trifunctional chemical structure of Irganox 1330 molecule
Irganox 1010 are of a lower molecular mass and higher polarity with three active hindered phenolic sites (numbered).
than Irganox 1010 and are therefore readily leached-out of a
geomembrane in contact with water or an aqueous solution, and In contrast to Irganox 1010 there are no ester groups in this
to a lesser extent out of a geomembrane in contact with moist or Irganox 1330 molecule. Careful selection of the primary
saturated soil. antioxidant is required to optimize both functionality (i.e.,
activity) and resistance to hydrolysis, as shown schematically in
Figure 2.7.
192
Figure 2.7 Functionality and resistance to hydrolysis of common
antioxidants used in HDPE geomembranes.
2.1.2.4 Hindered Amine Stabilizer (HAS/HALS) Figure 2.9 Chimassorb 119 molecular structure with 8 active methylated
Hindered Amine Light Stabilizers (HALS/HAS) have active piperidinyl sites (numbered) and no susceptible ester groups in this
stabilizer molecule.
tetramethylpiperidine group in their structure. The efficacy of 2.1.2.5 Acid Neutralizer
HAS in HDPE is a function of the number of active sites per An acid neutralizer such as hydrotalcite DHT-4A (Kisuma
HAS molecule. The permanence and resistance to extraction of Chemicals) is generally added to neutralize acidic catalyst
HAS is a function of the size and bulkiness of the HAS molecules residues and by-products of the catalytic polymerization
(a combination of their molecular weight and molecular volume). processes used to produce polyethylene. These acidic residues
A well-known HAS additive used in early geomembrane can reduce the effectiveness of basic HAS stabilizers and can
formulations was Tinuvin 770 (Figure 2.8). lead to corrosion of steel processing equipment. Thus, the
Tinuvin 770 has rather limited stabilizer activity compared to hydrotalcite additive is added to 'mop up' these acidic residues.
more modern HAS additives, and the ester groups in its structure
make it susceptible to hydrolytic degradation and rapid
extraction. In contrast, Chimassorb 119 has eight active sites per
molecule, and each active site is a methylated piperidinyl active
site with a nitrogen atom at its centre. There are no ester groups
susceptible to hydrolysis (see Figure 2.9). The methyl group
protects the 'basic' nitrogen atom of the active site from
protonation (i.e., adding a hydrogen atom) by acids (H+ donors).
Therefore, the methylated HAS such as Chimassorb 119 exhibit
excellent resistance to acids, and they are not deactivated (i.e.,
neutralized) in acidic environments. For this reason, methylated Figure 2.10 Chimassorb 944 molecular structure with two numbered
active piperidinyl sites within the repeat unit in the main structure and
119 are the stabilizers of choice for HDPE geomembranes two terminal active sites, making the total number of active sites around
intended for use in aggressive acidic environments such as acidic 10 per molecule.
pregnant liquor ponds, acid heap leach pads, and acidic tailings
dams such as those used in copper and uranium ore extraction.
Furthermore, Chimassorb 119 has a very bulky molecular
structure (Figure 2.9) with a high molecular weight (2285.6
g/mol), significantly decreasing the propensity to be extracted
from the HDPE matrix by liquid media.
Figure 2.8 Difunctional chemical structure of Tinuvin 770 molecule
with two active piperidinyl sites (numbered). Ester groups in this Figure 2.11 Functionality and acid resistance of common HAS
molecule are shown in dotted ovals. stabilizers used in HDPE geomembranes.
By comparison, Chimassorb 944, which is a regular N-H 2.1.2.6 Polymer Processing Aid
HAS with eight to ten active sites per molecule and excellent A fluorinated elastomer (fluoropolymer) is the most common
resistance to extraction, has a molecular weight of 2000-3100 type of polymer processing aid used in geomembrane resins.
g/mol (Figure 2.10). However, the active sites of Chimassorb Very small levels of a fluoro elastomer such as Dynamar FX
944 are relatively basic and thus are susceptible to acid attack 5911X substantially improve the melt processing of high-
and neutralization (i.e., deactivation) in acidic media. Therefore, molecular weight HDPE resins by establishing a low surface
careful selection of the HAS is required to optimize both energy between the polymer melt and the metal die surface of the
functionality (i.e., activity) and hydrolysis resistance, as shown extruder. This means that the HDPE can be processed by
schematically in Figure 2.11. extrusion at high speeds without melt fracture and avoid die
build-up that can otherwise lead to lacings, defects, and pin holes
in the geomembrane, such as those shown in Figures 2.12 and
2.13.
193
conversion processes (e.g., extrusion, internal mixing, milling,
calendaring) of the polymer to a finished product.
Antioxidants are used to prevent accelerated oxidative
degradation (i.e., autooxidation) of the polymers during
processing. The antioxidants retained after manufacturing
protect the geomembrane for much of the subsequent service life.
The Std-OIT (sometimes simply referred to as OIT) and HP-OIT
index tests provide complementary information regarding the
antioxidant and stabilizer systems in the post-manufactured
geomembrane. A standard OIT (Std-OIT) at 200oC (ASTM
D3895) measures the effectiveness of retained antioxidant levels.
To detect the retained stabilizer content, an HP-OIT at 150oC
Figure 2.12 Lacing hole defect in an HDPE geomembrane caused by die (ASTM D5885) is performed.
build-up. Std-OIT can only detect antioxidants that are effective at the
200oC test temperature, such as hindered phenolic and hindered
phosphites that are used to inhibit oxidation of the geomembrane
during the extrusion processing of the liner, welding of the liner,
and the geomembrane service life. However, stabilizers such as
hindered amines stabilizers (HAS) and thiosynergists are not
active at these test temperatures (i.e., around 200oC) and
therefore cannot be detected by the Std-OIT test. For this reason,
the HP-OIT test run at 150oC to detect these stabilizers (Scheirs
2009).
2.1.4 Specifications for Additives in HDPE Geomembranes
It is difficult to write a specification for HDPE geomembranes
Figure 2.13 Schematic of die build-up causing lacing hole defects in an
HDPE geomembrane
around a particular additive or group of additives, because the
additive package recipes used by manufacturers are proprietary.
Another advantage of a polymer processing aid in HDPE Moreover, due to the cost or availability of additives together
extrusion is that it greatly reduces polymer build-up in dead spots with ongoing R&D in geomembrane formulations, the additives
in the extruder, such as on the screw and barrel. Such dead spots are subject to change on a regular basis. Also, for reasons such
create polymer hang-ups that degrade and become carbonized as those discussed in the previous subsections, a geomembrane
due to their excessive residence time in the extruder. Periodically that performs well in one environment may perform badly in a
these degraded particles get swept back in the mainstream flow quite different environment depending on the compatibility of
of the melt and become embedded in the final geomembrane as the resin and additives to the two environments. Standard de
hard nodular defects such as the carbonized lump defect shown facto minimum requirements such as GRI-GM13 serve as a
in a microtomed section of HDPE geomembrane in Figure 2.14. useful guide in noncritical situations. However, since the end-
user is generally unaware of what additives are used in any
geomembrane, the only way of assessing the compatibility of a
geomembrane with a given environment or suitability for
applications where a long service life is essential is by
accelerated testing. For example, the selection of a geomembrane
for a 550-year service life in a low-level radioactive disposal
facility is described by Rowe et al. (2020) and that for a 20-year
service life in a pond for a highly acid liquid subject to high
temperature and UV exposure is described by Folwell et al.
(2021).
2.2 Geomembrane manufacture and design implications
HDPE geomembranes have good barrier properties due to their
Figure 2.14 Carbonized polymer defects in an HDPE geomembrane.
relatively high degree of crystallinity, where the crystalline
regions retard the diffusion of molecules significantly more than
2.1.3 Loss of Additives During Service
the amorphous regions. For this reason, HDPE is favoured as a
Different HDPE geomembrane manufacturers use different
landfill barrier over LLDPE which has a lower degree of
blends, ratios and levels of antioxidants and stabilizers. The
crystallinity. To achieve high crystallinity and high strength,
antioxidants are based on different hindered phenolic
HDPE or MDPE resin of relatively high molecular weight is used.
antioxidants, hindered phosphites and thioesters, while the
The melt flow index (MFI also known as the melt flow rate, MFR)
stabilizers are based on hindered amines (HAS). These
is an index for the average molecular weight of the polymers in
protective stabilizing additives are depleted by five main
the resin. Lower MFI implies higher average molecular weight.
mechanisms, namely:
The flow rate ratio (FRR) of the MFI obtained under two
(i) leaching (by liquids),
different pressures (e.g., extruded with an applied load due to a
(ii) migration (to interface),
mass of 21.6 and 2.16 kg) is an indirect measure of the breadth
(iii) consumption (by free radicals),
of the molecular weight distribution.
(iv) deactivation (by chemicals),
The two main processing techniques for manufacturing
(v) and neutralisation (by acid/base interactions).
HDPE geomembranes are flat bed (slot die) extrusion (FBE) and
A geomembrane's time to nominal failure (see §1.1) and
vertical blown film extrusion (BFE). The two processes require
ultimately, it’s service life depends on the specific resin and the
resin with different rheology, as reflected by the MFI and FRR
composition and amount of their antioxidant/stabilizer package
requirements. Both processes can produce high-quality
added. The potential for oxidative degradation of polymers is
geomembranes. However, different defects and challenges can
most pronounced during the high-temperature, high-shear
194
arise for the two processes, and these should be understood by The decrease in SCR from SCRo to SCRm is not degradation.
the design engineer. Instead, it is the geomembrane transitioning from an unstable
Flat bed extrusion (FBE) involves the molten mix of resin and thermodynamic state after its manufacture and shipping to the
additives being forced to spread laterally from the injection site to a stable state representing the real material property
points through two horizontal die lips. The die lip gap controls available to the designer before degradation. It is the value of
the thickness. Therefore, the rheology of the mix must be such SCRm that is relevant to the design. A reasonable estimate of
that it is spread uniformly (without segregation) across the die SCRm can often be obtained by ageing the geomembrane for 3
width from the injection points. This typically implies a lower months at 55 to 65°C. For some geomembranes, the SCR after 3
average molecular weight and narrower molecular weight months of ageing at this temperature will exceed SCRm.
distribution (i.e., higher MFI and lower FRR) than a BFE However, it is still a much better estimate of the representative
geomembrane. If the mix does not spread uniformly without SCR than SCRo.
segregation, it can differ in properties across the roll width. Although most of the geomembranes examined by Rowe et
Blown film extrusion (BFE) is one of the most used methods al. (2019) were blown film, the one flat bed geomembrane
of manufacturing HDPE geomembranes. The resin is first melted examined also experienced a significant decrease from SCRo to
by subjecting it to heat and pressure inside the barrel of an SCRm. The geomembranes tested had an average SCRm/SCRo =0.
extruder and then forced through a narrow annular slit in the die 37±0.1 (i.e., a standard deviation of 0.1; 70% of geomembranes
to control its thickness. Immediately after being extruded, the tested had 0.27 < SCRm/SCRo < 0.47) .
BFE process material must be pulled upwards from the die. The Another feature of the BFE process is the folding of the
resin requires a higher molecular weight (lower MFI) and geomembrane between the nip rollers before it is cut and placed
broader molecular weight distribution (higher FRR) than the on the roll for delivery to the site. This fold, located at the quarter
FBE process to support the weight of the vertical bubble. The points with half-a-roll-width between them, serves as a location
quality of the BFE product is controlled by the following: for the formation of shape wrinkles when the geomembrane is
Output rate: In blown film extrusion, screw speed most heated by solar radiation in the field (Figure 2.15). FBE
controls the output. Generally, higher screw speed implies geomembranes do not have this preferred location for wrinkling.
higher output. However, the coefficient of thermal expansion is essentially the
Blow-up ratio: The blow-up ratio is described as the ratio same for both BFE and FBE geomembranes, and hence the total
of bubble diameter per die diameter. It defines how much amount of wrinkling is also essentially the same. Therefore, there
the bubble expands in the transverse direction. is no advantage to either one or the other method of manufacture
Neck-point height: Neck-point height is usually expressed from the perspective of wrinkles. In both cases, the
geomembrane must be covered at a time when the wrinkles are
by the distance from the die surface to the point where the
minimal (e.g., evening or very early morning).
necked diameter of the bubble becomes the shortest.
Thickness is controlled by the take-up speed, which is
controlled by rotational speed primary nip rollers at the top
of the blown film column since the lip opening is fixed.
This BFE process may result in greater strength anisotropy
than the FBE process. However, this is usually not an issue for
design since HDPE must be designed for very small strains (well
below the yield strength) and break strength is a quality control
and not a design parameter. Figure 2.15 Aerial photo of wrinkles arising at quarter points from
A change in take-up speed and the consequent change in the fold in blown film geomembrane running parallel to the roll
cooling rate also affects crystallinity, yield strength, and internal
stresses in the geomembrane post-manufacture. Thus, changing 2.3 Problems with high levels of additives
take-up speed can result in different initial stress crack resistance,
SCRo, for the same resin. For example, four HDPE The ever-increasing drive to reach greater durability with HDPE
geomembranes made sequentially from the same resin with only geomembranes has meant increased loadings of antioxidants and
the take-up speed changing to change thickness had SCRo of 680 stabilizers are being incorporated into the polymer. While this
hours (1 mm), 800 hours (1.5 mm), 950 hours (2 mm) and 1100 leads to very high Std-OIT values and HP-OIT values and >90%
hours (2.4 mm) (Ewais and Rowe 2014). Rowe et al. (2019) retention of HP-OIT after oven ageing it has also resulted in
attributed the differences in SCRo in part to morphological increased problems with waxy surface layers and additive
differences but primarily due to different residual stresses – a blooming and, consequently, with fusion welding.
result of the different cooling rates. When the residual stresses Blooming, or exudation as it is technically known, is the
were allowed to relax to the representative value, there was no diffusion of the additives from the HDPE to the surface of the
statistically significant difference in the representative stress liner to form a greasy/waxy layer on the liner surface. Such
crack resistance, SCRm, of the four geomembranes with SCRm ~ surface residues can interfere with welding of the geomembrane.
290 hours (giving a ratio of 0.26 < SCRm/SCRo<0.43). Based on During welding, the additives and polymer waxes that have
a study of many HDPE geomembranes, Rowe et al. (2019) found bloomed from the HDPE accumulate on the hot wedges of the
that generally the initial SCRo does not provide a meaningful welding machines and then are pushed laterally to give excess
representation of the materials representative stress crack squeeze-out and abnormally large squeeze-out areas.
resistance, SCRm. From this study, the representative stress crack Excess blooming of HDPE geomembranes can be detected by:
resistance is between 17 and 52% of the initial post-manufacture (i) the appearance of a whitish visible line on upon
value SCRo (i.e., m/SCRo dragging a fingernail along the surface of the liner.
0.37). This is consistent with observations of a decrease in SCR (ii) observing excessive smoking and blackening of the
with time for a geomembrane stored away from sunlight at 20oC. brass wedges of the welding machine.
Rowe and Ewais (2015) reported that the SCR of an HDPE (iii) an abnormally wide and/or thick squeeze-out zone.
geomembrane reduced from SCRo=550 hours to SCR=330 hours (iv) a more prominent than normal amine peak in the FTIR
after 4 years of storage in the laboratory, and 190-195 hours after at 1560 cm-1.
6 years of field exposure without any evidence of degradation in The Blooming of HAS is due to their relatively low solubility
either case. in polyolefin matrices and the HAS reaching their saturation
solubility in HDPE starting when levels exceed 0.5 wt.% (5000
195
ppm) and approach 1 wt.% (10,000 ppm). Also, a base polymer exothermic peaks are typically observed during the thermal
with high levels of cyclohexane extractables can contribute to the oxidation in the Std-OIT test of conductive geomembranes
excess squeeze-out problem by increasing the level of oligomers (Figure 2.17, presumably due to the conductive layer that may
and waxes in the polymer. not be homogenously distributed across the entire geomembrane
Blooming and exudation are most prevalent with surface area). To minimize such local variations of stabilizer
geomembranes that are highly stabilized with AO and HAS. concentration and the difference in the formulation of the
Their limited solubility in the PE means that the excess diffuses different layers of these geomembranes, ASTM D5885 and
(blooms) to the surface of the geomembrane, particularly with D8117 recommend homogenization of the sample before OIT
hot storage. During wedge welding, the excess molten polymer analysis. Alternatively, individual layers may be separated from
mixed with waxes and additives is laterally extruded and the sample using thick microtomes to remove or separate a
deposited on the liner adjacent to the weld tracks. When particular layer as a whole and analysed for OIT separately.
blooming causes excess squeeze-out, the squeeze-out may retain Mechanical properties, especially tensile break strength/
high Std-OIT and HP-OIT values due to its excess additives elongation (ASTM D6693) and stress crack resistance (ASTM
despite more heating than the parent polymer. Regular cleaning D5397) are typically used to assess the initial properties of a
of the wedges reduces the incidence of excessive squeeze-out given geomembrane and for assessing the degradation beyond
due to blooming but at the expense of reduced welder antioxidant depletion.
productivity. The variable thickness resulting from the texturing in
Possible preventative measures to reduce excess squeeze-out coextruded nitrogen textured BFE geomembranes or the
arising from excess wax/additive migration include: variation of formulation in multilayered geomembranes can give
Testing the geomembrane before welding for the presence a large coefficient of variation in break properties.
of waxy layers of bloomed additives and waxes (see ATR- The SCR test involves preparing specimens with a prescribed
IR method below). notch that should produce a ligament thickness equal to 80% of
Using an ethyl acetate solvent to clean the flaps and the nominal thickness of the specimen (ASTM D5397). However,
remove excess bloom and waxes before welding. this procedure cannot be used to assess the SCR of textured
Regularly cleaning of the brass hot wedges but this is at the geomembranes due to the variation of the ligaments obtained
expense of welder productivity. when notching the textured part of the roll. Alternatively, the
Setting a specification for the maximum acceptable level SCR of textured geomembranes manufactured with a single base
of exuded polymer/additives for a geomembrane. resin in all layers is assessed using the roll's smooth edge to
Figure 2.16 shows the various infra-red absorption bands of minimize the SCR's variability. However, in this case, the skin
typical additives used in HDPE geomembranes. Since all these will be minimal or missing in the smooth edge sample relative to
additives have characteristic and distinct IR absorption peaks, the textured geomembrane.
the surface-sensitive analysis technique of ATR-IR can be used
to identify the presence and levels of migratory additives in
HDPE that have exuded from the geomembrane.
Figure 2.17 Differential scanning calorimetry thermogram showing
the multiple exothermic peaks in textured conductive geomembrane
.
For multilayered geomembranes coextruded with different
polymer formulations, the direction of the notch (top or bottom
Figure 2.16 IR Absorbance bands are detectable for blooming surface) can affect the SCR since the notch can eliminate one or
additives from HDPE geomembranes (courtesy of ThermoFischer more layers of the geomembrane, leaving the crack to propagate
Scientific). through the other layers of the geomembrane. Therefore, ASTM
D5397 test method recommends that the geomembrane be
2.4 Testing of multilayered geomembranes notched on the functionalized side in case of geomembranes
These layers have the potential to complicate traditional methods formulated with a single functionalized external layer or notched
of monitoring material behaviour. For example, standard on the side exposed to the sun in the field in case of
oxidative induction time (Std-OIT; ASTM D 3895) and high geomembranes with two functionalized external layers
pressure oxidative induction time (HP-OIT; ASTM D 5885) tests (Appendix X2 in ASTM D5397).
are typically used to assess the initial OIT values and An alternative method to assess the SCR, especially for
antioxidants/stabilizer depletion time for the aged geomembrane textured multilayered geomembranes, without eliminating any
specimens. For white and/or conductive geomembranes, skin layers is using plaques to homogenize the different layers
assessment of the OIT values can be more challenging relative according to ASTM D4703 Procedure C. In this method, two
to traditional black geomembranes. Rowe et al. (2020) showed a specimens of the geomembrane are faced together, then heated
notable variation of the OIT values between the textured white to above the normal melt temperature of the geomembrane, and
and conductive geomembranes and their smooth edges provided finally squeezed together under pressure. The plaque is then
for welding purposes in the field. In addition, multiple cooled and removed from the mould. Rowe et al. (2020)
196
compared SCR of samples taken directly from the roll to plaques 3.2 Puncture resistance of BGMs
of a smooth white using ASTM D5397. They found that the SCR
was more than 70% greater for the plaque specimens (>1000 The short-term puncture resistance is a key aspect of the BGM
hours) than the 590 hours obtained by the standard method by properties that governs the BGM performance as a hydraulic
the same laboratory. Such difference in the SCR can be attributed barrier in high stress applications. The cylinder test is typically
to the annealing of the geomembrane during the preparation of performed to assess gravel induced short-term damage in the
the plaques that may increase their SCR (Abdelaal et al. 2015). geomembranes (e.g., ASTM D5514; Shercliff 1998; Rowe et al.
This suggests that the plaque method does not give a 2013; Lupo and Morrison 2007). Previous investigators (e.g.,
representative SCR of the material taken directly from the roll Blond and Breul 2014; Lazaro and Bruel 2014; Clinton and
and tested (Rowe et al. 2020). Rowe 2017) used a similar setup to assess the compatibility of
While multilayered geomembranes can provide designers BGMs with the site-specific materials such as the gravel
with many desirable characteristics, assessing their properties drainage layer or subgrade by assessing their puncture resistance
(especially due to ageing) is an issue requiring further research. and/or the tensile strains induced under excessive overburden
pressures simulating different applications (Figure 3.2).
Blond and Breul (2014) examined different BGMs with
different thicknesses under both hydrostatic stresses (stresses
induced due to high water heads) and mechanical stresses
3 BITUMINIOUS GEOMEMBRANES (drainage layer is installed on top of the geomembrane with low
hydraulic heads). Depending on the BGM material, they
3.1 Introduction observed failure in the BGM under mechanical stresses between
2.7 MPa and 5.2 MPa while failure of the different BGMs took
Bituminous geomembranes (BGMs) comprise a thick nonwoven place under hydraulic stresses between 0.64 MPa and 1.0 MPa
geotextile impregnated with bitumen to give an essentially only. They concluded that puncture resistance of the BGM is
impervious yet flexible sheet (Figure 3.1). The bitumen used in significantly higher under mechanical stress than hydrostatic
the production of BGMs is either oxidized or, more recently, pressures. Another study by Clinton and Rowe (2017) examined
stabilized with elastomers such as styrene-butadiene-styrene punctures and strains for a 4 mm BGM under an overburden
(SBS). These bitumen modifications reduce its temperature pressure of 2.0 MPa using overliner and underliner materials
sensitivity and make it suitable for geomembrane applications simulating heap leaching applications. A direct comparison
(Peggs 2008; Touze-Folz and Farcas 2017). In some BGM was made between the BGM and a 1.5 mm HDPE and showed
products, a layer of glass fibre fleece is used to provide more physical damage to the BGM than the HDPE (in the form
dimensional stability during manufacturing and can increase the of holes). Under these conditions under these conditions, they
strength and puncture resistance (e.g., Bannour et al. 2013; observed hundred and 15 holes in the 60 cm diameter sample
Blond and Breul 2014). The surface layer of the BGM is usually (Figure 3.3). However, a subsequent leakage test showed that
sanded to increase the interface shear strength of BGMs to even though the BGM had more holes than the HDPE, its leakage
improve stability on side slopes (El-Dana and Breul 2010). The was lower, suggesting that the gravel may have sealed many of
bottom layer of the geotextile usually has an anti-root the holes in the BGM.
polyethylene backing film. BGMs are relatively thick (3–5 mm)
and are claimed to withstand installation-related mechanical
damage. They have also been used for cold regions applications
(e.g., lining Diavik Diamond Mine in Canada's Northwest
Territories).
Figure 3.1 BGM different layers (AXTER COLETANCHE, 2009)
Historically the use of BGMs has been largely limited to use
in dams, canals, potable water reservoirs, and other low risk
environmental applications. Due to their ease of field seaming
and suitable mechanical and physical properties (such as high
mass per unit area, high puncture resistance, relatively high
density, and low coefficient of thermal expansion; Peggs 2008),
BGMs are often considered as a strong candidate for the liner
materials in different geoenvironmental applications. Today, Figure 3.2 BGM puncture tests; (a) application of hydrostatic stresses;
BGMs have been used, or introduced as candidates, to line (b) Mechanical stresses (modified from Blonde and Breul 2014).
municipal solid waste (MSW), low-level radioactive waste
(LLRW) landfills (Breul et al. 2006), and more recently mine
tailing ponds and heap leach pads (Peggs 2008; Lazaro and Breul
2014).
197
to ensure their proper use in many different applications to
provide the desired environmental protection.
4 ENHANCED BENTONITE GEOSYNTHETIC CLAY
LINERS
4.1 Conventional geosynthetic clay liners
Geosynthetic clay liners (GCLs) are thin (~4- to 10-mm thick),
factory manufactured barriers commonly used for hydraulic and
chemical containment applications, including municipal and
hazardous solid waste landfills, surface impoundments (e.g.,
ponds and lakes, aeration lagoons, fly ash lagoons, tailings
impoundments), canals, storage tanks, and secondary
Figure 3.3 BGM sample after a cylindrical test at 2 MPa showing 115 containment of above-grade fuel storage tanks (e.g., Petrov et al.
holes in the 60 cm diameter sample (modified from Clinton and Rowe 1997a,b; Rowe 1998, 2005; EPA 2001; Bouazza 2002; Koerner
2017).
2005; Shackelford 2008; Bouazza and Bowders 2010; Scalia et
al. 2018a; Norris et al. 2020). GCLs often are used in lieu of
3.3 Chemical durability of BGMs
compacted clay liners (CCLs) due to lower cost, ease of
In the long-term, examining the chemical compatibility of BGMs installation, and minimal consumption of airspace (e.g., Koerner
with the service conditions in different applications is essential and Daniel 1995; Daniel et al. 1997; Bouazza 2002). The low
to ensure their longevity to meet the long service life requirement hydraulic conductivity of conventional GCLs to water, which
of waste containment applications. Ageing of the bitumen has typically is on the order of ~1×10-11 to 3×10-11 m/s, yields a
been extensively studied as the key factor to determine the barrier with a leakage rate that is equal to or lower than that of
lifetime of asphalt pavement and showed that oxidation can CCLs. The low hydraulic conductivity of conventional GCLs
happen as the result of photo-oxidation (UV) and/or thermal that do not include a geomembrane component is attributed
oxidation (Hofford 2015). However, the ageing of a BGM is exclusively to the natural bentonite within the GCL.
more complex since its different components may interact
differently with ageing conditions. This may have different 4.1.1 Swelling behaviour, of natural (unenhanced) bentonite
implications on the different properties of the geomembrane (i.e., Swelling of the montmorillonite mineral fraction of the bentonite
mechanical versus waterproofing properties). This was shown in upon hydration is the primary mechanism controlling the hydraulic
a study by Touze-Foltz and Farcas (2017) when investigating the conductivity of bentonite in GCLs. The bentonite granules (or
performance of aged elastomeric BGMs exhumed from two particles for GCLs with powdered bentonite) swell, filling the
ponds 6, 10, 15, 20, and 30 years after installation. While the intergranular pores that conduct the bulk of the flow (Tian et al.
tensile test results indicated that the mechanical properties of the 2016b, 2019). Bentonites in GCLs that are hydrated with liquids
core reinforcement were not affected after 30 years, that promote swelling have narrow and tortuous pathways that
specimensaged over 15 years showed complete consumption of yield low hydraulic conductivity (A in Figure 4.1), whereas
the polymer and oxidation of the bitumen within the first few hydrating liquids that suppress swelling lead to larger and less
microns from the surface of the BGM. Esford and Janssens (2014) tortuous flow paths and much higher hydraulic conductivity (B in
incubated BGMs in a diluted acidic solution (20% sulfuric acid Figure 4.1) (Jo et al. 2001; Kolstad et al. 2004; Scalia and Benson
+100 ppm kerosene) for 3 years at room temperature which was 2011; Tian et al. 2016a, 2019; Chen et al. 2018).
a much more aggressive chemical exposure. The results showed The swelling of bentonite is intimately tied to interlayer
that the BGM maintained its properties within acceptable limits, swelling in the montmorillonite mineral (Norrish 1954).
and hence they concluded that BGM could be suitable to be used Montmorillonite is a 2:1 aluminosilicate clay mineral, meaning
as a primary geomembrane for ponds. However, at the elevated that the crystalline structure of the mineral comprises sequential
temperatures (like those recorded for the geomembrane liners in
layers, also referred to as lamellae or platelets, comprising a single
some geoenvironmental applications), bitumen as a viscoelastic
sheet of aluminium (Al) octahedra sandwich between and
material is expected to exhibit significant changes in its
chemically bonded to adjacent sheets of silica (Si) tetrahedra, often
properties (Touze-Folz and Farcas 2017) while all components
of the BGMs may exhibit thermal degradation affecting the referred to as the TOT structure. As illustrated in Figure 4.2, these
overall properties of the BGM. This was shown by Addis et al. TOT layers stack to form individual particles or tactoids. Although
(2016), reporting a large-scale failure of the BGM liner in a the schematic depiction in Figure 4.2 implies that an individual
copper/gold mine disposal facility occurring several months after clay particle comprises only two layers, the number of layers for
installation. The thermal Imaging showed that the surface any given particle varies and can be substantially greater than two
temperature of the exposed BGM reached 86oC and the tensile (as will be demonstrated subsequently). For example, based on the
tests conducted on exhumed specimens showed a decrease in the fabric boundary surface (FBS) theory that considers the
tensile strength, especially at the seams. It was concluded that a microstructure of smectite-rich clay soils to be a dual-porosity
combination of the elevated temperatures and tensile stresses medium comprising both interlayer, inaccessible nanometer-size
caused BGM degradation and failure. Combining the effect of pores and intertactoid, accessible micropores, the average
elevated temperatures and chemical exposure, Samea and number of layers or lamellae per tactoid, Nl,AV, is given by the
Abdelaal (2019) examined the performance of BGMs using following expression (Manassero et al. 2018; Manassero 2020):
immersion tests in synthetic mining solutions at pH=0.5 and
pH=13.5, simulating the pregnant leach solution (PLS) from cs cs
heap leaching applications, and 55oC. They showed that the Nl , AV Nl , AV 0 1 em 1 exp (4.1)
BGMs exhibited more degradation in the high pH solution em c0 c0
compared to the low pH solution. However, further research is
still needed to investigate the long-term performance of BGMs where cs is the salt concentration of the equilibrium bulk solutio
198
n (pore water), c0 is the reference salt concentration (= 1 eq/L), em sheet, the distance between the source of the negative charge and
the location of the exchangeable cations is relatively large, such
that the bound cations are free to exchange with those in the bulk
pore water always provided that electroneutrality is maintained.
For example, two initially bound Na+ ions can be replaced by one
Ca2+ ion, which is the reason these electrostatically bound cations
are referred to as exchangeable cations. Also, the cation that
dominates the exchange complex of the montmorillonite dictates
the type of montmorillonite and, therefore, the type of bentonite
(e.g., Gleason et al. 1997). For example, when the predominant
exchangeable cation is Na+, the montmorillonite is referred to as
sodium montmorillonite (Na-montmorillonite) and the associated
bentonite is referred to as sodium bentonite (Na-bentonite).
Conversely, when the predominant exchangeable cation is Ca2+,
the montmorillonite is referred to as calcium montmorillonite (Ca-
montmorillonite), and the associated bentonite is referred to as
calcium bentonite (CaB).
In addition to the exchangeable cations, immobile water
Figure 4.1. Swelling of sodium bentonite (NaB) granules and
molecules also are bound to the surfaces of the mineral solid phase
corresponding hydraulic conductivity.
in the interlayer zone. Although the water molecule, H2O, is charge
is the bentonite micro-void ratio. The dimensionless parameters neutral, the polar nature of H2O imparts charged behaviour. The
Nl,AV0 ( 1), ( 0) and ( 0) have to be adjusted based on a volume of the bound water molecules dictates the thickness of the
given set of test results for the same clay soil. The FBS theory interlayer space and, therefore, the basal spacing, d001, between
(Eq. 4.1) provides an assessment of the microstructure under successive layers (see Figure 4.2). The hydration or crystalline
different chemico-mechanical boundary conditions and swelling phase, the extent of which is directly related to the
illustrates the primary parameters affecting Nl,AV. Individual clay magnitude of d001 (Norrish 1954; Jo et al. 2001), occurs as the
particles also can accumulate or aggregate to form larger granules montmorillonite is initially hydrated from a dry state and continues
or aggregates of clay particles. Note that the terminology used to until there are several monolayers of bound water molecules. The
describe the various components of individual montmorillonite osmotic swelling phase occurs subsequently and depends on the
particles and aggregates of particles as well as the associated pore cation concentration in the bulk pore water, provided that the
spaces between the layers, particles, and aggregates can vary exchangeable cations are predominantly monovalent. When the
significantly in the literature (e.g., compare Jo et al. 2006; Likos bulk pore water comprises a relatively low concentration of
and Lu 2006; Likos and Wayllace 2010; Acikel et al. 2018; monovalent cations, and the exchangeable cations are monovalent,
Manassero 2020). osmotic swell occurs such that the fraction of immobile, bound
water and the corresponding swell volume is relatively high,
Si Tetrahedral Sheet Layer resulting in relatively low hydraulic conductivity because the
Al Octahedral Sheet (Lamellae) or conductive interparticle (intertactoid) and intergranular pores are
001 Si Tetrahedral Sheet Platelet
Particle
smaller, and the flow paths are more tortuous. This situation occurs,
H2O + Exchangeable
Interlayer or for example, when Na-bentonite is hydrated and permeated with
Cations deionized water (DW).
Layer Tactoid
Si Tetrahedral Sheet Conversely, only the crystalline or hydration swelling phase
Al Octahedral Sheet (Lamellae) or
Platelet
occurs when the bulk pore water comprises a relatively high
Si Tetrahedral Sheet
concentration of cations, and the exchangeable cations are
Figure 4.2. Schematic illustration of 2:1 layered crystalline structure of multivalent. As a result, swelling is minimized, and the resulting
natural montmorillonite. hydraulic conductivity is relatively high. This situation occurs
when CaB is hydrated and/or permeated with electrolyte solutions
The interlayer space between consecutive layers is occupied by
with significant concentrations of multivalent cations. For example,
both water molecules (H2O) and exchangeable cations, which
the free swell of CaB upon hydration with DW typically is on the
typically include sodium (Na+), potassium (K+), calcium (Ca2+),
order of three millilitres per two grams of dry bentonite (3 mL/2
and magnesium (Mg2+) to varying extents. The cations are held or
g), whereas the free swell of Na-bentonite hydrated with DW
bound electrostatically to the surfaces of the layers due to a net
typically is on the order of 30 mL/2 g (Jo et al. 2001; Kolstad et al.
negative charge located within the Al octahedra sheet as a result of
2004).
the process of isomorphic substitution, whereby the Al3+ within an
aluminium octahedron is replaced by a lower charged cation, such 4.1.2 Chemical compatibility issue with bentonite
as Mg2+, during crystalline formation of the montmorillonite. The The aforementioned crystalline swelling occurs within the
magnitude of the negative charge or surface charge deficiency tactoids (i.e., within the interlayer pore space), and is mostly
increases with the number of substituted Al3+ in the Al octahedra controlled by the hydration potential energy of the interlayer
comprising the sheet, and generally, it varies on the order of 0.5 to cations, which is insensitive to the salt concentration of the
1.2 equivalents of charge (eq) per unit cell, with an overall average equilibrium bulk solution. In contrast, osmotic or double-layer
of 0.66 eq per unit cell (Mitchell and Soga 2005; Lee and swelling occurs between the tactoids (i.e., in the intertactoid pore
Shackelford 2005), which correlates to about 1.1 to 1.9 eq/m2 of space). This is driven by a difference in the osmotic pressure
surface area (Jury et al. 1991). The magnitude of this negative between the mobile pore water and the external bulk solution, the
charge dictates the number of equivalent positive charges magnitude of which depends on the concentration and valence of
associated with the exchangeable cations external to the crystalline the cations in the mobile pore water (e.g., Laird 2006). Thus,
structure required for electroneutrality. The total number of when Na-bentonite is hydrated or permeated with a liquid
equivalents associated with the sum of all exchangeable cations is containing a high concentration of ions, osmotic swelling of the
commonly referred to as the cation exchange capacity (CEC). montmorillonite is suppressed, increasing k. The increase in k can
Because isomorphic substitution occurs in the Al octahedral range from one to several orders of magnitude depending on
199
several factors, including the ionic strength, I cizi2, where ci interchangeably and sometimes “polymer modified” is used as a
and zi are the molar concentration and associated charge of ionic broad category to cover all modified bentonites, whereas
species, i, summed over all ionic species in solution), the ratio of “enhanced bentonite” represents a special category of modified
monovalent-to-multivalent cations of the hydrating/permeant bentonites. The latter approach is adopted in this paper because the
liquid, and the confining stress or void ratio, e, of the bentonite, role of the modification or enhancement may differ. In some cases,
where e = n/(1 – n) and n is the total porosity (Jo et al. 2001; low-quality bentonites are chemically modified by the addition of
Kolstad et al. 2004). polymers simply to improve the swell capacity or kw such that they
For example, consider the hydraulic conductivity ratio, k/kw, perform similarly to higher quality natural Na-bentonite in
shown in Figure 4.3 for two conventional GCLs permeated with standard index tests such as ASTM D 5890 and D 5887 (e.g., Haug
calcium chloride (CaCl2) solutions. As indicated, the ratio k/kw, and Boldt-Leppin 1994) and may be regarded as “substitution
defined as the ratio of k when the GCL is permeated by a particular bentonites” since the primary objective is to meet the standard
chemical solution to that when permeated by water, kw, increases specification for Na-bentonite. The term enhanced bentonites as
with increasing CaCl2 concentration for both GCLs, with k/kw used herein excludes these bentonites. Enhanced bentonites are
eventually reaching a value based on permeation with a 500 mM defined herein as natural bentonites that are chemically modified
CaCl2 solution ranging from 2.7 orders of magnitude for GCL1 to to achieve low k to solutions that alter the hydraulic conductivity
4.9 orders of magnitude for GCL2. Note that kw for GCL1 and of natural Na-bentonites.
GCL2 were 2.4×10-11 m/s and 7.0×10-12 m/s, respectively, such
that the respective k values based on permeation with 500 mM 4.2.1 Enhanced bentonites
CaCl2 were ~1.2×10-8 m/s and ~5.7×10-7 m/s. Chemicals that commonly have been used for bentonite
enhancement include anionic polymers such as polyacrylate and
2+ polyacrylamide (e.g., Heller and Keren 2002; Scalia et al. 2011;
Ca Concentration (mg/L)
Bohnhoff 2012; Scalia 2012; Bohnhoff and Shackelford 2013;
0.01 0.1 1 10 100 1000 10000
Bohnhoff et al. 2014; Chen et al. 2019; Prongmanee et al. 2018a,b;
w
100000
Hydraulic Conductivity Ratio, k/k
Scalia et al. 2018a,b; Prongmanee and Chai 2019; Chai and
10000 GCL1 Prongmanee 2020; Li et al. 2021), and carboxymethylcellulose
GCL2 (e.g., Di Emidio 2010; Di Emidio et al. 2015; De Camillis et al.
1000 2014, 2016; Fan et al. 2020), cyclic organic carbonates such as
propylene carbonate (Onikata et al. 1999; Katsumi et al. 2008;
100 MDL for Ca 2+ Mazzieri et al. 2010) and glycerol carbonate (Fehervari et al.
(~water) 2016a,b; Gates et al. 2016), and other organic compounds such as
10 ethylene glycol (Szczerba and Kalinichev 2016) and quaternary
ammonium cations (QACs) such as hexadecyltrimethyl-
1 ammonium and benzyltriethylammonium (Soule and Burns 2001;
0.00010.001 0.01 0.1 1 10 100 1000
CaCl Concentration, c (mM)
Burns et al. 2006).
2 Bentonites modified with QACs, known as organobentonites
Figure 4.3. Ratio of the hydraulic conductivity based on permeation with or organoclays, are used primarily as a sorptive medium for
calcium chloride (CaCl2) solutions, k, relative to that based on permeation organic compounds (e.g., Lee et al. 2012; Di Emidio et al. 2017)
with water, kw, as a function of the CaCl2 concentration for the permeant or as hydraulic barriers in GCLs or other configurations
solution for two GCLs [MDL = Method Detection Limit] (data from Lee specifically to contain nonaqueous phase liquids (Lee et al. 2012;
and Shackelford 2005). Benson et al. 2015). These enhanced organobentonites will not be
The k values shown in Figure 4.3 are based on a low effective discussed further herein.
stress (35 kPa) that may not represent the stresses encountered in A summary of some of the more commonly studied enhanced
many applications. Lower k values are anticipated at higher bentonites is provided in Table 4.1. However, the general category
effective stresses (e.g., see Petrov et al. 1997a,b; Petrov and Rowe of contaminant-resistant clays (CRCs) pertains to commercial
1997; Rowe 1998; Shackelford et al. 2000; Katsumi and Fukagawa products of which the modifying or chemical enhancement agent
2005; Thiel and Criley 2005; Bradshaw et al. 2013; Norris et al. and process are proprietary and, therefore, unknown. As a result,
2020). Nonetheless, the results shown in Figure 4.3 indicate that although CRCs are considered herein, little detail of these
the k of GCLs can be altered when permeated with aqueous enhanced bentonites can be provided.
solutions of moderate to high ionic strength and going beyond
Table 4.1. Some enhanced bentonites (modified from Scalia et al. 2018a)
Figure 4.3 the effect has been found to vary with the chemical
species in the solution as well as concentration (e.g., Fehervari et Name of bentonite Additive(s) or modification
al. 2016a,b; Chen et al. 2018; Xu et al. 2018; AbdelRazek and Bentonite polymer Sodium polyacrylic acid (Na-PAA)
Rowe 2019) and/or pH (e.g., Shackelford et al. 2000, 2010; composite (BPC) and other dry mixtures of bentonite
Benson et al. 2010b; Bouazza and Gates 2014; Athanassopoulos et and hydrophilic polymers
al. 2015; Liu et al. 2015; Ghazizadeh et al. 2018; Scalia et al. 2018a;
Dense prehydrated Sodium carboxymethylcellulose (Na-
Prongmanee and Chai 2019; Norris et al. 2020; Li et al. 2021). This
(DPH) CMC) plus Na-PAA
limitation of conventional GCLs comprising Na-bentonite has led
to the development and use of GCLs comprising enhanced HYPER clay (HC) Na-CMC
bentonites that have been modified to resist the potential for Multiswellable Propylene carbonate (PC) or glycerol
chemical attack resulting in hydraulic incompatibility (e.g., Di bentonite (MSB) carbonate (GC)
Emidio et al. 2011; Bohnhoff et al. 2013; Bohnhoff and
Shackelford 2014; Scalia et al. 2014, Scalia and Benson 2016, Di Contaminant Unknown (proprietary)
Emidio et al. 2017; Scalia et al. 2018a). resistance clay (CRC)
4.2 Polymer modified and enhanced bentonites The concept supporting the use of enhanced bentonites for
hydraulic and chemical containment applications is illustrated
The terms “polymer modified bentonite”, “polymer enhanced conceptually in Figure 4.4. As previously noted concerning Figure
bentonite”, or simply “enhanced bentonite” are sometimes used
200
4.3, the k of natural bentonite tends to increase with increasing Acikel et al. 2018; Tian et al. 2019; Norris 2021). The use of
concentration of dissolved chemicals, c, particularly multivalent chemical additives to prevent cation exchange was hypothesized
cations such as the metals. The objective of using enhanced early in the development of some enhanced bentonites (e.g., Flynn
bentonites is to provide greater resistance to hydraulic and Carter 1998; Trauger and
incompatibility, such that the relationship between k and c is supported in recent studies (Benson et al. 2014; Scalia et al. 2014;
shifted to maintain the lower k at a much higher c. Ideally, this Tian et al. 2016a,b; Chen et al. 2019). Therefore, no further
improved chemical resistance offered by the enhanced bentonite consideration of the prevention of cation exchange as the
would be sustainable such that k is independent of c. However, in mechanism underlying the hydraulic compatibility of EB-GCLs is
reality, there is an upper limit on the ability of the enhanced provided.
bentonite to sustain chemical resistance and maintain a suitably The concept of intercalation within the interlayer space of the
low k (e.g., see Scalia et al. 2014; Chen et al. 2019; Li et al. 2021). montmorillonite mineral is illustrated schematically in Figure 4.6.
Given the variability in the chemistries of liquids to be contained, Intercalation has been evaluated through determination of the basal
as well as the variability in bentonites and associated chemical spacing, d001, via x-
enhancements, the key practical consideration is to match a
specific type of enhanced bentonite to a specific liquid to be Fourier-transform infrared (FTIR) (Gates et al. 2016; Bokobza
contained such that the desired performance of the enhanced
bentonite GCL (EB-GCL) is achieved.
For example, consider the k values shown in Figure 4.5 for a amended with propylene carbonate (Onikata et al. 1999; Katsumi
conventional GCL, a natural Na-bentonite, and a polymerized
bentonite. The k values for the conventional GCL are the same as et al. 2016a,b; Gates et al. 2016) have been shown to exhibit
those for GCL2 shown in Figure 4.3 from Lee and Shackelford crystalline swelling over a wider range of solute concentrations.
(2005), whereas the results for the bentonite and polymerized Crystalline swell enhancement via intercalation is described in
bentonite are from Scalia et al. (2014). As indicated in Figure 4.5, detail in Onikata et al. (1999). Gates et al. (2016) note that other
the trend in the k of the Na-bentonite with increasing CaCl2 cyclic organic carbonates may provide similar crystalline swell
concentration of the permeant liquid is like that of the conventual enhancement. However, Onikata et al. (1999) showed that, in NaCl
GCL, with significantly higher k resulting from CaCl2 solutions with I
despite enhancement. This hydraulic compatibility enhancement is
maintained a low k over the entire range in CaCl2 concentrations consistent with the conceptual behaviour, shown in Figure 4.4.
examined.
Si Tetrahedral Sheet
Natural Al Octahedral Sheet
Bentonite 001 Si Tetrahedral Sheet
log k
Enhanced
Bentonite Si Tetrahedral Sheet
Al Octahedral Sheet
Si Tetrahedral Sheet
Figure 4.6. Schematic illustration of chemical intercalation of
log c montmorillonite crystalline structure.
Figure 4.4 Concept of improved resistance to chemical incompatibility in
hydraulic conductivity, k, with respect to the solute concentration, c, Polymer clogging of interparticle, intraparticle, and/or
resulting from the use of enhanced bentonite relative to a natural intergranular (interaggregate) pores, as illustrated schematically in
(unenhanced) bentonite. Figure 4.7, has been demonstrated to be a dominant mechanism
underlying improved hydraulic compatibility of some polymer
-6
enhanced bentonites in GCLs (Scalia and Benson 2016; Scalia et
10 al. 2018a; Tian et al. 2019; Li et al. 2021; Norris 2021). The
Hydraulic Conductivity, k (m/s)
10
-7 Conventional GCL polymer occupies the most conductive pores within the enhanced
Na-bentonite
10
-8 bentonite, forcing flow to more narrow and tortuous flow paths,
Polymerized bentonite
-9
resulting in lower k.
10
-10
10
-11
10
-12
10 MDL for Ca
2+
(~water)
-13
10
0.001 0.01 0.1 1 10 100 1000
CaCl Concentration, c (mM)
2
Figure 4.5. Effect of polymerization on the hydraulic conductivity of
and
Shackelford 2005; Scalia et al. 2014).
4.2.2 Mechanisms for enhanced bentonites
Scalia et al. (2018a) identified three potential mechanisms in
polymer enhanced GCLs that can prevent alterations in k, viz., (i)
prevention of cation exchange, (ii) intercalation or insertion of a Figure 4.7. Polymer gel clogs pores between sodium bentonite (NaB)
chemical additive within the interlayer space of the granules in a BPC-GCL (adapted from Gustitus et al. 2020).
montmorillonite mineral, and (iii) polymer clogging of
interparticle, intraparticle, and/or intergranular (interaggregate) Pore clogging was initially hypothesized as the mechanism
pores (see also Jo et al. 2006; Likos and Lu 2006; Salles et al. 2009; underlying improved hydraulic compatibility of BPC by Scalia et
201
al. (2014) based on observations of clogging of permeameter also have been used (e.g., Razakamanantsoa et al. 2012). In a
tubing by hydrogel and the lack of correlation between swelling polyanion, a negatively charged ionic group on the polymer
and k. Lack of correlation between swelling and k also has been repeating unit (e.g., negative carboxylate (COO-) and negative
observed by Tian et al. (2016a,b) and others. Further support of sulfonate (SO3-)) is satisfied by a cation to maintain
this pore-clogging mechanism was provided by Tian et al. (2016b) electroneutrality, similar to how bound cations satisfy the net
via SEM images of dehydrated polymer within the interparticle negative structural charge of the montmorillonite crystal. The
pore space of BPC permeated with DW (see Figure 4.8). Recent water solubility or degree of swelling of a polymer is attributed to
pore-scale Imaging of BPC-GCLs permeated with high ionic the solvation of these ionic groups and associated counterions
strength bauxite liquor solutions also indicates that the k of the (Buchholz and Graham 1998).
BPC-GCLs was controlled by pore-blocking with the polymer Polymers can be linear, branched, or crosslinked. Linear
hydrogel (Li et al. 2021). polymers comprise single chains of repeated monomers. Branched
A topic of current research is the degree of permanence of polymers comprise branches of repeated monomer chains.
pore-clogging polymer hydrogel within enhanced bentonites. Crosslinked polymers contain numerous repeated polymer chains
Specifically, under what conditions does pore-clogging polymer linked together at multiple crossing points. These crosslinks can be
hydrogel occur and attach to the mineral surface to render the physical (electrostatic interactions) and chemical (covalent), and
enhancement effectively permanent. Elution of polymer can be incorporated in the polymer during synthesis or may result
enhancements (both linear polymers observed as high viscosity when, for example, the negative charge from two anionic groups
effluent and hydrogels) has been reported for numerous enhanced are satisfied by a single divalent cation (termed cation bridging).
bentonites, including BPCs and CRCs (Scalia et al. 2014; Scalia Linear polycation and polyanion polymers will dissolve into water.
and Benson 2016; Tian et al. 2016a,b; Chen et al. 2019; Tian et al. Sufficient concentrations of dissolved polymers will increase the
2019). However, in some cases, this elution has only delayed the viscosity of the hydrating solution. Crosslinked polymers do not
eventual onset of high k. In contrast, in other cases, the polymer dissolve, but instead form hydrogels, which are three-dimensional
enhanced bentonite has maintained a low k after polymer elution networks of polymers that swell when hydrated and hold large
has ceased. It is important to recognize that the rate of polymer quantities of water (Wichterle and Lím 1960; Bahram et al. 2016)
elution may be controlled by the apparent viscosity of the polymer via osmotic forces similar to montmorillonite, except that the
at the shear rate imposed by the test conditions. This viscosity may montmorillonite platelets would be linked together into flexible
be orders of magnitude greater than water, giving a polymer networks that then respond to the same swell-controlling stimuli.
conductivity orders of magnitude lower than the hydraulic An example of a polyanion is polyacrylic acid (PAA) with
conductivity (to water) until water breakthrough occurs (Ejezie et negatively charged carboxyl groups. These carboxy groups are
al., 2018). Understanding elution behaviour is critical to typically satisfied by protons or sodium ions during synthesis. PAA
forecasting the long-term stability of polymer-enhanced bentonites. can be linear and aqueous miscible, and function as a flocculent
Based on several sources, Norris (2021) summarizes the three (among numerous applications) that accumulates numerous clay
potential mechanisms for anionic polymer adsorption to clay platelets via cation bridging until a floc of sufficient size is formed
platelets as (i) complexation via anion exchange, (ii) hydrogen to settle. PAA also can be crosslinked, often termed polyacrylate
bonding or ligand exchange, and (iii) cation bridging. A detailed (PA), and used as a superabsorbent in baby diapers and other
explanation of the adsorptive mechanisms between anionic applications (Buchholz and Graham 1998).
polymers and Na-bentonite, and hydrogel formation and stability, As indicated in Table 4.1, two principle polyanions used to
is provided by Norris (2021). produce enhanced bentonites in EB-GCLs are sodium polyacrylic
acid (Na-PAA) and sodium carboxymethylcellulos (Na-CMC). As
noted earlier, Na-PAA can be linear or crosslinked, although Scalia
et al. (2014) showed that the use of crosslinked Na-PAA was not
effective for enhancing the hydraulic compatibility of Na-
bentonite to high ionic strength CaCl2 solutions. These and many
other polymers have been evaluated for use in CRCs. Given the
range of different polymers available, the number of potential
polymers for use in enhanced bentonites is practically unlimited.
However, a matter that always must be considered is the potential
for degradation of the polymer, for example, by biodegradation.
Many naturally derived polymers (e.g., xanthan and
carboxymethyl-cellulose) are readily biodegradable. The
Figure 4.8. Scanning electron microscope image shows polymer hydrogel chemicals used and species released by elution and/or degradation
formation between bentonite particles/aggregates (from Scalia et al. 2018a). must be environmentally acceptable (for example, the release of
acrylamide will be unacceptable in many jurisdictions).
4.3 Enhancement chemicals
4.3.2 Cyclic organic compounds
4.3.1 Polymers Cyclic organic carbonate (COC) compounds, such as propylene
Polymerization is the process whereby small molecules referred to carbonate (PC) and glycerol carbonate (GC), have been proposed
as monomers are combined chemically to form much larger, as swelling activation compounds capable of being intercalated
chainlike molecules or macromolecules. Polymers may be and maintaining high swell in the presence of highly saline or
categorized as electrically neutral or nonionic and charged or ionic otherwise high ionic strength solutions (Onikata et al. 1996; Gates
depending on the polymerization process. Charged polymers et al. 2016), and PC is the chemical used for MSB (Onikata et al.
include cationic (positively charged), anionic (negatively charged) 1999; Katsumi et al. 2008), whereas GC has been used to resist
(e.g., Ashmawy et al. 2002; McRory and Ashmawy 2005; Ozhan alterations in the k of bentonite permeated with hypersaline
2018a,b), and zwitterionic (both positively and negatively solutions (Fehervari et al. 2016a,b). Unlike the large polymer
charged). Cationic polymers and anionic polymers also are macromolecules, there is considerable evidence that COCs are
referred to as polycations and polyanions, respectively (e.g., Breen intercalated and, through activation of crystalline swelling,
1999). Polyanions are currently the most prevalent polymers used maintain low k when subjected to permeation with chemical
in enhanced bentonites (Scalia et al. 2018a), although polycations solutions of extreme chemistry. However, as previously noted,
202
Onikata et al. (1999) showed that MSB did not maintain osmotic polymer intercalated in the interlayer region between the
swell when subject to NaCl solutions with I > 0.75 M. Thus, COCs platelets. Also, the interplatelet spacings increased with
appear to be promising organic compounds for enhanced increasing polymer dosage. Similar results were obtained by Di
bentonites but may not necessarily provide unlimited enhanced Emidio (2010) for the DPH-GCL (i.e., 0.016 nm). Qiu and Yu
performance as illustrated conceptually in Figure 4.4. (2008) demonstrated that CMC intercalated into the
montmorillonite interlayer, expanding the interlayer distance
4.4 Mixtures and processing from 1.25 nm to 1.47 nm as evidenced from the XRD pattern of
the enhanced montmorillonite. This shift was greater and,
The characteristics and properties of enhanced bentonites are a
therefore, more evident compared to that of the HC, probably due
complex function of several factors, including but not limited to
to the higher dosage of CMC in the study by Qiu and Yu (2008),
whether the constituent materials are mixed in a dry state or a
which was 2-to-1 CMC-to-montmorillonite weight ratio or 100
wet state, the control exercised on the temperature and pH in the
times greater than that for HYPER clay with 2% polymer
processing, and any resulting milling or other post mixing
modification. Qiu and Yu (2008) also show results of a peak
procedures that are implemented to control such factors as the
broadening that, according to the authors, most likely indicated
particle (tactoid) and/or aggregate (granule) size of the final
a disordered intercalated structure. However, further research is
product for use in the GCL. To illustrate some of these factors,
needed to validate these assumptions, namely the adsorption of
the general preparation procedures for the enhanced bentonites
Na-CMC within the interlayer pore space based on a measured
listed in Table 4.1 are described.
increase in the basal spacing of less than 0.1 Å and the
BPC (Table 4.1) was created by vigorously mixing dry Na-
improvement in the resistance against polymer elution as a result
bentonite with an aqueous solution containing acrylic acid
of oven-drying at 105 °C.
monomer and sodium persulfate as the thermal polymerization
MSB (Table 4.1) is produced by mixing dry Na-bentonite
initiator, leading to the formation of sodium polyacrylate (SAP).
with liquid propylene carbonate (PC), an aprotic polar organic
Based on the analogies between the powder XRD patterns of
solvent with a relative permittivity ranging from 66 to 71 at
sodium bentonite and BPC, Scalia et al. (2014) and Scalia and
temperatures between 20 and 0 °C, respectively (Simeral and
Benson (2017) concluded that the long-chain molecules of SAP
Amey 1970). Although relatively high if compared to the typical
do not intercalate between the montmorillonite platelets. They
values measured for non-polar and other polar organic solvents,
hypothesized that the low hydraulic conductivity of BPCs
these values of the relative permittivity for PC are lower than the
against concentrated solutions of multivalent cations should be
range of 80 to 88 associated with pure H2O. Based on the powder
ascribed to the partial occlusion of the conductive pores between
XRD patterns presented by Onikata et al. (1999), PC is believed
montmorillonite tactoids rather than an improved crystalline
to intercalate between individual montmorillonite platelets and
swelling. This finding has led to the commercial BPCs
promote crystalline swelling as a result of increased solvation
manufactured today, which comprise a dry mixture of granular
energy of the exchangeable cations (Mazzieri et a. 2010; Di
sodium bentonite and dry polymer particulate (Tian et al.
Emidio et al. 2017; Scalia et al. 2018a).
2016a,b, 2019; Chen et al. 2019; Gustitus et al. 2020). When
As previously noted, commercially produced CRCs have been
contacted by the permeant solutions, the bentonite and polymer
used extensively in practice for EB-GCLs (e.g., Ashwamy et al.
particulate hydrate, the latter forming a polymer gel that blocks
2002; Scalia et al. 2018a). However, as proprietary materials,
the pores between the bentonite granules. SEMs presented by
CRCs are generally produced by unknown methods and/or
Tian et al. (2016b, 2019) illustrate how the polymer gel fills the
materials, including in-situ polymerization, wet mixing, or dry
intergranular pore space.
mixing. Therefore, the proprietary nature of these materials
DPH-GCL (Table 4.1) is manufactured by vigorously mixing
precludes the ability to provide further details as to how these
dry sodium bentonite with an aqueous solution containing the
materials are produced.
SAP, Na-CMC, and methanol to promote montmorillonite
exfoliation and extrusion under vacuum until a bentonite sheet 4.5 Properties of enhanced bentonites
of reduced thickness and high bulk dry density is obtained.
Although the exact mechanisms by which Na-CMC is adsorbed 4.5.1 Hydraulic conductivity
on the surface of montmorillonite crystals are still the subject of The k of EB-GCLs versus conventional GCLs containing natural
debate (Theng 2012), a consensus is that the aforementioned Na-bentonite is compared in Figure 4.9 as a function of I (Figure
polymer can intercalate between the montmorillonite platelets, 4.9a) and pH (Figure 4.9b). Sources and screening methodology
thus acting as a prop to hold open the interlayer pore space and for the data shown in Figure 4.9 are provided in Scalia et al.
to prevent flocculation or aggregation in the presence of (2018a). Enhanced bentonit and EB-GCLs generally exhibit lower
aggressive liquids (Kolstad et al. 2004; Qiu and Yu 2008; k than natural Na-bentonite to permeant solutions with I greater
Mazzieri and Di Emidio 2015). than about 20 mM following the conceptual trend illustrated in
HCs (Table 4.1) are bentonites modified with the anionic Figure 4.4. In contrast, CRCs generally exhibit k similar to those
polymer, Na-CMC. However, instead of being vacuum-extruded for Na-bentonite over the same range in I. However, some CRCs
as with DPH-GCL, the slurry of the bentonite, polymer, and exhibit lower k than Na-bentonite, similar to other enhanced
water is oven-dried at 105°C. As a consequence of oven drying, bentonites. Of note, greater than 95% of the data shown in Figure
the polymer is irreversibly adsorbed through the fixed cations to 4.9 are from k tests performed at low effective stresses (<100 kPa).
the negative surface of the clay, thereby promoting a dispersed In general, the detrimental effect of hydraulic incompatibility is
structure of the clay lamellae resulting in more tortuous flow paths reduced or eliminated with an increase in effective stress,
and lower hydraulic conductivity (Di Emidio 2010). This depending on the magnitude of the effective stress and the
dehydration at high temperatures is hypothesized to improve the chemistry of the permeant liquid (Shackelford et al. 2000). Scalia
long-term containment performance against aggressive permeant et al. (2018a) provide additional discussion on the influence of
solutions due to the reduced tendency of the polymer to be eluted. effective stress on the k of enhanced bentonites based on limited
XRD analyses performed by Di Emidio et al. (2015) show an available data.
increase (i.e., 0.006 nm) in the basal spacing, d001, of HCs with Although data for the k of polymer EB-GCLs to extremely high
increasing polymer dosage. The characteristic peak of the and low pH solutions are limited, most of the available data are
enhanced montmorillonite broadened, shifting to the left, which
corresponds to greater basal spacings and indicates that the k is primarily controlled by I. Low k at both low pH (< 1) and high
203
pH (>13) are shown for BPC, illustrating the potential for as anion exclusion (or negative adsorption). Kang and Shackelford
enhanced bentonites to provide enhanced hydraulic compatibility (2011) showed that consolidation of GCLs (e.g., due to waste
in extreme pH environments. DPH-GCLs, BPC-GCLs, and some loading) could significantly enhance solute restriction and
CRCs also show promise in hyperalkaline solutions (pH >12) semipermeable membrane behaviour. Since solute restriction
(Tian and Benson 2019; Li et al. 2021). Additionally, linear and enhances the containment function of GCLs used as barriers,
cross-linked polymers are well-known to be pH sensitive. semipermeable membrane behaviour can have important practical
However, the results of a recent study reported by Norris et al. implications (Scalia et al. 2018a,b).
(2020) indicated that a CRC permeated with a copper mining Malusis and Shackelford (2002a) measured the chemico-
process solution with a pH of 1.2 resulted in a high k (>1.0 10-8 osmotic or membrane efficiency coefficient, , as a function of the
m/s). These results highlight the need for site-specific salt (KCl) concentration difference established across specimens
compatibility testing and additional research on the effects of pH of a conventional GCL using the closed-system apparatus
on the hydraulic compatibility of different enhanced bentonites. described by Malusis et al. (2001). Values of represent the
The significant variability in the data for CRCs and Na- degree or extent of solute restriction and generally range from zero
bentonite shown in Figure 4.9 illustrates the need for k testing for for a porous medium that does not exhibit solute restriction to unity
specific combinations of CRCs and conventional GCLs with the for a porous medium that exhibits complete solute restriction (i.e.,
actual liquids to be contained. Recent findings have shown that the The results from Malusis and Shackelford (2002a) for
current generation of enhanced bentonites that have been designed the case where the porosity of the GCL specimen, n, was 0.74 are
for specific applications (e.g., containment of coal combustion shown in Figure 4.10 in the form of as a function of the average
residuals) can provide necessary hydraulic compatibility for KCl concentration established across the specimen, cave. As
containment of high I leachates (e.g., Chen et al. 2019). However, indicated by these results, the GCL specimen exhibited significant
given current limitations on screening methods for enhanced solute restricting capability (i.e., even at a
bentonite hydraulic compatibility, hydraulic compatibility testing relatively high n of 0.74, although decreased approximately
using the actual liquids to be contained as permeant liquids is semilogarithmic linearly with an increasing cave. The reason for
required for CRCs. In addition, compatibility testing must continue this resulting trend is that, as cave decreases, the distributions in the
until all reaction fronts have passed through the test sample negative electrostatic surface potentials or diffuse double layers
(Jefferis 1992). (DDLs) associated with the surfaces of the individual clay particles
(tactoids) are extended and overlap, resulting in a net negative
potential between the particles such that anions are prohibited from
10-6
entering the micropore. The associated cations are also prohibited
Na-B
10-7 MSB from entering the micropore due to electroneutrality in the pore
HC water. However, as cave increases, the negative electrostatic surface
Hydraulic Conductivity (m/s)
BPC
10-8
DPH-GCL
potentials become increasingly more compressed, such that the
CRC overlap is increasingly reduced and eventually eliminated to the
10-9
extent that the micropores become accessible to the anions and the
10-10 associated cations. This trend of decreasing with increasing cave
reflects chemical incompatibility of the natural Na-bentonite
10-11
within the GCL with respect to semipermeable membrane
10-12 behaviour.
(a) 1
10-13 BPC (n = 0.92)
0.9
0.1 1 10 100 1000 10000
Membrane Efficiency Coefficient,
BPC (n = 0.80)
Ionic Strength (mM) 0.8
GCL (n = 0.74)
0.7
10-6
0.6
10-7 0.5
Hydraulic Conductivity (m/s)
0.4
10-8 0.3
0.2
10-9
0.1
0
10-10 1 10 100
Average Salt Concentration, c (mM)
ave
10-11
Figure 4.10. Improvement in semipermeable membrane behaviour of
10-12 bentonite polymer composite (BPC) relative to that for a GCL containing
(b) Na-bentonite as a function of specimen porosity, n (data from Malusis
10-13 and Shackelford 2002a and Bohnhoff and Shackelford 2013).
0 2 4 6 8 10 12 14
pH
Figure 4.9. Hydraulic conductivity of natural Na-bentonite (Na-B),
Bohnhoff and Shackelford (2013) used the same apparatus as
multiswellable bentonite (MSB), and polymer enhanced bentonites, that employed by Malusis and Shackelford (2002a) to measure
including HYPER clay (HC), bentonite-polyacrylic-acid composite (BPC), for two specimens of BPC subjected to differences in KCl
dense prehydrated geosynthetic clay liner (DPH-GCL), and proprietary concentrations over a similar range in cave. Their results, also
contaminant resistant clays (CRCs) versus (a) ionic strength and (b) pH of shown in Figure 4.10, indicated that both BPC specimens
permeant liquid for all stress and prehydration conditions (from Scalia et al. generally resulted in improved membrane behaviour, relative to
2018a,b). that exhibited by the GCL specimen with natural Na-bentonite,
4.5.2 Membrane behaviour and solute diffusion even though the porosity of the two BPC specimens was greater
Malusis and Shackelford (2002a) reported that GCLs containing than that for the GCL specimen. Thus, the BPC specimens
natural Na-bentonite could perform as semipermeable barriers that reflected greater resistance to chemical incompatibility than the
restrict the migration of solutes via the process commonly known conventional GCL tested by Malusis and Shackelford (2002a),
204
showing the advantage of using BPC in GCLs for chemical in Figure 4.12. Also shown in Figure 4.12 is the empirical
containment applications (e.g., also see Shackelford and Scalia relationship between m and n based on the following exponential
2016). However, Bohnhoff and Shackelford (2013) also noted that function adopted by Olsen and Kemper (1968), among others, to
the trend of decreasing with an increasing cave for the two BPC investigate the changes in diffusive tortuosity for both saturated
specimens was similar to that for the GCL, indicating that the and unsaturated porous media:
improved resistance to chemical incompatibility for the BPC was n
limited. m 10 (4.3)
n
The semipermeable membrane behaviour of bentonites
resulting from solute restriction affects the transport of solutes where the dimensionless parameters and were calibrated using
(ions) following three different mechanisms, i.e., hyperfiltration, the aforementioned experimental data. Regardless of the type of
chemico-osmotic counter advection, and restricted solute diffusion testing apparatus and solutes contained in the solutions, the
(Malusis et al. 2013, 2015, 2021). All three of these mechanisms values of m follow a non-linear trend versus porosity, suggesting
work to reduce the solute mass flux relative to the case where there that the degree of interconnectivity of the conductive pores in the
is no solute restriction (i.e., = 0), with restricted diffusion being enhanced bentonites is similar to that of unenhanced bentonites
the most relevant flux reduction mechanism when the hydraulic at a given porosity. However, the m values determined for the
conductivity is sufficiently low such that diffusion is the dominant BPC tested by Bohnhoff and Shackelford (2015) tend to be lower
transport process. In this case, an osmotic effective diffusion than the those based on the overall trend for the other bentonite
coefficient, D* , can be defined for a given diffusing solute as specimens, suggesting that the enhancement mechanism for the
follows (Dominijanni et al. 2018): BPC can be attributed to partial clogging of the conductive pores
such as previously indicated in Figure 4.8 and, therefore, to more
D* 1 Ds,0 meandering channels, greater tortuosity (lower m), and lesser
m (4.2)
solute diffusion.
Even though the fitting parameters used for the trends
where m is the dimensionless matrix tortuosity factor, which
described in Figure 4.12 based on Eq. 4.3 were derived from the
accounts for the tortuous nature of the solute diffusive pathways
results of diffusion measurements, Al-Tarawneh et al. (2009)
through the porous medium, and Ds,0 is the free-solution (aqueous-
observed that the differences in the geometric tortuosity factors
phase) diffusion coefficient of the salt as defined by Shackelford
estimated either via diffusion or k tests are negligible when the
and Daniel (1991). For a given m and Ds,0, the relationship uniformity coefficient, Cu, is close to unity (i.e., narrow grain-
between D* and given by Eq. 4.2 is linear as first postulated by size distribution), as typically is the case with bentonites used in
Manassero and Dominijanni (2003). geoenvironmental containment applications. Thus, Eq. 4.3 likely
For example, Mazzieri et al. (2010), Malusis and Daniyarov can be extended to provide an estimate of the matrix tortuosity
(2016), Di Emidio et al. (2015), and Bohnhoff and Shackelford factor to be used in the Kozeny-Carman equation (Kozeny 1927;
(2015) conducted laboratory tests for the simultaneous Carman 1956) for the evaluation of the bentonite k.
measurement of and D* on specimens of MSB, DPH-GCL, HC 0.35
and BPC, respectively, subjected to single salt aqueous solutions 0.30
= 0.00157
(i.e., solutions of KCl or CaCl2). The experimental results of these = 2.4012
(-) m
studies have been re-evaluated herein in accordance with Eq. 4.2 0.25
Matrix tortuosity factor,
based on a single value of m for each enhanced bentonite and in 0.20
terms of the restrictive tortuosity factor r D* m Ds,0 , which 0.15
accounts for the effect of anion exclusion in semipermeable clays 0.10
(Malusis and Shackelford 2002b), and the resulting values of r are 0.05 = 0.0002473
shown as a function of in Figure 4.11. The generally good 0.00
= 2.9905
agreement between the experimental data and the relationship 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Porosity, n (-)
given by Eq. 4.2 suggests that the proposed theoretical framework Malusis and Shackeford (2002b) Van Loon et al. (2007)
upon which Eq. 4.2 is based is reasonable for interpreting the Dominijanni et al. (2013) Malusis et al. (2015)
Shackelford et al. (2016) Musso et al. (2017)
influence of the pore-scale electrical interactions on solute Dominijanni et al. (2018) Mazzieri et al. (2010)
Bohnhoff and Shackelford (2015) Di Emidio et al. (2015)
diffusion in enhanced bentonites. Malusis and Daniyarov (2016)
1.0 e = 11.5
e=11.5 (Bohnhoff
(Bohnhoff andand Shackelford,
Shackelford, 2015)
2015)
e = 4.0
e=4.0 (Bohnhoff
(Bohnhoff andand Shackelford,
Shackelford, 2015)
2015)
0.9 e = 18.2
e=18.2 (Bohnhoff
(Bohnhoff andand Shackelford,
Shackelford, 2015)
2015)
e = 4.7
e=4.7 (Bohnhoff
(Bohnhoff andand Shackelford,
Shackelford, 2015)
2015) Figure 4.12. Matrix tortuosity factor, m, as a function of the porosity, n,
(-)
0.8 e = 2.55
e=2.55 (Di(Di Emidio
Emidio et al.,
et al., 2015)
2015)
for natural Na-bentonite (closed symbols) and enhanced bentonites (open
r
e = 1.55
e=1.55 (Malusis
(Malusis andand Daniyarov,
Daniyarov, 2016)
2016)
Restrictive tortuosity factor,
0.7 e=2.53 (Mazzieri
e = 2.53 et al.,
(Mazzieri 2010)
et al., 2010)
symbols) (note: the continuous lines represent the interpolation curves
0.6
based on Eq. 4.3).
0.5
0.4
0.3
4.6 Current and future directions
0.2
The current and future directions of the study of enhanced
0.1
0.0
bentonites are focused on resolving questions and potential
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 issues related to using these materials in GCLs and optimizing
Membrane efficiency coefficient, (-)
materials and methods for production. These directions include
Figure 4.11. Measured values of the restrictive tortuosity factor versus providing a better understanding of the mechanisms associated
measured membrane efficiency coefficient, , compared with the
with the long-term performance of EB-GCLs and resolving
theoretical interpretation based on Eq. 4.2 (continuous line) [e = void
ratio]. potential issues related to polymer elution, wet-dry cycling, and
non-standard temperatures (Scalia et al. 2018a). Also, there is a
The aforementioned values of m determined for the enhanced need to develop screening methods for enhanced bentonites that
bentonites are compared with the corresponding values that were provide a quick and reliable indication of the expected
determined for natural Na-bentonite and conventional GCL performance of a specific enhanced bentonite for a specific
specimens by Dominijanni et al. (2018) and Manassero (2020) containment liquid. Some discussion of ongoing efforts to address
these issues follows.
205
4.6.1 Mechanisms of Na-CMC in favouring a dispersed microstructure upon
Critical to predicting the long-term hydraulic behaviour of exposure to harsh chemical environments. Furthermore,
enhanced bentonites and EB-GCLs is understanding the although the preparation of HC involves a polymeric solution
mechanisms governing the hydraulic compatibility of these similar to that used for DPH-GCLs, HCs are oven-dried. As such,
materials (e.g., see Norris 2021). For example, the long-term k the contribution of densification to the long-term barrier
tests performed by Katsumi et al. (2008) using single-salt performance is negated.
permeant solutions with specimens of MSB and DPH-GCL, by As the contribution due to densification may be significant
Di Emidio et al. (2015) with a HYPER clay (HC) specimen, and (Kolstad et al. 2004), further study is needed to investigate
by Scalia et al. (2014) with BPC specimens are interpreted herein whether treatment with Na-CMC is adequate for aggressive
to gain insight into the microstructural arrangement of these leachates in the absence of the preconsolidation effect. In this
enhanced bentonites and EB-GCL using a modified Kozeny- regard, Fan et al. (2022) have shown recently that the k of HC is
Carman equation as described by Guarena et al. (2020) and 1-2 orders-of-magnitude lower relative to that of natural Na-
Manassero (2020). In this case, except for the HC specimen, a bentonite when permeated with landfill leachate laden with
value of Nl,AV was determined from each measured value of k
heavy metals. Di Emidio et al. (2019) showed that the k of HC
based on the known values for the CEC, e, and the specific
(non-prehydrated) to seawater (I = 0.77M) under low confining
gravity (Gs) for the tested enhanced bentonites, and m was
stress (14 kPa) remained on the order of 10-11 m/s (2 orders-of-
assessed via Eq. 4.3. For the HC specimen that was tested to
determine both the chemico-osmotic and diffusion properties, a magnitude lower than untreated clay) in the long-term after 80
pore volumes of flow over about 8.5 years of permeation.
m of 0.137 was obtained via extrapolation of the apparent
tortuosity factor a r m to = 0 corresponding to r = 1 400 Deionised
DeionizedWater
Water
(Malusis and Shackelford 2002b; Shackelford and Moore 2013; NaCl
NaCl200
200mN
mN
Average number of lamellae per tactoid, Nl,AV (-)
NaCl
NaCl500
500mN
mN
350
Dominijanni et al. 2018). NaCl
NaCl700
NaCl
700mN
NaCl1000
mN
1000mN
mN
NaCl
NaCl2000
2000mN
mN
The resulting experimental values of Nl,AV are shown in Figure 300
CaCl2
CaCl2200
200mN
mN
CaCl2500
CaCl2 500mN
mN
4.13 versus the micro-void ratio (em) and are compared to the 250
CaCl21000
CaCl2 1000mN
mN
corresponding iso-concentration curves based on the FBS 200
represented by Eq. 4.1 (Nl,AV0 = 1.56; = 8.82; =10.01) as 150
calibrated by Manassero (2020) using the experimental results 100 cs = 0 - 200 - 500 - 700 - 1000 - 2000 mN
provided by Petrov and Rowe (1997) for k of a conventional 50
needle-punched GCL prehydrated with distilled water. Note that 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
the salt concentration used in Eq. 4.1 was expressed as Micro-void ratio, em (-)
CaCl2
CaCl2500
500mN
mN
millinormality (mN) instead of millimolarity (mM) (e.g., 50 mM 110
CaCl2
CaCl21000
1000mN
mN
Average number of lamellae per tactoid, Nl,AV (-)
100 CaCl22000
CaCl2 2000mN
mN
= 100 mN for CaCl2 whereas 50 mM = 50 mN for NaCl). 90
As shown in Figure 4.13a, all of the calculated Nl,AV values 80
for the MSB specimens based on permeation with the 70
60
concentrated electrolyte solutions are greater than the model 50
predictions based on Eq. 4.1 as calibrated with respect to the 40
results from Petrov and Rowe (1997), likely due to the absence 30
20
of prehydration by exposure or permeation with DW prior to 10
permeation with the electrolyte solutions (e.g., Shackelford et al. 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
2000) in the case of the MSB specimens tested by Katsumi et al. Micro-void ratio, em (-)
(2008). The reduction in k for the MSB specimens permeated 110 Deionised
DeionizedWater
CaCl2
CaCl210
10mN
Water
mN
Average number of lamellae per tactoid, Nl,AV (-)
100
with electrolyte solutions containing NaCl at intermediate 90
concentrations (200 < cs < 1000 meq/L) relative to that of the 80
natural Na-bentonite can be attributed to the influence of the PC 70
60
on the extent of osmotic swelling. Even though crystalline 50
swelling (i.e., increased basal spacing) due to the PC cannot be 40
30
excluded, the reduction in k also may be attributed to the lower 20
relative permittivity of PC compared to that of molecular water 10
(H2O), which can be effective in blocking the negative electric 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
charge carried by the associated montmorillonite crystals Micro-void ratio, em (-)
DeionizedWater
Deionised Water
110
resulting in a closer intertactoid distance under the same effective CaCl210
CaCl2 10mN
mN
Average number of lamellae per tactoid, Nl,AV (-)
100 CaCl240
CaCl2 40mN
mN
confining stress. This postulated decrease in intertactoid distance 90
CaCl2100
CaCl2 100mN
CaCl2400
CaCl2
mN
400mN
mN
is reflected by the low em values in Figure 4.13a, which correlate 80
with the low k values based on the modified Kozeny-Carman 70
60
equation. 50
Unlike MSB, the ability of DPH-GCL to maintain a dispersed 40 cs = 0 - 10 - 40 - 100 - 400 mN
bentonite fabric and, hence, a low k upon permeation with 30
concentrated CaCl2 solutions, as demonstrated by Nl,AV values 20
10
that are close to the iso-concentration curve for DW (Figure 0
4.13b), is attributed to the treatment procedure for DPH-GCL, 0 1 2 3 4 5 6 7 8 9 10 11
Micro-void ratio, em (-)
which involves prehydration with the aqueous solution of Na-
PAA and Na-CMC followed by preconsolidation (densification) Figure 4.13. Comparison between the iso-concentration curves given by
through vacuum-extrusion. In this regard, low hydraulic Eq. 4.1 (continuous lines), as calibrated on the hydraulic conductivity
conductivity also was evident for the HC specimen, which again test results by Petrov and Rowe (1997), and the values of the average
is consistent with low Nl,AV values (Figure 4.13c). However, number of lamellae per tactoid, Nl,AV, obtained for: (a) the MSB tested by
unlike the tests performed with DPH-GCL by Katsumi et al. Katsumi et al. (2008); (b) the DPH-GCL tested by Katsumi et al. (2008);
(2008), the salt solutions used by Di Emidio et al. (2015) for HC (c) the HYPER clay tested by Di Emidio et al. (2015); (d) the BPC tested
were not sufficiently concentrated to indicate the effectiveness by Scalia et al. (2014).
206
The relatively high Nl,AV values observed for the BPC 4.16, wherein the comparison between the iso-concentration
specimens tested at the highest CaCl2 concentrations (Figure curves associated with both the prehydrated GCL tested by
4.13d) can be attributed, similar to the MSB specimens, to direct Petrov and Rowe (1997) and the non-prehydrated MSB tested by
permeation with the chemical solutions, i.e., without Katsumi et al. (2008) reveals that the most significant differences
prehydration. Furthermore, the reduction in k noticed by Scalia in the Nl,AV parameter occur at medium-to-high salt
et al. (2014) in comparative tests performed on natural Na- concentrations. In contrast, less marked differences are observed
bentonite specimens (e.g., see Figure 4.5) is related to the low m at low salt concentrations.
values based on Eq. 4.3 and the low em values evident in Figure The iso-concentration contours (Figure 4.16) from Eq. 4.1
4.13d. However, in contrast to MSB, the decrease in the pore based on the results for the prehydrated GCLs are in good
volumes accessible to the solvent and solute transport with agreement with the data for the enhanced bentonites treated with
increasing CaCl2 concentration is caused by the polymer Na-CMC (i.e., DPH-GCL and HC), and corresponding Nl,AV values
hydrogel partially occluding the bentonite pores, as previously for DPH-GCL and HC are practically independent of the salt
noted with respect to Figure 4.8. concentration of the permeant solution. This result means that,
Given the wide range of void ratios and salt concentrations of from a practical perspective, the and parameters in Eq. 4.1 are
the permeant solutions used by Katsumi et al. (2008), an close to zero. Therefore, for this kind of polymeric additive (i.e.,
evaluation of the reliability of the theoretical framework based Na-CMC), the related physico-chemical mechanism likely is
on the FBS (Eq. 4.1) can be undertaken by adjusting the Nl,AV0, intercalation that maintains separation between the individual
and parameters using the dataset pertaining to the MSB lamellae resulting in a dispersed microfabric. Thus, the MSB and
specimens. As shown in Figure 4.14, ordinary least-squares BPC maintain low values of k at a high salt concentration (Figure
regression of experimental versus theoretical Nl,AV values 4.15) due to pore clogging and/or a reduction in the intertactoid
resulted in Nl,AV0 = 1.00, = 27.51, and = 77.42 with a pore volume (microscopic pores) even with flocculation, whereas
coefficient of determination (R2) of 0.9151, which is considered the Na-CMC polymer apparently maintains a dispersed fabric of
satisfactory given the complexity of the mechanisms governing the enhanced bentonite, which is characterized by low values of
bentonite flocculation/dispersion. The resulting good agreement Nl,AV.
shown in Figure 4.15a suggests that the theoretical framework is As a practical consequence, in the case of MSB and BPC, even
reliable for describing the macroscopic effects of the if a low k can be maintained in the long-term in the presence of
microstructural arrangement of smectitic clays for both concentrated (high I) permeant solutions, the persistence of
prehydrated and non-prehydrated specimens. osmotic efficiency and the related membrane behaviour is doubtful,
as reflected by the results in Figure 4.10. In contrast, the Na-CMC
400 enhanced bentonites appear to maintain a low k in the presence
of concentrated permeant solutions and a significant osmotic
Experimental number of lamellae per tactoid, Nl,AV (-)
350
efficiency when exposed to solutions with up to the medium salt
300 concentrations. Nonetheless, further experimental and theoretical
studies are needed to validate these preliminary conclusions.
250
400 Deionised
DeionizedWater
Water
200 NaCl
NaCl200
200mN
mN
Average number of lamellae per tactoid, Nl,AV (-)
NaCl
NaCl500
500mN
mN
350 NaCl
NaCl700
700mN
mN
150
NaCl
NaCl1000
1000mN
mN
300 NaCl2000
NaCl 2000mN
mN
100 CaCl2200
CaCl2 200mN
mN
250 CaCl2500
CaCl2 500mN
mN
cs = 0 - 200 - 500 - 700 - 1000 - 2000 mN CaCl21000
CaCl2 1000mN
mN
50
200
0
150
0 50 100 150 200 250 300 350 400
Theoretical number of lamellae per tactoid, Nl,AV (-) 100
Figure 4.14. Experimental versus the theoretical average number of
50
lamellae per tactoid, Nl,AV, using the hydraulic conductivity test results
provided by Katsumi et al. (2008) for non-prehydrated MSB specimens 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
and the related calibration of the FBS parameters (Eq. 4.1).
Micro-void ratio, em (-)
DeionizedWater
Deionised Water
As shown in Figure 4.15b, despite the relatively limited 110
CaCl210
CaCl2 10mN
mN
Average number of lamellae per tactoid, Nl,AV (-)
CaCl240
CaCl2 40mN
mN
available data, interpretation of the hydraulic conductivity tests 100
CaCl2100
CaCl2 100mN
mN
90
performed by Scalia et al. (2014) for BPC specimens also yields CaCl2400
CaCl2 400mN
mN
80
Nl,AV values that agree reasonably well with those from Eq. 4.1
70
calibrated using the results for the MSB specimens tested by
60
Katsumi et al. (2008). This consistency suggests that the 50
microstructures of both MSB and BPC are not affected 40
significantly by the presence of the PC or the long-chain Na-PAA 30
(Table 4.1), respectively, which impact the accessible pores 20
cs = 10 mN
and/or the degree of interconnectivity of such pores, whereas a 10 cs = 0 mN
more prominent role is likely attributable to the order in which 0
the bentonite specimens are permeated with the electrolyte 0 1 2 3 4 5 6 7 8 9 10 11
Micro-void ratio, em (-)
solutions. Thus, only the treatment procedure developed for HC Figure 4.15. Comparison between the iso-concentration curves of the
and DPH-GCLs is effective in maintaining a dispersed fabric FBS given by Eq. 4.1 (continuous lines), as calibrated using the results
regardless of the ionic strength of the permeant solution (Figure of the hydraulic conductivity tests performed on MSB specimens by
4.13b), thus supporting the hypothesis that Na-CMC is Katsumi et al. (2008), and the values of the average number of lamellae
intercalated and acts as a prop between the montmorillonite unit per tactoid, Nl,AV, for (a) the MSB tested by Katsumi et al. (2008), and (b)
layers. the BPC tested by Scalia et al. (2014).
The detrimental impact on the bentonite barrier performance
due to the absence of prehydration is illustrated further in Figure
207
400 MSB (Katsumi et al., 2008) (iii) retention of a greater fraction of polymer corresponds to a
Average number of lamellae per tactoid, Nl,AV (-)
DPH-GCL (Katsumi et al., 2008) greater discrepancy between swelling and k, i.e., low k associated
350
HYPER Clay (Di Emidio et al., 2015) with low swelling. Lower k also has been shown to correlate to
300 BPC (Scalia et al., 2014) higher initial polymer loading (Zainab and Tian 2020; Li et al.
250
2021). These observations illustrate that, for the polymer to reach
intergranular pores and affect k, sufficient polymer must be present,
200 the polymer must be evenly distributed, and the polymer must be
150
retained in the bentonite pores. Additional study is needed to
understand the mechanisms that lead to polymer redistribution and
100
retention within polymer EB-GCL pores and to understand better
50 the mechanisms governing hydraulic compatibility of polymer
cs = 0 mN EB-GCLs. Methods to measure polymer loading in EB-GCLs are
0
0 1 2 3 4 5 6 7 8 9 10 11 described by Gustitus et al. (2020) and Norris et al. (2022).
Micro-void ratio, em (-) The eluted polymer may be in the form of the crosslinked
Figure 4.16. Iso-concentration curves based on Eq. 4.1 calibrated using hydrogel (e.g., Figure 4.17) or high viscosity (sticky) effluent from
the results of the hydraulic conductivity tests performed by Katsumi et linear polymer. Unfortunately, current standard methods for
al. (2008) on non-prehydrated MSB specimens (continuous lines) and by hydraulic compatibility tests (e.g., ASTM D6766) do not consider
Petrov and Rowe (1997) on prehydrated GCL specimens (dashed lines). polymer elution. Thus, additional care should be taken to ensure
that the measured k is not the result of clogged porous stones or
4.6.2 Polymer elution
tubing. This possibility can be evaluated by reassembling the
Clogging of conductive intergranular or interaggregate pores by
permeameter and passing the permeant liquid through the tubing
polymer hydrogel is the mechanism responsible for the low k of
under the same differential head between the platens. Tian et al.
BPC-GCLs, and the reason k of these materials is not directly
(2019) used total organic carbon analysis (ASTM D4839) to
related to swelling (Scalia et al. 2014; Scalia and Benson 2017;
measure effluent polymer content, although the study of additional
Tian et al. 2019; Li et al. 2021). Scalia et al. (2014) observed
methods is ongoing. The necessity to continue permeation until no
elution of polymer hydrogel from k tests performed on BPC with
further polymer is eluted (mobile) requires further research.
high polymer loading, viz. 28.5% by mass (Figure 4.17). This
Understanding polymer elution behaviour is critical to
eluted polymer clogged the small (3.2 mm) inside diameter
understanding the long-term behaviour of EB-GCLs and ensuring
effluent tubing of the flexible-wall permeameter such that the
that laboratory hydraulic compatibility results are accurate
apparent k was observed to decrease to a value on the order of
(Gustitus and Benson 2020).
1.0×10-13 m/s (see Figure 4.5). Upon cleaning the permeameter
tubing, the apparent k increased approximately 2-3 orders of 4.6.3 Wet-dry cycles in polymerized clays
magnitude, and after multiple cleaning cycles, eventually Conventional GCLs used as containment barriers can be exposed
stabilized. Similar observations have been made by Scalia and to changes in the degree of hydration, such as periodic
Benson (2017), Chen et al. (2019), Tian et al. (2019), and Li et al. precipitation events (e.g., daily, weekly, monthly, seasonal) as well
(2021), albeit generally to a lesser extent, commensurate with the as seasonal temperature variations, especially when the GCL is
lower polymer loading of the later generation polymer EB-GCLs placed close to the ground surface as in the case of cover systems.
that were tested. Scalia et al. (2014) hypothesized that the same For these reasons, several studies have evaluated the effects of wet-
clogging effect observed in the flexible-wall permeameter tubing and-dry cycling on the hydraulic performance of conventional
also was occurring within the BPC pores. When permeant liquids GCLs (e.g., Lin and Benson 2000; Bouazza et al. 2007; Rowe et
with high I prevent osmotic swelling of the bentonite, the polymer al. 2011; Hoai and Mukunoki 2020).
hydrogel may migrate to occupy intergranular pores yielding low Although natural Na-bentonite used in GCLs has exhibited
k. considerable self-healing capability upon rehydration after being
subjected to defects (e.g., Sari and Chai 2013; Rowe and Li 2020),
several studies have indicated significant detrimental impacts
resulting from wet-dry cycling (e.g., Hoai and Mukunoki 2020),
especially in the case where hydration or wet-dry cycling is
accompanied by cation exchange of divalent cations (e.g., Ca2+) in
the pore water for initially bound monovalent cations, primarily
the Na+ associated with the Na-bentonite, in the GCL (e.g., Lin and
Benson 2000; Bouazza et al. 2006, 2007; Thiel et al. 2006; Benson
and Meer 2009; Rowe and Abdelatty 2011; Tang et al. 2011;
Bradshaw et al. 2013; Rowe and Hosney 2013; De Camillis et al.
2014; Mukunoki et al. 2014; De Camillis et al. 2016, 2017; Yu and
El-Zein 2019; Yu et al. 2020; Lieske et al. 2020; Rowe and
Hamdan 2021; Khan et al. 2021). In this latter case, the effect of
the cation exchange is to reduce osmotic swell such that upon
rehydration, the k can be detrimentally impacted, resulting in poor
performance on the GCL (e.g., James et al. 1997; Meer and Benson
Figure 4.17. Mobile polyacrylate hydrogel crosslinked with Ca2+ eluted
from flexible-wall permeameter (note: scale is in mm).
2007; Benson and Meer 2009; Benson et al. 2010a). In particular,
thermal cycles combined with cation exchange have been shown
As previously noted, the presence of polymer hydrogel within to significantly affect the hydraulic performance and kw of some
polymer EB-GCL pores was documented using SEM images by NaB-GCLs due to the loss of swell capacity and ability to seal
Tian et al. (2016, 2019) (see Figure 4.8). Recent studies have needle-punched bundles (Rowe et al. 2017, 2019, 2021).
shown that: (i) elution behaviour, is sensitive to the chemical The use of EB-GCLs for increased resistance to downslope
nature of the permeant liquid (Scalia et al. 2014; Tian et al. 2019); erosion of GCLs in exposed composite liners and chemical attack
(ii) elution of a greater fraction of the polymer from an EB-GCL and improved hydraulic performance in situations where the
correlates with a higher k (Chen et al. 2019; Tian et al. 2019); and barrier is subjected to wet-dry cycling has been suggested.
208
However, only a limited number of studies have evaluated the difference between the different samples is clear, i.e., the cracks in
impact of wet-dry cycling on the hydraulic performance of EB- the untreated Na-bentonite are larger compared to those for the HC
GCLs. specimens. The cracks in the HC specimens occur mainly in the
De Camillis et al. (2016, 2017) studied the performance of upper parts of the chosen subsection and did not propagate
polymer modified GCLs when cation exchange was combined downwards. During the wetting cycle, the bentonites swell,
with wet-and-dry cycling. The performance of untreated Na- thereby closing the drying cracks. The CT scans revealed more
bentonite after wet-and-dry cycling in seawater (SW) was unhealed cracks for the untreated Na-bentonite relative to the HC
compared to that of a polymer treated Na-bentonite. The clay was specimens. This observation is more evident for the HC+8%
treated with 2% and 8% by dry mass of Na-CMC anionic polymer specimen, wherein almost all cracks are closed.
following the HYPER clay (HC) procedure (Di Emidio 2010). In
De Camillis et al. (2016), the results of swell index tests indicated
that the swelling ability increased as the polymer content increased.
Similar results were obtained from one-dimensional swell tests. As
shown in Figure 4.18, the swelling of HC with 8% polymer
(HC+8%) after the 6th wet-dry cycle was still considerably greater
relative to that for the Na-bentonite under the same conditions and
surprisingly similar to the swell of the Na-bentonite in contact with
DW. Also, as shown in Figure 4.19, the k of the untreated Na-
bentonite increased sharply within three wet-dry cycles. Cracks
formed in the Na-bentonite by drying remained intact, resulting in
preferential flow paths and high k. In contrast, the HC specimens
Figure 4.20. 3D rendered results of the CT scan of untreated bentonite,
maintained a k to seawater lower than 1.0 10-9 m/s. During the 4th HYPER clay with 2% or 8% Na-CMC anionic polymer after the 3rd dry
wet-dry cycle, the k of untreated bentonite was three orders-of- cycle and during the 4th wet cycle (note: cracks are displayed in red).
magnitude higher than that of the HC with 2% polymer (HC+2%).
25 The untreated Na-bentonite developed cracks up to 6.0% of the
20 Na-bentonite total initial volume. In contrast, the volume of cracks in the HC+2%
Vertical Swell (mm)
HC+2% and HC+8% specimens were lower than 1.0% of the initial volume,
15
HC+8% i.e., 0.94% and 0.12%, respectively (Figure 4.21). During the
10 wetting cycle, the volume of the cracks decreased for all clays.
5 After wetting, the volume of the cracks in the untreated Na-
0 bentonite specimen was 1.2% of the total volume, whereas those
for the HC+2% and HC+8% specimens were 0.45% and 0.05%,
-5
respectively.
-10
0 1 2 3 4 5 6 7 7
No. of Cycles 6 Na-bentonite
Figure 4.18. Swell of untreated Na-bentonite and HYPER clay with 2%
5 HC+2%
Percent of Cracks
or 8% Na-CMC anionic polymer (HC+2% and HC+8%) at the end of
each wetting cycle (De Camillis et al. 2016). 4 HC+8%
10
-6
3
Hydraulic Conductivity, k (m/s)
-7 Na-bentonite 2
10
HC+2%
10
-8
HC+8% 1
-9
10 0
-10 Dry Cycle 3 Wet Cycle 4
10
-11
Figure 4.21. Percent volume of cracking in untreated Na-bentonite and
10
HYPER clay with 2% or 8% Na-CMC anionic polymer (HC+2% and
10
-12
HC+8%) at the end of the 3rd dry and 4th wet cycles.
0 1 2 3 4 5
No. of Cycles
Results of several studies involving enhanced bentonites or EB-
Figure 4.19. Hydraulic conductivity of untreated Na-bentonite and Hyper GCLs subjected to wet-and-dry cycling using electrolyte solutions
clay with 2% or 8% Na-CMC anionic polymer (HC+2% and HC+8%)
are shown in Figure 4.22. Mazzieri and Pasqualini (2008) and
permeated with seawater after each wetting cycle.
Mazzieri et al. (2016) reported the performance of DPH-GCL and
MSB subjected to wet-and-dry cycling with a CaCl2 solution and
De Camillis et al. (2017) evaluated cracks by means of micro
seawater (SW). The DPH-GCL tested by Mazzieri and Pasqualini
computed tomography ( CT) scanning analysis on untreated Na-
(2008) was subjected to four wet-dry cycles with 0.0125 M CaCl2
bentonite and the HC+2% and HC+8% after six wet-dry cycles
(I = 0.075 M) and dried at 35°C until the weight was constant. The
with seawater. CT is a non-destructive three-dimensional (3D)
DPH-GCL and MSB tested by Mazzieri et al. (2016) were
Imaging and analysis technique, which is frequently used in the
subjected to five wet-dry cycles with SW (I = 0.71 M) and dried at
study of geomaterials (Cnudde and Boone 2013). The CT
35°C until the weight was constant. The GCL with HC (GCL
analysis demonstrated a better self-healing capacity (smaller
w/HC) is the needle-punched GCL prototype with HC tested by
volume of cracks) generated on the HC specimens relative to the
De Camillis et al. (2016, 2017) and De Camillis (2017), which was
untreated Na-bentonite.
subjected to four wet-dry cycles first and then permeated with SW
The 3D visual results of the CT scans for the different samples
(I = 0.77 M). The hydration was either first with DW and then for
(Na-bentonite, HC+2% and HC+8%) after the 3rd drying stage and
three cycles with SW or simply with four wet-dry cycles using SW.
during the 4th wetting stage are shown in Figure 4.20. The analysis
The GCL w/HC specimens were dried at 60°C until the weight was
also allowed determining the volume percentage of the cracks
constant.
compared to the coherent zone of the samples. Visually, the
The GCLs tested by De Camillis (2017) were exposed to the
209
worst conditions, as the hydrating liquid was the most aggressive associated laboratory testing have not explored the effects of cation
(i.e., I = 0.77 M vs. I = 0.075 M and I = 0.71 M) and the drying exchange and polymer elution with repeated wet-dry cycles in the
temperature was the highest (60oC vs. 35oC), resulting in severe field. To test this hypothesis, in 2011, Rowe et al. (2016b)
desiccation. In addition, the needle-punched GCL produced more constructed one section of a full-scale field liner at the Queen’s
cracks relative to the MSB and DPH-GCL. Overall, the GCL University Environmental Liner Test Site (QUELTS II) with an
w/HC specimens provided the lowest k values, suggesting a good EB-GCL containing a polyacrylamide enhanced bentonite.
hydraulic performance in contact with SW. After 15 months of field exposure, there was significant
evidence of polymer elution to the surface of the EB-GCL, which
o
DPH-GCL: CaCl @ 35 C (Mazzieri & Pasqualini 2008)
2
had become very slick and slippery. In addition, there was
o evidence that evaporated, and subsequently, condensed water had
DPH-GCL: SW @ 35 C (Mazzieri et al. 2016)
o
been flowing down the slope in rivulets. As a result, there was 132
MSB: SW @35 C (Mazzieri et al. 2016) mm of panel overlap shrinkage (35% greater than the comparable
o
GCL w/HC: DW&SW @ 60 C (De Camillis et al. 2017) NaB-GCL with traditional Na-bentonite), although no downslope
o
GCL w/HC: SW @ 60 C (De Camillis et al. 2017) erosion was observed. Thus, the EB delayed downslope erosion
relative to the companion product with NaB. However, 15 to 28
-6
10 months of flow over the GCL washed the eluted polymer away and,
Hydraulic Conductivity, k (m/s)
-7
after 28 months of exposure, downslope erosion was observed for
10 the EB-GCL. This erosion included both irrecoverable and
10
-8 extreme erosion features (e.g., an eroded zone 130 mm wide, as
shown in Figure 4.23).
-9
10
-10
10
-11
10
-12
10
0 1 2 3 4 5 6
No. of Cycles
Figure 4.22. Hydraulic conductivity of several enhanced bentonites
permeated with different liquids at different temperatures as a function of
the number of wet-dry cycles (DW = deionized water; SW = seawater).
Lieske et al. (2020) addressed the cracking and self-healing
capacity of GCLs subjected to drying and wetting in CaCl2 salt
solutions. The evolution of cracks and the cracking patterns were
evaluated via X-ray Imaging, and soil-water retention behaviour
was tested on the raw bentonite using micro-cells and a chilled-
mirror hygrometer. The morphology of the crack patterns in the
GCLs was highly affected by the pore fluid, which was attributed Figure 4.23. Extreme downslope bentonite erosion and irrecoverable
to reduced tensile strength caused by the salt induced aggregation. erosion in a 720-mm-long segment of an EB-GCL after 28 months in field
Khan et al. (2021) investigated the suction of GCLs prototypes exposure at the QUELTS. Composite of two photographs taken in the field
containing HC+8% compared to untreated GCLs. The matric using a thin portable light table slid below the GCL and taken in a
suction of the GCLs was measured by filter paper test and lightweight (cardboard) field “darkroom”. The band in the middle is the
humidity sensors. Different electrolytic solutions were used as the edge of the portable light table. Other black areas are hydrated bentonite,
permeant liquids. The soil-water retention capacity was greater for and the leopard-skin areas are highly desiccated bentonite with light
shining through the desiccation cracks (modified from Rowe et al. 2016b)
the GCL with HC+8% relative to the GCL with the untreated Na-
bentonite. The aforementioned full-scale field study only examined one
The aforementioned studies suggest that enhanced bentonites, EB-GCL. However, the findings initiated several studies (e.g.,
particularly HC, can provide improved resistance to the Rowe and Hamdan 2021a,b) involving periodic water flow over
detrimental impacts resulting from wet-and-dry cycling in the five commercially available GCL products from three continents,
presence of electrolyte solutions relative to that for unenhanced including three EB-GCLs (one of which was used at QUELTS as
Na-bentonites. However, the unanswered questions from all of a control for which the field behaviour was known). All five GCLs
these studies are: (a) for how long does the enhancement of the were hydrated with a simulated porewater comprising 10 mmol/L
bentonite provide positive effects, and (b) what are the CaCl2 to induce the cation exchange commonly observed with
consequences of elution of the polymer? There is a paucity of data hydration from natural soils, including at QUELTS. All five GCLs
addressing these questions; however, the available data, discussed were subjected to similar wetting periods but daily, weekly, and
subsequently, raises some questions about long-term performance fortnightly interludes between wetting cycles at room temperature.
of at least three enhanced bentonites. The EB-GCLs initially had a slightly lower hydraulic conductivity
Leaving a composite liner comprising a geomembrane over a to tap water than that for the NaB-GCLs. However, over time, the
GCL exposed to solar radiation will subject the GCL to repeated NaB-GCL performed far better than the EB-GCLs. For example,
hydration and dehydration (i.e., wet-dry) cycles. This is known to one NaB-GCL with a scrim-reinforced nonwoven carrier
induce significant shrinkage of some Na-bentonite GCLs (Thiel experienced an order-of-magnitude increase in hydraulic
and Richardson 2005; Thiel et al. 2006; Bostwick et al. 2010; conductivity, kw, that was expected based on cation exchange and
Rowe et al. 2011, 2013b, 2018; Brachman et al. 2018) as well as a low at 15 kPa stress. The second NaB-GCL with a woven carrier
downslope erosion of the bentonite in the GCL (Rowe et al. 2014, experienced an up to a two orders-of-magnitude increase in kw. The
2016; Ashe et al. 2015; Brachman et al. 2015; Rentz et al. 2016a,b). increase was partially due to cation exchange and partially due to
Enhanced bentonite in an EB-GCL has been hypothesized to mini-wrinkles that developed due to GCL shrinkage during the
prevent downslope erosion. However, this hypothesis and wet-dry cycles. In contrast, the three EB-GCLs experienced up to
210
3 to 4 orders-of-magnitude increases in kw. study by Rowe and Hamdan (2021a,b), with one GCL (denoted
GCL2 in Rowe et al. 2016a,b) being a conventional GCL with
4.6.4 Temperature effects
unenhanced Na-bentonite whereas the other GCL was a
Organic materials associated with solid waste containment
polyacrylamide EB-GCL (GCL7 in Rowe et al. 2016a,b). The
facilities decompose biologically, resulting in increases in the
influence of temperature on the desiccation of the two types of
temperatures of the waste (Rowe 2005, 2012). GCLs used as
GCLs was investigated by monitoring water loss and desiccation
barriers in mining, variant, and industrial applications may be
cracking. Some desiccated specimens were subjected to direct k
exposed to elevated temperatures ranging from 60 °C to 80 °C
tests with distilled water, while others were permeated with
when the GCL has been in contact with the wastes for more than
synthetic brine. With a moderate temperature of 40 °C applied to
five years (Thiel and Smith 2004; Bouazza et al. 2014). Also,
the top of the liner, the EB-GCL experienced less cracking than the
containment facilities that expose GCLs to high temperatures and
unenhanced GCL, but this advantage disappeared when
aggressive organic and/or inorganic chemicals (e.g., landfills, solar
temperatures increased. Both desiccated specimens of the
ponds, brine ponds) present a challenge (Ghavam-Nasiri et al.
unenhanced GCL and the EB-GCL exhibited significant self-
2019; Rowe 2020; Yu and El-Zein 2019). In facilities such as brine
healing when permeated with distilled water.
ponds and solar ponds where temperatures up to 90 °C have been
-4
reported during daytime (El-Sebaii et al. 2011; Bouazza et al. 10
Hydraulic Conductivity, k (m/s)
Unenhanced GCL (a)
2014), the lower overburden stresses on the liner system, 10
-5
1% AP GCL
compared to solid waste landfill systems, correlate with greater -6
2% AP GCL
10
risks of desiccation (Southen and Rowe 2005; Rowe and Verge 1% CP GCL
-7
2013; Hoor and Rowe 2013). For these and other reasons, 10 2% CP GCL
evaluating the potential effects of elevated temperatures on the 10
-8
properties and performance of conventional GCLs and EB-GCLs 10
-9
may be important. -10
Ratkievicius et al. (2016) and Geng et al. (2016) performed 10
-11
viscosity tests on specimens of bentonite mixed with cationic and 10
10 20 30 40 50 60 70
anionic polymers. Polymer addition increased the viscosity of the
mixtures. Low viscosity was associated with a contracted polymer Temperature (oC)
conformation, whereas high viscosity was attributed to the -1
10
formation of an extended polymer conformation. Contracted
Hydraulic Conductivity, k (m/s)
-2 (b)
10 Unenhanced GCL
polymers (low viscosity) prevent the formation of effective -3
10 1% AP GCL
bridging between bentonite particles and polymer chains (Lage et -4 2% AP GCL
10
al. 2009). -5 1% CP GCL
10
Ozhan (2018b) investigated the effects of temperature on the -6
2% CP GCL
10
hydraulic conductivity of polymer EB-GCLs. Hydraulic -7
10
conductivity tests were performed on specimens of EB-GCLs that -8
10
were treated with either 1% or 2% by mass of an anionic polymer -9
10
(AP) or a cationic polymer (CP) and an unenhanced GCL at -10
10
temperatures of 20°C, 40°C and 60°C using 0.1 M and 0.5 M 10 20 30 40 50 60 70
MgCl2 solutions as the permeant liquids. The results, shown in
Temperature (oC)
Figure 4.24, indicated that the k increased with increasing
temperature, which was attributed to a decrease in the viscosity of Figure 4.24. Effect of temperature on the hydraulic conductivity of
unenhanced GCL and GCLs enhanced with an anionic polymer (AP) or a
the permeant liquids with increasing temperature (Rowe 2005;
cationic polymer (CP) based on permeation with MgCl2 solutions: (a) 0.1
Bouazza et al. 2014; Vryzas et al. 2017). However, the hydraulic M MgCl2; (b) 0.5 M MgCl2 (data from Ozhan 2018b).
conductivity of the polymer EB-GCLs was up to two orders-of-
5
magnitude lower than that of the unenhanced GCL, which suggests 10
that the dry polymer mixing procedure used for the EB-GCLs may Unenhanced GCL
4
10 1% AP GCL
have partially clogged the pores of the EB-GCLs and/or reduced 2% AP GCL
the sizes of the conductive pores (Li and Zhang 2009; Haase and 3 1% CP GCL
0.1
10
Schanz 2016). Generally, the EB-GCL with 2% CP resulted in the
k /k
2% CP GCL
0.5
lowest k. Also, as indicated in Figure 4.25, the ratio of the 100
hydraulic conductivity based on permeation with the 0.5 M MgCl2
solution, k0.5, to that based on permeation with the 0.1 M MgCl2 10
solution, k0.1, or k0.5/k0.1, was greater than unity in all cases,
indicating some degree of hydraulic incompatibility with all 1
10 20 30 40 50 60 70
specimens at all temperatures. Overall, the values ranged from 6.3
for the 2% AP GCL at 60 oC to 232 for the 1% CP GCL at 60 oC. Temperature (oC)
However, unlike the k results shown in Figure 4.24, the 2% AP Figure 4.25. Effect of temperature on the hydraulic conductivity of
GCL performed the best with respect to k0.5/k0.1, with values in the unenhanced GCL and GCLs enhanced with an anionic polymer (AP) or a
k0.5/k0.1 4. cationic polymer (CP) based on permeation with a 0.5 M MgCl2 solution,
k0.5, relative to that based on permeation with a 0.1 M MgCl2 solution, k0.1,
Yu et al. (2020) assessed the desiccation and healing potential
or k0.5/ k0.1 (data from Ozhan 2018b).
of GCLs with and without polymer modification when exposed to
thermal gradients and then permeated with distilled or highly
The soil-water retention characteristics of the unenhanced GCL
saline water. Two types of GCLs were subjected to thermal and the EB-GCL were tested along the drying path by filter paper
gradients in a series of instrumented column tests that replicated
and a vapour sorption analyser. The experimental results are shown
conditions like those found in brine ponds with low overburden
in Figure 4.26. For volumetric water contents greater than about
pressure, high thermal gradients, and subsequent permeation with
10%, the total suction for the EB-GCL is noticeably greater than
a salt solution (brine). These were the same GCLs as used in the that for the unenhanced GCL. The inflection point, air entry
field study reported by Rowe et al. (2016a,b) and the laboratory
211
suction, and the slope of the curve also were greater for the EB- (data from Yu et al. 2020).
GCL. Yu et al. (2020) concluded that the addition of
polyacrylamide appears to have enhanced the water retention As previously noted, De Camillis et al. (2016, 2017) studied the
properties of the GCL. effect of wet-and-dry ageing on untreated, and polymer treated
1 HYPER clays. The HYPER clays exhibited three orders-of-
magnitude lower hydraulic conductivity relative to the untreated
Volumetric Water Content,
GCL clays after four wet-dry cycles in seawater. Khan et al. (2021)
0.8 EB-GCL
postulated that a possible reason for this improved behaviour is due
0.6 to the improved soil-water retention capacity of HYPER clay (HC)
compared to the untreated clay (UC). As a result, Khan et al. (2021)
0.4 evaluated the matric suction of prototype GCLs containing HC and
UC using the filter paper technique under both wetting and drying
0.2 conditions in the presence of different electrolytic solutions. The
filter paper technique is suitable to measure the suction due to its
0
1 10 100 1000 10
4
10
5
10
6 low cost, and the ability to measure a wide range of matric suctions,
typically from 10 to 105 kPa (Acikel et al. 2015; Hanson et al. 2013;
Total Suction, (kPa)
Barroso et al. 2006). The resulting measured soil-water retention
Figure 4.26. Soil-water retention curves of conventional (untreated) GCL curves (SWRCs) for specimens of both UC and HC are shown in
and enhanced bentonite GCL (EB-GCL) treated with polyacrylamide
Figure 4.28.
polymer (Yu et al. 2020).
Also, in the study by Yu et al. (2020), the crack patterns of the
Gravimetric Water Content, w (%)
600
unenhanced GCL and EB-GCL specimens were evaluated after the
UC-Wetting
specimens were subjected to different temperatures, and the results 500
UC-Drying
from the quantitative image analysis are shown in Figure 4.27. The 400 HC-Wetting
results indicated that the cracks became denser at higher HC-Drying
temperatures for both specimens, consistent with the water content 300
variations, which was expected since higher temperatures lead to 200
faster and more severe desiccation. However, the cracks in the EB-
100
GCL were less significant for heating temperatures of 40 °C and (a)
50 °C. For almost all results shown in Figure 4.27, the EB-GCL 0
4 5
experienced less desiccation than the unenhanced GCL. Also, the 1 10 100 1000 10 10
average crack width for the EB-GCL specimen at 40 °C heating Matric Suction, (kPa)
m
was considerably smaller than that for the unenhanced GCL,
Gravimetric Water Content, w (%)
which could result in a better self-healing potential for the EB- 600
GCL upon rehydration of distilled water (Yu and El-Zein, 2019). UC-Wetting
500 UC-Drying
The critical unanswered question of this study, as with other HC-Wetting
similar research showing positive results for EB-GCLs relative to 400
HC-Drying
NaB-GCLs, is for how long will this benefit persist, and what will 300
be the performance over the longer term?
35 200
(a)
30 GCL 100
Crack Proportion (%)
(b)
EB-GCL
25 0
4 5
1 10 100 1000 10 10
20 Matric Suction, (kPa)
m
15
Gravimetric Water Content, w (%)
600
10
UC-Wetting
500
5 UC-Drying
400 HC-Wetting
0 HC-Drying
35 40 50 60 300
Temperature (oC) 200
1.4 100
Average Crack Width (mm)
GCL (b) (c)
1.2 0
EB-GCL 1 10 100 1000 10
4
10
5
1 Matric Suction, (kPa)
m
0.8
Figure 4.28. Soil-water retention curves of untreated clay (UC) and
0.6 HYPER clay (HC): (a) distilled water; (b) 10 mM CaCl 2; (c) seawater.
0.4 For all liquids, the gravimetric water content for the HC at a
0.2 given suction for both wetting and drying SWRCs was greater than
that for the UC. In the case of distilled water (Figure 4.28a), the
0 HC demonstrated higher suction for the same water content
35 40 50 60
suggesting the presence of more microvoids (intra-aggregate
Temperature (oC)
pores), resulting in increased matric suction. The intercalation of
Figure 4.27. Effect of temperature on cracking from tests on specimens of the Na-CMC polymers within the montmorillonite platelets in HC
unenhanced GCL and enhanced bentonite GCL (EB-GCL) treated with (Di Emidio 2010) results in a decrease in pore diameter, which
polyacrylamide polymer: (a) crack proportion; (b) average crack width
212
correlates with an increase in the matric suction (Khan et al. 2021). swell pressure test are more complex than those for the standard
In addition, the increase in gravimetric water content also can be index tests. Also, the testing durations for the swell pressure test
attributed to the absorption of water by the Na-CMC anionic typically are longer than testing durations for the standard index
polymer. In the case of the wetting SWRC curve, the decrease in tests. For example, the testing durations of the swell pressure tests
gravimetric water content with increasing electrolyte conducted by Di Emidio et al. (2013) lasted from 5 to 10.5 days.
concentration (i.e., distilled water <10 mM CaCl2 < SW) was In contrast, those for the SI testing procedure, including the
sharper for the UC relative to the HC, which can be attributed to a preparation of the specimen, typically are is
compression of the DDLs in the UC (Khan et al. 2021). The typical still a need to develop rapid and simple screening tests relevant to
hysteresis effect between wetting and drying curves is also EB-GCLs and reflect the unique mechanisms underlying the
apparent for HC and UC specimens. The hysteresis effect hydraulic behaviour of this class of materials.
decreases with an increase in the ionic strength of the solution.
However, the hysteresis effect was more distinct in the HC 4.7 Summary of current status and remaining questions
specimens, which was attributed to the possibility that polymer
This section of the paper presents a state-of-the-art review of the
intercalation decreased the contact angle and diameter of
development and evaluation of enhanced bentonites in
conducted pores, thereby increasing the capillary pressure and
geosynthetic clay liners (GCLs), referred to as enhanced-bentonite
water retention (Khan et al. 2021).
GCLs (EB-GCLs), for use as barriers in chemical containment
4.6.5 Screening methods applications. Enhanced bentonites (EBs) are natural bentonites that
Lee et al. (2005) showed that the k of unenhanced GCLs are chemically modified to achieve greater resistance to hydraulic
permeated with DW and CaCl2 solutions could be correlated incompatibility and low hydraulic conductivity when permeated
with the index properties of the bentonite portion of the GCL, with liquids of extreme chemistry, including high ionic strength
including the liquid limit (LL) test (ASTM D4318), sedimentation liquids as hypersaline solutions and/or highly acidic or highly
volume, and swell index (SI) test (ASTM D5890), when alkaline solutions. Chemicals commonly used for the enhancement
performed using the same permeant liquids. Thus, these index of natural bentonites include anionic polymers such as polyacrylic
properties may help provide a quicker and less time-consuming but acid (PAA) and carboxymethylcellulose (CMC), cyclic organic
preliminary indication of the potential for a specific GCL to carbonates such as propylene carbonate (PC) and glycerol
achieve a suitably low k for a specific liquid. However, Scalia et al. carbonate (GC), and other organic compounds such as ethylene
(2018a) demonstrated that indicator tests used for conventional glycol (EG). Prevalent EBs include bentonite polymer composite
GCLs generally are not effective for predicting the k of polymer (BPC), dense prehydrated (DPH) bentonite, HYPER clay (HC),
EB-GCLs. Tests shown to be ineffective include LL, SI, and fluid multiswellable bentonite (MSB), and a wide variety of
loss conducted with the permeant liquids. This lack of correlation commercially available EBs with proprietary (unknown)
results from the different mechanisms underlying hydraulic enhancements collectively referred to as contaminant resistant
compatibility of polymer enhanced bentonites relative to natural clays (CRCs). These enhancements may be grouped into three
NaB-GCLs (viz., swelling). More recently, Gustitus and Benson categories, viz. polymer enhancements (e.g., PAA and CMC),
(2021) introduced the flow-swell index test for BPC-GCLs that organic compound enhancements (e.g., PC, GC, and EG), and
addresses the swelling of bentonite granules and retention of unknown enhancements (i.e., CRCs). Low-quality bentonites that
polymer gel in the pore space of a BPC. They illustrate that the are beneficiated to perform similarly to higher quality bentonites
flow-swell index can differentiate between BPC-GCLs with low and bentonites modified with quaternary ammonium cations
and high k and indicate that more extensive testing is needed to known as organobentonites or organoclays have been excluded
develop a threshold flow-swell index that applies to a broad from consideration because these modifications typically are
range of BPC-GCLs and permeant solutions. performed for purposes other than the achievement of hydraulic
Also, Christian et al. (2020) have recently shown that one of compatibility when permeated by liquids of extreme chemistry.
the reasons the SI of polymer-enhanced bentonites does not Two primary mechanisms have been postulated for improved
correlate with k is because the standard specimen preparation chemical resistance and hydraulic compatibility of EBs and EB-
procedure for the SI test based on ASTM D5890 requires sieving GCLs based on permeation with liquids of extreme chemistry, viz.
of the material, such that a portion of the polymer component of polymer clogging of interparticle, intraparticle, and/or
the enhanced bentonite can be removed before the SI is measured. intergranular (interaggregate) pores, and intercalation of polymers
They hypothesize that the correlation can be improved when the or organic compounds within the interlayers of the
standard swell index testing procedure is modified such that the montmorillonite crystals comprising the montmorillonite mineral
polymer component is not removed via sieving. component of bentonite particles. Preliminary evidence based on
Due to these limitations with the SI test, Di Emidio et al. (2013) XRD analysis and theoretical evidence based on the fabric
recommended using one-dimensional (1D) swell pressure tests boundary surface (FBS) theory supports intercalation as the
because the mechanisms governing swell pressure are different primary mechanism for MSB and HC, whereas both SEM Imaging
from those governing swell index. Hydrated bentonite adsorbs a and the FBS theory support pore-clogging as the primary
large amount of water. Because of this water entering the mechanism for BPC. In the case of DPH-GCL, both intercalation
interparticle pores and the interlayers in individual of CMC and densification via vacuum preconsolidation likely
montmorillonite crystals comprising the bentonite particles, a contribute to improved chemical resistance, although the relative
swell pressure is developed, leading to an increase in volume and effect of each contribution remains unknown. Overall, further
accompanying stress on the surrounding materials that are experimental and theoretical studies are required to confirm these
manifested as the swell pressure. In contrast, the macropore preliminary conclusions for wider range of conditions, including
formation in the SI test leads to an increase in volume but does not variations in the degree of pre-hydration of these EBs, effective
increase the interstitial stress between clay particles or the stress stresses, and permeant liquid chemistries. Also, since the long-
on the surrounding materials. Furthermore, because of the term persistence of these mechanisms is largely unknown, longer-
hydrophilic nature of polymers used for polymer enhanced term experimental studies should be conducted considering these
bentonites, the measured swell pressures for these bentonites are broader ranges of conditions. Finally, for polymer enhanced EBs
higher than those for the unenhanced bentonites, which overcomes and EB-GCLs such as BPC, additional study is needed to
the aforementioned technical limitations of the SI test. understand the mechanisms that lead to polymer redistribution and
However, the testing procedures and equipment for the 1D retention within the pores of these materials and better understand
213
the mechanisms governing hydraulic compatibility for polymer author of this paper has observed biodegradation of a polymer
EB-GCLs. claimed by the supplier to be firmly bound when stored under
The polymer clogging mechanism for BPC has been attributed moist, anaerobic conditions. Polymer degradation is well
to the formation of polymer hydrogel within the pores of the BPC, recognized as a problem for drilling fluids and other excavation
which is largely a function of the polymer loading and the pore support fluids – especially for naturally derived polymers. For
water chemistry during permeation. Under certain conditions, such polymer enhanced GCLs, relatively little research has been
as relatively high polymer loadings, the hydrogel may be eluted reported on the amount of polymer that can be bound. This may be
from the BPC, resulting in inaccurate laboratory measurement of very small compared to the amounts used in some enhanced GCLs
the hydraulic conductivity, especially for long-term behaviour. (Lam and Jefferis, 2019). If the polymer is not bound but merely
Thus, understanding polymer elution behaviour is critical to provides a viscous pore blocker (Ejezie et al. 2020), an appropriate
understanding the long-term behaviour of polymer EB-GCLs and question to ask is why use clays such as bentonite with their
ensuring that laboratory hydraulic compatibility results are intraparticle porosity. Could silts or sands give similar (if equally
accurate. Also, the necessity to continue permeation until no time-limited) performance? Also, what are the properties of the
further polymer is eluted (mobile) requires further research. polymer-porewater system (solution?) when prepared at the
The evidence to date suggests that EBs and EB-GCLs can offer expected in-GCL concentration?
improved resistance to hydraulic incompatibility and improved
performance with respect to solute transport via diffusion and
semipermeable membrane behaviour, relative to that of natural
bentonites when exposed to relatively high ionic strength liquids.
However, hydraulic conductivity test results based on permeation 5 GEOMEMBRANE AND GCL IN A BARRIER SYSTEM
with extreme pH liquids (acidic or alkaline solutions) are limited
and mixed. Also, the results of hydraulic compatibility tests for 5.1 Design life
CRCs are generally more variable than those for the more standard Barrier systems should be designed to control contaminant escape
EBs and EB-GCLs, likely due to the variability in, and the to a negligible level or the ingress and egress of water or gases to
unknown nature of, the enhancement materials and procedures for an acceptable level. The contaminating lifespan of a waste disposal
CRCs. Thus, for practical applications, prudence dictates that facility is, by definition, the period during which the barrier system
hydraulic conductivity, membrane, and/or diffusion testing be must control contaminant escape to a negligible level (MOE 1998;
conducted using the specific EB or EB-GCL with the specific Rowe et al. 2004). Thus, the design life for a facility containing
liquid(s) to be contained. contaminants will be either when the source of contaminant is
The results of limited studies performed using HCs and HC- removed (e.g., for a leachate lagoon) or the facility reaches its
GCLs, DPH-GCL, and MSB indicate that EBs and EB-GCLs are contaminating lifespan (e.g., a MSW landfill). The design life for
more resistant than natural bentonites to the detrimental effects of a lagoon, pond, or dam is the time the barrier system needs to
wet-and-dry cycling with medium-to-high ionic strength liquids control the escape of fluids to an acceptable level. While the design
and elevated temperatures. However, additional studies to evaluate life of the cover is the length of time the cover is required to control
these effects for broader ranges of EBs and EB-GCLs and testing the ingress and egress of liquids and gases to an acceptable level.
conditions are needed. The design life of ponds and lagoons may be a few decades. The
Reliable and accurate screening methods are needed to provide design life of a liner for a dam is commonly about a century. The
more rapid and less costly assessments of the potential for contaminating lifespan of a waste disposal facility depends on its
chemical compatibility between site-specific EBs and EB-GCLs size (predominantly the thickness of waste) and the nature of the
and containment liquids. In this regard, the recently developed waste. It typically ranges from about 100 to 300 years for
flow-swell index test shows promise. municipal solid waste, to over 500 years for low-level radioactive
Most of the research described in this section has been based waste, and to millennia for mine waste and long-lived low-level
on well-conceived and conducted laboratory experiments that have radioactive waste. Thus, the service life of the barrier system
largely focused on the fundamental behaviour of the material components is a critical consideration since the system requires
under essentially ideal conditions. The typical type of studies as service life greater than the contaminating lifespan or design life
described herein, while extremely valuable at providing insight as appropriate.
into the material behaviour under the conditions examined, need
to be viewed with considerable caution in terms of providing 5.2 Designing liners as part of a system
recommendations or comments regarding their likely good 5.2.1 Geomembrane-clay composite liners
performance in the field. Performance in the field depends on The previous sections have primarily considered the performance
many variables that are difficult to simulate under laboratory of geomembranes and GCLs alone. However, there are many
conditions but may significantly impact the performance in the applications where HDPE geomembranes and GCLs are used
field. Even the performance of traditional NaB-GCLs commonly together to form a single composite liner. In these applications, the
used as low permeability liners is complex, as will be illustrated in strength of one component offsets the weaknesses of the other. In
§5. EB-GCLs are at least one additional step more complex. While particular, while intact HDPE is practically impermeable to
there is still plenty of good laboratory-scale research that can be aqueous solutions and many gases, the leakage through even
conducted, the future major insights into EB-GCL behaviour and relatively small holes can be significant. On the other hand, the
performance likely will come from well-documented, full-scale GCL, while still having a relatively high hydraulic resistance,
field trials of EB-GCLs in applications for which there are cannot match the intact geomembrane for resistance to fluid flow
currently being marketed and used. but is highly effective in limiting the leakage through holes in the
The long-term performance of enhanced GCLs will depend on geomembrane (e.g., Giroud and Bonaparte 1989; Rowe 1998,
how much polymer can be bound to the bentonite. The unbound 2006, 2012, 2018, 2020). Therefore, the geomembrane and GCL
polymer will be sensitive to elution, especially as, over time, the form a single composite liner (Figure 5.1a). Similarly, a
polymer is diluted by influent permeating liquid. It will be geomembrane in intimate contact with a compacted clay liner
vulnerable to degradation (both anaerobic and aerobic) though (CCL) is a single composite liner.
bound polymer will not be immune. Laboratories may be relatively The importance of the hydraulic conductivity, k, of a GCL and
sterile with little opportunity to initiate biodegradation. The second the effect of permeant and stress on k are well-recognized, as
214
highlighted in §4. However, it is generally not well recognized that
combining a geomembrane and GCL (or CCL) to form a
composite liner introduces a new and important parameter: the
GMB-GCL interface transmissivity. This interface transmissivity
controls the ability of fluid to spread laterally at the interface
contact between the GMB and GCL (or CCL). With a few notable
exceptions, relatively little attention has been paid to the
importance of the interface transmissivity between the
geomembrane and GCL. It is often implicitly and incorrectly
assumed that the interface transmissivity is proportional to the k.
However, there is little correlation between the GCL k and GMB-
GCL interface transmissivity (Rowe 2012, Abdelrazek and Rowe
Figure 5.1. Barrier system with a single composite liner: (a) GTX filter
2019). The interface transmissivity can be affected by bentonite,
over coarse granular drainage layer over GTX protection layer over
but it is controlled by the geometry and the path of least resistance HDPE GMB with a hole over GCL over subgrade (b) GTX filter over
at the GMB-GCL or GMB-CCL interface. coarse granular drainage layer over sand protection layer over HDPE
Rowe (2020a) provides a recent overview of GMB-GCL GMB with a hole over GCL over CCL over subgrade.
interface transmissivity. Rowe (2018) discusses the effect of
leaving a geomembrane exposed over compacted clay and the
significant impact the desiccation of even a few millimetres of the
clay at the clay surface can have on interface transmissivity.
Design calculations should consider the potential implications of
desiccation of the clay surface because it can effectively decouple
the performance of the geomembrane from the CCL to the point
that it no longer acts as a composite liner but they act as two
separate liners. One approach to addressing this concern is to have
a composite liner comprised of a GMB/GCL/CCL, as shown in
Figures 5.1 and 5.2b. In this scenario, the self-healing capacity of
the GCL minimizes the effect of desiccation of the clay liner
surface (Rowe 2018). Thus, when controlling leakage is a critical
consideration, and the advantages of the thickness of the CCL are
desired, the composite liner system may be comprised of an HDPE
geomembrane over a GCL over a CCL (Figure 5.2b; Priyanto et al. Figure 5.2. Barrier system with a double composite liner: (a) GTX filter
2019; Rowe et al. 2020). Here, the GCL is less prone to cracking over coarse granular drainage layer over GTX protection layer over
and has self-heal capacity even if it does crack. The use of the GCL HDPE GMB with a hole over GCL over granular drainage layer over
provides a geomembrane-GCL interface transmissivity orders of GTX protection layer over HDPE GMB over GCL over subgrade (b)
magnitude lower than that of the compacted clay (Rowe 2012). GTX filter over coarse granular drainage layer over sand protection layer
over HDPE GMB with a hole over GCL over granular filter over granular
5.2.2 Geomembrane-tailings composite behaviour drain over sand protection layer over GMB over GCL over CCL over
Another example of how adjacent materials can interact to subgrade.
perform better than either material alone by forming a composite
liner arises when silty sand or finer tailings are placed over a
geomembrane in a tailings storage facility (Figure 5.3). The
leakage through a geomembrane hole is substantially reduced by
the head loss that occurs in the tailings within about 5 hole-
diameters of the hole (Joshi et al. 2017; Rowe et al. 2017; Chou
et al. 2018, 2021; Rowe and Fan 2021; Fan and Rowe 2021,
2022a). The subgrade also plays a critical role with respect to the
potential for piping in the shorter-term and tensile strains that
control the service life of the geomembrane in the longer-term
(Fan and Rowe 2022). Figure 5.3. A single geomembrane covered by silty sand tailings forms a
5.2.3 Barrier system component interactions composite system with the tailings above and within the hole,
Rowe (2011) emphasized that while consideration of the substantially reducing leakage through the hole. The nature of the
characteristics of the individual components of the design (e.g., subgrade can also control the service life of the geomembrane.
geomembrane or GCL) is necessary, it is not sufficient to ensure a For example, it is now well established that coarse uniform
good design. Liners do not operate in isolation. They are part of a gravel provides the best drainage system for leachate in a
larger barrier system (Figures 5.1 and 5.2). Therefore, it is essential municipal solid waste landfill (Rowe 2005; Rowe and Yu 2010,
to consider the potential interactions between different system 2013). However, that excellent drainage layer may compromise
components and how they may affect the entire system's the service life of the barrier system if the strains induced in the
performance. It is particularly critical when seeking to “improve” underlying geomembrane give rise to premature stress cracking
or optimize a system component that one also looks carefully for (Abdelaal et al. 2014; Ewais et al. 2014b). Thus, in addition to
the potential unintended consequences of that optimization. Rowe improving the leachate collection system, one must also introduce
(2011) demonstrated how an action that may enhance the a system to adequately minimize the strains in the geomembrane
performance of one part of the system might negatively affect (Rowe et al. 2013; Rowe and Yu 2019; Rowe et al. 2019c). More
other parts of the system. The design objective should be to ensure generally, the potential interaction between the soils or other
optimal performance of the entire system, not the components in materials above or below a component in a liner system must be
isolation. considered in any design.
215
5.3 Impact of the environment on barrier system performance appropriately qualified expert independent of the contractor. The
critical importance of appropriate CQA has been well-recognized
When designing a barrier system, consideration must also be given
for more than 35 years, as highlighted in the classic paper by
to the environment in which it is constructed and the environment
Giroud and Fluet (1986). These days CQA should include a leak
in which it will serve its design function. For example, when a
protection/drainage layer is placed, the weather conditions will location survey before and after cover soil is placed on the
significantly influence subsequent leakage through a composite geomembrane (Giroud 2016; Gilson-Beck 2019; Rowe 2020).
liner. This is because the weather conditions influence the liner
temperature. The hotter the geomembrane when covered relative 5.5 Barrier system Service life
to when it was placed and welded, the greater the thermal 5.5.1 Geomembrane service life
expansion and, consequently, the number of wrinkles and the area When landfills were first being built with geomembrane liners 50
beneath interconnected wrinkles in the geomembrane. The height years ago and for the next 40 years, the effect of leachate on the
of wrinkles and area beneath wrinkles directly affect the service life could have been regarded as a known unknown.
probability that a hole intersects a wrinkle and the area over which Research over the last 20 years has changed that, and now the
leachate can spread once it flows through that hole (Figure 2.15; service life of a geomembrane liner can be reasonably estimated
Giroud and Morel 1992; Chappel et al. 2012a,b; Rowe et al. and regarded as known if one simply asks the question (Rowe
2012a,b; Rentz et al. 2017). Wrinkles can have a detrimental
2020). Service-life of a modern geomembrane depends on
impact on system performance for multiple reasons. These include
1. geomembrane used – (polymer and additive package), and
the facts that:
2. The exposure conditions:
(i) the wrinkle acts as a target for damage. A wrinkle is more
Elements (UV; rapid changes in temperature),
likely to have a hole than a flat geomembrane sheet
Chemical composition of fluid in contact with a
(Gilson-Beck 2019; Rowe 2020),
geomembrane,
(ii) a hole in the wrinkle would allow leakage orders of
Temperature,
magnitude greater than the hole alone if it were in direct
Nature of exposure,
contact with the clay liner (Rowe 1998; 2005; 2012;
Sustained tensile strains in geomembrane, and
2020),
Seams (welds)
(iii) the local tensile strain in the geomembrane at a wrinkle
makes the wrinkle a location for early stress cracking, HDPE geomembranes have evolved substantially over the last 30
especially when a weld passes over a wrinkle, and years. Today's high-quality HDPE geomembranes are highly
(iv) the wrinkle acts as a point for the accumulation of humid engineered materials, and one can select a geomembrane with an
air beneath an exposed geomembrane and subsequently a appropriate resin and additive package for a given application (§2).
point for condensation of the water vapour i. focusing the With appropriate testing to confirm the compatibility of the
flow beneath the wrinkle as the geomembrane cools in the geomembrane with the fluid to be contained, appropriate design to
evening after a warm day (thereby potentially contributing minimize tensile strain in the geomembrane, and appropriate
to downslope erosion; §5.5.2.2). construction and CQA to ensure that it is constructed following the
(v) the wrinkle creates an area of reduced stress into which design, a modern HDPE geomembrane used as a landfill liner
bentonite in an underlying GCL could migrate, could have a service-life of many hundreds of years and in some
(vi) GCL below a wrinkle will have a higher k than the GCL cases millennia (Rowe 2020, Rowe et al. 2020).
in direct contact with the geomembrane.
Unfortunately, the level of attention being paid to providing
(vii) the wrinkle increases the risk of internal erosion when the
premium geomembranes with a long service life has not been paid
GCL is above a drainage layer in double liners (Figure 5.2;
to the other geosynthetic components of the barrier systems such
§5.5.2.2 (ii)).
Thus, the cumulative effect of a transient event, such as the rise as the geotextiles shown in Figures 5.1 and 5.2. Unlike modern
and fall of the sun each day with the consequent rise and fall of HDPE geomembranes, geotextiles and geocomposite drains are
geomembrane temperature throughout the period that the generally not well stabilized against thermo-oxidative degradation.
composite liner is left exposed, can cause permanent damage to the While they may have service lives up to decades, there is a paucity
GCL beneath the geomembrane. of data on longer-term performance. This should be an essential
Similarly, the wrinkles that were present when the cover soil consideration in designing facilities with a century or more design
will be placed will be locked into the system. While the locked-in life. This is especially relevant to landfills and mine waste.
wrinkles may get smaller with loading, they generally do not 5.5.2 GCL service life
disappear (Stone 1984; Soong and Koerner 1998; Brachman and
Gudina 2008). Wrinkles provide a conduit for enhanced leakage 5.5.2.1 Factors affecting the service life of a GCL
through any hole in the geomembrane on or near a wrinkle (Rowe The majority of GCLs on the market comprise a layer of bentonite
1998, 2012). In both cases, transient events not considered in sandwiched between two geotextiles and needle-punched together.
design can have a permanent and substantial impact on the leakage However, the description is deceptively simple. The Canadian
through a barrier system and, consequently, on the surrounding GCL manufacturer produces 50 different GCLs that meet this basic
environment. definition and, except for the multicomponent GCLs, they all have
the same k 5x10-11 m/s (ASTM D5887) given on the
5.4 Need for Construction Quality Assurance (CQA) manufacturer’s product sheets, However, they have significant
variance in price and field performance. The factors affecting the
In barrier system construction: ‘you get what you inspect, not what field performance of the GCL include:
you expect’ (see §15.17). For every important project, an effort the type and granularity of the bentonite,
should be made to minimize the effect that construction activities the mass of bentonite per unit area,
will have on leakage through the liner. This should include the amount of needle-punching,
minimizing the number of holes in the geomembrane when it goes the mass of the nonwoven needle-punched cover geotextile
into service and minimizing leakage through any undetected holes (upper during manufacture),
or holes that subsequently develop. Doing so will require high- the type and mass of the carrier geotextile (lower during
quality, full-time and knowledgeable observation of every step of manufacture),
liner installation and covering as part of the construction quality whether the needle-punched fibres are thermally bonded to the
assurance (CQA) program. The CQA should be conducted by an carrier geotextile,
216
whether there has been a plastic layer and if so context, the bentonite in the GCL may be expected to have a very
was the layer a coating applied in a molten state so that long service life (thousands of years) provided that:
it is impregnated into and onto carrier geotextile (known (i) there is no significant loss of bentonite during installation
as a coated GCL), or of the liner,
was it produced separately and glued to the carrier (ii) there is no significant lateral movement (thinning) of
geotextile (a laminated GCL), bentonite or gaps/holes post-installation,
what polymer is used for the coated/laminated layer, and (iii) there is no significant internal erosion,
what is the mass per unit area of the coated/laminated (iv) there is no change in macrostructure due to wet-dry or
layer? freeze-thaw cycles or this is accounted for in the design k,
The needle-punched structure of these GCLs serves three potential (v) the k was selected based on a realistic assessment of the
purposes: field conditions (chemistry of fluid to be contained, pore
i. keeping the bentonite contained between the geotextiles fluid adjacent to the GCL, effective stress etc.), and
following GCL manufacture through to and after it is (vi) it is installed with appropriate seams.
installed, It follows from the preceding that the service life of a GCL will
ii. providing restraint to the GCL during hydration at low stress also be of the order of millennia provided the six factors above
levels, thereby reducing the swelling and improving the have been addressed in design and construction and provided that
hydraulic characteristics of the GCL at low effective stress. the GCLs long-term performance is not contingent on the service
iii. providing internal shear strength. The needle-punched fibres life of the geosynthetic components.
increase the apparent friction angle from about 8-10° for The factors affecting the bentonite above, together with the
hydrated bentonite alone to a higher value. The magnitude dependence of the GCLs service life on the long-term performance
of the increased friction angle and apparent cohesion will of the geosynthetics, will be discussed in the following subsections.
depend upon the amount of needle-punching and how the
needle-punched fibres are anchored to the carrier geotextile. (i) Avoiding loss of bentonite during installation of the liner
A coating or laminate applied to GCL reduces the k by about Good CQA is essential to ensure that appropriate measures are
an order of magnitude from less than 5x10-11 m/s (ASTM D5887) taken in moving and placing the GCL while avoiding loss of
to typically less than 5x10-12 m/s (ASTM D5887) for bentonite or significant movement during installation.
multicomponent GCLs with either a coating or laminated plastic (ii) Avoiding significant lateral movement (thinning) of bentonite
layer. In addition to reducing k, the coating can also protect the or gaps/holes post-installation
GCL from hydrating or losing moisture in a manner that could Potential causes of zones with inadequate bentonite post-
impact its ultimate performance. installation include:
Although GCLs have a typical ASTM D5887 hydraulic • voids in subgrade that are bridged by the GCL,
conductivity, as stated in the previous paragraph, the actual k in the • traffic (human or machine) over the GCL, especially if
field can vary from an order of magnitude lower than that indicated permitted to hydrate prematurely,
above to 4 to 5 orders of magnitude higher depending on the • protrusions in the subgrade,
factors listed above together with the exposure conditions. A • stones that are not initially in contact with the liner but
discussion of how these factors affect GCL field performance is migrate into contact post-installation due to the
provided by Rowe (2020). consolidation of compacted till or movement of stones due
5.5.2.2 Discussion of factors affecting the service life of a GCL to freeze-thaw,
The following discussion of GCL service life is primarily based on • gravel/stone in the cover soil directly above the liner,
Section 12.10.2 of Rowe et al. (2004) but has been updated. • significant internal erosion (see (iii) below),
The primary active component of a GCL is bentonite. • wrinkles in a geomembrane (see below),
Bentonite is a naturally occurring swelling clay formed by the • premature wetting and shrinkage of the GCL, and/or
deposition of volcanic ash that was consolidated into • downslope erosion (see below).
shale/mudstone. The nature of the bentonite may depend on The foregoing can all be mitigated by good construction quality
whether the ash settled in saltwater or freshwater. The lowest assurance. The last four can also be mitigated, to some extent, by
permeability bentonite is natural sodium bentonite deposited in appropriate design, as discussed subsequently.
saltwater and was first discovered in the Cretaceous Benton shale Wrinkles in the geomembrane should be minimized for the reasons
found at Fort Benton in Montana but subsequently found in given in §5.3. Wrinkles can be mitigated by the designer specifying
Wyoming, North Dakota, South Dakota, Colorado, Kansas, and that the protection geotextile and granular layer shall only be
Nebraska, USA. Most of the bentonite used in high-quality GCLs placed over the geomembrane when:
manufactured in North America is natural sodium bentonite The area beneath wrinkles is less than 6% of the
obtained by mining the shale and grinding it into either coarse geomembrane area covered when the overlying material is
grains (coarse sand size), fine grains (medium to fine sand size), or placed (i.e., no visually interconnected wrinkles and any
powdered form (silt size). Most high-grade sodium bentonite is wrinkles present are hydraulically isolated from other
Wyoming bentonite, with Turkey also having deposits of high- wrinkles).
quality sodium bentonite. Mixed sodium/calcium bentonite is All wrinkles are less than 50 mm in height.
mined in Australia, Greece, India, Russia, and Ukraine. Calcium The longest wrinkle is less than 10 m in length.
bentonite is mined in southern USA, China, Germany, Greece, and having good CQA to ensure compliance with the
India, New Zealand, and Turkey. China is now the world’s largest specifications. While minimizing wrinkles is desirable, it is
producer of bentonite, followed by the USA and India. essential that the geomembrane is not covered while in a state of
The minerals in bentonite (predominantly montmorillonite) tension either due to thermal contraction on very cold day or due
have been around for millennia and are likely to remain for to efforts to avoid wrinkles.
millennia under typical field applications. However, its k can be
affected by many factors, particularly the chemistry of the water Shrinkage of the GCL due to wet-dry cycles
with which it is hydrated and permeated and the effective stress Manufacturers’ installation guidelines recommend covering a
under operating conditions. Thus, in design, consideration must be GCL with at least 0.3 m of soil shortly after the liner is installed
given to the operative k under the expected field conditions. This k (i.e., before the GCL can hydrate). If this is followed, then
will, as noted above, depend both on the details of the GCL shrinkage and downslope erosion is highly unlikely.
manufacturer itself and the exposure conditions. Within this Panel shrinkage occurs when GCLs are left exposed without
217
at least 0.3 m of cover soil for a prolonged period if, during this without a change in apparent k, which remained at about 1.5x10-11
period, they can experience wet-dry cycles with hydration m/s. With a silty sand subgrade having 25% fines, none of the
followed by water evaporation. These repeated cycles can result in GCLs experienced any internal erosion at a head of up to 49 m. In
shrinkage of some GCLs both longitudinally and/or transversely, landfill projects, the service life of the geotextile component is
reducing the overlap and ultimately creating gaps between the likely to be less than the contaminating lifespan of the landfill. In
overlapped panels (Thiel and Richardson 2005; Koerner and these situations, a suitable granular filter should be placed between
Koerner 2005; Thiel et al. 2006; Bostwick et al. 2010; Rowe et al. the GCL and any drainage layer unless it can be demonstrated that
2010, 2011a, 2013, 2018, Brachman et al. 2018). for the highest reasonable head, no erosion is likely to occur for
GCL panel shrinkage can be mitigated by preventing wet-dry the proposed GCL with the hydrated bentonite in direct contact
cycles when there is negligible ballast soil over the GCL. This can with the drainage layer.
be best achieved using a multicomponent GCL. Encapsulation of
the bentonite between the overlying geomembrane and the coating (iv) Avoiding or accounting for macrostructure due to wet-dry or
applied to the carrier geotextile can prevent moisture uptake and freeze-thaw cycles
hence no wet-dry cycles. This will also prevent downslope erosion Wet-dry cycles and freeze-thaw cycles can change the structure of
discussed below. The downside of this approach is that if there is the bentonite and increase k. The effect on k is a function of the
a hole in the geomembrane, either in the shorter- or longer-term, level of cation exchange that has occurred before or during the wet-
the bentonite will be hydrated by the leachate in the environs of the dry or freeze-thaw cycle. For example, Table 5.1 summarizes the
hole. This is most likely and most critical below a wrinkle where results for two NaB GCLs and three EB-GCLs subject to the same
the effective stress is zero. If the leachate-bentonite interaction cation exchange from the hydrating simulated pore water prior to
increases k more than the subgrade pore water-bentonite being subjected to wet-dry cycles by Rowe and Hamdan (2021a)
interaction, then an alternative approach for minimizing shrinkage in the swell index values of 9-12 mL/2g. The k values reported
and downslope erosion is to use the multicomponent GCL with the were obtained with 8-hour wetting cycles and drying cycles
coating facing up in contact with the geomembrane. In this way, ranging from 16 hours to 14 days for each GCL for one year.
the GCL can hydrate from the subsoil. Although the exposure conditions were identical, the impact on k
If a multicomponent GCL is not available or suitable for a varied substantially with k increasing from 1-5x10-11m/s prior to
given application, the magnitude of the thermal cycles that cause cation exchange to 0.9-4 x10-10 m/s with cation exchange.
the generation of both wrinkles and the evaporation of moisture However, after the wet-dry cycles there was no significant change
from the GCL can be reduced with the use of a white in k for the GCL with NaB bentonite and a SRNW carrier
geomembrane (Take et al. 2016; Rentz et a; 2016, 2017). A white geotextile. For the GCL with NaB bentonite and W carrier
geomembrane will reduce the magnitude of, but not prevent, geotextile, k increased tenfold to 4 x10-9 m/s with this increase due
eventual shrinkage and downslope erosion. to mini-wrinkles that developed in the woven geotextile caused
If a multicomponent GCL is not available or suitable for a local thinning of the bentonite. For the three EB-GCLs, the wet-
given application, the magnitude of the thermal cycles that cause a dry cycles increased k by between a factor of 10 and 200 relative
generation of both wrinkles and the evaporation of moisture from to that with cation exchange alone (i.e., to between 5 x10-8 m/s’ and
the GCL can be reduced with the use of a white geomembrane 2x10-7 m/s). Thus, the natural sodium bentonite was more resistant
(Take et al. 2016; Rentz et al. 2016, 2017). The white to wet-dry cycles and freeze-thaw cycles than alternative enhanced
geomembrane will reduce, but not prevent, eventual shrinkage and bentonites. However, there was still an impact and the first
downslope erosion. objective of design, where practical, is to avoid wet-dry or freeze-
thaw cycles. If this is not possible, the effect of freeze-thaw can be
Downslope erosion occurs when a GCL in a composite liner is mitigated by minimizing cation exchange before and during the
exposed for a prolonged period without at least 0.3 m of cover soil. wet-dry or freeze-thaw cycles and adjusting the design k to account
Typically, the GCL in the composite liner will uptake moisture for the change in macrostructure due to the cycles.
from the subgrade, followed by water evaporation from the GCL For a composite liner where freeze-thaw or wet-dry cycles can
into the airspace between the GCL and geomembrane, followed by not be avoided, one design approach is to encapsulate the bentonite
condensation of that moisture and the consequent distilled water between the geomembrane above and the coating/laminate bonded
running down even a gentle slope (e.g., 3%). The distilled water to the carrier geotextile below. Provided that the geomembrane is
running over the GCL can cause bentonite erosion to the bottom of placed without allowing the GCL to hydrate (e.g., from rainfall),
the slope. The bentonite can be partially or totally eroded and can these two layers serve to prevent hydration of the GCL. Without
leave two layers of geotextile with effectively no bentonite hydration, neither wet-dry cycles nor freeze-thaw cycles will
between them (Ashe et al. 2014, 2015; Brachman et al. 2015; have any effect on the hydraulic performance of GCL. If a small
Rowe et al. 2016a,b). This mechanism only occurs with distilled hole in the geomembrane or coating permits minor local hydration,
water. It does not occur with tap water flowing over the GCL. It the general effect will be no consequence.
has only ever been reported when the composite liner is left
Table 5.1. k permeated with tap water at 15 kPa after cation exchange
exposed for a prolonged period to solar radiation, contrary to the and 24-315 wet-dry cycles (modified from Rowe and Hamdan 2021a)
manufacturers’ recommendations.
GCL downslope erosion can be mitigated by preventing wet- Bentonite Ma Carrier k x 1010
dry cycles when there is negligible ballast soil over the GCL. The granularity (g/m 2
) GTX (m/s)
approaches described above on mitigating panel shrinkage will be NaB Fine 5527 NWSR_TT 0.9-3
equally effective at mitigating downslope erosion. NaB Powder 4557 W_TT 4-20
EB Fine 5257 NWSR_TT 2-1000
(iii) Avoiding significant internal erosion.
EB Powder 5160 W_TT 20-500
Rowe and Orsini (2003) examined the potential for internal
EB Powder 5775 NWSR 2-2000
erosion for several subgrades and GCLs. A GCL with a woven
NWSR= scrim reinforced nonwoven; W= woven, TT= thermally treated
carrier over the pea gravel subgrade experienced internal erosion
at a head of 27 m. Bentonite erosion caused a two order of For a GCL used alone, the GCL is protected from hydration
magnitude increase in apparent k and permittivity by the time of from the subgrade and hence cation exchange by the plastic layer
test termination. A GCL with a nonwoven carrier experienced one and using a clean sand layer with minimal cations as the ballast
order of magnitude increase in apparent k and permittivity at a head layer above the GCL. In this scenario, it is likely that the bentonite
of 28.5 m. A GCL with a scrim-reinforced nonwoven carrier did will be hydrated from above by rainfall or snowmelt. Therefore, at
not experience any internal erosion. It sustained a 41 m head least 0.3 m and preferable 0.6 m thick soil layer should be placed
218
quickly after placement of the GCL and before hydration of the the specifications should be permitted without the expressed
GCL. Generally speaking, since increasing thickness of fill above written approval of the designer of record.
the GCL minimizes the effect of wet-dry or freeze-thaw cycles, the
sooner the soil is placed the better. Even with wet-dry and/or (vi) GCL is installed with appropriate seams.
freeze-thaw cycles, GCL k of a multicomponent GCL is likely to To provide appropriate seams, the GCL specification should
be less than 5x10-12 m/s for the service life of the coating since the generally require that (Brachman et al. 2016; Joshi et al
coating controls the hydraulic resistance. However, some impact ):
on k may be expected once the service life of the coating is reached. • GCL panels be overlapped by 300 mm along the roll length
(600 mm at ends) and supplemental bentonite be placed
(v) Realistic selection of design hydraulic conductivity, k between the panels at a rate of a 400 g/m in a continuous 150
The design GCL k must be selected based on a realistic assessment mm wide fillet,
of the field conditions. Important considerations in selecting the k • Alternative seaming using impregnated bentonite or a
are: manufactured groove are not acceptable substitutes for the
• the choice of the GCL and the factors listed in §5.5.2.1. With supplemental bentonite applied as indicated above. These
respect to these points, consideration should be given to the approaches work under ideal conditions but are far more
following when a long service life is required: prone to failure in real field situations than appropriately
A scrim reinforced carrier geotextile with a mass per unit applied supplemental bentonite. Inappropriate bentonite at
area . the overlap leads to excessive leakage at the overlaps.
A nonwoven cover geotextile with a mass per unit area • Wrinkles and crimps should be avoided at overlaps.
. Good CQA will be required to ensure these conditions are met.
A dry bentonite mass unit area
g/m² for critical projects.
A multicomponent GCL with the coating having a mass 5.5 Design such that GCLs long-term performance is not
per unit area when the contingent on the service life of the geosynthetic
coating is particularly important. Serious consideration components.
should be given to using a multicomponent GCL for The service life of the GCL is only contingent on the factors
pond/lagoon applications. affecting the bentonite as itemized above provided that:
No more needle-punching than is essential for internal • needle-punching (for internal shear strength),
shear strength (needle-punched bundles can become a • bonding of the coating/laminate to the carrier geotextile (to
pathway for fluid flow when they are large and the GCL provide adequate shear strength),
subject to cation exchange (e.g., Rowe et al. 2017). • coating/laminate (reduce k and/or to prevent hydration from
Unless the GCL remains perpetually submerged, subgrade),
powdered bentonite generally performs better than fine • carrier geotextiles (to prevent internal erosion),
granular bentonite, which generally performs better than are not needed for the design life of the facility.
coarse granular bentonite in real field situations. These four factors are discussed below.
Natural sodium bentonite is preferred for most important (a) The needle-punching is not required for long-term internal
applications (see §4 for extreme applications). Sodium shear strength in the GCL if the design features are such that
bentonite that has been generated from calcium either the slope is sufficiently small that the shear strength of
bentonite by processing with soda ash or poor-quality the bentonite at the applied normal stress provides a
bentonite to which polymers have been added to reduce sufficient factor of safety or there is a reinforcing layer to
its k and/or increase its swell index are more prone to a carry the shear forces on slopes and prevent excessive shear
substantial increase in k (e.g., from 3x10-11 nominal to stress transfer to the GCL on the side slopes. Under either of
1x10-7 m/s after several years in the field) and decreasing these circumstances, the service life of needle-punched
service life under less than ideal field conditions. fibres will not affect the service life of the GCL.
If polymer EB-GCLs are being considered, wet-dry or (b) The bonding of the coating/laminate to the carrier geotextile
freeze-thaw or potential polymer elution represents a is not required for the long-term to prevent sliding at the
limitation to the service life. interface between the two layers for the same two reasons
• cation exchange due to the adjacent soils and permeation discussed in (a) above. Under either of these circumstances,
with the fluid to be retained. the service life of the bonding to the carrier geotextile will
• the effective stress. Note that for a GCL in a composite liner, not affect the service life of the GCL.
the effective stress on the GCL below a wrinkle is essentially (c) The coating/laminate is not required for the long-term if the
zero. Hence, k below a wrinkle will be substantially higher coating is not required to maintain a low k or to prevent
than the k where the GMB is in direct contact with the GCL. hydration or cation exchange after the liner system is
operational. Often the coating may be relied upon to prevent
Caveat emptor hydration and cation exchange during construction when the
Unfortunately, it is not uncommon for a designer to specify an GCL could be subjected to wet-dry or freeze-thaw cycles
appropriate GCL considering all the factors above and then, during before it is covered with waste or liquid to be contained.
construction, the contractor substitutes an alternative GCL to that Under these circumstances, the service life of the
was specified. The alternative is usually proposed because it is coating/laminate will not affect the service life of the GCL.
cheaper (i.e., it has less mass per unit area of geotextile, and/or (d) The carrier geotextile and coating/laminate are not required
coating, and/or bentonite mass per unit area) and has the same k on to prevent internal erosion of the bentonite for the long-term
the manufacturer’s product information sheet. This type of if the GCL is resting on a filter compatible material.
decision is usually made without realizing that the k on the product In situations where the geotextile is required to provide shear
information sheet does not represent field conditions. While the strength or to prevent internal erosion or the coating is required to
quantities listed in §5.5.2.1 may not affect k in a standard index provide a low k or isolate the bentonite from an aggressive solution
test where the GCL is permeated by deionized water under an then the service life of the GCL will be controlled by the service
effective stress ~ and differential head of ~1.5 m (e.g., life of the geotextile and/or coating.
, they have been shown to substantially (by orders The service life of the geotextile components of typical GCLs
of magnitude) affect the long-term performance and service life of and that of the coating/laminates are presently unknown (although
the GCL. No substitution of the material that does not fully meet both are currently subject to investigation at Queen’s University in
219
Canada). von Maubeuge and Ehrenberg (2000) have suggested a construction, it is reasonable to anticipate that the system in
“conservative” estimate of a service life of 25-50 years for the Figures 5.1a and 5.2a will have a service life of more than 50 years
polypropylene fibres in the GCL. Assuming that this estimate is and hence be sufficient. However, if the contaminating lifespan
for normal operating temperatures of about 25°C or below, and in and hence design life were 200 years, and one could not rely on
the absence of any other published information, this is probably a the geotextile protection layer to prevent excessive strain in the
reasonable guesstimate for the geotextile and coating in typical geomembrane once the protection layer geotextile has reached the
applications where the fluid is not aggressive. end of its service life, which by definition is when it no longer
serves its design function of protecting the geomembrane. This
5.5.2.4 GCL service life summary would then reduce the service life of the geomembrane to an extent
Provided that the design GCL k is selected considering the dependent on the resin that was used, and the tensile strains
potential effects of leachate interaction and low effective stress developed when the protection layer is no longer effective.
below wrinkles, and with appropriate consideration to ensure no Similarly, the geotextile component of the GCL is unlikely to last
potential internal erosion, then the bentonite component can be 200 years. If stability is dependent on the internal shear strength of
considered to have a service life of millennia provided the the geotextile, the system's capacity to perform effectively is
performance of the bentonite is not dependent on the longevity of limited by the service life of the needle-punching.
the geosynthetic components of the GCL. However, if the Similarly, if the subgrade is not filter compatible with the
geosynthetic components are critical to performance, then the bentonite, then there is potential for internal erosion of the
service life of the GCL is dictated by the service life of the critical bentonite into the subgrade when the lower (carrier) geotextile has
geosynthetic component(s). reached its service life, thereby reducing the service life of the
5.5.3 Composite liner service life summary GCL. These considerations apply to the single and double
Subject to appropriate consideration of the issues raised in the composite liner systems in Figures 5.1a and 5.2a. If these designs
foregoing subsections including the effects of exposure conditions are not appropriate, alternative designs are shown in Figures 5.1b
in the selection of compatible materials, the service life of a and 5.2b.
composite liner will be equal to the shorter of the service life of the
geomembrane or the GCL. Thus, with appropriate design 5.7 Monitoring for and managing the unexpected
(including the selection of materials; Rowe et al. 2020; Rowe A potential failure modes and effects analysis can effectively deal
2020a, 2020b) and construction, HDPE underlain by a GCL can with the known knowns and the known unknowns. The greatest
provide very long service life and is suitable for use in dams, ponds, challenge is, of course, the unknown unknowns. The unknown
or lagoons, or in covers for closure of landfills or mine waste unknowns are unexpected, and monitoring is needed to detect
facilities, or as a bottom liner for landfills or mine waste facilities. them. When the unknown unknowns manifest themselves and
Depending on the exposure conditions, the likely service life may become known, they are usually found to be due to an
range from decades to millennia. unanticipated complex interaction of many factors that influence
5.5.4 Drainage layer service life the system's performance. Once known, these issues can be
With appropriate design, the gravel drainage layers (leachate resolved with time and appropriate resources. The following
collection and leak detection systems) in Figures 5.1 and 5.2 can paragraphs discuss one such example of the field revelation of
be designed to have a service life of centuries to millennia (MOE unexpected outcomes and the flow-through consequences.
1998; Rowe et al. 2004; Rowe and Yu 2013). Similarly, sand Following more than 15 years of extensive field use of NaB
protection layers and granular filters can be designed to have a GCLs and research into their behaviour, Richardson and Thiel
service life of centuries to millennia. (2005) made public the problem of GCL panel shrinkage. This was
There is very little information published on the longevity of shortly followed by a more detailed paper (Thiel et al. 2006)
geocomposite drains. Unpublished work by Reinert and Rowe has documenting several field cases and illustrating the mechanism
shown that there can be a substantial variability between products with laboratory tests. This prompted two field trials at different
with some having an estimated service life of only a few decades sites to monitor panel shrinkage. These unpublished trials failed to
while others may have an estimated service life in excess of a identify any significant problem. Thus, while field cases had
century as a liner for a water retaining structure. This topic is demonstrated significant shrinkage and panel movements of up to
subject to considerable ongoing research. 1.2 m, trials designed to emulate this yielded negligible shrinkage,
thereby raising the question as to why this unexpected behaviour
5.6 Need for Potential Failure Modes and Effects Analysis occurred in some cases and not others? In these circumstances,
how could a design engineer know whether there is likely to be a
The need for a potential failure modes and effects analysis is well- problem with a GCL hidden below the geomembrane? The search
established in dam engineering. However, this approach could go for an answer to this question has involved 15 years of coordinated
far beyond dam engineering and extend to any significant research at several institutions, as outlined below. This case
geotechnical engineering project. For example, in the field of illustrates the complexities of in-field performance and the
geoenvironmental engineering, any unacceptable environmental multiplicity of factors influencing that performance.
impact arising from a facility intended to prevent that impact may The shrinkage reported by Thiel et al. (2006) was
be regarded as a failure. In this context, many failures occur due to predominantly for GCLs from one US manufacturer. Another
known factors whose consequences are known but were not major manufacturer had no reports of any problems. To understand
considered in design due to a failure to consider all potential failure the mechanisms, the observed variability in the likelihood of
modes. shrinkage, and to assess whether it was good luck or good
A failure mode that is commonly overlooked in the design of management by the manufacturer with no reported problems, 18
barrier systems is the service life of the components of the system. full-scale test sections were constructed using four GCL products
Consequently, little consideration is given to the potential effects at the Queen’s University Environmental Liner Test Site
of the service life of the component being reached. For example, (QUELTS) in 2006. Some sections were with a smooth
consider a design for a construction and demolition waste landfill geomembrane, and some with textured to resolve whether the
with a contaminating lifespan of 50 years and a maximum liner texturing was the cause of the problem. A composite liner was
temperature of less than 25°C. In this case, the design profiles exposed on six relatively flat (3% base) sections and six on a 3:1
shown schematically in Figures 5.1a and 5.2a for a single and (horizontal:vertical) side slope facing south. Another six north
double composite liner barrier system may be appropriate. Subject facing sections involved GCLs covered by about 0.7 m of silty
to the usual caveats of appropriate design, material selection, and sand.
220
During the 5 years monitoring of the QUELTS I sections side- onto the GCL and runs down the GCL in rivulets, eroding
by-side on the same subgrade and exposed to identical the bentonite. Repetition of this process creates areas with
climatological conditions, it became apparent that: (a) little or no bentonite. The loss of bentonite is difficult to
geomembrane texturing was not an issue, (b) covering a GCL with detect visually.
the silty sand effectively eliminated shrinkage, and (c) there were Downslope bentonite erosion was more extensive on a
notable differences in the performance of the four GCLs in terms 3H:1V slope than a 33H:1V (3%) slope; however, it also
of the exposed composite liner shrinkage on both the side slope occurred on a 3% (2o) base slope.
and base (Rowe et al. 2018; Brachman et al. 2018). The maximum The extent of downslope erosion and the time required to
shrinkage for the four products in the exposed composite liner over occur appeared to be related to the GCL capacity to retain
4.8 years was 30, 75, 415, and 665 mm. These differences could moisture when heated. Two potential factors contributing to
not be fully explained based on laboratory shrinkage tests, the reduced rate of downslope erosion progression were: (i)
although it became apparent that the drying rate was a factor in a higher mass per unit area of bentonite and (ii) the use of
explaining the magnitude of the shrinkage (Bostwick et al. 2010; powdered bentonite.
Rowe et al. 2010, 2011a). It also became apparent that the uptake It was confirmed that the best way of avoiding both
and loss of moisture were different for different GCL products with shrinkage and downslope erosion was to cover the
the same bentonite. The shrinkage was affected by the nature of composite liner in a timely manner with a suitable protection
the geotextile component of the GCL. These findings led to a series
of studies to evaluate the differences in water retention A multicomponent GCL with an intact polypropylene
characteristics of these four GCLs as a function of their coating facing up to minimize moisture loss into the airspace
geosynthetic structure (Beddoe et al. 2010, 2011). The work was between the geomembrane and GCL prevented downslope
then extended into a broader study of GCL water retention curves erosions and reduced, but did not eliminate, panel shrinkage
for different GCLs (Acikel et al. 2015, 2018a,b, 2020, 2021, 2022; after 28 months of exposure.
Tincopa et al. 2020). An EB-GCL with polyacrylamide-based polymer
During the monitoring of panel shrinkage between 2006 and enhanced bentonite experienced widening desiccation
2011 at QUELTS I, it also became evident that shrinkage was not cracks in laboratory tests, indicating minor bentonite
the only problem. All four GCL products were experiencing erosion. No holes had formed in the field at QUELTS II in
downslope erosion (Brachman et al. 2015; Take et al. 2015; Rowe 15 months of field exposure. However, irrecoverable erosion
et al. 2016a,b). Laboratory tests designed to examine downslope was observed on the 3:1 slope at QUELTS II after 28 months
erosion (Rowe et al. 2014; Ashe et al. 2015) reproduced the of field exposure. Thus, the benefits of polymer
downslope erosion observed at QUELTS I. The downslope erosion enhancement may be time-limited for EB-GCLs.
was only observed for GCLs that had experienced a hydration and The EB-GCL experienced the most panel shrinkage of all
drying cycle. Laboratory tests indicated that downslope erosion GCLs at QUELTS II. As discussed in the previous section,
only occurred with distilled water. About 40 ppm calcium in tap even without downslope bentonite erosion, polymer elution
water was sufficient to prevent downslope erosion for at least up and loss with water running down the slope over the GCL
to 365 cycles, and it was inferred to be unlikely to occur with combined with drying cycles can increase the k of EB-GCLs
leachate leaking through a hole in a wrinkle in a geomembrane. by 1000-10,000-fold.
The QUELTS I field study, combined with growing interest in The investigations outlined above highlight the value of an
the hydration and desiccation of GCL under a constant thermal integrated, multi-institutional approach that brings together
gradient (Southen and Rowe 2004, 2005a,b, 2007, 2011; Azad et different laboratory techniques and field studies to address a
al. 2011,2012; Rowe and Verge 2013, Hoor and Rowe 2013; complex issue (an unknown unknow until earlier this century) such
Bouazza et al. 2017b; Ghavam-Nasiri et al. 2017, 2019, 2020) as the performance of GCL in real-world applications.
served as the impetus for co-ordinated studies of:
thermal conductivity of GCLs (Ali et al. 2016).
gas permeability, water vapour movement and uptake by
GCLs (Rouf et al. 2016a,b,c). 6 CONTAINMENT OF EMERGING CONTAMINANTS
the uptake and loss of moisture from the soil (Rayhani et al. OF CONCERN
2011; Siemens et al. 2012; Anderson et al. 2012; Rowe et al.
2011b, 2019b, 2021; Carnero-Guzman et al. 2021). 6.1 The issue
the effects on subgrade mineralogy (Bouazza et al. 2014,
2017a). Municipal solid waste (MSW) landfill regulations and generic
subgrade particle size, density and water content, and the role designs commonly in use have been largely developed for the
they played in shrinkage (Rowe et al. 2019b, 2021). contaminants of concern of the last century. They generally
potential for self-healing of holes in GCLs, such as due to proved very effective at containing those contaminants over the
downslope erosion (Li and Rowe 2020, 2021; Rowe and Li past few decades. However, the situation has become more
2020). complex with the recognition that synthetic chemicals used for
QUELTS I lead to a search for a means of minimizing the risk many decades because of their numerous advantages over
of downslope erosion. Therefore, based on input available in 2011- existing natural chemicals have three significant drawbacks.
2012 from the studies cited above, QUELTS II was constructed in First and foremost is recognizing actual or potential health and
2012 with a control section identical to one of the sections at environmental impacts from these chemicals in the products in
QUELTS I and the remaining six sections exploring means of which they have been used and are now in landfills. Second,
minimizing or eliminating downslope erosion (Rowe et al. 2016a; many of these synthetic chemicals are not readily biodegradable
Rentz et al. 2016,2017). by existing bacteria and hence are persistent organic pollutants
QUELTS II provided the following new insights and guidance (POPs) “forever chemicals”. Third, many of the chemicals are
for design (Rowe et al. 2016a and §5.5.2): lipophilic and tend to accumulate in fat and, in doing so in
When a geomembrane composite liner is heated by the sun, progress through the food chain to be consumed by humans. This
moisture evaporates from the underlying GCL. The water creates the need to re-evaluate the safety of our existing landfills.
vapour accumulates at geomembrane wrinkles generated by Landfills that adequately contain low concentrations of volatile
the sun’s heat. At night, the water vapour condenses on the chlorinated solvents and heavy metals cannot be assumed, a
underside of the geomembrane. This distilled water drips priori, to be adequate for containing the contaminants of
emerging concern. Their suitability can only be established, in
221
advance of a problem, by appropriate modelling. Also, the fact Table 6.1. Probability, P, that the leakage, Q, through a single composite
that these contaminants liner will be exceeded (based on New York State landfills, Beck 2015)
(a) can be in landfill leachate at concentrations well above Q P Q P Q P
allowable levels in drinking water, and (lphd) (%) (lphd) (%) (lphd) (%)
(b) do not readily biodegrade 10 91 75 48 300 8
raises questions about the whether service life of the engineered 20 82 100 38 400 6
facilities is long enough to contain them for their contaminating 30 74 150 24 500 5
lifespan. 40 67 200 16 750 4
Landfill liners are intended to act as barriers to the diffusive 50 61 250 11 1000 4
advective transport of contaminants present in the waste and
leachate (Rowe 1988; Rowe et al. 2004). Therefore, a
geomembrane with no holes would provide a complete barrier
from advective transport of contaminants resulting in diffusion
being the dominant means of transport.
HDPE geomembranes provided an excellent diffusion barrier
Figure 6.1 Two common PFAS with a tail comprised of 8 carbon atoms
for most contaminants of concern when regulations were
bonded to fluorine with different heads: (a) PerFluoroOctanoic Acid
developed. The exceptions were volatile organic compounds
(PFOA), and (b) PerFluoroOctane Sulfonate (PFOS)
such as benzene, toluene, ethylbenzene, and xylenes (BTEX)
together with chlorinated compounds such as dichloromethane Studies by US Centre for Disease Control (CDC) suggests
(DCM), trichloroethylene (TCE), and tetrachloroethylene (PCE) that 98% of Americans have PFAS in their blood serum, with
that could easily diffuse through HDPE geomembranes (August PFOS being the most common form of PFAS identified in human
and Tatzky 1984; Park and Nibras 1993; Sangam and Rowe 2001; blood (Benskin et al. 2012; EPA 2015; Lindh et al. 2012; Zeng
Rowe et al. 2004; Divine and McCray 2004; Joo et al. 2005; et al. 2015; ATSDR 2015; CDC 2015). A more recent but less
McWatters and Rowe 2009a,b, 2010, 2014,2018, McWatters et extensive study in Germany identified 37 PFAS with one or more
al. 2019; Park et al. 2012, Shackelford 2014, Jones and Rowe in every human plasma sample tested (Göckener et al. 2020).
2016; DiBattista and Rowe 2020a,b, DiBattista et al. 2020). Even Furthermore, studies indicate that PFAS are either known or
these compounds were only likely to be problematic for a barrier suspected to (a) be carcinogenic, (b) affect human reproduction
system over unsaturated soil underlain by an aquifer where the and development, (c) affect liver and kidney function, (d) affect
contaminants could diffuse through the typical composite liner immune system function, (e) elevate cholesterol levels, (f) cause
and then migrate in a gaseous phase through the unsaturated soil thyroid hormone disruption etc. (ATSDR 2015, EPA 2016,
to an aquifer. This could be mitigated by several metres of OECD 2002; EPA 2017; Alexander et al. 2003; Alexander and
saturated or near-saturated attenuation layer below the composite Olsen 2007; Mandel and Johnson 1995; Thomford 2002; Eriksen
liner. et al. 2013; Frisbee et al. 2010; Nelson et al. 2010; Apelberg et
Advection through a composite liner will be controlled by the al. 2007; Chen et al. 2015; Darrow et al. 2013; Maisonet et al.
number and size of holes in the geomembrane and whether the 2012; Washino et al. 2009).
holes coincide with wrinkles. Studies have shown that even with Significant sources of PFAS include areas used for fire
good construction quality assurance (CQA), 5 holes/ha may be training purposes, military bases, airports, industrial areas,
expected and 20 or more holes/ha with casual CQA (Giroud and wastewater treatment facilities including sewage sludge and
Bonaparte 1989, Giroud 2016). However, the critical factor effluent water, and landfills and landfill leachate (EPA 2017;
controlling leakage is the number of holes coincident with ITRC 2020; Bouazza 2021; Moody and Field 2000; Baduel et al.
wrinkles that allow the fluid to spread out over a much larger 2017; Cousins et al. 2016; Gallen et al. 2017; Hamid et al. 2018).
area than would occur had the hole in the geomembrane been in As a result of their ubiquitous nature, their persistence and
direct with the clay liner. their known or suspected health impacts, PFAS are contaminants
Based on monitoring data of leakage through the primary of emerging concern. However, despite the relatively high
liner in double-lined New York State landfills with very good concentrations in landfill leachate relative to allowable drinking
construction quality assurance but no leak location survey, the water concentrations, there are no specific guidelines regarding
probability of the leakage exceeding a given value is shown in suitable landfill barrier systems for the containment of PFAS.
Table 6.1. This needs to be considered in assessing impact.
6.2.1 Modelling transport of PFAS through liners
6.2 Per- and polyfluoroalkyl substances (PFAS) Using an approach developed by Rowe and Abdelrazek (2019),
Rowe and Barakat (2021) published the first study of the
Per- and PolyFluoroAlkyl Substances (PFAS) are a group of over potential implications of liner systems for PFAS. They focused
4700 synthetic compounds with a carbon chain where fluorine their attention on PFOS in a landfill with an average waste
atoms have replaced some or all of the hydrogen atoms. In large thickness of 25 m and a barrier system with a leachate collection
part, PFAS are highly stable due to the strong chemical bond system, a geotextile protection layer, and a composite liner
developed between the carbon and fluorine atoms. Thus they are comprised of a 1.5 mm thick HDPE geomembrane on a 7 mm
very persistent organic pollutants and hard to degrade, being thick GCL on a 3.75 metres thick attenuation layer overlying a 3
resistant to hydrolysis and different degradation mechanisms m thick aquifer. The paper considered pure diffusion and
(e.g., thermal, microbiological and photolytic; Milinovic et al. advective transport of leachate through a hole in a wrinkle of
2015, Bouazza 2021). As a result, they have been extensively length Lw.
used in industrial and consumer goods for applications such as To calculate the diffusive transport through the HDPE
treating and protecting textiles including clothes, carpets, coated geomembrane with no advection, Rowe and Barakat (2021)
paper, and coated food and cardboard packageing or leather adopted a permeation coefficient Pg ~ 6 × 10-16 m2/s and a
products (ATSDR 2015, CELA 2019, CONCAWE 2016, EPA partitioning coefficient, Sgf ~1 for PFOS based on data available,
2003, EPA 2017, NTP 2016, Ahrens 2011, Cousins et al. 2016). at that time, from ongoing studies at Queen’s University.
In addition, PFAS characterized by a lipophilic tail in a Continued testing has lowered the value to Pg < 2x10-16 m2/s and
hydrophilic head (Figure 6.1) have been widely used as industrial this will be periodically revised as more data becomes available.
surfactants and insecticides (EPA 2017; Benskin et al. 2012). To put the numbers in context, in a diffusion test where benzene
diffuses from a source chamber through a 1 mm thick HDPE
222
geomembrane to receptor chamber, the benzene is detected Leakage was calculated assuming very good [VG] GCL k
within 10 days of the start of the test. A reasonably accurate and GMB-GCL interface transmissivity, , representative of
estimate of the diffusion coefficient can be made based on about excellent CQA, and more typical parameters representative of
one month's data. At the same time, the test approaches average CQA. The concentration of PFOS in the leachate
equilibrium between source and receptor in a little over 2 months concentration was assumed to be a maximum reported in the
(Figure 6.2). With more than a year of testing of 0.26 mm thick literature (4,800 ng/L =4.8 g/L = 4.8x10-6 g/L). The calculated
HDPE film, PFOS in the receptor is barely detectable without leakage for the different conditions and the corresponding
any clear trend of an increase in the receptor concentrations or calculated peak impact in the aquifer assuming the service life of
decrease in the source concentrations with time. One cannot the barrier system exceeds 425 years (a shorter service life would
obtain a definitive permeation coefficient until there is a clear indicate greater impact) are given in Table 6.3.
trend and, at most, one can estimate an upper bound based on the
detection limit and the length of time the test has been running. Table 6.3. Probability, P, that the leakage, Q, and peak concentration in
When data is maintained at or below the detection limit, the the aquifer, cp, will be exceeded for a single composite liner with a 3.75
upper bound estimate of Pg becomes a function of the length of m thick attenuation layer between the liner and a receptor aquifer (based
on data from Rowe and Barakat 2021)
time the test has been running and the thickness of the specimen.
Thus, the estimate of Pg decreases the longer the test runs until a Holed
Leakage Probability Ok in?
clear trend of increasing concentration in the receptor and wrinkles cpeak
Note Q greater than Juris-
decreasing concentration in the source are observed. This Lw (ng/L)
(lphd) Q dictions
difference in behaviour between benzene and PFOS is reflected (m/ha)
in Table 6.2, which summarizes the diffusion coefficient, Dg, 0 PD 0 99.9% 8 8
partitioning coefficient, Sgf, and permeation coefficient, Pg, for a 50 VG 9 90% 30 6
number of contaminants of emerging concern discussed in this 100 VG 18 88% 60 5
section as well as for benzene and chloride. The upper bound 200 VG 35 76% 110 3
estimate of Pg, for PFOS is more than 5 orders of magnitude 300 VG 53 67% 180 2
lower than that for benzene through HDPE. 50 Typ. 71 57% 260 2
100 Typ. 140 30% 520 2
200 Typ. 280 10% 810 0
300 Typ. 430 6% 950 0
NOTES: PD = pure diffusion;
VG = very good: = 3x10-11 m2/s, ka = 2x10-11 m/s, kb = 2x10-10 m/s
Typ. = typical: = 3x10-9 m2/s, ka = 2x10-10 m/s, kb = 2x10-9 m/s
To provide a frame of reference for the calculated impacts,
the allowable PFAS concentration in 4 of the most populous US
states, Australia, Canada, and Europe, are summarized in Table
Figure 6.2 Variation in benzene concentration in the source and receptor 6.4. A comparison of the results shown for the peak impact in the
compartments adjacent to a 1 mm thick HDPE geomembrane. Note that aquifer cp in Table 6.3 with the allowable levels in the 8
benzene is detected in the receptor within less than 10 days of the start jurisdictions in Table 6.4 indicates that all the regulatory limits
of the experiment an equilibrium is approached after about 2 months. considered would be met with pure diffusion. However, based on
leakage data, with a single composite primary liner and very little
To provide further context, a test examining the diffusion of natural attenuation, there is a 99% probability that Michigan’s
chloride through 2 mm thick HDPE geomembrane has now been guidelines would not be met.
running continuously for 28 years and after this time chloride is Rowe and Barakat (2021) concluded that:
still undetectable above the detection limit in the receptor. For 1. A 1.5 mm HDPE geomembrane was an excellent diffusion
contaminants like PFOS and chloride the partitioning coefficient barrier to PFOS, and a single composite liner would be
Sgf 1 and is very difficult to measure. Assuming an Sgf ~ 1 for adequate for the jurisdictions considered if one could
both chloride and PFAS, the upper bound estimate of Pg after one guarantee no holes. However, based on past experience, this
year of testing a 0.26 mm thick film is very similar to, and is extremely unlikely. There is more than 99% probability
marginally less than that of, chloride through a 2 mm thick that both Michigan and New York’s proposed limits would
geomembrane. This highlights the advantages of testing a very be exceeded with a single composite liner of the form
thin specimen when the diffusion coefficient is extremely low. examined. However, it is noted that New York State requires
Based on available data, it is estimated that Sgf ~ 0.008 for
a double composite liner, which very substantially increases
chloride and HDPE. Interpreting the same 28 years of data using
the probability of their proposed limit being met.
this value of Sgf , Pg < 7x10-18 m2/s. In any event, HDPE is an
excellent diffusive barrier to both chloride and PFOS.
Table 6.4. Allowable PFOS concentrations in jurisdictions (Based on
Table 6.2. Diffusion, partitioning and permeation coefficients for a Rowe and Barakat 2021)
number of contaminants through HDPE geomembranes.
Allowable Allowable
Dg Sgf Pg = Dg x Sgf Jurisdiction PFOS conc. Jurisdiction PFOS conc.
Chemical
(m2/s) (-) (m2/s) (ng/L) (ng/L)
-13
Benzene 2x10 140 3x10-11
PCB 1x10 -14
100,000 1x10-9 Michigan 8 Europe 100
PDBE 4x10 -15
2,000,000 8x10-9 New York 10 Ontario 150
BPA < 3x10-15 ~1 < 3x10-15 California 40 Texas 560
PFOS < 2x10-16 ~1 < 2x10-16 Australia 70 Canada 600
-16
Chloride < 3x10 ~1 < 3x10-16
2. The suitability of a single composite liner for the other
Chloride < 9x10-15 0.008 < 7x10-18
jurisdictions considered depended on the peak concentration
of PFOS in the landfill leachate, the level of construction
223
quality assurance, and hence the number of holed wrinkles particularly for PFOS and PFOA. In addition, LOR also depends
per hectare. For Australia, California, Texas, Ontario, and upon the instrumentation used. Nonetheless, these ongoing
British Columbia, either a low peak leachate concentration results highlight that a few shorter carbon-length PFAS
and/or a very high level of construction quality assurance commonly present in modern landfill leachate are highly mobile.
would be required to meet PFOS limits with a single
composite liner. However, in Ontario, a double-lined barrier
system would be required for landfills of the size considered
in the example.
3. Analyses similar to those performed by Rowe and Barakat
(2021) are recommended for all landfills with a single liner
to ensure the barrier system's suitability for containing PFOS
in a MSW landfill.
4. Given the persistence of PFAS and the relatively high
concentrations relative to allowable levels in many
jurisdictions, more consideration needs to be given to the
geomembrane's service life in MSW landfills containing
PFAS.
6.2.2 Modifying GCL bentonite to retard PFAS migration Figure 6.3 Hydraulic conductivity, k. As a function of pore volumes of
Preliminary laboratory trials reported by Bouazza (2021) flow of PFAS-containing landfill leachate (modified from Gates et al.
indicated that a standard needle-punched GCL with ~ 4 kg/m2 2020) and Bouazza 2021).
NaB had kw = 2.8 x 10-11 m/s when permeated by deionized (DI)
water and k =4.4 x 10-11 m/s when permeated by landfill leachate
containing about 35 g/L total PFAS (Figure 6.3).
Another GCL with activated carbon powder mixed with the
bentonite (ACA GCL in Figure 6.3) had kw = 6.5 x 10-11 m/s when
permeated by DI water and k = 9.0 x 10-11 m/s when permeated
with landfill leachate containing about 35 µg/L of PFAS. While
the addition of the activated carbon increase k by 30% compared
to kw, it substantially reduced the PFAS concentrations in the
effluent from the GCL (Bouazza, 2021). The latest results (after
30 pore volumes of flow, PVF; Figure 6.4) show that total PFAS
35 µg/L
to the limit of reporting (LOR) within 13 PVF (except for one
measurement at 10.9 PVF). However, the sorptive capacity of
the carbon black appears to have been reached by about 16 PVF
and total PFAS had increased
30 PVF. These results highlight the need to conduct such trials
for sufficiently long to attain more than 20 PVF.
The total PFAS measured in both the initial landfill leachate
and effluent after passing through the activated carbon modified
GCL was dominated by two short-chained PFAS analytes, PFBA
and PFBS, which exceeded the adsorptive capacity of the Figure 6.4 Landfill leachate PFAS concentrations in the effluent after
reactive medium. PFBS concentrations initially decreased permeation through a modified GCL containing powdered activated
exponentially to LOR within 5.3 PVF but then increased carbon as an additive. Initial PFAS concentrations are depicted at the “0”
exponentially in concentration from 21.6 PVF. PFBA pore volume. The limit of reporting (LOR) is 100 ng/L (0.1 µg/L)
concentrations decreased exponentially to LOR by 5.2 PVF but (modified from Gates et al. 2020).
revealed a more sporadic albeit higher concentration from 16.5 Furthermore, while these preliminary results, perhaps
PVF. Significantly for PFBA, the concentrations in the effluent unsurprisingly, show that a GCL modified with activated carbon
from 16.5 PVF onward exceeded the initial PFBA concentrations may only provide temporary respite against some shorter carbon-
in the landfill leachate, indicating that perhaps some degradation chain length PFAS depending on the level of exposure (e.g.,
reactions had occurred or, more likely, that PFBA attenuated volume, hydraulic head, and initial concentration) to the liquid,
during initial penetration had broken through. PFHxA, PFPeA they do indicate that significant retention can occur, particularly
and PFHxA largely dropped to LOR (an exception being a single for the regulated PFAS (PFOA, PFOS and PFHxS). Of concern
anomalous result for PFHxS after 5.2 PVF, then decreasing to its is the relatively high temporal variability in the PFAS analyses.
LOR up to 30 PVF). PFPeA evidenced breakthrough by 16.4 For example, PFOS was measured as high as 4000 ng/L at 5.2
PVF and increased to over its initial values by 28 PVF, returning pore volumes, more than 200-times its concentration in the
to around a third of its initial value by 30 PVF. Levels of PFOA, original leachate. Hypothetically, this could result from analyte
and PFOS, were below the LOR of 0.1 µg/L up to 30 PVF except concentrate associated with its breakthrough when sorptive
for a single anomalous result for PFOS at 5.2 PVF, then also capacity within the modified GCL is attained. However, this is
returning to LOR up to 30 PVF. unlikely given the (i) number of PVF and estimated saturated
Concentrations of the three currently regulated PFAS (PFOS, pore volume of the specimen, (ii) return in measured values to
PFOA and PFHxS) were reduced to LOR (Figure 6.5) with the LOR and (iii) variability in total PFAS analysis between 7.7 and
ACA GCL, but only PFOS was reduced to LOR by a standard 13 PVF. Instead, this is most likely associated with an as-yet-
GCL. Given that these three PFAS largely remain at LOR after unidentified experimental error, particularly with the samples
30 PVF, the ACA GCL appears to provide a reasonable degree taken at 5.3 PVF. Further work on this and other systems is
of resistance to their migration. Note, however, that due to the required to fully understand the efficacy by which GCL-based
method of direct injection and single dilutions conducted at liner components can mitigate PFAS mobility.
many commercial laboratories, the LOR does change with time,
224
Concentration in effluent (ng/L) 1000
Original Leachate
800 GCL concentration of PBDEs in human blood and tissue has doubled
ACA GCL
600
400
Wastes containing PBDEs are commonly disposed
200 landfil
LOR
- and
0
PFOA PFOS PFHxS conducted laboratory
PFAS Compounds PBDEs (commercial name: DE-71) reported congener-
Figure 6.5. Residual concentrations of PFOS, PFOA and PFHxS after specific partitioning and diffusion coefficients and
permeation of 2 pore volumes of flow in different GCLs. The limit of partitioning and diffusion coefficients for an HDPE
reporting (LOR) is 100 ng/L (0.1 µg/L) (modified from Bouazza, 2021). geomembrane he partitioning coefficients
generally increased and the diffusion coefficients generally
6.3 Polybrominated diphenyl ether (PBDE) decreased
partitioning coefficient for PBDEs in DE-71 around 3-
Polybrominated diphenyl ethers (PBDE) are a family of orders of magnitude higher
organobromine compounds (C12H Br orders of magnitude higher than for PCBs
diffusion coefficients
multplying
high partitioning coefficient Sgf, by the ffusion coefficient
PBDE congeners depending on the number and location of Dg permeation coefficient Pg Pg Sgf g
around 300-fold greater than
potential implications for a single composite liner like that
either as a result of manufacturi performed for PFO Diffusion parameters for the geomembrane
based on
calculated impact in the
e composition of a
the
concentrations (Environment Canada 2013) for pure diffusion and
es/ha but only 1 hole in a ha (adapted
from
Diffusion Diffusion
(No GMB) (GMB) GMB/GCL GMB/GCL
Congener 100
Peak/ Peak/ Peak/ Peak/
Guideline Guideline Guideline Guideline
47
Left: General structure PolyBrominated Diphenyl Ether 30 10
100 106 4 36
-
)
Composition of commercial PBDE studied and analysed by
c)
Percent c Dg Guideline
Congener gf co
composition (ng/L) (m2/s) (ng/L)
- 6
concentration in drinking
47 24 23
- 6
-
- 6
1000
- 6
100 210
- 6
is primarily due to the high Sgf essentially locking up most of the
160 120
- 6
- Pg
- 6
- Consider
go
PBDEs are lipophilic and hydrophobic persistent organic the impact
the guideline 4-fold for
Had the hole been in a their
PBDEs through high- -
225
10 m wrinkle, and Cogener 47 was now approaching the using the diffusion characteristics given in Table 6.2, and GCL
regulatory limit. Like PFAS, it is found that a single composite properties given by Saheli and Rowe (2016). Once again,
liner is unlikely to be adequate for the long-term containment of consideration was given to pure diffusion with no geomembrane,
PBDE unless there is also very good natural attenuation available pure diffusion with a geomembrane, and diffusion and leakage
at the site. This example highlights the need for more attention for 5 holes/ha, where one of those holes aligned with a 10 m or
to potential PBDE release in landfill design and the evaluation 100 m long wrinkle.
and assessment of the site's suitability for expansion. Again, each Table 6.7 summarizes the results for the pure diffusion
site needs to be considered based on its size and the local assuming no geomembrane and (i) a GCL and 3.75 m attenuation
hydrogeology to assess the likely impacts. layer (denoted GCL/AL in Table 6.7), or (ii) a 0.75 m thick
compacted clay liner underlain by a 3 m thick attenuation layer
6.4 Bisphenol A (BPA) (denoted CCL/AL in Table 6.7). For these two cases the peak
concentration in the aquifer was 11-fold above the allowable
Mass-produced for over 60 years, BisPhenol A (BPA; Figure 6.7) level with no significant difference between the cases with the
was, until recently, in the top 2% of chemicals produced GCL and CCL.
worldwide (Nataraj and Aminabhavi 2010). More than 70 % of A geomembrane was then placed above the two clay liners
this large mass of BPA was used in polycarbonate production and and another pair of analyses were perfocmed assuming no holes
more than 20% in epoxy resin production (Nataraj and in the geomebrane (pure diffusion). For these two cases the peak
Aminabhavi 2010; US EPA 2010). concentration in the aquifer was only 30% of the allowable level
with no significant impact for either the GCL or CCL.
Assuming the geomembrane had 5 holes/ha with one in a
wrinkle 10 m long, the impact was still acceptable with the GCL
but almost twice the permissible concentration for the CCL. The
difference in impact between the two different liner systems
Figure 6.7. BisPhenol A (BPA): (CH3)2C(C6H4OH)2
primarily arises from the much lower interface transmissivity
BPA has been extensively used in construction (roofing, that can be achieved between the geomembrane and GCL than
internal safety shields for stadiums, and road signs), safety can then be achieved between a geomembrane and CCL.
equipment (e.g., goggles, helmets, and safety shields for police), Assuming the length of the wrinkle with a hole was increased
optical media (e.g., CDs and DVDs), electrical and electronic to 100 m, the impact on the aquifer significantly exceeded
devices (e.g., cell phones, cameras and TVs), beverage bottles, allowable levels by 5.7 fold and 10-fold for GCL and CCL,
infant feeding bottles, food containers and microwave ovenware, respectively (Table 6.7). Once again, demonstrating that minimal
the interior lining of food and drink cans and metal lids on glass leakage can be tolerated through a single composite liner if it is
jars and bottles (Nataraj and Aminabhavi 2010; FAO/WHO the primary line of defence. For the impact to be acceptable with
2010). Most of these items are eventually disposed in (MSW) a single composite liner, it would be necessary to also have a very
high level of hydrogeologic protection with a substantial
landfills where the BPA can leach into landfill leachate
attenuation capacity. Also, as in the other cases, to maintain an
(Yamamoto and Yasuhara 1999; Mansilha et al. 2013). Reported
acceptable impact with no more than one hole in a wrinkle with
maximum BPA concentrations in MSW landfill leachate range
Lw 10 m, the service life of the geomembrane would have to
from 12 g/L to 17,200 g/L (Yamamoto et al. 2001) with an exceed 270 years for the case considered.
average of 4,740 g/L. The average value in leachate is almost
1000 fold higher than the 5 g /L allowable in drinking water and Table 6.7. The calculated peak concentration of BPA in the aquifer
almost 4000 times higher than the maximum concentration normalized to water quality guideline (cpeak/cguideline) [all number
allowed to be released from a landfill into groundwater in rounded to 2 significant digits] (adapted from Saheli and Rowe 2016)
Ontario. Type of Diffusion Diffusion 5 holes/ ha 5 holes/ ha
BPA has emerged as a contaminant of concern because of its clay liner (No GMB) (GMB) Lw= 10 m Lw= 100 m
potential impact on human health and the environment. BPA Peak/ Peak/ Peak/ Peak/
mimics estrogen, raising concerns regarding the potential effects Guideline Guideline Guideline Guideline
on the prostate gland in fetuses, infants, and children (NTP- GCL/AL 11 0.3 0.7 5.7
CERHR 2008). Available evidence also suggests that BPA may CCL/AL 11 0.3 1.9 10
have an effect on the brain including affecting the prenatal
development of the brain and behavioural problems in children The studies outlined above all show that HDPE
(NTP-CERHR 2008; FAO/WHO 2010; Erler and Novak 2010). geomembranes are likely to be excellent diffusion barriers for the
Saheli et al. (2016) studied the diffusion of BPA through a 0.5 three contaminants of concern (PFOS, PBDE, and BPA)
mm-thick HDPE geomembrane. Over the 4.2 years that the examined even though they are very different compounds
diffusion tests were conducted continuously, BPA did not covering a vast range of diffusion properties with Pg values
experience any measurable partitioning or decrease in the source (2x10-16 Pg -9 m/s) spanning 7 orders of magnitude.
concentration. Correspondingly, the concentration in the
receptors showed no trend with time and remained essentially 6.5 A reality check: PolyChlorinated Biphenyls (PCBs)
identical to that in the blank cell, indicating no measurable
diffusion through the 0.5 mm geomembrane in over 4 years. An Polychlorinated Biphenyls (PCBs) are a group of 209 individual
upper limit to Pg was obtained by assuming the receptor chlorinated hydrocarbons congeners. Each congener is unique,
concentration was at the detection limit at 4.2 years, giving the comprised of chlorine atoms substituted onto different positions
value in Table 6.2. The inferred permeation and diffusion of a biphenyl ring (Figure 6.8). Historically they would typically
coefficient were obtained assuming Sgf~1. As in the case of PFOS be used as a mixture of PCB congeners and given the commercial
and chloride, Sgf is not known with any certainty but is less than name Aroclor. For example, Aroclor 1242 is a mixture of over
one. This, combined with the calculation method, means that the 100 PCB congeners containing 42% chlorine by mass. Similarly,
value of Pg in Table 6.2 for BPA is conservative, and HDPE is a Arclor 1254 was a mixture with 54% chlorine and Arcolor 1260
highly effective diffusion barrier for BPA. with 60% chlorine by mass. Arcolor mixtures (1242, 1254, 1260)
As with PFOS and PBDE, the effectiveness of a single were used in the large-scale manufacturing of electrical
composite liner, like that previously described and studied, was equipment. For example, all three were used in electrical
evaluated by modelling the impact on an underlying aquifer transformers.
226
material for incineration. During decommissioning, the
geo and soil samples were taken from the
cover and baseliner. In addition, soil cores were taken through
the base layers to obtain the variation in PCB concentration with
depth ( ).
Using the independently established parameters for the
geomembrane and soil, the diffusion profiles for PCB's through
Figure 6.8 PolyChlorinated Biphenyls (PCBs): C12H Cl the barrier system was predicted using the same computer model
; Rowe and ; Rowe et al.
Non-volatile and hydrophobic, PCBs do not readily degrade 2004) that was used to analyse the composite liner system for the
and are persistent organic pollutants. PCBs have been reported contaminants of concern in the previous sub-sections. The model
to cause several adverse health effects and have been classed as results were in very encouraging agreement with the observed
probable human carcinogens (USEPA 2013a). Because they are PCB migration.
harmful to human health, many countries restrict their incidence Based on the detailed investigation of the long-term (22-
in drinking water to very low levels. he United year) application of geosynthetics in a landfill barrier system
designed to contain PCBs, reached the
drinking water (USEPA 2014). The drinking water objective in following conclusions.
Ontario, Canada, , although the 1. Over the 22-25 year operating period, the composite liner
allowable landfill impact on cover system limited percolation through the cover to an
the health-related average of less than 12 lphd in an area with an approximate
PCBs were recognized as hazardous, persistent pollutants in annual precipitation of 27,000 lphd (1 m/year of rain
the environment and their industrial use was banned in Canada equivalent) and where a typical soil cover will typically
permit 4100 to 5800 lphd percolation through the cover.
2. The base barrier successfully minimized PCB transport from
years after they were banned, PCBs are still found the landfill to no more than a few centimetres in over 20
years.
permitted 3. PCBs remained bound to the geomembranes and GTXs on
impacts on groundwater from a landfill. either side.
elevated PCB levels in whole fish samples 4. After more than two decades in service, no PCB
were identified concentrations were found in the groundwater.
Canada). Sampling identified PCB 5. Previously published laboratory HDPE geomembrane -PCB
ppm in the sediments and soil in the Pottersburg Creek watershed diffusion parameters combined with independently obtained
, the contaminated soils clay liner parameters gave predictions in encouraging
were and placed in four engineered landfill cells (one agreement with the observed PCB migration.
per year) for “temporary” storage until destruction of PCBs was 6. The storage facility could have continued operating,
possible (Environment Canada 2013). All four cells had a provided that leachate collection prevented outward
composite liner (HDPE geomembrane a final advective transport. Cover integrity was maintained
cover system. The bottom barrier system for Cells 1 and 2 indefinitely (likely at least many centuries and possibly
longer) without impacting groundwater under these low-
temperature operating conditions. This finding cannot be
barrier system for Cells 3 and 4 comprised of a leachate generalized to MSW landfill barrier systems operating at
collection system, HDPE geomembrane, sand protection layer, substantially higher temperatures.
Rowe et al. (2016) used the same parameters used in the analysis
The system was designed to operate
of the field case described above to predict the long-term
with hydraulic containment, such as the only mechanism for
outward PCB migration was diffusion. The diffusive properties
the scenario as described in previous sections. the findings
of liners were established. The partitioning
coefficient, Sgf, for PCBs (and even more so for PBDE) was in the previous subsections, a single composite barrier system
ed by a very rapid complete would provide adequate protection to the underlying
loss of PCB in a source compartment of a conventional diffusion groundwater provided the migration of contaminants was
test and unmeasurably small migration into the receptor as the
PCB's preferentially partitioned to the HDPE geomembrane required to limit the leakage to a very low acceptable level to
(Rowe et a. 2016). Sgf inferred from sorption tests was between maintain PCB concentrations below regulatory limits. Also, a
– 110,000, with an average of 100,000 (Table 6.2). The very long service life would be required for the HDPE
diffusion coefficient was obtained by microtoming the HDPE geomembrane.
geomembrane and obtaining a PCB diffusion profile through the Rowe et al. (2016) also concluded that the potential impact
geomembrane, which gave a best of Dg = 1×10-14 m2 of contaminants with high Sgf values should not be based on a
10- m2 close to the bulk value simplistic steady-
inferred for PBDE (Table 6.2). Although Sgf permeation coefficient (Pg) or using a computer program that
PCB's (almost three orders of magnitude higher than for
benzene), it was more than an order of magnitude lower than Sgf geomembranes. Using Pg as the diffusion coefficient will give
for PBDE (Table 6.2), and consequently, Pg for PCB's was almost very misleading and incorrect results. For contaminants with
an order of magnitude lower than for PBDE (Table 6.2; Rowe et high Sgf, consideration must be given to modelling Sgf and Dg as
al. (2016). independent entities.
In 2008, PCB Regulations required all 6.6 Summary of current status and remaining questions
to be destroyed by incineration. Thus, in the PCB
storage site was decommissioned Based on the preceding review of research related to the four
3) of contaminated synthetic or synthetic groupings of persistent organic pollutants
227
of concern (PCBs, PBDEs, BPA, and PFAS), it was found that a The growth of mineral production in the world results in the
typical 1.5 mm HDPE geomembrane was an excellent diffusion generation of large volumes of tailings from the ore beneficiation
barrier to all the compounds examined. However, with typical process and the consequent need for tailings storage facilities in
construction quality assurance, there is high probability that the the mining footprint. Robertson (2011) reports that the volume of
leakage of many of these contaminants through a single tailings from mining activities has increased significantly, in
composite liner will eventually exceed allowable levels in the excess of 10-fold every 30 years. As a result, more and more
environment unless one or more of the following apply. structures are needed in shorter periods of time. In 1900 the daily
(a) there was an electrical leak location survey before and production of tailings was around 100 tons, which increased to
after the geomembrane was covered, 100,000 by 2000. Currently the tonnages are approximately
(b) there is a high-level of hydrogeologic protection and 670,000 tons per day. It is estimated that by 2030 the volumes of
attenuation, or waste will be of the order of 1,000,000 tons per day. Robertson
(c) there is an extra level of engineering in the form of a (2011) also shows that the average height of tailings dams has
secondary leachate collection/ leak detection system. doubled every 30 years.
It may take decades before the problem is identified. The International Commission on Large Dams (ICOLD)
The impacts associated with the contaminants of emerging carried out a study that concluded that the frequency and severity
concern can be kept to an acceptable level for new facilities if of failures were increasing globally, the majority of which would
they are designed and constructed to satisfy (a) and either (b) or be preventable if diligence on the part of dam owners and
(c) above. operators was observed (ICOLD 2001). Chambers and Bowker
Based on experience, the level of CQA required to achieve (2017), analysing the risk of failures, considered that between
negligible leakage is extremely unlikely for existing landfills, 2016 and 2025, about 30 significant tragedies related to tailings
and there is an 80-99% probability that allowable limits on dams should be expected.
groundwater would be exceeded for one or more of the four Despite the context described above, as far as the authors are
groupings of contaminants of concern if they were lined with a aware, regulations regarding the use of geosynthetics in Latin
single composite liner of the form examined. Therefore, analyses America are incipient or non-existent. For example, the Brazilian
similar to those performed by Rowe and Barakat (2021) are legislative system regarding dam safety promulgated its first law
recommended for all landfills with a single liner to ensure the (Law No. 9,433) in January 1997, when the National Water
barrier system's suitability for containing PFOS (and likely other Resources Council (CNRH) was created. Currently, the country
PFAS and contaminants of emerging oncer) in a MSW landfill. has 34 national and state legislations, including laws, decrees,
Considering the persistence of the four groupings of ordinances, normative deliberations, resolutions and provisional
compounds and the ratio of leachate concentrations relative to measures. Of this total, 15 laws were produced after the Fundão
allowable levels in many jurisdictions, more consideration needs dam failure, Mariana, State of Minas Gerais, Brazil, in
to be given to the service life of the geomembrane in particular November of 2015. Law No. 12334, of September 20th, 2010,
the barrier system in general for MSW landfills containing PFAS, established the National Dam Safety Policy (PNSB), whereas
BPA, PCBs, and/or PBDE. law No. 14,066 of September 2020 is the most recent law dealing
with dam safety policy.
In Africa, the development of water resources infrastructure
reflects an increasing rate of consumption with time. Only in
7 REGIONAL CONTEXT: OF AFRICA, ASIA, AND recent decades has an appreciation of water conservation with
SOUTH AMERICA demand management and resource quality grown significantly.
The process and practice followed in pursuit of authorization for
There are 48 countries in Asia, most of which are developing waste management facilities have evolved significantly over the
countries, including China, India, Iran, Saudi Arabia etc. The past three decades with changes in philosophy, technology and
Asia continent covers an area of around 44 million km2, which performance standards. The need for and process of change with
accounts for 29.4% of the total land area of the world. The Asian time is reported along with some of the challenges to which the
continent, with a population of 4.5 billion, is the most densely industry has been exposed and innovative solutions to overcome
these challenges.
populated area in the world. The economic development levels
The intent of this section is to present the challenges faced in
for the Asian countries vary widely, so do the regulations and
the planning, design, construction, and operation phases of
technologies for waste management and contaminant
contaminant barrier systems outside of North America and
containment. For example, Singapore and Japan have some of the Europe. The range of experiences in municipal solid waste,
most stringent environmental regulations worldwide, while some mining and industrial waste and pollution control facilities
other Asian countries are still struggling with proper solid waste reported on addresses changes in performance standards over
collection and disposal systems. The latest publication of the time. It is not intended to be an all-embracing report on state-of-
World Bank on waste management in 2018 concluded that the-art technology and practice outside Europe and North
significant investments are still needed for most of the Asian America. Still, it is intended to give insight into the reasons for
countries to establish integrated waste management systems to diverse practices and unique approaches. Reconciling water
meet good international hygiene standards (Kaza et al. 2018). availability and increasing demand in the near future requires
Other recent and notable articles related to waste management today’s decisions to provide sustainable mining practices in
and containment in south-east Asia are authored by Jambeck et several regions.
al. (2015) and Lebreton et al. (2017). Both articles consider the
south-east Asian region as one of the “leakiest” regions for 7.1 Overview of policy and legislation
marine plastic waste.
Africa is the second largest continent with a population of 1.3 A brief overview of water resources development legislation and
billion and along with Asia it is forecast to be the area of highest commensurate regulatory structures and objectives in the
world population growth in coming decades. South America is African region gives significant insight into the rate of change of
the 4th largest continent and 5th on the world population list with development with time, with the associated generation of waste
431 million people living there. Latin America is world renowned or pollution and significance attributed thereto.
for mining activities including at extreme altitudes. Today South Africa is a relatively young developing country
in which the industry matured following a foundation in both
228
agricultural expansion and mining practices in the hinterland. A
century ago the results of early mining raised very limited
concerns with respect to pollution, and legislation was
introduced in 1935 to assist with the management of water
resources in the upper Vaal River catchment, which is one of the
major rivers of the country and situated to the south of the gold
mining reserves. Cities such as Johannesburg developed around
these mining practices. For many years, pollution due to other
sources such as municipal solid waste was not considered to be Figure 7.1 The 1990s’ philosophy on pollution generation and range of
a serious threat to the water resources nor environment. mitigation (Ball 2002): (a) The extreme philosophies of dilution and
As recently as the 1970s a senior official within the regulatory dispersion relying on attenuation through a range of standards to
system had stated publicly that landfill waste was not considered containment; (b) The concept of water balance showing water surplus
a source of pollution to be of concern. The practices for and water deficit areas based on precipitation and evaporation in South
landfilling were thus unregulated and municipal solid waste was Africa
disposed of in unlined abandoned quarries, random valley fills,
and in any other accessible positions including as fill material to Classifying sites as hazardous or general, large medium or
reclaim wetlands. small, and water surplus or water deficit led designers to
It was during the 1980s that an academic institution minimum requirements for containment barriers beneath such
established a research program to look into whether or not facilities. However, it is recognized that these prescriptive quasi
landfills were indeed a source of pollution. This program looked containment barrier standards did not pursue containment to the
at two existing facilities, which had been operating for many extent that was reasonably achievable but allowed for some
decades. One facility known as Waterval was a site serving the dilution and dispersion of pollutants. Thus, only hazardous waste
early gold mining city of Johannesburg. The other waste facility facilities in recent years generally employed a double liner
was situated at the other end of the country, just outside the city system of which at least the primary liner was a composite liner
of Cape Town and known as Coastal Park. The investigation into made up of both a geomembrane and a compacted clay liner.
these waste bodies included aspects of conventional geotechnical This compacted clay liner was made up of multiple layers with a
engineering such as drilling auger holes through the waste body total thickness of at least 600 mm. In all cases where double
and profiling the holes with depth, keeping a record of moisture liners were specified as a minimum requirement, they were
content, extent of waste degradation, type of waste, extent of ash, separated by a leak detection system. There was also a minimum
and layering, along with the presence of free liquid. This requirement that, for all containment barrier systems
investigation concluded that landfill is indeed a source of incorporating a geomembrane, there should be a protection layer
pollution and that the main drivers of leachate development were between geomembrane and the overlaying leachate collection
related to drainage of surface and surrounding water; the system. For the general waste facilities, clay-only barriers were
moisture of the waste being deposited including whether co- employed with thickness in proportion to size classification, and
disposal of liquids with solids was taking the place; and the leachate collection systems were required only for those sites
climatic water balance at the site. The research process had also situated in a positive water balance area.
included monitoring of groundwater with distance from the This range of documents gave waste facility owners guidance
waste facility and it was noted that in the case of Coastal Park beyond design and construction and included operation and
landfill, a pollution plume was moving off site at about performance monitoring. In addition, the minimum requirements
4 m/annum. were used extensively when the country was going through
significant change in legislation and a growing environmental
7.1.1 Development of initial waste management philosophy awareness with recognition of the limits to natural resources.
This 1980s research led to the belief or philosophy that it was Thus, while these documents were employed for over a decade
only necessary to contain the most hazardous of waste material, leading into significant law reform, the regulators and
whereas for small sites or low risk point sources of pollution the professional societies involved in waste management used the
groundwater regime could employ the philosophy of attenuation opportunity to ascertain the suitability of the philosophy and
for protection. During the early 1990s, the regulatory body associated prescriptive standards for pollution control. As a
responsible for water resource protection developed a suite of result, these prescriptive standards were embraced at least in
guidelines that became the accepted standard against which the principle by neighbouring countries of Namibia, Botswana,
permitting of landfills took place. These standards were based on Mozambique, Swaziland and Lesotho, and seen as guiding
documents in many other parts of Africa. In addition, the
the philosophy that mitigation of pollution impacts can be
guideline documents were used to motivate the development of
achieved by containment for the most severe pollution source
standards in parts of Asia in the late 1990s. It is however
and attenuation for the low significance pollution sources, with
recognized that although prescriptive in nature rather than
a range of standards between these two extremes as shown in performance based, they were seen to be one of the lowest
Figure 7.1(a). The waste classification was thus addressed by standards of containment in the worldwide survey of regulatory
toxicity and concentration, leading to the classification of waste standards (Koerner and Koerner 2007).
as either hazardous or general. The size of a site was taken into There were several lessons to be learnt from applying these
consideration based on the rate of waste disposal, leading to a requirements or standards in the local framework. It was
second parameter of large, medium, small, or communal recognized early on in their application that even in the driest of
facilities. Finally, to give effect to the research result which regions, falling well within the water deficit classification,
identified available water as a critical factor, it was accepted that leachate was indeed produced within the waste body and was as
a third factor would be water balance at the site, with the leachate a result a source of pollution as seen in Figure 7.2(a). Similarly,
producing facilities conceptualized as being those situated at it became quite apparent that the waste stream to landfill varied
positions where, for the wettest six months of a year, the rainfall with time and position, and this had an influence on the
would exceed evaporation as shown in Figure 7.1(b). Sites were performance of leachate collection systems in the water surplus
thus classified with the third parameter as either water surplus or areas. These leachate collection systems were found to block
water deficit. readily as seen in Figure 7.2(b), as they were not protected by
filter systems defined within the standards and thus not
229
employed by the owner or designer. So too was significant and performance assessment of geosynthetic products was
experience gained in the use of geomembranes as liner materials. undertaken by academic institutions to a large extent and to a
lesser extent by the regulator itself. With the limited resources
available for research, the use of international literature, access
to material producers from around the world, and cooperative
studies were pursued.
Developments in African countries, particularly South of the
Sahara Desert, in which a waste disposal facility was required,
tended to adopt the standard of the investor's host country if
European or North American or alternatively followed the
standard employed by South African legislation of the time.
Figure 7.2. Experiences from applying the Minimum Requirements for 7.1.3 State of the art in the African Region
waste disposal (DWAF 1998) philosophy: (a) Excessive leachate is Regulations employed in foreign countries were considered for
generated even in a desert environment (Google Earth image of their suitability to the South African landscape. In particular,
Kupferberg, Namibia (Alphabet Inc. 2019); (b) Aggregate leachate Australia, Canada, Germany, and the USA, along with the
collection systems clog readily if not protected by a filter (Shamrock and supporting technology, were studied.
Msiza 2015) An analysis of international standards led the regulator to
believe that South Africa had one of the lowest standards of
The use of geomembranes in South Africa had initially been containment for waste amongst those countries in the world that
for applications where water containment was required, often were regulating waste containment barriers systems at that time.
associated with inter basin transfers, or delivery of raw water A worldwide survey (Koerner and Koerner 2007) confirmed that
over significant distance from its source and hence competent the world standard for waste containment, irrespective of
storage was required for this valuable asset. whether the waste was hazardous or general, was the
employment of a composite (geomembrane plus compacted clay
7.1.2 Development of a modern philosophy in intimate contact) liner within a barrier system. In contrast,
The new-found international relations with the world allowed South Africa was largely relying on clay-only liners for partial
South Africa to exchange technology with countries willing to containment and, in many cases relying on rudimentary drainage
do so, and the advent of the Internet facilitated an even greater layers and clay-only liners for controlling the rate at which
rate of sharing of knowledge and experience. groundwater pollution took place.
The new legal framework in the Republic of South Africa Although it was tempting to adopt a standard from another
(RSA) placed environmental protection and human rights high country with experience in modern containment standards for
on the legislative agenda. The Constitution of RSA, 1992 and barriers systems, the South African situation of a large industrial
1996, Bill of Human Rights demanded a rethink of waste and developed economy in a water-scarce environment had to be
management and pollution control measures. A hierarchy of considered. In addition, the water availability per capita per
waste management strategies emerged with a priority to address annum needed consideration to reflect the acceptable
the rapidly deteriorating limited water resources as seen in assimilative capacity of our ecosystems within watercourses.
Figure 7.3. It is widely known that with the long history of The quantity of renewable water per person per annum varies
mining practices in certain parts of the country, that salts are a widely across the planet. It was observed that regions with large
significant problem in the groundwater and river systems (Brink quantities such as Canada, the North Americas, and Europe
et al. 2009). The assimilative capacity of many watercourses and tended to have established legislation that relied extensively on
rivers was exceeded bringing about the de-oxygenation and the containment of contaminants achieved by composite liner
anaerobic conditions with associated loss of biota. Many human systems. Although there is no known register of regulatory
lives and the health of communities or persons who made use of standards, it became apparent that numerous countries outside of
runoff river water supplies were placed at risk. North America and Europe did not have a standard for
contaminant containment barriers (Msiza, 2020) despite less
renewable water per person per annum being available. The
predictions are that Chile, South Africa, and Australia will be
among the world's first countries to have used between 80% and
100% of their available water resource by the year 2040, as
shown in Figure 7.4, with other countries following not far
behind in growing water scarcity.
Figure 7.3. Visual evidence of waste and mining pollution in rivers: (a)
The confluence of the Vaal and Klip Rivers (Google Earth Image of
confluence the Vaal and Klip River); (b) The confluence of the
Steelpoort and Klip River (photo courtesy RB Martin)
This attention to the significance of RSA's water resources
was escalated largely by the World Commission on Dams South
African Initiative, which recognized the scarcity of water in this
and other regions of the world. Similarly, the dependence of poor
persons on natural resources drew attention to the water quality
of river systems which was further emphasized by acceptance of
international treaties and the importance of biodiversity and
wetlands in the remediation of water resources impacts.
Despite the privilege of a domestic geomembrane Figure 7.4. Water deficit regions for the world by 2040, as predicted by
manufacturer, several lining installation contractors, and a local the World Resources Institute (2015)
geotextile manufacturer with several suppliers, the investigations
230
While South Africa has a lean economy, it has an obligation (Government of Ghana 2002, 2009). These policy documents
to fulfil multiple responsibilities such as the alleviation of similarly pursue pollution containment.
poverty; the provision of basic water of an acceptable quality to The transition from the prescriptive standards based on the
all its inhabitants; the right to a safe environment; and the philosophy of pollution dilution and dispersion which relied on
protection of its environment for future generations. The attenuation as mitigation, to the performance-based criteria,
prioritisation of needs and regulation required careful which rely on containment as mitigation and allowing extensive
consideration to maintain a healthy economy. Furthermore, use of alternative geosynthetic materials as partial or full
regulatory actions should be fair and just to participants within replacement of natural materials, has led to a multitude of
the social and economic framework. The opportunity to build challenges and some innovative solutions to meet contaminant
more dams is limited, with water delivered becoming containment objectives. While some challenges are common to
increasingly expensive, especially given the long-distance over Africa, Asia and South America, there are regional influences
which transfers are required. The development of new dams must that bring about distinctive challenges, as discussed in the
be evaluated in terms of impact, economics, and alternative remainder of this section.
options. Water quality deterioration through pollution from
agricultural, industrial, and mining activities and poor urban 7.2 Challenges in barrier system planning, design,
wastewater management, is arguably the most serious threat to construction, and operation: Africa
the region’s water resources. While treatment and reuse of water
is one option in meeting growing demand, the prevention of 7.2.1 Changes in Philosophy, Legislation and Commensurate
pollution and water conservation by use of high-performance technology (influenced by attitude; knowledge,
containment barrier systems was proposed as a more cost experience, and trust)
effective and broader benefit option. A change in philosophy and commensurate legislation over time
It was thus decided to embark upon a consultative procedure has been necessary for sustainable development of available
from 2009 to 2013 with proposals to refine the historic minimum water resources in a region where (a) they are finite although
requirements leading to national norms and standards for the renewable annually and (b) there is a diminishing volume of
classification and containment of waste (following a revised renewable water per person per annum. This has led to improved
waste assessment process) within the waste management standards of contaminant containment over time, with regulatory
hierarchy, which prioritised waste reduction and re-use only then standards being influenced by research results that have
followed with disposal to a facility designed for containment. demonstrated the advantages of geosynthetic materials when
This strategy would clearly lead to a change in waste stream and used in conjunction with or as partial or complete replacements
a change in resultant behaviour of the waste body. The intention of natural materials.
was to move away from numerous small sites and move towards The regulations used in the Southern African region require a
regionalisation of municipal solid waste facilities. An example contaminant containment barrier system comprising drains and
of the change in waste stream was the regulator’s introduction of composite liners. The predicted performance of the design and
paying for plastic bags, so readily used in society at that time. construction quality assurance protocol forms the basis of the
This change brought about a reduction in the plastics content of authorization. The regulation makes provision for the use of
new municipal solid waste streams, which changed strength geomembranes and geosynthetic clay liners as full and partial
characteristics and resultant angle of repose of waste and hence replacement of soil liners, allowing for geotextiles as cushion
the outer slopes. layers to geomembranes and as filters, and allows for geodrains
Today’s regulations for pollution control barrier systems in as a replacement for sand and aggregate drains, provided
South Africa are based on a risk assessment of the pollutant equivalent performance is shown. The challenge lies in
informed by the concentration of toxic elements. Then for each demonstrating the equivalent performance of an alternative
waste risk type, a commensurate containment barrier standard is element in a barrier system where it is not just the primary
required (NEMWA Regulations 2013). All containment barrier function required to be equivalent. Still, the equivalence within
standards require a composite liner and drainage system. A the context of the barrier system must be demonstrated. For
double composite liner separated by a leak detection drainage example, a GCL may be seen as equivalent to a CCL in respect
layer is required for the highest risk waste type disposed of at an of seepage by a particular permeant such as water. Still, the
engineered facility. Although the barrier system of drainage barrier system requires that the performance recognizes all
layers and composite liners is prescribed by material type and attributes of the alternative material including the interface and
thickness, the regulations simultaneously require the internal shear strength, the performance under low and high
performance of the design to be defined and thus also allow for confining stresses, and the performance over the service life (of
the use of alternative materials provided the equivalent operation plus post closure polluting period) for the specific
performance thereof can be demonstrated. This performance chemical and physical exposure conditions.
prediction is based on the design itself and the associated Designers often fail to thoroughly investigate and evaluate the
construction quality assurance plan (which includes quality equivalence of the system, relying on product supplier’s
control). It is to quantify service life for the barrier system and information on a particular element of geosynthetic materials,
assess the total solute transport (seepage) anticipated post- usually brand related without an appreciation of design
construction of the barrier system. In addition, the authorities complexity and uncertainty. It thus falls upon the regulator to
consider acceptability based on predicted performance for have state of the art knowledge and experience to assess the
containing contaminants at a particular site and whether the appropriateness and correctness of a performance-based design,
predicted performance is acceptable to the receiving especially when influenced by material suppliers pursuing brand
environment, be it the water resource, natural or physical specification over performance requirements. Examples of non-
resource, including air, and to society at present and for performance-based specifications apart from brand name or
generations to come. product code include excessive descriptions or markings such as
The principle of pollution containment is not unique to for the use of a GCL in a composite liner for a hazardous waste
Southern Africa. For example, Ghana had established Landfill facility where the specification called for non-toxic ink markings
Guidelines in 2002, which followed on from the 1999 to define the GCL overlap or in reinforcing grids defining the
Environmental Sanitation Policy, which was revised in 2009 to colour of the grid. Neither a marking ink nor the to be buried
broaden the application of the polluter pays principle product colour will influence the systems performance but will
influence the cost of the development.
231
The anti-competitive practice of brand specification without The regulator is on occasion challenged on fairness and
performance evaluation or specifications has led to substantial consistency of decision-making. However, there does is need to
resistance to the use of geosynthetic materials in contaminant be a clear separation of roles and responsibilities between the
containment or pollution control barrier systems by facility designer and the regulator. This is emphasised particularly in
owners to the extent that both the private and public sector have forensic analysis following geotechnical structure failures such
called for an abandonment of the geomembrane liner as of a tailings dam (Morgenstern et al. 2015). The reliance on
requirement. This resentment by developers of designers’ observational methods and government inspections alone is
behaviour has been led by the mining industry (Chamber of inadequate to avoid many incidents because no amount of
Mines 2016) and could have had significant detrimental impacts inspection will reveal the hidden flaws. Hence, further actions
on society and the greater economy. This led to an assessment of may be prescribed from time to time. Nevertheless, it is
the socio-economic impacts of competent containment systems axiomatic that the regulator cannot regulate its own activities,
which found that despite their site-specific cost, the benefit and if the regulator were to usurp the role of the designer it would
outweighed the impact. Furthermore, it has not been the usurp its own role. Thus, reviews result in an assessment of
geomembrane inclusion per se which has created the escalated acceptability, whereas the designer is responsible for the
unit costs of containment barrier systems, but (a) the cost of the approval of reports and drawings.
drainage aggregate, which is the most expensive element, along
with (b) anti-competitive geosynthetic material specifications
that lead to alignment between some designers, contractors, and
product suppliers. The latter has escalated costs unnecessarily
and resulted in post construction substandard performance.
To overcome this, voluntary organisations such as the
International Geosynthetics Society have pursued training
exercises for the benefit of society and awareness of appropriate
technology. Similarly, technical advisory notes on design and
construction have been made available to the public to counter
the misuse of otherwise sound geosynthetics materials. There are
alleged instances where designers have specified unjustifiably
thick leachate collection systems and larger aggregate size than
the normal 53 mm to require a larger mass per unit area
geotextile cushion layer only available from one supplier and
restrict product supply to a subset of the available lining
contractors. It is alleged that this behaviour has led to escalated
costs of single composite liner barrier systems, for the benefit of
the designer when contracted on a fee system related to a
percentage of the capital cost. Similarly, there are alleged
instances of a comfortable relationship having developed
between designer having brand orientated specifications, the
brand of supplier, and the lining contractor to whom the product
supplier is aligned that has led to the installation of unacceptable
geomembranes (which have suffered separation in plane and
been supplied with factory faults including regular holes, despite
having ISO 9000 certification (Figure 7.5).
The change in philosophy from relying on attenuation to the
philosophy of containment relying on competent containment as
mitigation of pollution is reflected in the modern regulations
which require the use of geosynthetic materials and allow for still
further geosynthetic substitution of natural materials is
underpinned by the state-of-the-art technology and training
provided by the International Geosynthetics Society, its regional
Chapters and other voluntary organizations. The challenge posed Figure 7.5. (a) Separation in-plane for a geomembrane having 0.9 mm
by some designers alleged misuse of the standards and selective micro-spike height and (b) Geomembrane manufactured with
technology for unjustifiable financial gain through partial discontinuities including ridges and holes installed despite the designer's
specifications or construction quality assurance plans is being specification requiring a 0.9 mm micro-spike height and minimum
resolved by exposing the practice in public forums and thickness of 1.5 mm.
simultaneously sharing appropriate technology. The latter
validates the benefit of competent barrier systems in the minds Despite this separation of roles and responsibilities, it is not
of developers and reduces the antagonism of parties opposed to uncommon for a developer to seek guidance from the regulator
geosynthetic materials used in regulations. or seek regulatory approval based on a report containing a
disclaimer in which the designer absolves him or herself of all
7.2.2 Role Players and Responsibilities (relying on clear responsibility. In such cases, if the designer cannot trust its own
separation, avoiding disclaimers and alignment) work, the regulator certainly cannot trust the design report either
The developer of a waste disposal facility (WDF) relies on the and such an application should not receive authorization. The
designer, usually, an engineer, to provide a specific solution situation has also arisen in which once a developer has an
meeting the developers needs and complying with regulatory authorization or a licence from the regulator, it is believed that
norms and standards. The role of the developer’s team and of the such a licence and conditions are the fixed property of the
regulator is quite different in that the former is the proponent developer and it is the responsibility of the regulator should there
with a site-specific view whereas the latter has a broader be an incident associated with the design or construction of the
responsibility to society. containment barrier system. However, in such situations reliance
is made on the regions human rights entitlements and over-
232
arching law for which a judgement has been made as follows to 86°C at midday, and temperatures of 89 to 90°C measured in
“licences are subject to administrative withdrawal and change. the Northern Cape when the ambient temperature was near 40°C.
They are never absolute, often conditional, and frequently time- A limited drop in temperature of the order of 10°C was measured
bound. They are never there for the taking, but instead are subject between the upper and lower surface of geotextiles deployed
to specified pre-conditions. In time, a licence holder may cease over geomembranes. However, when covered with 100 mm of
to be suitable to hold the licence and they are also not freely soil, the temperature of the liner and the number of associated
transferable” (Froneman et al. 2015). wrinkles reduced significantly. This data emphasizes the need to
Other means by which such challenges are resolved include consider time-related activities and the consequences thereof,
the mobilisation of public interest in the technology of such as the diurnal expansion and contraction affecting the
containment barrier systems. This is partly achieved by geomembrane/compacted clay liner interface performance,
addresses to public meetings during environmental impact especially when textured geomembranes are employed, and the
assessment processes prior to a development taking place, and clay surface is effectively combed and soil particles displaced. It
partly by addresses to the monitoring committees of public is similarly argued that designers should consider the time-
representatives which oversee developments where containment temperature curve during operation and post-operation life,
barriers are utilised. The results of social media commentary can along with the pre-utilization period, to give a fair assessment of
be extremely beneficial to the developer and regulator in service life. The use of average temperatures at leachate outlet
rectifying designer and contractor deviations from accepted points does not necessarily reflect the containment liner
practice. The recent growth in geosynthetic technology temperature beneath the cell. Such temperatures do not reflect
availability through webinars addressing aspects of design and the effect of leachate mounding or energy dissipation and
construction technology has been particularly valuable in radiation for which models exist (Hao et al. 2017). The time-
resolving disagreements between designers and the interested temperature service life curve for an HDPE geomembrane being
and affected parties (IandAPs) of a particular development ahead hyperbolic will show that a short period at a high temperature
of construction, when these recorded webinars by international can significantly reduce performance far more than the
experts are made available to the IandAPs to substantiate equivalent average temperature of exposure over the same
engineering norms and rebut concerns about deficient designs period. Similar consideration of time of exposure to temperature
and unsubstantiated claims. and pH should be afforded to geotextiles when considering the
acceptability of polymer (Mathur et al. 1994).
7.2.3 Designers reliance on material supplier’s data sheets To avoid such shortcomings in the design, a regulators'
There are other unusual approaches outside the norm that affect checklist of performance parameter results is called for by the
the development consideration which takes a systems approach. facility's developer. This checklist is aimed at assessing the
Some designs specify performance based on product data sheets relevance and diligence with which the designer interrogates
without validating the parameters relevance to a particular facts and assumptions made and requires the predicted
design, without ascertaining how the parameters change with performance to be specified for the assessed service life and
time or temperature, and occasionally without ascertaining quantified total solute transport through the barrier system for the
whether such materials are still available or whether technology specific development. The developer and designer are both
has surpassed the specified test method, performance standard, required to certify the check list as true and correct prior to
material type or interaction with other materials. submission to the authority. This management technique has led
An example of parameter change with time is the reported to an improved appreciation of the benefits and constraints of
interface friction between a smooth geomembrane and particular geosynthetic materials by owner’s representatives other than the
soil such as flexible polypropylene geomembrane and Ottawa engineering profession and led to a reduction in the potential
sand reported as 29.5o (Montell brochure 1997), let alone the misrepresentation of predicted performance by designers and
change with other soil interfaces such as clays, or time related product suppliers.
change to interface shear strengths due to polymer creep or Perhaps the most valuable innovation in resolving the
effects of leachate irrespective of texturing. This lack of attention application of inappropriate data sheet parameters is the statutory
is especially true in respect of filter design for both granular and end of construction site inspection by the regulator in the
geotextile filter compatibility with adjacent soils and, in presence of the facility owner and design engineer provided the
particular, fine grained cohesive soils as used in CCLs (Sherard regulator has adequate insight. The site inspection goes beyond
and Dunnigan 1989). comparing the construction completion report and its supporting
Seldom is sufficient consideration given to the difference in evidence with the accepted design and specifications in that on
performance between the office specimen and the material site observations and physical measurements are made and
installed with respect to project specific specifications. The samples collected for the regulator’s independent assessment.
method of post-production storage also can influence These activities take place under the scrutiny of the developer
geosynthetic performance prior to and during installation, and design engineer with constant communication on reasons for
especially for geotextiles (ICOLD Bulletin 55, 1986), and for the evaluation and implications of the assessment being made.
GCL carrier components. The potential implication of the revised licence conditions on
Observations on a particular day in a supplier’s stockyard volume or type of waste that may be disposed or similar
using an infrared thermometer revealed the following: black amendments to remedy designer’s reliance on invalid data sheet
plastic sheet wrapping at 70 to 75°C; unwrapped geotextile rolls parameter values have financial implications that are significant
at 55 to 69°C with colour varying from pale grey to black and motivator for owner and designer to accept appropriate
the darker the geotextile the higher the temperature; and even a technology and a motivator for the regulator to be diligent in
red plastic sheet wrapping was cooler (than the geotextiles) at applying state of the art technology to avoid litigation and
57°C, as was the temperature of unwrapped black pipe surfaces associated costs. This remedial behaviour has in recent years
which allowed ventilation through the stack (Legge, 2019). generated an eagerness for knowledge and application thereof in
This makes for interesting service life assessments when the design and construction of contaminant containment
considering the rate of UV degradation, especially on systems. The knock-on effect has been far reaching as designers
polypropylene (PP) and polyester geotextiles in relation to revert to product suppliers on inappropriate data sheets and/or
cushion layer performance and thermal conductivity. Similar misleading guidance to the extent that product data sheets are
measurements on geomembranes in Limpopo Province, RSA at amended to reflect specific product and date references as well
various sites have shown the upper exposed surface to be at 83 as conditions under which parameters were derived such as at
233
room temperature. The further consequences include the revision the below geomembrane foundation layer influences by showing
or updating of national standards to reflect current technology the 32% strain for a 50 mm aggregate above a GM plus CCL
and in particular the move by designers to base designs on site layer is reduced to 17% for a geomembrane/GCL composite
specific evaluations including test results rather than reliance on liner, for all other variables being held constant. The influence of
data sheets and publications. aggregate or stones within the soil layer beneath a geomembrane
should thus not be underestimated. The study further confirmed
7.2.4 Performance based Designs and alternative materials that a sand layer under the loading conditions reduced the
Barrier system designs are required to recognise the waste risk geomembrane strain to 0.2%, whereas for all geotextiles tested
to inform the commensurate containment standard and reduce as cushion layers, the strain remained above the 3% limit, but for
risk to the environment. Understanding the waste risk also a particular geotextile of 2,200 g/m2, the geomembrane strain is
requires an understanding of how the barrier system performance limited to 6%. However, the authors provide guidance on strain
will be influenced by the waste material deposition, the way in limitation by way of flattening slopes to reduce down drag,
which the waste mass itself will change with time as leaching or reducing above liner interface shear or increasing below liner
decay takes place if at all, and the consequences of such actions. interface shear or combinations thereof, and noting the effect of
Most mining waste streams are however relatively uniform and wrinkles on performance reduction.
well-understood, which should lead to cost-effective There is a danger in extrapolating one study’s results to all
containment by skilled and experienced designers. The use of designs without recognizing the conditions under which the
relatively modern geosynthetic materials in barrier systems, investigation was undertaken. Laboratory test results are
relies on modern technology which was not available in formal influenced by the loading rate, as for interface shear, the duration
educational structures until the last decade and even then, it is of load applied and the temperature at which the test is
limited to postgraduate engineering studies. Many designers are undertaken.
thus faced with limited knowledge or experience and may be Further to this, a geotextile protection layer performance will
misled by guidance provided by materials manufacturers or be influenced by its mass per unit area, the type of polymer(s)
product suppliers which is brand specific and not necessarily used in filament manufacture, the length and thickness of
relevant to the system being designed. filaments, including the cross-sectional shape, the form of
There are distinct perceptions generated in the civil and needling and needling density and the extent of in situ
mining industry which must be tested for relevance in each deformations before loading. To emphasize the difference in the
design. Some of these aspects are addressed below. performance of various geotextiles, the parameter of
transmissivity or in-plane drainage is used to demonstrate the
7.2.4.1 Perceived performance of geotextile cushion layers effect of normal load on transmissivity reduction for a range of
above geomembranes to limit tensile strain. geotextile variables. Figure 7.6 confirms that for the same mass
The regulations pertaining to pollution control barriers provide a per unit area, there is a significant difference in geotextile
schematic layout and suite of requirements to demonstrate the performance for materials made of continuous filament PET or
performance of a barrier system, which is to be taken into staple fibre PP. Similarly, mass per unit area has a significant
consideration when adjudicating the suitability of a barrier influence on performance for the same form of geotextile, and
system design. Therefore, using geotextile cushion layers as an for the same resin, there is a significant difference between staple
alternative to soil protection layers provided equivalent fibre and continuous filament performance. Geotextiles having
performance is demonstrated as acceptable. In practice, however, multiple differences in various contributing elements can have
a perception has arisen that a geotextile having a mass of significant differences in performance. Specifications should
1,000 g/m2 provides equivalent protection to a 100 mm thick define the performance required and not be limited to a select
silty sand layer, which is not true for coarse grained and similar group of geotextile parameters as this could be misleading.
puncture-producing material under high compressive stress. This The use of aggregate material in drains for tailings and similar
monofill waste disposal facilities is usually limited in extent, but
perception is further aggravated by product suppliers suggesting
critical to performance of the barrier system and stability of the
a hierarchy of needs in decision-making on geotextile selection
structure. It has been noticed that size of aggregate alone is not
should be driven by the type of polymer used in geotextile
the only parameter influencing strain and geomembrane damage,
manufacture, be it polypropylene (PP) or polyester (PET). but the aggregate shape has a marked influence due to both the
The assessment of maximum allowable tensile strain in type of point load and the workability during placement.
geomembranes has been addressed extensively in the literature, Aggregate shape can be influenced by the type of crusher used
with early suggestions that strains as high as 6 to 8% in HDPE in manufacture be it a jaw breaker or rotary crusher for the
geomembranes could be acceptable (Peggs, 2005). The limit set particular type of rock. However, the post crushing treatment by
by the German authority, Federal Institute for Materials rolling with a smooth drum roller compactor can significantly
Research and Testing (BAM) (BAM, 2015) of 3% is generally reduce particle sharpness and improve roundness. Similarly, the
accepted as the appropriate upper limit, as advocated at the IGS method of placement of aggregate drainage layers is a significant
specialty workshop on barrier systems held in Munich, Germany contributor to construction damage leading to tears and folds
in 2017. It has been shown that the vertical load on aggregate rather than the commonly assumed holes and wrinkles, which if
causes tensile strains within the geomembrane, which can be undetected would render assumed or assessed leakage rates
measured by short-term laboratory tests using lead plates, misrepresentative. These incidents can, however, be reduced by
aluminium plates or tin foil to record the deformation (Rowe et defining the means of placement and spreading of aggregate to
al. 2004; Hornsey, 2013) or by means of calculation. The method ensure aggregate is dumped before being distributed by a tracked
advocated by Tognon et al. (2000) is based on a combination of vehicle rather than low pressure wheel vehicle (to avoid turning
membrane and bending theory and has been shown to be most induced torsional tears) and post placement electric leak location
reliable (Rowe and Yu, 2019). Although the method advocated surveys (Beck 2015; Rowe and Yu 2019). It is postulated that the
by Eldesouky and Brachman (2018) shows promise and is of fines or dust on aggregate that washes into the geotextile
significant value in that it acknowledges the increased strain protection layer receiving face leads to a stiffening of the
effect due to aggregate rotation on slopes. Rowe and Yu (2019) geotextile layer and improved performance as applied load
show that the geomembrane peak tensile strain is related to increases over time; however, this is still to be quantified.
particle size under particular loading conditions, with 16% strain Figure 7.7 shows typical damage induced during the spreading
for a 25 mm particle and 32% for a 50 mm particle above of protection layers or drainage aggregate, as well as the fines
geomembrane without protection layer. They further emphasize
234
and dust component that may accumulate at the geotextile and may be expanded to incorporate longer loading periods to
aggregate interface when unwashed aggregate is used in drainage assess the effects of creep.
applications.
The specification of drainage system aggregate by particle
size limits alone can thus lead to poor performance of the barrier
system due to highly angular material increasing tensile strain or
poorer drainage performance as the material breaks down due to
either mechanical damage or chemical weathering or
combinations thereof. This unintended poor behaviour is noted
in the metallurgical furnace waste disposal facilities where
coarse slag is specified without assessing crushing strength nor
appreciating the effects of wetting and drying on breaking down
the slag used as drainage media.
To reduce the risk of substandard performance by drainage
systems and their impacts on the liner component of the barrier,
the design specifications should also include performance
requirements related to the strength and durability of the
drainage material. So too, should the CQA protocols include
records of quality control tests on the aggregate. This, along with
photographic records of the aggregate in-situ with a scale
provided by either a tape measure or standard-sized items such
as a golf ball, assists even inexperienced personnel in identifying
outsized material as shown in Figure 7.8 taken from the primary
drainage layer or leachate collection system of a hazardous waste
facility by the regulator during a one-off inspection, despite full-
time supervision by the design engineer and a CQA monitor.
It was the designer’s “belief” that the larger size aggregate
specified would be superior as a drainage layer (Jewaskiewitz
2019); however, the effect on geomembrane strain had not been
quantified. The reliance on test results provided by the designer’s
preferred geotextile supplier relied on a test undertaken in a UK
laboratory using up to 53 mm aggregate on the geotextile cushion
layer and geomembrane as specified in the design, but on a stiff
foam aimed at simulating the soil. This yielded a peak strain of
2.8% at 900 kPa load, whereas the evaluation, when repeated in
an Australian laboratory using 53 mm nominal diameter
aggregate from the site and the same geotextile and
geomembrane as before but using a compacted soil also from the
site, yielded a peak strain of 2.7% at a mere 675 kPa load. This
shows that with nearly half the load, the strain did not halve for
the more representative laboratory test, both of which were
undertaken at ambient temperature and for which loading was
applied over a 24-hour period – which does not replicate the field
condition anticipated to be at 40oC and loaded over the years.
The above shows that the performance of geotextile cushion Figure 7.6. Transmissivity of various geotextiles under increasing load
layers to limit total tensile strain in a geomembrane can be misled
by reliance on use of sample aggregate not from the actual source 7.2.4.2 Perceived performance of geosynthetic clay liner (GCLs)
to be used in the development or by laboratory tests which layers as an alternative to compacted clay liners (CCLs).
attempt to use a stiff foam as a replacement of the CCL. Such The potential benefit presented by GCLs as a partial or total
tests may be adequate for comparative indicators, but not for replacement of CCLs has been advocated for two decades, and
performance-based design. The promotion of such multiple investigations into the performance of GCLs have been
misrepresentation by product suppliers to influence designers published to address the concerns of designers who have had
increases distrust in advice from manufacturers, especially when confidence in their knowledge and use of clay liners, mainly
the short-term tests do not take into consideration the effects of when related to origins in dam engineering. The performance of
creep and elevated temperature to favour a particular brand of such composite liners has been evaluated and assessed for a
geotextile. range of conditions (Foose et al. 2001; Rowe 1998, 2005, 2011,
To overcome the advocation of inappropriate technology, 2012, 2020a). However, today there are some perceptions that a
which has been embraced by designers who have not recognized GCL offers only beneficial use without recognizing the
the complexity of the loading and exposure conditions and
limitations or a detailed investigation of the site-specific
consequences thereof, design reviewers have promoted the use
performance requirements. Recent developments in which GCLs
of a field test undertaken in the trial pad or early phase of
are improved by adding polymers or bentonite modifiers to
construction of the containment barrier itself to confirm the
designer’s assumptions (e.g., Rowe et al. 1993). This evaluation compete with sodium bentonite-based GCLs have had a further
follows the form of laboratory evaluations without long waiting influence on perceptions (see §4).
times induced by sample recovery and transport to a suitable lab, It is noted that the inclusion of a blanket drain beneath the
laboratory sample preparation time and queuing time for access clay component of composite liner subjected to elevated
to laboratory equipment. This innovation has also allowed for temperatures, such as in heap leach pads and some mineral
field tests at more representative temperatures than a laboratory deposits, will probably lead to desiccation and reduced
performance of mineral clay liners (Doll 1997), and similarly for
235
GCLs (Acikel et al. 2018; Lu et al. 2018) further compounded applications, to ensure adequate hydration of the GCL for
by the effect of salts and cations in the permeant. equivalent performance to a 600 mm thick CCL. However, this
is not possible in composite liners above capillary breaks or
drainage layers nor in double composite liner facilities (Legge et
al. 2007) and may not be true for applications where the GCL is
in direct contact with adjacent soils that have not been compacted
with adequate moisture content or soils containing smectite. In
such cases, the degree of GCL saturation may be even lower than
the 30 to 70% typical of application in contact with non-cohesive
soils (Rowe 2020a).
It may be presumed that the advective performance of a GCL
is defined by the bentonite content and not the method of
manufacture This may be true irrespective of particle size
(granular or powdered) provided that the GCL is (1) fully
hydrated, (2) under high confining stress, (3) the bentonite is
from the same source, and (4) the products have the same
bentonite mass per unit area. However, it is most unlikely to have
these four parameters constant for different GCLs. Rowe (2020a)
notes that the GCL structure can significantly influence the
moisture content of the bentonite at the low confining stresses
typical of the hydration phase, and thus finer may be better. Even
this suggestion may not hold true for facilities in which
containment is required soon after installation – because the
initial wetting front advance through a GCL is higher for
granular than powered bentonite. Thus, it may be preferable in
some situations where the pollutant receiving face is on the
opposite side of the GCL to the wetting face to make use of less
finely ground bentonite.
Installation guidelines from GCL manufacturers usually
recommend covering of a GCL within a timely manner to
overcome panel shrinkage, the risk of which can be further
reduced by specifying GCLs incorporating a scrim in their
structure. The timely manner is ill defined and may be influenced
by a range of factors including the moisture retention curves for
the particular bentonite, local climatic conditions, and nature of
materials adjacent to the GCL. While it is good practice to cover
a GCL with a material that provides confining stress (and also
usually reduces induced temperature and moisture loss as well)
on the same day as placement, this may not always be practical
in which case a more careful selection of GCL and alternative
construction methodology may be required to limit loss in
performance.
The influence of bentonite powder loss from a GCL on the
performance of underlying thin drainage systems, especially if in
the form of a geosynthetic drainage layer of net or cuspated
sheet, can significantly affect drainage performance. The loss of
gel from a modified GCL can have a similar detrimental
influence on thin drainage systems, especially at the bottom of
Figure 7.7. Typical construction related problems: (a) Tear induced by side-wall slopes where the bentonite or gel would accumulate
material shearing (platinum tailings storage facility); (b) Rotational tear within the thin geosynthetic drain and partially or wholly block
induced by wheel turning; (c) Dust penetration of a geotextile protection the thin drain.
layer interface A further challenge arises fail to consider unintended
consequence of all potential failure modes (see §5.6). For
example, concerned with the potential of hydration of NaB from
adjacent soil or groundwater containing cations or salts that
would detrimentally affect NaB GCL performance, a designer
may specify a GCL with a polyethylene coating in a composite
base liner without recognise the implications of the polyethylene
backing on structural stability. Consideration must also be given
to the interface shear strength between polyethylene backing and
adjacent soil or the influence on consolidated undrained
conditions developing within the GCL bentonite layer between
the polyethylene backing and overlying geomembrane should
the bentonite be hydrated prior to the application of confining
Figure 7.8. Oversized aggregate recovered from the leachate collection
system specified as 38 mm to 75 mm aggregate in a 300 mm thick layer
stress. A further consideration is the potential initial hydration of
above a geotextile cushion or protection layer aimed at limiting the bentonite by the leachate or pollutant itself at holes in the
geomembrane strain. geomembrane increasing k. As discussed in §5, the designer
must consider all these factors; unfortunately, they often fail to
It is commonly perceived that there is adequate moisture in do so.
the soil above or below a GCL, such as in capping and base liner
236
It is not unusual to expose bentonite samples from various such as these with either a sacrificial geotextile which is removed
GCL suppliers to a modified swell index test in which the fluid immediately before waste placement or preferably to cover the
medium is replaced by leachate from site or a simulation thereof. geotextile with a compatible buffer layer of unsaturated tailings
The results of such tests have shown significant differences in to avoid the risk posed by excessive fine mobile material. A
the performance of various bentonite GCLs subject to the same further innovation in mining waste deposits of gold and coal is
leachate that could not be predicted from the manufacturer’s the specification of a P-trap on the outlet pipes, which are
product description. For example, in a recent case two GCLs employed as a means of operating drains in an anaerobic
each had initial properties given in the manufacturer’s sheets condition to inhibit the rate of ferric oxide precipitate forming
indicating >80% montmorillonite, swell index SI > 24 mL/2g, within the drains. It is recognized that micro-invertebrates
and kw < 5x10-11 m/s. However, when permeated with sewage develop and form a biological clogging risk under such stable
lagoon water at a low stress applicable to a sewage lagoon, the habitat conditions. Thus, the P-trap is removed during the
Wyoming bentonite in one GCL had k~1x10-10 m/s (SI=12) operational phase to induce a shock or sudden change in
while the bentonite in the second GCL, said to have been natural environmental conditions within the drain and reduce the volume
NaB from India, had k~ 2x10-8 m/s (SI=6). For the sewage of biological material as aerobic conditions develop before
lagoon with 2 m of head initially on the GCL, in 7 days this replacing the P-trap. This practice has also evolved in coal
corresponded to a decrease in liquid level in the lagoon by 0.02 mining to reduce the risk of air intrusion into drainage layers and
m and 1.8m, respectively. For a small community in remote reduce the risk of spontaneous combustion.
area, this difference represents a difference between acceptable The above reflects that performance-based designs which
leakage and failure. make use of geosynthetic materials as alternative elements to
natural materials may be detrimentally affected if the barrier
7.2.4.3 Perceived performance of geosynthetic filters and drains system is not considered as a whole by considering the
as opposed to granular filters performance of elements and their interaction with other
It is believed that a geotextile may be used as a replacement to components of the barrier as well as the change in loading
filter sand in tailings and return water dam facilities or similar. conditions with time (see also §5). The assurance of predicted
This may be true provided the filtration and drainage system as performance is improved by relatively simple, inexpensive, and
a whole remains functional throughout the operational life and short-term, site-specific nonstandard “test and measure”
post-closure service life requirements, including incidents of evaluations of components or the barrier system to confirm
earthquake loading or tremors and liquefaction. But, again, there aspects of chemical compatibility, filter compatibility, and
are many caveats. Post-construction filter and drainage tensile strain limitation.
performance does not have the benefit of performance evaluation
before use that geomembrane liner systems have by way of 7.2.5 Project specifications related to containment which may
electric leak location surveys (Beck 2015). So too, is it important affect the stability
to recognize the extent of filter compatibility between waste Design reports have been submitted claiming safe structures
material or tailings and the underlying material, particularly at based on a single interface shear test. Still other designs have
holes or discontinuities in geomembrane-lined facilities. Chou et been submitted in support of licence applications with a claim
al. (2018) identify a critical stress condition where piping of that stability will be addressed only after tender award or during
tailings occurs through geomembrane defects where the construction. The failure to recognise the importance of interface
underlying foundation or embankment soils do not meet shear limitations, and the effect of saturated conditions and strain
conventional filter criteria, which could lead to underestimation compatibility poses an unfair risk and liability on the facility
of leakage and of pore pressures within an embankment. owner – the facility being either unsafe or having excessive
The use of data sheets for designing geocomposite drains fails conservatism.
to recognize the effect of intrusion into synthetic drainage cores In a particular facility in the coal mining industry, a low-risk
(Legge et al. 2014, 2015), which could lead to similar waste stream was provided with a single geocomposite (LLDPE-
unacceptably high pore water pressure and reduced stability S and GCL) liner with substantial overlying granular drainage
Factor of Safety. The loss of stability may be further aggravated layer. The authority’s rejection of the design based on unproven
in designs failing to take cognizance of the site-specific stability resulted in confirmation of a Factor of Safety of 1.0 for
probability of seismic events for dams (Singh et al. 2010), and the pollution control dams (PCDs) under construction. The
the effect of dynamic displacement on geosynthetic interface challenge faced then was that the facility was under construction
shear strengths which may be greater or lesser than static state in advance of authorization or licensing. Hence, the regulator
values for various seismic events (Pavanello et al. 2018). advised the applicant to profile the base of the intermediate walls
Empirical and theoretical design criteria are available for in a saw-tooth cross-section so as to mobilize the base material
geotextile filter design. These may be augmented by short-term shear strength. The consulting engineer embraced the concept
tests such as the gradient ratio, interface flow capacity, or fine and optimised the profile with increasing wave length between
fraction filtration tests. However, these tests do not reflect the saw-teeth as the embankment wall height reduced above the
medium to a long-term reduction in filter performance brought composite liner. It is considered good practice to include
about by precipitate or biological clogging within the highly drainage trenches and berms in liner systems so as to increase
porous geotextile structure or porous granular drainage material. stability. Caution must however be taken in assessing the
The chemical and biological clogging of geotextile-wrapped consequences of such drainage trenches on inducing excessive
drains in gold tailings dams is well known. total tensile strain, particularly at positions where the drainage
The development of filter performance tests that recognize line changes direction which requires the intersection of three
the influence of the fine fraction from the base soil on geotextile planes in the geomembrane and related geomembrane welding at
filter capacity has been a significant advancement in providing these critical containment points of maximum pollutant or
competent drainage components of contaminant containment leachate flow concentration.
barrier systems, in particular in mining applications where Recognition of interface shear limitations and managing the
windblown or water-born fines may lead to a build-up of the fine slip surface above the containment liner in designs incorporating
fraction at the geotextile interface or within the geotextile geosynthetics is not adequate to ensure stability. The
structure in the early phase of operation, which leads to construction of composite (geomembrane plus CCL) liners in
substandard performance. As a result, it has become good which the clay component compaction specification is 98%
practice to either cover geotextile filters in drainage applications Standard Proctor maximum dry density (MDD) at minus 1% to
plus 2% OMC can be compromised if construction takes place at
237
excessive moisture contents. The pursuit of particular The challenges brought about by designs based on product
specifications requires justification. specifications may be overcome by regulations and facility
The multiple failure of various HDPE geomembrane products owners who demand performance-based designs that rely on
at a particular water storage reservoir, emphasises the need to evaluating the actual materials and their interaction. To enhance
consider the nature of the material to be contained, especially the this performance-based approach, a check list requiring the
presence of oxidising agents before construction. results of a performance assessment or determination must be
The same can be said for specifying drainage material and signed off by both the waste management facility owner and the
interface shear strength. The specification of an asperity height designer responsible for the containment barrier needed to
does not define interface shear for geomembranes. The texturing protect the environment and human life. This focus on
through micro-spikes or other features is dependent on more than accountability by sign-off significantly improves the quality of
just height, and is influenced by the density, stiffness and nature design confidence and predicted performance.
of the texturing in relation to the adjacent material. Table 7.1
provides a comparison of two series of interface shear results for 7.2.6 Inadequacies and virtues of construction quality
a range of geomembranes having different micro-spike height assurance (CQA) plans
and density of distribution. The nature of soil selected for the test It is increasingly common to include material specifications in
series was a low permeability clay typically used in CCLs, and the CQA part of a design report. While these specifications on
an ash as these are typically found in large facilities. It is critical occasion contradict the specifications on the engineering
that interface shear tests be conducted on the materials that will drawings, there are occasions where particular specifications
actually be used on a project to obtain the strength parameters amend standards and test methods without technical
and the concept of multi-layer shear box tests (Khilnani et al. substantiation. For example, the specification of particular ash
2017) to reflect not only interface but also internal stress-strain content standard and/or high-pressure oxidation induction time
relationships of materials requires more consideration by (OIT) for a geomembrane does not necessarily add value to the
designers. It is erroneous to assume a particular height of micro- design (Ewais et al. 2014). Still, it tends to increase costs by
spike will provide the highest interface shear or even adequate excluding material suppliers whose products comply with
interface shear strength. national or international standard specifications. Furthermore, no
justifiable reason is provided for deviating from the rationale
Table 7.1 Interface shear results for five geomembranes (having different behind the GRI GM13 standard specification for geomembranes
micro-spike heights and distribution) with a compacted clay and ash, (Hsuan and Koerner, 2015) in almost all particular specifications
under different normal stresses (acknowledgement H Venter, NAKO of deviation. In addition, these limited specifications are
LBE (Pty) Ltd for sponsoring the test series). indicators and do not inform the design with respect to polymer
or resin performance in the specific containment application.
Textured Clay at interface Ash at interface
The differential scanning calorimeter (DSC) test provides
Test geomembrane MDD 1,509 kg/m3 MDD 1,130 kg/m3
invaluable insight into the resins and antioxidants/stabilizers
number micro-spike height OMC 26.5% OMC 40.7%
used to formulate a black geomembrane (§2). For example, the
(mm) c (kPa) c (kPa)
temperature peak or spike at a transition such as at melt
1 0.40* 12.8 8.3 16.5 27.8
temperature is characteristic of a polymer (e.g., PE or PP etc.).
2 0.65 7.5 9.4 40.8 36.2
By observing the DSC thermogram of energy absorption versus
3 8.4 3.7 39.9
temperature for a particular polymer sample, one can identify
4 13.1 7.2 7.8 35.8
with some certainty from the peak or multiple peaks the nature
5 6.8 8.5 0 40.8
of the polymer that has been used (e.g., a blend of PE and PP).
6 1.1 18 9.4 13.7 38
The area under the bell curve gives the comparative percentage
Shear tests were undertaken at normal stresses of 20, 200 and 400 kPa.
of makeup of the specimen. The position of the peak reflecting
*Texturing is a random full surface treatment, not spaced micro-spikes.
the thermal transition temperature may give insight into the
makeup of resin blends. However, the DSC will not identify the
The above results confirm the specification that micro-spikes presence of inclusions such as polycarbonate nor polystyrene.
height alone does not define maximum interface shear strength The interpretation of such DSC results requires careful
benefit. This has been recognized in the 2019 amendment of the consideration with respect to the particular application. It is
Geosynthetics Research Institute (GRI) standard specification known that some thin (less than 1 mm-thick) blended
for geomembranes known as GRI GM13. geomembranes have served as satisfactory lagoon liners for
There is an increasing number of polymer enhanced bentonite decades under exposed African conditions to hazardous
GCLs (EB-GCLs) being marketed on the basis of improved substances including hydrocarbons. On the other hand, HDPE
hydraulic performance (§4). However, insufficient consideration geomembrane liners have failed under similar exposure
has been given to the potential unintended consequences of conditions when serving for potable water containment.
polymer biodegradation, polymer leaching, and polymer elution. Although in these applications the geomembrane is accessible
If polymer elution occurs, then open pores will remain within the for repair, they are included for an appreciation of the complexity
bentonite and result in a substantially higher long-term the in assessing liner suitability for particular applications. The type
hydraulic conductivity of EB-GCLs (§4.6.3 and §5). Even if of geomembrane and its specific formulation is critical to its
elution does not occur water soluble polymers will be leached longevity (Koerner and Koerner 2017; Rowe 2020b; Rowe et al.
and over time under field conditions and they may be 2020). It is recommended that the owner of a facility makes
biodegraded. Polymer elution also reduces interface shear provision for geosynthetic performance evaluation at the
strength (Chen et al. 2017). Furthermore, the effect of anions on discretion of themselves and/or reviewers, including the
increasing permeability of bentonite-polymer GCLs (Tian et al. statutory review in particular, to enhance assurance in the
2017) has not been addressed in the cases referenced previously, materials being procured for projects. Such truly independent
nor has the effect on diffusion through such EB-GCLs been considerations may include analysis using DSC, gas or liquid
adequately considered (Sample-Lord et al. 2017). Until these chromatography mass spectrometry, infrared scanning or the EU
issues are resolved, extreme caution is advised in the use of EB- 10/2011 test. Resin containing calcium carbonate has been
GCLs and such use should only be considered after a careful marketed without consideration of its significant detrimental
failure modes and effects analysis (§5.6). effects on service life. The value of thermogravimetry tests as
used for determining carbon black but undertaken at even higher
238
temperatures (approximately 900°C) should not be does not always yield the optimum containment performance or
underestimated. cost-effective solution for the type of waste material to be
It is generally accepted that the higher the crystallinity (and contained due to chemical incompatibility or similar site-specific
associated transitional phase temperature), the more chemically conditions.
resistant the geomembrane, however, this may also be coupled
with lower stress crack resistance (§2 and Rowe et al. 2004). 7.2.7.1 Brine and acid lagoons
Significant variation has also been recorded in OIT test results, In the gold and coal mining industry, liquid waste streams with
both prior to and post installation (Morsy and Rowe 2017). The a high concentration of salts or strong acids are occasionally
importance of OIT testing is reasonably well recognized for produced. Such liquid waste streams are generally considered
geomembranes but received scant attention with respect to difficult to contain due to the reaction between the liquid waste
geonets and geotextiles. In recent evaluation the long-term and clay or similar material.
performance of two physically very similar geonets, one product It is noted that the brine discharge into the pollution control
is Std-OITo~ 100 minutes the standard deviation of 7 minutes dam (PCD) or lagoon is not the concentrated condition
and data ranging from 80 – 110 min. This is at the minimum that representing the worst case rather, the end of the evaporation
would be expected for the geomembrane to be used conjunction period represents the concentrated waste type to be contained for
with the geonet however it greatly exceeds that of the second pollution prevention.
geonet which had Std-OITo~70 minutes the standard deviation The historic lagoon containment standard of 1998 or modern
of 25 minutes and data ranging from 20 – 120 min. post-2013 standard barrier system for such waste of double
It is thus recommended that the CQA report demonstrating composite liners separated by a leak detection drainage layer
compliance of the as-built facility with the accepted design could be made acceptable provided the performance of
should include not only the as-built drawings, but also the type alternative barrier system recognizing hydraulic gradient across
of test, number of those tests undertaken; the maximum, the alternative to the primary composite liner is demonstrated as
minimum, mean values and standard deviation for the test having equivalent performance to that of a conventional system.
method and an indication of where the tests were undertaken. It is known that sodium bentonite GCLs react negatively to
This record should be for all materials – natural and geosynthetic cation exchange (Elges 1985; Scalia and Benson 2011) and to
(e.g., geomembranes, geonets, geotextiles), and include the salts (Petrov and Rowe 1997; Petrov et al. 1997a,b).
aggregate leachate collection and detection systems where Furthermore, the hydraulic gradient across thin geocomposite
relevant. Similarly, an evaluation of geomembrane performance liners (2 mm thick geomembrane plus 6 mm thick GCL) or
relative to standard or project specific specifications of OIT geomembrane only liners is orders of magnitude higher than
across a composite (2 mm thick geomembrane plus 600 mm
(standard and/or high pressure) should be sensitive to variation
thick CCL). It is critical to address such gradient by mitigation
in results for the same material tested in a single machine or
thereof through ensuring at least direct/intimate contact and
single laboratory or different laboratories. The results may be
normal loading for composite liners, or by alternative means.
influenced by the machine, the operator or any contamination of Although direct contact between geomembrane and GCL or CCL
the extremely small specimen (5 mg) by hydrocarbons or even is pursued by particular construction techniques and ballast
finger oils. Reliance on ISO standards for manufacture is not layers to overcome wrinkling etc., the low interface
adequate to define performance, as shown over the past five transmissivity affecting barrier performance is achieved only
years during which ISO 9001 geomembranes have been supplied when loading exceeds about 50 kPa.
to RSA from European Manufacturers which upon installation Although it is normal practice to have a double composite
have not met performance specifications, including visually liner separated by a LDS for hazardous liquid waste containment
obvious discontinuities. alternative designs of the same performance can be achieved
Descriptions provided by an independent CQA monitor are without the use of a GCL or clay component to the primary
inadequate and have been known to not reflect the offsite stored composite liner when the nature of waste and loading requires
data. This aspect has brought the definition of independence into so. This is achieved, typically for acids and concentrated brine
question. The value of electrical leak location survey cannot be liquid waste, by a triple geomembrane liner (Figure 7.9)
understated. pioneered in South Africa in the 1990s as a cost-effective
It is with some disillusionment that design reports and alternative in the gold mining industry for strong acid
drawings do not reflect instrumentation to the extent typical of containment, and later adopted for coal mine wastewater
other engineering structures. For example, the absence of site treatment works (reverse osmosis plants).
boundary definition, strain gauges, thermistors and flow gauges Although the upper two geomembranes separated by a
place the facility owner at risk in many situations. Similarly, the drainage layer as shown in Figure 7.9 are accepted as serving
failure to address critical elements of barrier systems such as equivalent performance to a primary composite liner, the
drainage pipes, drainage cores, sump liners, and filter materials underlying leak detection system (LDS) must still be above a
to the same level of surety as geomembranes, clays, or GCLs bottom (tertiary) liner (often just a single geomembrane). The
introduces weaknesses in the design and performance of barrier lowest (tertiary) liner often rests on a bed of sand or geotextile
systems. which renders its composite effect as null and void. This lower
To illuminate such shortcomings in CQA protocols critical to geomembrane/ geomembrane interface should be drained and
the performance of barrier systems, a standard CQA protocol has preferably under gravity from the lowest point to a monitoring
been developed for use by designers that address the above sump at which the performance of the overlying system is
principles and procedures. This CQA document is made confirmed. It should be noted that a drainage layer is only
available by the CQA regulator for use by designers of public effective when there is a low permeability layer beneath it.
and private-owned barrier systems to enhance knowledge and In this design, the upper two geomembranes with separating
skills development, although its use is not compulsory. drainage layer serve to reduce the hydraulic gradient across the
tertiary liner to approximately unity irrespective of the depth of
7.2.7 Distinctive waste streams and particular barrier systems
the brine.
The conventional barrier system of a filter protected aggregate
There is often no ballast layer above the primary liner. Such
drainage layer for leachate collection and management,
ballast layer would normally protect the geomembrane from UV
overlying a cushion layer of soil or geotextile to protect the light and heat which accelerate degradation, as well as protect
composite containment liner of geomembrane in intimate contact the liner from wind and/or mechanical damage. It is inconsistent
with a compacted clay liner or GCL (or combinations thereof) to have a concrete access ramp into the lagoon for vehicles and
239
then allow traffic of vehicles over the unprotected geomembrane 2004) and by the partition coefficients of the compounds being
liner. considered.
To mitigate this diffusive characteristic, designers may
consider alternative materials (McWatters and Rowe 2010;
McWatters et al. 2016a), alternative systems or combinations
thereof to either contain the VOC stream to a greater extent or to
remove the diffused VOCs from further diffusion potential and
treatment (Legge et al. 2007). Interested persons are referred to
methods of treatment of VOCs such as at Casey Station in
Antarctica (McWatters et al. 2016b, 2019) and a petrochemical
facility containing sludge (Meyer et al. 2014). In the latter case,
the leak detection layer is subject to a negative pressure to draw
Figure 7.9 Triple geomembrane liner for brine evaporation barrier
moist air through the system to commensurately maintain
The question of thermal conductivity across barrier systems hydration of the clay component to the primary composite liner
has been studied by various investigators over the past decade. and remove the VOCs that have diffused through that barrier.
Early work in the Department Water and Sanitation, RSA The exhaust fumes are passed through a treatment process
laboratory in Pretoria West showed thermal gradients across incorporating carbon scrubbers to capture the VOCs for further
double geocomposite liners to be about 2°C irrespective of the treatment or disposal.
temperature, but note this primary liner had a moist GCL in
contact with the overlying geomembrane. This research 7.3 Challenges in barrier system planning, design,
development was published at Landfill Interest Group (LIG) construction and operation: Asia
Durban 2005 and Sardinia 2007 (Legge et al. 2007).
7.3.1 Influence of MSW variants on stability, leachate
When two liners are separated by an air void induced by a
spacer of geonet, or cuspated drain, or geotextile, or similar collection and gas management
material, the air void acts as a thermal barrier (see Singh and This section will focus on challenges to managing the
containment of municipal solid waste (MSW) in most of the
Bouazza 2013 for thermal conductivity of HDPE
geomembranes, wet and dry GCLs, and wet and dry geotextiles.) Asian countries. Table 7.2 shows physical compositions of fresh
As the primary geomembrane heats up due to say solar radiation MSWs collected from the Asian developing countries. In most
developing countries like China, India and Indonesia, the food
in exposed conditions, so the enclosed air void reaches the same
temperature and heats the drainage layer and secondary liner. waste content is much higher than that in developed countries in
The extent to which the drainage layer and in particular, in RSA Europe and North America. MSW with food waste making up
over 40% can be classified as high-food-waste-content (HFWC)
the usually thinner, secondary geomembrane liner cools largely
depends on the moisture content and conductivity of what lies MSW, otherwise it is low-food-waste-content (LFWC) MSW.
beneath it. If it is moist/wet GCL or soil, then the heat in the As shown in Table 7.2, the water content in the HFWC MSW is
more than 50% (wet weight basis), greater than that in the LFWC
secondary liner can be conducted away from the liner as quickly
as it passes through that geomembrane without an increase in MSW. When compared with the containment of LFWC MSWs
temperature of the materials above. The same can be said for in Europe and North America, the containment of HFWC MSWs
has more technical challenges with respect to landfill stability,
designs with geotextiles beneath the secondary liner; however,
the elevated temperature tends to drive off moisture from GCLs landfill gas collection, and control of leachate-related pollution
(Rowe 2020a). Obviously, the same drying applies to geotextiles (Chen et al. 2014). Any innovative solutions to the challenges
posed by HFWC require an understanding of the degradation
well. If the air void remains trapped, such as in double glazed
windows of northern Europe, then the temperature of the upper process of HFWC MSWs and the associated liquid-gas
liner and barrier system rises even higher than the norm due to interaction at wet landfills.
the lower thermal conductivity of underlying layers inhibiting Table 7.2. Physical components and initial water content of fresh
heat escape (Take et al. 2014, 2015). The temperature difference MSWs in different countries (by wet basis, %)
of geomembranes at wrinkles is typically 10°C higher than areas Country GDP per capita Organic fractions Inorganic fractions Initial water
adjacent to the wrinkle where the membrane is in direct contact (US $)
Food waste Paper, Textile, Wood Plastics Metals Others
content
with soil or hydrated GCL. China 8643 57.91
cardboard
10.55
leather
2.72 1.69 11.85 0.76 14.52 52.44
In the case of double geomembrane only liners, the primary India 1983 42.00 6.00 4.00 0.00 4.00 2.00 42.00 NA
Philippines 2976 41.60 19.50 0.00 0.00 13.80 4.80 20.30 NA
liner could be cooled by overlying liquid, but more important is Indonesia 3876 74.00 10.00 2.00 0.00 8.00 2.00 4.00 NA
to avoid the air temperature in the spacer voids from rising Sri Lanka
Nepal
4085
834
76.40
80.00
10.60
7.00
0.00
0.00
0.00
0.00
5.70
2.50
1.30
0.50
6.00
10.00
NA
NA
excessively. By moving the air or ventilation, the separating Thailand
Malaysia
6591
9813
48.60
40.00
14.60
15.00
0.00
3.00
0.00
0.00
13.90
15.00
3.60
3.00
19.30
24.00
NA
NA
structure or drainage spacer does not increase in temperature to Brazil
U.K.
9895
39735
42.90
25.00
19.70
31.00
4.50
5.00
5.20
0.00
18.70
8.00
1.50
23.00
7.500
32.00
50.00
32.00
the same extent. Even a 10°C drop in liner temperature will U.S. 59501 13.60 35.50 4.60 3.40 13.20 6.90 22.80 18.00
substantially increase the service life (Rowe 2005, 2012, 2020b).
Alternatively, the liner could be covered with, say, cement 7.3.2 Degradation and stabilization of HFWC MSWs
stabilized soil layer if compatible with the waste stream, or a The behaviour of the HFWC MSWs was investigated by
high-performance geomembrane specifically developed for such performing a large-scale bioreactor experiment (CELL1, 5m×5m
conditions could be used to partially mitigate this exposure ×7.5m, 91.3 tons of wastes) in Zhejiang University (Zhan et al.
impact. The nature and extent of cover layers in lieu of ballast 2017a). Based on the experimental results and field observation,
bags would be influenced by the shape of the floor area and the a degradation–consolidation model was developed to simulate
nature of the sediment or precipitate to be removed. the stabilisation process of HFWC MSWs at landfills (Chen et
al. 2020). The model incorporated the biological degradation,
7.2.7.2 Wastes containing volatile organic compounds skeleton deformation, two-phase flow, and solute transport. The
The performance of composite liners with respect to total solute experimental and simulation results showed that the stabilization
transport is widely published (Foose et al. 2001). Similarly, it is process of HFWC-MSW landfills could be divided into three
widely known that the rate of diffusion of volatile organic stages: rapid degradation, slow degradation, and post-
compounds (VOCs) even though perfectly intact liner systems is stabilization, as shown in Figure 7.10. Rapid degradation occurs
influenced by temperature (Sangam and Rowe 2001; Rowe et al. to the HFWC MSWs just after being disposed of at a landfill.
240
the exudation coefficient of landfilling operation unit, C L 2 is
Biodegradation
the exudation coefficient of intermediate cover unit, C L 3 is the
of HFWC-MSW exudation coefficient of final cover unit, M d is the daily
landfilling of MSW (t/d), M c is the initial water content of
MSW (%), Fc is the field capacity of completely degraded
Release of intra-particle Generation of landfill gas MSW (%), and w is the density of water (t/m3).
water within two months (LFG) from half to two years
In addition, a two-stage LFG generation model was proposed
for HFWC MSWs to capture the distinct behaviour of LFG
Large leachate Significant waste Gas entrapped generation during the rapid and slow degradation stages:
production compression underwater
High leachate
level Reduction of
hydraulic conductivity where Qn is the LFG production for the nth year of operation
(m3/year); n is the time since the start of operation of landfill
Low collection Low drainage (year); t is the time since MSW is filled (year); M t is the filling
efficiency for LFG efficiency for leachate
amount of MSW in the nth year (t); is the conversion
Figure 7.10. Stabilization stages of HFWC-MSW landfills (Chen et al. coefficient of degradable organic carbon (%); LR 0 is the
2020) potential CH4 generation capacity of rapidly degradable
This is because food waste contains a large proportion of components, including food waste (m3/t); LS0 is the potential
rapidly degradable chemical components (e.g., fast cellulose, CH4 generation capacity of slowly degradable components,
sugar, protein and lipids). The rapid degradation results in a including paper, textile, wood, and so on (m3/t); k R is the CH4
release of intra-particle water embodied in the food waste, and a generation rate of rapidly degradable components (year-1); and
large quantity of leachate is produced. It is observed from the k S is the CH4 generation rate of slowly degradable components
large-scale bioreactor experiment that 26.8 m3 of leachate was (year-1). Calibration of LFG pumping test results from the
collected from the 91.3 tons HFWC-MSW within the first two landfills in China indicates that the value of k R ranges from 1.1
months after waste filling (Zhan et al. 2017a). The first two
to 1.4, and the value of k S from 0.09 to 0.11.
months of leachate production represented nearly 30 % of the
To assess the degree of degradation of HFWC MSWs, the
total mass of the waste. The rapid degradation of food waste also
normalized ratio of cellulose to lignin -RC/L/R0 was
resulted in a rapid generation of landfill gas (LFG). The time lag
proposed as the degradation degree index. T and the index is well
for LFG generation is only 0.5-1 year at the MSW landfills in
correlated with static respiration index, fine fraction content, and
China. The rapid degradation stage generally lasts about 2–3
leaching characteristic (Chen et al. 2018).
years. Within that stage, 80% of the solid mass is hydrolyzed,
As shown in Figure 7.10, the value of corresponding to the
100% of the intra-particle water is released, 60% of the LFG is
end of the rapid and slow degradation stages is 0.6 and 0.83,
generated, and 80% of the compression is completed (Figure
respectively. When the value of is greater than 0.83, the
7.11). During the subsequent slow degradation stage, which lasts
landfilled MSW is basically stabilised. The food waste content
about 20–30 years, over 90% of the solid mass is hydrolyzed,
and degradation degree would significantly influence the
90% of the compression is completed, and 80% of the LFG is
engineering properties of MSW. It is found that the primary
generated.
compression ratio and bio-induced compression ratio of HFWC-
MSW are both larger than those of low food waste content
(LFWC) MSW (Xu et al. 2019, 2020), and the hydraulic
conductivity and gas permeability of HFWC-MSW are both one
order of magnitude smaller than LFWC-MSW, and the shear
strength of MSW changes significantly with degradation degree
(Gao et al. 2019).
7.3.3 Liquid-gas interaction at wet landfills of MSWS and its
counter measures
The liquid-gas interactions are very complex at HFWC-MSW
landfills, as shown in Figure 7.11. The great quantity of leachate
production will cause a buildup of leachate or leachate mound at
landfills with inadequate drainage conditions. Under such
circumstances a large amount of LFG generated from the waste
below the leachate level cannot be collected effectively (Zhan et
Figure 7.11. Liquid-gas interactions at HFWC-MSW landfills (Zhan et
al. 2017b) al. 2015) but is accumulated in the waste piles below the phreatic
line. It is observed from another large-scale bioreactor
After that, the HFWC MSWs are almost stabilized. Due to experiment (CELL2, 5m×4.2m×7.5m) performed in Zhejiang
the significant contribution of self-released leachate of HFWC University that the gas accumulation zone was distributed
MSWs during the rapid degradation stage, a modified leaching randomly and the gas pressure in the zone increased and
coefficient method was proposed to predict the leachate dissipated in cycles. The gas-breaking value was up to 10.5 kPa
production rate at landfills (Chen et al. 2018): greater than the pore liquid pressure (Zhan et al. 2019). A similar
phenomenon was observed at the Xingfeng landfill in China,
where a maximum gas pressure of 80 kPa was measured at
depths below 15–20 m (Ma et al. 2019). The gas accumulated
where Q is the daily leachate production (m3/d), I d is the underwater would further increase the leachate mound, and this
daily average precipitation (mm/d), A1 is the area of contribution was estimated to be 21%–28% of the total leachate
landfilling operation unit (m2), A2 is the area of intermediate level (Zhan et al. 2017b). The combined effect of high leachate
cover unit (m2), A3 is the area of final cover unit (m2), C L1 is levels and high gas pressure was the critical cause of frequent
241
landfill failure in Asian countries (Chen et al. 2017; Ma et al.
2019). Centrifuge model tests revealed that the ratio of critical
leachate level and landfill height was 0.75-0.92 when the slope
ratio was 1:1-1:3, and it increased with an increase in waste
degradation (Chen et al. 2017). In addition, the gas accumulated
underwater would result in a reduction of hydraulic conductivity,
decreasing by one order with an increase in gas content from
13% to 21% (Zhan et al. 2017b). Under the condition of low
hydraulic conductivity, the produced leachate is not easy to drain
out of the waste pile, causing an inadequate drainage situation.
The above complex liquid-gas interactions tend to develop into Figure 7.12. Configuration of three-dimensional drainage system at a
an adverse condition, resulting in great difficulty in managing wet landfill
wet landfills. Thus, effective measures should be taken to tune
the interactions to solve the technical difficulty. 50000
46472 m3/h
LFG collection for LFWC may be via either vertical or
LFG collection (m3/h)
horizontal gas collection wells (Zheng et al. 2018, 2019a,b, 40000
2020). However, for HFWC landfills, understanding the coupled Enhanced temporary
30000 cover system +
leachate and gas flow in the waste body is needed for the leachate drawdown
scientific design of leachate drainage and LFG collection. Given 20000
Leachate
that leachate and gas migration in a waste body is dominated by drawdown 10169 m3/h
the preferential flow, a dual-porosity model for coupled leachate 10000
2000 m3/h
and gas flow has been established, in which the waste body is
divided into fracture and matrix domains (Hu et al. 2020a). It has 0
2007/03/19 2009/05/04 2011/06/20 2013/08/05 2015/09/21 2017/11/06 2019/12/23
been verified by pumping test data that the dual-porosity model Time (data)
performed better than the single-porosity model. Pumping Figure 7.13. Total LFG collection rate measured at the Xiaping landfill
vertical wells is commonly used to lower the leachate level in
wet landfills. However, the efficiency in the well pumping tends 7.3.4 Use of capillary barrier soil covers in the Asian monsoon
to decrease with a drawdown of leachate level, which can be regions
explained by the gas volume of waste underwater increases with Last century, the design of final covers in China followed the
the drawdown of leachate level and therefore significantly standards and procedures of developed countries, incorporating
reduces the hydraulic conductivity of waste (Rowe and a geomembrane as an impervious barrier as specified in the
Nadarajah 1996; Zhan et al. 2017b)—also clogging of the wells. current national standards. This century has seen the
The horizontal well is a potential alternative for the drawdown development of a new understanding of the degradation of
of leachate level at wet landfills. The trial construction and HFWC MSW and identified the critical role of rapid hydrolysis
operation of horizontal wells at Tianziling and Xingfeng landfills leading to large leachate production and fast gas generation
in China showed that the leachate level could be lowered to near (Zhan et al. 2017a). After two or three years, the LFG generation
the elevation of horizontal wells (Hu et al. 2020b). Thus, a three- rate decreases sharply to a low level. Additionally, much of
dimensional drainage system was proposed to control high China is subjected to monsoons and the consequent heavy rain
leachate mounds in wet landfills (Figure 7.12). The drainage coincides with the period of high temperature and vegetation
system involves a combination of layered horizontal trenches or growth. These factors have led to a revision in the design of final
wells and deep vertical extraction wells and serves in addition to covers in China to accommodate the nature of HFWC MSW and
the leachate collection system (LCS) at the base. Each layer of the climate. With the successful application of soil covers
horizontal trenches or wells is set up to capture part of the incorporating the capillary barrier effect (capillary barrier soil
leachate generated above the present layer and so as to relieve covers, CBSCs) in arid and semi-arid areas of North America,
the leachate drainage pressure on the LCS. Vertical extraction many countries have explored the feasibility of CBSCs in
wells are to collect LFG as well as to draw down the leachate response to their own climates. In recent years, research on the
level and so enhance the leachate drainage after the failure of the application of CBSCs in China has drawn much attention.
LCS. Thus, the leachate-gas interaction can be separated and the Theoretical and experimental methods were used to assess the
efficiency in draw down of leachate levels and collection of LFG performance of two-layer CBSCs (see Figure 7.14a) in
can be enhanced. The configuration of a three-dimensional northwestern China in arid, semi-arid, and semi-humid climates
drainage system was successfully implemented at many landfills (Jia et al. 2010; Zhan et al. 2016, 2017c; Long et al. 2018; Zhang
in China. For example, the drawdown of leachate level resulted et al. 2014, 2016). Zhan et al. (2016) constructed the first full-
in an increase of LFG collection rate from 2,000 to 10,169 m3/h scale final cover test site (30 m × 20 m × 1.2 m, see Figure 7.15)
at the Xiaping landfill in Shenzhen of China (Figure 7.13; Zhan at Jiangcungou landfill in Xi’an city, China, where the humidity
et al. 2015). With consideration of the rapid generation of LFG index (i.e., the ratio of annual precipitation over potential
for HFWC, an enhanced temporatory cover system was evaporation) is higher than 0.5. The test site adopted a two-layer
implemented at the Xiaping landfill. The cover system CBSC that used local loess as fine-grained layer material and
comprised: a 1 mm-thick HDPE geomembrane, a soil protection gravel as coarse-grained layer material and was well-
and gas diffusion layer, and gas collection pipes with a spacing instrumented. Artificial rainfall events were implemented to the
of 20 m. Implementation of the enhanced temporatory cover test site before and after the site was vegetated to measure the
system resulted in a further increase of gas collection rate from water storage capacity of the cover. Results demonstrated that
10,169 to 46,472 m3/h at the Xiaping landfill. Half of the the loess/gravel CBSC increased the available water storage by
collected LFG was used to generate 38 MW of electricity per 41% compared with the monolithic loess cover due to the
hour by using LFG power generators, and the other is used to capillary barrier effect between the two layers (Zhan et al. 2017c).
recover automobile fuel. Vegetation had an insignificant influence on water storage
capacity. Field tests were performed to investigate the on-site
methane oxidation capacity of the cover. Results indicated that
the maximum methane oxidation was 93.3 g CH4/m2/d,
242
confirming effective methane mitigation and control for the
HFWC-MSW that has completed the fast gas generation phase
(Zhan et al. 2020a). 2-year monitoring of the hydrologic
response of the test site indicated a good performance of the
loess/gravel CBSC at limiting percolation. The accumulated
percolation was around 8.0 mm/year, which was less than the
recommendation of the USEPA alternative cover assessment
program of 30 mm/year (Zhan et al. 2020b). The CBSC was also
shown to be effective in reducing LFG emissions. This was in
large part because notable water storage decreased the gas
permeability in the loess near the fine-coarse interface by up to
Figure 7.15. Conceptual diagrams of CCBEs: (a) two-layer CCBE; (b)
five orders of magnitude. The above observations facilitate the newly proposed three-layer CCBE in China (Ng et al. 2016)
use of two-layer CBSCs at landfills in non-humid areas of China.
7.4 Challenges in barrier system planning, design,
construction, and operation: South America
The primary waste problem in South America is mining waste.
There is a diversity of geotechnical problems within a mining
operation including stability of open pit and underground
excavation, static and seismic stability of heap leach pads,
drainage systems for tailings dams and ore stockpiles, reservoir
waterproofing barriers for the disposal of fine tailings and
waterproofing of tailings dam slopes, ore pile covers,
reinforcement of associated infrastructure (roads, railways etc.)
and dewatering and consolidation of fine tailings.
Geosynthetics use has been increasing in mining due to
Figure 7.14. Conceptual diagrams of CBSCs: (a) two-layer CBSC; (b)
improved technical specifications, excellent manufacturing
newly proposed three-layer CBSC in China (Ng et al. 2016)
quality control and ease of installation. Their use has improved
environmental protection by saving exploitation and usage of
A new three-layer CBSC has been proposed for landfill
natural construction materials. Nevertheless, mining countries
closure in humid areas of China (Figure 7.14b). The cover
like Brazil, one of the largest ore producers of the world,
system involves the addition of a fine-grained soil below a two-
Argentina, Chile, Peru, and others, still have no regulations with
layer CBSC (Ng et al. 2015a, 2015b, 2015c 2016, 2019). The
regard to the use of geosynthetics in mining, particularly as
added layer has a lower water permeability at low suctions in
components of tailings dams and high piles of sterile rock and ore.
rainy seasons. Thus, it can intercept the infiltrated water through
Designs of facilities incorporating geosynthetic products have been
the upper two layers and reduce the percolation when the upper
based on the knowledge of the designers on classical
two layers lose their effectiveness. The bottom layer can
geotechnical engineering, similar geosynthetics applications
maintain a high volumetric water content in dry seasons
where the performance of such products is well established, and
attributed to the protection of the upper two layers as they have
specifications provided by manufacturers. To obtain relevant
low water permeability at high suctions, indicating that the
geosynthetic design parameters, specific laboratory tests have
bottom layer can serve as a LFG emissions barrier. Results
been carried out in national and overseas laboratories.
obtained from flame model tests and back-analysis indicated that
The challenges of managing mine waste in South America
the coarse-grained layer changed from being a capillary barrier
are sufficiently large that §8 is devoted to their discussion.
in a conventional CBSC cover system to being a lateral diversion
passage after the breakthrough of the upper two layers due to the
7.5 Synopsis of challenges and innovations to containment
presence of the bottom layer. The saturated hydraulic
barriers outside North America and Europe
conductivity of the bottom fine-grained layer should be kept
below 1 × 10-8 m/s to minimize percolation (Ng et al. 2015a). 7.5.1 Regulations and Standards:
By comparing the LFG emissions from three types of covers The dearth of regulations, policies and national guidelines has
under dry weather conditions, including a single clay layer cover, led to the embracing of the South African regulations in much of
a two-layer CBSC and a three-layer CBSC, it is found that the the Sub-Saharan African region or shadowing of standards with
LFG emission rate through the three-layer CBSC was the lowest which the designer is familiar where such obligatory standards
and this rate was lower than the criterion set by the Carbon are not regulated. This gives rise to large variations in design
Farming Initiative (Ng et al. 2015b). A full-scale test site (20 m philosophy which is tempered by the sharing of appropriate
× 12 m × 1.8 m) was constructed and instrumented at Xiaping technology through voluntary organizations, when making use
landfill in Shenzhen city, China, which was located in a humid of electronic media.
climatic region with a humidity index greater than 0.75, to
validate the field performance of a three-layer CBSC using 7.5.2 Variations in contaminant significance
recycled construction waste. One section of the test site was The climatic conditions, population density and other factors
planted with Bermuda grass while the other section was left bare. influence the significance attributed to particular contaminants
Results obtained from about 1.2-year monitoring showed that the and associated containment technologies with particular
accumulated percolation was about 27 mm and 20 mm for the aberrations being the food content of MSW in Asia, mine waste
bare and grass covered landfill cover respectively, both of which disposal at high altitude sites having steep and rugged founding
meet the recommended criterion of 30 mm/year (Ng et al. 2019). conditions in Latin America, and the limited availability of
The LFG emissions data from the test site have not been reported renewable water resources per capita with associated constraints
in the literature yet, but the recorded high degree of saturation on assimilative capacity of the environment in the Southern
(higher than 80%) at the bottom layer suggests a good Hemisphere in particular. These challenges have been addressed
performance of the cover serving as a LFG gas emissions by adopting first-world technology and developing region-
controller (Ng et al. 2015d). specific theory and practice for municipal, mining, and industrial
contaminants. These aspects include addressing the higher-than-
243
normal leachate and gas generation from MSW in humid parts of the hydration or rehydration of adjacent clay components, the
Asia, the development and implementation of regulations and removal of VOCs, and/or the removal of heat so as to extend
standards modelled on North American and European practices service life and enhance performance. The efficiency of this
with adaptations to accommodate domestic socio-economic and unique system is influenced by multiple factors but notably the
climatic conditions, and the pursuance of appropriate technology type of fluid drawn through the system and retention time within
here and elsewhere through the support of regulators and the cell.
voluntary organizations. The reduction and reuse of waste by
recycling of polyester bottles to form continuous filament
nonwoven geotextiles used as filters or cushion layers to
geomembranes in waste facilities requiring protection from a 8 MINING: A SOUTH AMERICAN PERSPECTIVE
relatively short polluting period due to the putrescible nature of
the particular waste is a novel advance of containment barrier Geomembranes used as liners for heap leach facilities (Figure
performance and economy, provided the performance of the 8.1) are probably the single biggest geosynthetics use in the
geotextile for the type of load and adjacent materials are mining industry in South America (Fourie et al. 2010). Fresh ore
confirmed as limiting tensile strain in the geomembrane to a total is transported directly from the pit, or crushed to a specific size,
of less than 3%. and stacked over the geomembrane before leaching with acidic
or basic solutions. During the long leaching process, which could
7.5.3 Containment barrier design and construction quality take many years, metals are dissolved and the leachate (pregnant
assurance solution) is transported by gravity to the HDPE drainage pipes
The deviation from norms and standards in regulations or above the geomembrane and conveyed to solution ponds. The
standard specifications occurs from time to time, even for pregnant solution is pumped from the ponds to the process plant
materials having ISO 9000 certification; however, the for the recovery of metals such as gold, silver, copper, and nickel.
development and implementation of domestic techniques to Uranium is also a metal that is obtained by heap leaching but this
confirm the performance of materials and construction quality has still not been exploited in South America. The biggest heap
leads to rectification of deviations or omissions in many cases. leach pad in the world is the ROM Sulfide Leach project of
Indicator tests have a limited value in assessing the consistency Escondida mine located in the Atacama Desert in Chile, where
of a particular material. Hence, a suite of site-specific materials currently more than 1,000 ha of LLDPE geomembrane has been
and techniques are used to assure the quality of the barrier used to stack 5,000 Mt of ore (= average 5M t/ha =500 t/m2 or
system. Such techniques include pre and post construction >200 m3/m2) and additional expansion is planned.
confirmation of “swell index” of the bentonite in a GCL when HDPE geomembrane is usually used for the process and
exposed to site hydration media, groundwater or of leachate; the emergency ponds. These are essential facilities for a heap
use of real time electronic data tracking of geomembrane leaching project. Process ponds use double liner systems with a
seaming or welding parameters which are uploaded onto a cloud geonet or geocomposite between the two geomembranes for leak
based system which can be viewed at any time; the in-situ detection and recovery system. Emergency ponds are lined with
confirmation of geomembrane strain limitation in a a single liner.
geomembrane by fitting the laboratory approach to a field test Geomembrane liners are now also being used more frequently
which allows for the influences of site aggregate shape and in tailings storage facilities (Figure 8.2) to protect the
placement method, CCL and foundation conditions, and elevated environment from polluting elements (tailings with acid drainage
temperature with longer loading periods to overcome in part the or metal leaching), as well as to comply with environmental and
concerns of creep and material softening; and the evaluation of social commitments and to obtain operating permits.
geotextile filter performance by test methods which recognise Geomembranes are also used extensively in the potassium
the influence of a build- up in fine fraction from base soil or and lithium extraction industry from brine reservoirs in Chile,
waste (due to that being the fraction which controls Argentina, and Bolivia, where a series of very large evaporation
permeability). ponds, lined with HDPE or PVC geomembrane, contain salt-rich
7.5.4 Operational phase influence on barrier performance water and where evaporation occurs over months. When the
The operational phase influences of waste stoichiometry and lithium chloride in the evaporation ponds reaches an optimum
biology can lead to chemical or biological clogging of filters and concentration, the solution is pumped to a recovery plant. The
drains to barrier systems, and/or the development of elevated Salar de Atacama in Chile is home to the largest PVC
temperatures in gas and leachate adjacent to liner components geomembrane installation globally, with more than 150 ha
with commensurate decrease in liner performance due to utilized in mining operations since 1996 (Berube et al. 2007).
polymer degradation, diffusion rate increases and moisture loss In situations where mine waste rock has the potential for acid
of clay or geosynthetic elements which in turn reduces thermal drainage or metal leaching, geosynthetics are used to minimize
conductivity and aggravates the rate of barrier performance the generation of acid drainage and comply with environmental
degradation still further. The influence of chemical precipitate restrictions.
and biological clogging in drainage systems designed to operate Geosynthetic clay liners (GCL) are used in mining as
under gravity and atmospheric pressure is partially mitigated by hydraulic barriers in composite liner systems for heap leach pads,
the incorporation of a “P-trap” on the outlet pipes of leachate tailings dams, waste rock dumps and ponds. In some cases, the
collection systems in gold mines, which reduces air intrusion into GCL is used instead of compacted clayey soil.
drainage systems and hence reduces the rate of precipitate Geotextiles are used extensively in the mining industry to
formation. Similarly, the removal of the outlet P-traps to construct drainage systems or solution collection systems. Since
suddenly change drains from anaerobic to aerobic conditions is several cases of clogging of geotextiles have been observed over
employed to reduce the build-up in biological growth of bacteria the years, the geotextile must be used with care, previously
which flourish in a habitat having fixed environmental verifying its compatibility as a filter of the soils surrounding the
conditions. The reduction in containment barrier performance drainage systems. When used in the collection system of a heap
and service life due to influences of elevated temperature on total leach pad, in addition to verifying its filtering characteristics,
solute transport is mitigated by the development of novel consideration must be given to the potential for clogging due to
drainage systems adjacent to liners to facilitate the application of chemical precipitates, especially when dealing with acid
a negative pressure to draw fluid through the permeable drainage leaching. Geotextiles are also used as protection of the
layer, which has allowed for options of introducing moisture for geomembrane to avoid puncturing.
244
Geonets, a synthetic drainage material manufactured of 8.3 Some technical considerations for the use of GCL
HDPE resin to transmit fluids (and gases), and geocomposite
drains (typically a geonet with a geotextile on one or both sides There are several technical considerations commonly associated
to act as a separator/filter) are used as a leak detection layer in with the use of a GCL:
double-lined ponds. Also, in South America, geocomposites are GCLs may be used on steep slopes subject to verification of
used to protect geomembranes and GCLs from damage when heap leach pad stability.
installed on irregular/rocky slopes in heap leach pads. The effect of hydration of the GCL on shear strength must
HDPE pipes are used extensively for underdrain systems to be based on project specific laboratory testing. During heap
collect underground water below heap leach pads, waste rock leach pad operation, the GCL is likely to be hydrated by
storage facilities, tailings storage facilities, solution ponds, and solution leakage through the liner system or by underground
emergency ponds. Underdrain systems may also be used as a leak water, making the shear strength even lower.
detection system below mining facilities. In general, Renken et al. (2005) recommend the use of
needle-punched GCLs with a nonwoven geotextile on top
and a scrim-reinforced nonwoven geotextile on the bottom
on slopes steeper than 3H:1V. The experience in several
heap leach pad projects suggests that this recommendation
is appropriate.
Minimum overlap of the 25 cm in flat areas and 45 cm on
slopes to reduce the risk of GCL separation by shrinkage as
shown in Figure 8.4 (see also §5.5.2.2).
Figure 8.1: Heap leach pad (gold) in South America 8.4 Geosynthetics design and installation in difficult and
steep slopes
For heap leach pad facilities with very steep slopes, extra cuts
are required to achieve the slope (usually less than 2.5H:1V or
2H:1V) so that it is possible to place and compact low
permeability soil and facilitate liner installation. The slope will
be lined with low permeability soil or GCL before a
geomembrane is put in place, subject to a check that there is
adequate stability.
Very steep rocky surfaces are often shaped by blasting,
leaving irregular surfaces with sharp edges or cavities that could
Figure 8.2: Tailings storage facility in South America
damage/puncture the GCL and geomembrane during operations
(Figure 8.5). Figure 8.6 shows some techniques used on irregular
8.1 Geosynthetics in difficult terrain slopes and final surfaces to protect the liner system.
The topography of much of South America (e.g., Argentina, 8.5 GCL design and installation
Bolivia, Colombia, Ecuador, Peru, and parts of Mexico and
Chile), is very unfavourable for the design and construction of Special attention is required in areas where the final slope after
mining facilities that involve the use of geosynthetics. Mining grading is steeper than 2H:1V (or 2.5H:1V in some places). A
projects in the Andes region typically operate at altitudes of more scrim reinforced nonwoven needle-punched GCL with the upper
than 3,000 meters above sea level with moderate to high edge buried in an anchor trench and the roll deployed down the
precipitation and in an area of high seismicity. Usually, the only slope is commonly used. The slope and needle-punching
place available for constructing the mining facilities is a valley. should be such as to avoid significant bentonite migration under
The topography, the environment, and the scale of the projects gravity. The GCL should only be installed in areas of the leach
invariably necessitate design criteria different to those used in pad where overall stability is not compromised. These zones
flat areas or at lower altitudes (César et al. 2013). In areas of must be identified as part of the design process and verified
difficult terrain special consideration must be given to the during construction. Figure 8.7 shows the installation of
optimization of earthworks, use of soil liner or GCL together geocomposite for protection (black) and GCL (white) on an
with the geomembrane to form a composite liner or a secondary almost vertical slope, with the installers hung on the slope with
liner, geosynthetic installation, geosynthetic protection, etc. strict safety measures. It is essential that the installers are
experienced and knowledgeable to ensure installation quality
8.2 Selection of soil liner or GCL and safety.
A decision as to whether to use low permeability clayey soil liner 8.6 Geomembrane selection
or GCL will involve consideration of the slope upon which the
liner is to be placed. Generally speaking, A soil liner, a GCL, or a zoned combination of both will be
placed as a second containment layer on a prepared subgrade.
Soil liner may be placed on slopes up to 2.5H:1V or even
The first and main containment will be the geomembrane liner,
2H:1V. which will be installed over the soil liner or GCL.
Steeper slopes must be lined using a GCL. Single side textured linear low density polyethylene (LLDPE)
Some projects in South America have been developed where geomembranes are currently popular for use in heap leach pads
there are no borrow sources suitable for a soil liner close to the (Parra et al. 2014). Typically, the textured side is placed in
project. Under these circumstances, a GCL may be used on contact with the soil liner or GCL to improve the shear strength
slopes flatter than 2.5H:1V (e.g., as part of the barrier system for of this interface. High density polyethylene (HDPE) is still
heap leach pad as shown in Figure 8.3). extensively used in single or double lined ponds. PVC has also
been used in the past mainly in interlift liner applications in
copper heap leach projects in Chile. Figure 8.8 shows a
245
geosynthetic installation in difficult terrain in a valley heap leach
pad located in the Andes.
Figure 8.3: Optimized area for GCL utilization
Figure 8.6: Typical protection with heavy geotextile (upper) and
geocomposite (lower) for geosynthetics installation in difficult slopes
Figure 8.4: Separation by shrinkage in two panels of GCL installed
Figure 8.7: Installation of geocomposite for protection (black) and GCL
(white) on an almost vertical slope
8.7 Geomembrane integrity test (puncture test)
The effect of the ore heap weight on the liner system on the
Figure 8.5: Very steep rocky slope with irregular shape after blasting integrity (e.g., punctures) is assessed by means of laboratory tests.
The tests involve first placing the subgrade material (typically a
Although the geomembrane thickness determination is soil subgrade or GCL) inside a square or cylindrical test device.
typically based on puncture tests, César et al. (2013) presented a A sample of the geomembrane proposed for the specific project
practical rule for thickness selection based on the design and is then placed over a subgrade material and covered with either
construction experience of dozens of these kinds of facilities: an overdrain or ore. The entire system is gradually loaded to
1.5 mm for heaps up to 100 m, or reach a pressure equivalent to the proposed height of the heap
2.0 mm for higher heaps. plus a factor of safety. The maximum load is held for 48 hours.
The geomembrane sample is exhumed and subjectively
If the heap leach pad is very high, then a geotextile protection inspected to assess whether it has experienced minor, moderate,
is better and cheaper than a thicker geomembrane (2.5 mm); or severe yielding or puncture. A vacuum box is used to
however, the final thickness to be used will depend on laboratory physically confirm whether a hole has occurred. Figure 8.9
puncture testing. The focus here should be on minimizing short- shows a puncture testing device and geomembrane samples after
term puncture and long-term stress cracking due to excessive testing. This specimen shows evidence of pronounced yielding
tensile strains, which is not generally considered yet. and holes. In South America, puncture testing has been carried
out up to 4,400 kPa (about a 250 m heap).
246
shows the geotextile already placed and partially covered with
the overliner and the ore of the heap.
The practice described above to limit punctures addresses the
very short-term issue of puncture under static vertical load with
the immediate objective of minimizing the loss of pregnant
liquor to the environment. It presumes that a test with only
vertical loading of the geomembrane and protection layer is
indeed adequate for assessing the likelihood of punctures.
Control studies of the potential for puncture and strain in a
geomembrane have shown that the potential for puncture is
highly dependent on both the grain size distribution of the
overliner material and the compressibility of the underliner
(Rowe et al. 2013a; Brachman et al. 2014). Tests that do not
adequately represent these two conditions will not provide a
good indication of the likelihood of puncture. Also, with respect
to the difference between high-quality LLDPE and HDPE, it has
Figure 8.8: Complex geomembrane and GCL liner installation on very been found that for geomembranes of the same thickness there is
difficult terrain a little difference in the puncturing (Rowe et al. 2013a) when
tested under identical conditions. Consideration is only given to
vertical load. However, holes can develop in the geomembrane
at much lower vertical loads if the interface is also subject to
shear (Abdelaal and Solanki 2021).
At present, the focus on geomembrane selection in South
America is on minimizing puncture. While minimizing puncture
is essential, it is not sufficient for providing long-term
environmental protection. If the geomembrane is to provide
environmental protection in the long-term, there must also be a
limit on tensile strains that will result in stress cracking and
potential environmental impact over time. This applies to both
HDPE and LLDPE. Stress crack tests are generally not
performed for LLDPE because either there is no yield strength
or, if there is, the SCRo is typically more than 30,000 hours and
a single test can take more than 3½ years. However, this does not
mean that stress cracking is not an issue. With ageing, once
thermo-oxidative degradation commences the LLDPE stress
crack resistance and tensile break strength decrease very rapidly
and the LLDPE can become very brittle (Figure 8.12). While not
a current priority, there is growing concern about environmental
Figure 8.9: Puncture testing device and damaged geomembrane impacts of mining and the future failure of these geomembranes
due to stress cracking and this is likely to become an issue as
8.8 Geomembrane protection more people become aware of the potential consequences of the
For very high heap leach pads, holes are possible even in a 2 mm excessive strain developed in LLDPE liner in heap leach pads.
LLDPE geomembrane. In this case, there are two options Once the LLDPE becomes brittle, this excessive strain will cause
available to the design engineer to improve geomembrane extensive stress cracking and the liner will have reached its
behaviour: service life. As with HDPE, there can be a substantial difference
in the performance of different LLDPE’s in terms of service life.
Increase the geomembrane thickness to 2.5 mm, or
Use a geotextile to protect the geomembrane. 8.9 Bituminous geomembrane
In both cases, the area to improve will be that in which
damage (severe puncture or holes) is expected to occur. The first Bituminous geomembranes (BGM; Figure 8.13) have also been
option is a little complicated to carry out because the installation used as a liner in heap leach pads in South America (Lazaro and
must deal with two different geomembrane thicknesses. In Breul 2014) but to a lesser extent. This geomembrane has some
addition, the second option is preferred because geotextile is easy advantages over HDPE and LLDPE but has its own share of
to install and much cheaper than a thicker geomembrane. Tests challenges and large-scale failures in mining applications (e.g.,
have shown that placing a geotextile over the geomembrane liner Addis 2016), particularly in terms of creep failure of seams and
reduces yielding and in this case, it is inferred (often incorrectly) performance in hot and cold climates. The interested reader is
that it's use achieves the design objective of adequate referred to §3 for a fuller discussion of BGMs.
geomembrane protection. Based on this experience, Parra et al.
(2014) recommend the following for preliminary design: 8.10 Solution collection pipes
Use at least 270 g/m2 nonwoven geotextile or heavier if the
Robust drainage systems to collect leach solutions are required
heap is very high. to improve mineral recovery, reduce leaks through the liner
Geomembrane protection must be provided in zones where system, improve slope stability, and reduce liquefaction potential.
the heap height exceeds 120 m or 130 m. ROM or crushed ore heap leach pads are irrigated with solvent
However, suitability of the proposed geotextile for these heights solution (dilute alkaline cyanide for precious metals and dilute
must be verified by testing. sulfuric acid for base metals). This solution, along with any
Figure 8.10 shows an example of a 160 m height heap leach stormwater, snowmelt, and cumulative season surplus water is
pad and the area in the pad base (blue hatching zone) where the collected at the base of the leach pad through the solution
2 mm-thick LLDPE single-side-textured geomembrane was collection system which typically consists of dual wall HDPE
protected using a 270 g/m2 nonwoven geotextile. Figure 8.11 perforated pipes (smooth inner and corrugated exterior).
247
Solution collection system design in difficult terrain is
governed by the following considerations:
Collection pipes are only used in areas where the slope is
less than 4H:1V since collection will not be efficient on
steeper slopes.
It is generally good practice to use intermediate benches for
evacuating solutions from subsequent phases directly into
the process pond, or even to a process plant if possible.
Figure 8.14 shows a plan view of a solution collection system in
very steep/difficult terrain with intermediate benches.
The behaviour of the collection pipes and their durability,
when subjected to the loads of the heap, depends not only on the
pipe stiffness but, more importantly, on the surrounding soil.
Generally, the larger the pipe diameter the lower its stiffness.
Thus, very large diameter pipes are generally not used because
of the risk associated with large deflections, buckling, joint
separation, and collapse that would decrease the pipe flow
capacity and the efficiency of the collector system. The required
Figure 8.10: Very high heap leach pad and area where the geomembrane flow capacity is achieved by increasing the number of pipes
requires protection rather than their diameter. It is important to have adequate
collection capacity to reduce the risk of an increase in the
solution level inside the heap leach pad, which would generate a
potential static instability, saturated ore liquefaction, and seismic
instability; the last condition is not very common but is possible
(Castillo et al. 2005).
8.11 Pipes under high loads
Underdrain HDPE pipes beneath heap leach pads, tailings
impoundments and waste rock facilities are sometimes exposed
Figure 8.11: Nonwoven geotextile placed for geomembrane protection to high loads because of the material weight above them (ore,
tailings or waste rock), the thickness of which may exceed 180 -
200 m. Figure 8.15 shows a solution collection pipe from a heap
leach pad facility that has undergone over 20 percent deflection
but is still within the design deflection criteria for the collection
and conveyance of leach solutions.
Current, design practice for secondary collection systems is
to use 100 mm dual wall HDPE pipe based on test results (Smith
2004). Primary collection pipes are at least of 300 mm and
commonly 450 mm diameter. These pipes are less stiff than the
smaller secondary collection pipes. Usually, the HDPE pipes are
placed directly on the geomembrane, however, when the pipes
are used in heap leach pads with a thickness of 120 m or greater,
Figure 8.12: Variation in geomembrane normalized break strength with the larger diameter primary collector pipes may be placed into
time after ageing in high pH heap leach solution at 85 oC (Abdelaal and trenches that will facilitate compaction of the drainage gravel
Rowe 2022) around the pipe (Figure 8.16; Parra et al. 2014) to allow arching
in the gravel when the pipe deforms, thereby minimizing load
transfer into the pipe.
8.12 Geomembrane overstressing
Based on finite element models run and calibrated to both small-
and large-scale laboratory tests, Leduc and Smith (2004)
reported a significant reduction in the load (as compared to the
overburden stress) immediately below the pipe. But the load
increased to about 125% of overburden at a distance of one pipe
diameter, and the zone of overstress extended to about 4 pipe
diameters. These results mean that there are narrow strips of
geomembrane on either side of each collection pipe that are
Figure 8.13: Construction equipment on bituminous geomembrane receiving a higher stress than commonly considered. The authors
note that the overstress will vary depending on the pipe stiffness,
The pipe diameter depends on the catchment area, irrigation rate, the type and the degree of compaction of the material
and the pipe slope. Bigger diameters are used for the main surrounding it, and the height of the heap. This is confirmed by
headers and primary collection pipes. Diameters up to 600 mm the numerical modelling performed by Castillo (2005) for the
have been used in the industry, but in general the design of the case shown in Figure 8.16, which indicated that for a single pipe
primary collection system uses pipes up to 450 mm because the vertical stress on the liner close to the pipe increases in about
bigger pipes will show more deflection, buckling, and potential 14% for 130 m heap and 13% for a 180 m heap.
for collapse. On the other hand, laterals or secondary (or tertiary)
pipes are smaller. These are usually 100 mm, although in
dynamic leach pads, 50 mm pipes are used.
248
Figure 8.17. Primary collector pipe installation with geotextile for
Figure 8.14: Solution collection system in very difficult terrain geomembrane protection
8.13 Raincoats
Raincoat terminology is used in the mining heap leach industry
in reference to a temporary exposed geomembrane cover that
minimizes rain stormwater infiltration into the stacked ore and
diverts cover surface storm runoff to a dedicated pond or natural
drainage (Breitenbach and Smith 2007).
Some gold and silver mining operations, which use heap
leaching technology, located in rainy regions use raincoats on the
top of the ore heap to reduce the amount of rainwater that enters
the heap. The raincoats are used in inactive areas of the ore heap
still under irrigation. Various experiences of heap leach pads on
an industrial scale indicate that the entrance of rainwater into the
system causes solution dilution, making metal recovery less
efficient. It also produces surplus contaminated water that
requires recirculation or treatment before it can be discharged
into the environment, causing significant additional costs. Figure
Figure 8.15. Interior of a solution collection pipe below a heap leach pad 8.18 shows a heap leach pad in Peru using raincoat system.
In addition, raincoats can be used in dry regions, where water
Based on the findings cited above, geomembrane protection
inside the trench using a nonwoven geotextile is also needed, in is scarce, to avoid evaporation, thus reducing the water demand.
a similar way to that suggested for areas in the heap where the
loads are very large. In Figure 8.16 the geotextile used for
geomembrane protection is shown as well as the secondary pipes
connecting to the primary pipe. Figure 8.17 shows the pipe
installation in trenches with geotextile for geomembrane
protection. As noted earlier, large strains can be anticipated in the
geomembrane at these loads with a 250 g/m2 geotextile as a
protection layer (Rowe et al. 2013; Brachman et al. 2014) and
can be expected to significantly influence the service life of the
geomembrane, particularly if there is any shear or generation of
heat from exothermic reactions or during heap leaching, and/or Figure 8.18. Gold heap leach pad in Peru covered almost entirely with
the leach solution has a high pH or contains any surfactant. raincoats.
8.14 Fazenda Brasileiro Mine and Jacobina Yamana Gold
Mine, Brazil
Two field experiments to investigate the performance of HDPE
and PVC geomembranes were conducted at gold tailings dams.
Santos (2014) conducted one study at four gold tailings dams
built between 2006 to 2010 at Fazenda Brasileiro Mining (FBM).
The second experiment was conducted by Sampaio (2015) at
Jacobina dam located in the state of Bahia, Brazil, to verify the
performance of the 1.5 mm thick textured HDPE geomembrane
installed in the dam reservoir and of the nonwoven geotextile
Figure 8.16. Primary collection pipe placed into trench for pipe
protection used as a filter of the drainage system. These studies evaluated
the physical properties of HDPE and polyvinyl chloride PVC
geomembranes intended as barriers for the slopes and
foundations of the tailings dams. The system operation basically
consists of pumping tailings from the processing plant to Lake
II. The underflow is disposed as a pile in Lake II and the
249
overflow is released to Lake III, which also receives the despite having a tendency for increased rigidity. However, both
supernatant of Lake II. Error! Reference source not found.19 geomembranes were considered to have exhibited satisfactory
shows a view of Lake II. properties for the project, even under severe exposure conditions
Table 8.1. Locations of samples exhumation (Santos, 2014).
Figure 8.19. A view of Lake II (FBM) - PVC geomembrane 0.8 mm thick
(Santos 2014).
Note: (1) Only this sample was exhumed from Jacobina mining (JYGM);
Error! Reference source not found.20 shows a view of the others were exhumed from Fazenda Brasileiro Mining (FBM).
Jacobina, Yamana Gold Mining (JYGM) dam. The tailings from
the gold ore processing have damaging environmental
contaminants such as arsenic and cyanide, and they should be
stored appropriately.
Figure 8.21. Exhumation of HDPE samples – FBM Security dike
(Santos, 2016).
The results obtained by MEV showed good mechanical
Figure 8.20. Gold tailings dams – Jacobina Dam (JYGM) – 1st step resistance for HDPE geomembranes. However, PVC
(Sampaio, 2015). geomembranes showed a much greater loss of strength than
HDPE. The PVC geomembrane installed in 2006 showed a
Santos (2014 and 2016) evaluated the physical properties of structure with a totally cracked surface with deep cracks
HDPE and PVC geomembranes with thicknesses of 1.5 and 0.8 indicating a compromise of its waterproof barrier property.
mm, respectively, that were installed for four years, from 2006
to 2010, by Fazenda Brasileiro Mining (FBM). Tests were 8.15 Maracás Menchen Mine
carried out on four types of samples: virgin samples (GMV),
samples aged in the laboratory (GMD), samples aged in a The Maracás Menchen Mine property totals 17,690 hectares
weathering chamber (GMC) and samples exhumed from the and is located in the eastern region of Bahia State, Brazil. The
dams (GMEX) after 18 months. The JYGM geomembrane No 4 (BNM-04) non-magnetic tailings deposit basin of
samples were tested in an ageing chamber under the same field vanadium ore was built in three stages, with an estimated final
conditions (immersed in cyanide fluid from the dam, temperature storage capacity of 1,047,532 m³ of non-magnetic tailings.
of 40oC and subjected to ultraviolet radiation) for a period of 6 Vanadium is an essential ore used for the processing of steel for
years (Gardoni and Morais 2019). This method made it possible aerospace, oil and gas industries, and for the production of
to compare the durability of aged/exhumed samples to that of surgical materials. The Maracás Menchen Mine has one of the
virgin samples and to assess the geomembrane’s degree of highest-grade vanadium resources in the world (1.34%) and is
ageing and durability. one of the lowest cost primary producers of vanadium in the
Exhumed geomembrane samples were extracted from the world (Botelho 2019). Figure 8.23 shows a view of Maracas’
tailings lake and security dikes which were part of the dam dam.
structure. Exhumed sample details are given in Table 8.1 and The barrier system on the bottom and slopes of this basin
sampling at one location is shown in Figure 8.2 comprised four layers of geosynthetics (from top to bottom),
All samples were subjected to tensile tests, thermo- namely: a 1.5 mm HDPE geomembrane, a 5.5 mm thick
gravimetric analysis, Oxidative Induction Time (OIT) tests, polyethylene geonet leak detection layer, a second 1.5 mm
Scanning Electron Microscopy (MEV), and Atomic Force HDPE, and 400 g/m2 nonwoven polyester geotextile on a
Microscopy (AFM). Figure 8.22 shows some MEV results. The compacted soil layer to protect the geomembrane (Figure 8.24).
following mechanical properties were evaluated as a possible
way to measure degradation: tensile strength, deformation, and 8.16 Behaviour of geotextile filters
stiffness. The results showed that geomembranes have 8.16.1 Performance of geotextile filters under severe conditions
differentiated behaviour over time: PVC geomembranes become
more rigid whereas HDPE exhibited a more stable behaviour,
250
Geotextile filters have been extensively used in mining, with the systems tailings FA-G1 and FA-G3 for a system gradient
commonly under severe conditions where mechanical damages, (isystem, total hydraulic gradient along the system height) equal to
in addition to clogging potential, may compromise the filter 1. The results show low values of gradient ratio (GR), rather
performance. Figure 8.25 (Beirigo 2005; Palmeira et al. 2010) constant with normal stress, indicating piping mechanisms. This
shows examples of geotextile filter applications in the Germano behaviour may be attributed to the potential internal instability
and Fosfertil tailings dams, in Brazil. of the tailings. The same behaviour can be noted in the test with
geotextile G3, which has a smaller filtration opening size.
However, the holes left by the needle-punching process were
clearly visible in that geotextile, which may have influenced the
results. Under field conditions it is likely that these needle holes
will have their sizes reduced due to distortion of the geotextile
during deposition of the tailings and increasing vertical stress.
The tailings-geotextile systems investigated remained stable
throughout the durations of the tests.
(a) Cross section of the dam
Figure 8.22. Scanning electron microscopy results (Gardoni and Morais,
2019).
(b) Double barriers construction
Figure 8.24. Double layer barrier for upstream waterproofing at
Maracas’ dam.
The variations of GR with normal stress obtained in tests with
tailings SA and geotextiles G1, G2 and G3 are shown in Figure
8.27 (for isystem =1). In this case, the GR values were very close
to unity and rather constant or slightly increasing with normal
stress. The same pattern of variation was observed for the system
SL-G1, as shown in Figure 8.28.
Figure 8.23. A view of vanadium Maracás dam.
Table 8.2. Properties of the tailings tested.
Results of tailings-geotextile compatibility tests are reported Property(1) FA SA SL
by Palmeira et al. (2010) using samples of tailings collected in D15 (mm)(1) 0.066 0.033 0.007
the Germano and Fosfertil dams (Figure 8.25). The main D50 (mm) 0.128 0.065 0.041
properties of the tailings tested are shown in Table 8.2. Tailings D85 (mm) 0.251 0.105 0.099
FA was collected in the Fosfertil dam, whereas tailings SA and Cu 3.7 2.6 9.2
Cc 0.9 1.0 0.8
SL were collected in the Cava do Germano dam and in the drying % < 0.074mm 29 66 72
bays of the Germano dam, respectively. The properties of the 3
s (g/cm ) 3.46 2.97 3.12
geotextiles tested are listed in Table 8.3. The performance of 3 wL (%) NP(2) NP 14
nonwoven, needle-punched geotextiles, manufactured from wP (%) NP NP 7
polyester was investigated, with masses per unit area varying PI (%) NP NP 7
between 200 and 627 g/m2 and filtration opening sizes varying emax 1.13 0.93 0.91
between 0.06 and 0.23 mm. A gradient ratio apparatus capable of emin 0.67 0.46 0.50
performing tests with vertical stresses of up to 2000 kPa was Notes: (1) Dn – particle diameter for which n% of the remaining particles
have smaller diameters, Cu = coefficient of uniformity, Cc = coefficient
used.
of curvature, wL = liquid limit, wP = plastic limit, PI = plasticity index,
Figure 8.26 shows the results of gradient ratio tests (as per emax = maximum void ratio, emin = minimum void ratio, (2)NP = non plastic.
ASTM D5101) versus normal stress on the soil sample obtained
251
2.0
Geotextile filter
FA-G1
1.5
Gradient ratio
Initial dyke FA-G3
Tailings 1.0
1.5
1 1.5
1 0.5
Compacted fill
0
0 400 800 1200 1600 2000
Drainage blanket
Normal stress (kPa)
(a) Drainage system-Germano tailings dam.
Figure 8.26. GR vs. normal stress for tests with tailings FA.
22m
Vertical drain 3
Gradient ratio
SA-G1 SA-G2 SA-G3
Geotextile filter 2
1
Fine tailings 0
Sandy tailings 0
0 400 800 1200 1600 2000
Drainage pipe
(b) Drying bays of the Germano tailings dam. Normal stress (kPa)
Figure 8.27. GR vs. normal stress for tests with tailings SA.
Germano dam Drying bays 2.0
SL-G1
Gradient ratio 1.5
1.0
0.5
(c) Drying bays – Germano tailings dam. 0
0 400 800 1200 1600 2000
Fine tailings Normal stress (kPa)
82m Coarse tailings Figure 8.28. GR vs. normal stress for tests with tailings SL.
2.0
FA-G1 SA-G2 SL-G2
Geotextile filter
1.5
Gradient ratio
Rockfill
1.0
0
0.5
Initial compacted Drainage blanket
soil dyke 0.0
0 2 4 6 8 10
(d) Drainage system -Fosfertil tailings dam. i system
Figure 8.25. Geotextile filters in Germano and Fosfertil tailings dams
Figure 8.29. Variation of GR with system hydraulic gradient.
(modified from Palmeira et al. 2010).
Table 8.3. Properties of the geotextiles.
Satisfactory performance was also observed for the systems FA-
Property G1 G2 G3 G1, SA-G2 and SL-G2 in terms of values of GR for system
MA (g/m2) 200 400 627 hydraulic gradient varying between 1 and 10 in tests without
tGT (mm) 2.3 3.7 4.5 confinement (no vertical stress on the sample top), as depicted in
n 0.93 0.92 0.90 Figure 8.29. The measured values of GR were close to 1 even for
FOS (mm) 0.13-0.23 0.09-0.16 0.06-0.13 large hydraulic gradients. The same pattern of behaviour can be
kn (m/s) 0.004 0.004 0.004 seen in Figure 8.30, which shows the variation of GR with
(s-1) 1.9 1.1 0.9 vertical stress for a system hydraulic gradient of 10. The values
df (mm) 0.027 0.027 0.027 of GR for the 3 systems tested remained rather constant or
Notes: MA = mass per unit area; n = porosity; tGT = geotextile thickness
decreased with increasing vertical stress, reaching a value
under 2kPa normal stress; FOS = Filtration Opening Size (hydrodynamic
sieving); kn = normal permeability; = permittivity; df = diameter of the
slightly smaller than 1 for the largest vertical stress applied in the
fibres. tests.
252
1.5
FA-G1 SA-G2 SL-G2
Gradient ratio
1.0
0.5
0.0
0 400 800 1200 1600 2000
Normal stress (kPa)
Figure 8.30. Variation of GR with vertical stress for isystem = 10.
Samples of the geotextile filter were exhumed from the
Germano tailings dam. Some images from microscopy are
presented in Figure 8.31, where different levels of impregnation
of the filter voids by tailing particles can be identified. Figure
8.31(a) shows an image of a specimen with open voids and very (a) Low impregnation level.
few particles inside the textile matrix. Figure 8.31(b) shows that
the tailings deposition process or the liquid flow could cause
intrusion of large particles into the geotextile voids. Significant
impregnation of the filter by tailings particles can be observed in
Figure 8.31(c).
The amount of impregnation of the pores of a geotextile filter
can be quantified by the impregnation level ( ), defined as the
ratio between the mass of particles in the geotextile voids per unit
area and the mass of geotextile fibers per unit area (Palmeira et
al. 1996). The entrapment of particles in the geotextile voids can
have a marked effect on some of its properties, such as
permeability coefficient, compressibility, and filtration opening
size (Palmeira et al. 1996; Palmeira and Gardoni 2000; Palmeira
and Trejos Galvis 2017). Figure 8.32 and Table 8.4 show values
of obtained in 9 specimens from a sample of a geotextile filter
with 431 g/m2 mass per unit area which was exhumed from
vertical drains (Figure 8.25b) of the Germano tailings dam
(Palmeira et al. 2010).
Figure 8.32 also presents the values of obtained at the end (b) Large tailings particle in the geotextile void.
of gradient ratio tests carried out in the laboratory. Some values
of in the field ( ranging from 2.0 to 10) were significantly
greater than those obtained in the laboratory, indicating the
influence of the tailings disposal. Table 8.4 also presents the
fraction of geotextile open voids (FOV), which can be obtained
as a function of the values of density of geotextile fibers, density
of the particles entrapped in the geotextile, , and geotextile
porosity (n). Values of FOV are presented for different values of
n. Under virgin and unconfined conditions, the porosity of the
geotextile was 0.92 and this corresponded to FOV values ranging
from 62.1 to 92.4%. However, if the geotextile was compressed
to the extent of reaching a value of n equal to 0.6, the results of
FOV obtained suggest that the exhumed sample of geotextile
filter would be completely clogged by the entrapped particles in
some locations, which is consistent with the image in Figure
8.31(c). The variability of values of and FOV in the geotextile
layer is certainly a consequence of the heterogeneity of the (c) High impregnation level.
tailings mass and segregation process during tailings deposition.
Hence, these results suggest the worst conditions expected for Figure 8.31. Different levels of geotextile filter impregnation by tailings
the geotextile filter should be simulated in laboratory particles (Beirigo 2005).
performance tests.
8.16.2 Geosynthetics for the reduction of ore moisture content
Serra de Carajás mine (Figure 8.33), which belongs to the Vale
Mining Company, produces one of the best iron ores in the world,
but it has one undesirable characteristic: high moisture content.
This characteristic is significantly aggravated by the wet sieving
process utilized to obtain the adequate grain size distribution of
the final product.
253
12 collected is appropriately discarded at the port. Numerical
analyses were performed to evaluate the efficacy of the
10 geosynthetic drainage system at the base of ore piles and in the
train wagons (Figure 8.35). For the wagon transport, the
8 numerical analyses examined moisture movement and collection
Lambda
over a period of 12 hours, which was the approximate time for
6 transporting iron ore from the Carajás mine (Pará state) to the
Port of São Luis (Maranhão state). For the iron ore piles, a typical
4 pile cross-section with a height of 20 m and a base width of 56
m was considered. The results showed that after 12 hours the
2 geocomposite for drainage would be able to decrease the ore
water content to 6% (Figure 8.35, Gardoni et al. 2010 and
0 Moreira 2011), with favourable implications for shipping costs.
0 1 2 3 4 5 6 7 8 9
Specimen No.
Exhumed specimens SL-G2, i = 1 to 10
SA-G2 i = 1 Range of values obtained in the
SL-G2, i = 1 laboratory for all tailings and
SA-G2, i = 1 to 10 geotextiles tested
Figure 8.32. Impregnation level in specimens of an exhumed geotextile
filter layer (Palmeira et al. 2010).
Table 8.4. Impregnation levels in geotextile specimens exhumed from
Germano tailings dam (modified from Palmeira et al. 2010).
Specimen No. FOV (%)
n = 0.92 n = 0.75 n = 0.6 Figure 8.33 General view of Carajás mine.
S1 4.6 82.6 33.2 0
S2 4.0 84.8 41.9 0
S3 2.0 92.4 70.9 41.9
S4 6.6 75.0 4.1 0
S5 2.2 91.7 68.0 36.1
S6 3.4 87.1 50.6 1.2
S7 10.0 62.1 0 0
S8 5.6 78.8 18.6 0
S9 2.5 90.5 63.7 27.4
Average 4.5 82.8 39.0 11.8
Standard deviation 2.4 9.7 26.9 17.8
In the mine, the sinter feed typically leaves the sieving
process with a moisture content of 12% and it is transported on
a conveyor belt to the storage yard, where it is deposited on piles
that may reach 20 m in height (Figure 8.34) to be exposed to
drying for 2 to 3 days. The ore water content may cause the
instability of the piles in the storage yard. The ore transportation
by Vale’s railway allows some drainage of the water from the ore
pores through drainage outlets in the wagons. However, piped
ore fines contaminate the railway ballast and the subgrade and Figure 8.34. View of ore piles showing the excess of moisture.
therefore decrease the balast’s life. At the port located in São Luis,
MA, Brazil, the iron ore is stocked in the storage yards and its Additional studies were carried out in the laboratory using the
moisture content can increase because of the constant rains in Gradient Ratio test. Ferreira (2009) investigated the use of
northern Brazil. After shipping, the presence of free water geosynthetics for the reduction of moisture content of sinter feed
complicates the removal of the bottom part of the iron ore from and pellet feed materials from Carajás mine. The objective was
the ship cargo compartments (note the presence of significant to assess the efficiency of drainage layers to be used at the base
free water can lead to dangerous liquefaction during the shipping of the piles and in the train wagons that transported the ore
of ores). This fact results in a high cost of transportation and departing from the mine, as mentioned above. Gradient ratio
unloading at the destination. Final moisture contents ranging tests on unconfined and confined ore-geosynthetics (geotextiles
from 10 to 12% imply the transport of a large volume of water and geocomposites for drainage) systems were carried out as part
with the iron ore, given 35 million tons per year of ore and water of the research programme. The same equipment used by
transported. Palmeira et al. (2010) was employed, with some adaptations to
The use of geosynthetics to enhance the drainage of fluids accommodate the geocomposites specimens at the base of the
from the iron ore (piles and wagons) produced by Carajás mine cell. Tests with the application of vacuum at the base of the
was investigated. Based on this study, a prototype drainage system were also performed to assess the increase in drainage
system was designed consisting of a nonwoven geotextile and a efficiency.
geonet installed on the side walls and bottom of the freight train Table 8.5 presents the main properties of the ore materials
wagons that transported the iron ore to the São Luis port. The tested, which consisted of a pellet feed (code PF) and a sinter
drainage system conducts the water drained from the ore to a feed (code SF). The pellet feed has a value of D85 (diameter for
reservoir installed at the wagon bottom, and then the water which 85% of the remaining particles are smaller than that value)
254
of .095 mm, whereas the sinter feed has a D85 value of 1.70 mm, Table 8.5. Properties of the ore materials.
with dry density of each material equal to 22.6 kN/m3 and 24.1 Property PF SF
kN/m3, respectively. The ore specimens in the GR tests were D85 (mm) 0.095 1.70
prepared with a relative density of 50%. D50 (mm) 0.048 0.18
D10 (mm) 0.0081 0.007
Cu 6.8 44.3
% finer than 0.074 mm 80 30
G (g/cm3) 4.94 4.95
emax 1.40 1.17
emin 0.89 0.85
3
d (kN/m ) 22.6 24.1
k (m/s) 8.4 x 10-7 3.0 x 10-5
Notes: (1) PF = pellet feed, SF = sinter feed; (2) Dn – particle diameter
for which n% of the remaining particles have smaller diameters, C u =
coefficient of uniformity, G = particle density, emax = maximum void ratio,
emin = minimum void ratio, d = dry unit weight, k = permeability
coefficient.
Table 8.6. Properties of the geosynthetics tested.
Property GT1 GT2 GC1 GC2
MA (g/m2)(1) 400 600 950 900
(a) Without geosynthetic drainage system in the wagon. tG (mm) 1.6 2.2 7.8 11.7
kn (m/s) 0.0035 0.0035 0.0011 0.003
(s-1) 2.0 1.5 NA 3.0
(m2/s) NA(2) NA 0.4 NA
FOS (mm) 0.13-0.23 0.11-0.19 0.145 0.16
q (l/s/m) NA NA NA 0.41
Notes: (1) MA = mass per unit area, tG = thickness, kn = permeability
coefficient normal to the plane (ASTM D4491), = permittivity (ASTM
4491), = transmissivity (ASTM D 4716), FOS = Filtration Opening
Size (AFNOR G 38017), q = discharge capacity (for a hydraulic gradient
of 1 and under a vertical stress of 100 kPa); (2) NA = not available or not
applicable.
Figure 8.36 shows results of gradient ratio tests on unconfined
sinter feed-geosynthetic systems in terms of GR versus system
(b) With geosynthetic drainage system in the wagon. hydraulic gradient (isystem). Very low GR values were obtained
for the geotextiles tested. The value of GR was below 0.4 for
GT1 during the entire duration of the test, whereas for GT2 GR
values of approximately 0.5 were measured for isystem greater
than 2. Significantly low values of GR were also observed for
geocomposite GC1. These results indicate severe piping of fines
from the ore through the geotextile filters. A different behaviour
was observed for geocomposite GC2, for which large values of
GR (between 2.5 and 3) were observed up to a isystem of 2 and
values of GR close to 1 for greater gradients. This behaviour may
be a consequence of an initial blinding mechanism of the
(c) Ore pile moisture content diagram without basal drainage. geotextile filter of GC2 for lower values of isystem. With the
increase of the hydraulic gradient, the fine particles accumulated
on the geotextile filter must have been washed out through the
geotextile, yielding to values of GR close to 1.
After a value of system hydraulic gradient of 4 was reached
in unconfined GR tests with sinter feed, the vertical stress on the
system was increased up to a value of 2000 kPa. The variation of
GR with vertical stress in tests with sinter feed and an isystem equal
to 4 is depicted in Figure 8.37. It can be noted that the values of
GR remained quite constant for tests with GT1 and GC1,
whereas it increased for GT2 and GC2. Despite the initial severe
(d) Ore pile moisture content diagram with basal drainage. piping observed in some tests, the systems remained stable until
Figure 8.35. Results of numerical analysis – Ore water content in the the end of the tests. Large entrapped ore particles were also found
wagons and in piles (Moreira 2010). in the voids of the geotextile filter at the end of the tests (Figure
8.38), consistent with what was also observed in geotextile
The properties of the geosynthetics tested are listed in Table samples exhumed from tailings dams, as reported by Gardoni
8.6. Geosynthetics GT1 and GT2 are polyester nonwoven, and Palmeira (2002), Beirigo (2005) and Palmeira et al. (2005
needle-punched, geotextiles with masses per unit area of 400 and 2010).
g/m2 and 600 g/m2, and filtration opening sizes in the ranges The drainage capability of the draining geosynthetics for the
0.13-0.23 mm and 0.11-0.19 mm, respectively, according to the reduction of ore moisture content was also investigated using the
manufacturer’s catalogues. Geosynthetics GC1 and GC2 are gradient ratio test. In these tests, the ore was placed in the testing
geocomposites for drainage consisting of a geomat between cell and the system was subjected to water flow as in the GR tests
nonwoven, needle-punched, geotextile layers in the case of GC1 under vertical stress of 2000 kPa. After the stabilization of the
and of a geonet also between layers of nonwoven, needle- flow rate, the water flow was halted, and the system was allowed
punched, geotextile in the case of GC2. to drain through the base (with and without vacuum application).
255
The moisture content at different elevations of the specimen was
measured after water stopped exiting the system. Despite the
difference in scales, this test may be useful to evaluate the use of
draining geosynthetics to reduce the moisture content of ore piles
in the mining footprint. Tests without and with the application of
vacuum at the drainage layer were carried out to investigate to
what extent the application of vacuum would increase the
dewatering efficiency. Under gravity conditions, the water flows
towards the drainage layer until the capillary forces in the ore
pores halts its movement. Thus, vacuum in the drainage layer
should increase its dewatering action. Figures 8.39 and 8.40
show results of variation of ore moisture content along the ore
specimen height in tests with geocomposites GC1 and GC2 and
sinter feed, with and without the application of vacuum. Under
gravity conditions only (no vacuum) the reduction in moisture
content regarding its initial value under saturated conditions Figure 8.38. Large ore particle in the geotextile filter void – 600 X
magnification (Ferreira 2009).
varied between 18% and 37% for the system with CG1 and
between 21% and 45% for the system with GC2, depending on
the elevation (Figures 8.39 and 8.40) considered. Greater
20
reductions in moisture content along the specimen were caused Moisture content at saturation
Moisture content, w (%)
by the vacuum and ranged from 53% to 67% for GC1 and 53%
to 62% for GC2. If the moisture content reduction without
vacuum is taken as the reference, the application of vacuum
caused reductions between 44% and 48% in the test with GC1 10
and between 31% and 40% in the test with GC2. Despite the
encouraging results, it is acknowledged that the application of
Without vacuum With vacuum
vacuum to the drainage layer under field conditions may not be y
With vacuum (repetition)
simple or economically feasible. The cost-effectiveness of such 0
solution must be analysed on a case-by-case basis. 0 25 50 75 100
Elevation, y (mm)
Figure 8.39. Moisture content profile – Geocomposite GC1/sínter feed.
Figure 8.41 shows the results of tests with vacuum on the
system consisting of pellet feed and geotextile G1. Less
efficiency can be noted in the case of the fine-grained pellet feed
material. Nevertheless, if the moisture content at saturation is
taken as a reference, the application of vacuum caused a
reduction in moisture content between 22% and 33%, depending
on the elevation considered. The lower efficiency in this case
may be a consequence of the characteristics of the ore material
tested and lower drainage capability of the geotextile in
comparison to those of the geocomposites, particularly under a
vertical stress of 2000 kPa.
Figure 8.36. GR versus isystem in tests on unconfined sinter feed- 8.17 Performance of geotextile tubes (geotubes)
geosynthetics systems.
Silva (2017) conducted a study to evaluate the use of woven
geotextile tubes to dewater, confine and stack gold ore tailings.
A stacked geotextile tube structure was built in the Fazenda
2
i system = 4 Brasileiro tailings dam, Bahia, Brazil, owned by Yamana Gold
mining company. Figure 8.42 shows a view of the tailings dam.
1.5 The investigation started by submitting the gold ore tailings to
geotechnical tests. The specific gravity of the tailings particle (G)
obtained was equal to 2.85 and the particles size distribution
GR
1
showed that 75% of the tailings consisted of fine particles. The
woven geotextile used had a tensile strength of 100 kN/m, mass
0.5 per unit area of 554 g/m2, and a thickness of 1.8 mm. Dewatering
efficiency was evaluated by means of jar-tests, geotextile cone
0 tests and geotextile dewatering tests (GDT). From the
0 500 1000 1500 2000 dewatering tests results, it was defined that the flocculant
Vertical stress (kPa) Magnafloc LT27 should be added to the tailings to avoid
excessive losses of fine particles. Once the flocculant dosage was
GT1 GT2 GC1 GC2 defined, the experimental geotextile tubes were assembled in the
Figure 8.37. GR versus vertical stress for isystem = 4. mining footprint (Figures 8.43 and 8.44).
The tests were conducted inside the mining area, on the top
of a tailings pile since there were no scheduled works to take
place in that area in the coming years. Thus, the geotextile tubes
were installed and expected to be monitored for at least two
years. The dimensions of the test area were 7 m x 9 m and a
256
drainage layer was placed on a geomembrane liner installed at
the base of the stack to collect the fluid during the dewatering
process. The test setup was designed for a three-level structure,
being 9 geotextile tubes units in the first level, 4 in the second
level, and 1 in the third level of stacking. Every geotextile tube
unit was 2.0 m wide, 6.0 m long and was designed to have a
maximum height of 1.0 m during the filling process. Figure 8.43
and Figure 8.44 show the geotextile tube experimental setup
(dimensions in meters).
20
Moisture content at saturation
Moisture content, w (%)
Figure 8.43. Geotextile tube experimental setup (Moraes et al. 2020).
10
Without vacuum With vacuum
y
With vacuum (repetition)
0
0 25 50 75 100
Elevation, y (mm)
Figure 8.40. Moisture content profile – Geocomposite GC2/sinter feed.
Moisture content at saturation
20
Moisture content, w (%)
Figure 8.44. Geotextile tubes stacking – Transversal direction
A numerical analysis of the geotextile tubes was carried out
using the computer code Abaqus version 6.14. A parametric
10 analysis was performed to verify the influence of the filling
pressure and the woven geotextile’s tensile stiffness, as well as
which elements of the finite element program used would be
y most representative of the behaviour of the geotextile tube under
field conditions. In addition, this analysis provided insights on
0
0 25 50 75 100 how suitable the computer code is for such simulations.
Elevation, y (mm) Predictions from analytical methods were also obtained. The
Figure 8.41 Moisture content profile – Geotextile G1/pellet feed. analytical methods proposed by Plaut and Suherman (1998) and
Guo (2014) were employed in the research, and the results were
The solids concentration of the tailings confined in the calculated using a numerical algorithm developed by Silva
geotextile tubes was evaluated by taking samples close to the (2019). The results predicted by the analytical methods were
tube inlet port. After the last filling stage, samples were taken consistent with the ones found in the literature, as well as with
after 24h, 48h, 72h, 7 days, 21 days, and 28 days. Figure 8.45 the results obtained in the field study. The numerical modelling
shows the variation of solids concentration with time. The was also able to represent the behaviour of the geotextile tube
greatest increase in solids content occurred in the first 24 hours during its filling. Figure 8.46 shows some images of the field
when the solids concentration increased from 53% to 75%. These experiment and of the numerical modelling.
values correspond to a water content reduction from 100% to 25% It was also found in the research that the correct choice of the
in a day. After the first 24 hours, the desiccation rate decreased finite element to be used in the numerical analysis is a
and the increase in solids content was approximately 10% after determining factor for the simulation of the stress–strain
13 days. Each geotextile tube unit was able to dewater and behaviour of the geotextile tube. The filling pressure has a direct
confine an amount of 5.4 m3 of tailings. influence both on the shape of the tube cross section and on the
stresses generated during the filling process. Despite the
limitations of the analysis, the analytical method and the
numerical modelling were able to represent the geotextile tube
behaviour during filling.
After filling and stacking the geotextile tubes, the structure
remained in the field for 24 months prior to the exhumation of
samples to be examined in the laboratory test program (Moraes,
2020). Figure 8.47 shows the geotextile tubes after two years.
Figure 8.48 presents the extraction of disturbed and undisturbed
samples of consolidated tailings, as well as samples of the
geotextile tube.
After 24 months of sludge consolidation inside the tubes,
initially containing 50% solids, it was found that the apparent dry
Figure 8.42. A view of Fazenda Brasileiro tailings dam (Moraes, 2020)
density was 1.8 t/m3. This value corresponds to a final percentage
of solids of 83%. According to Bittar (2017), the apparent dry
density of tailings from gold ore in the field reaches values
around 1.2 to 1.4 t/m3. Therefore, the result obtained by the
257
consolidation of fine tailings in geotextile tubes proved the the alternative subballast reinforced with geogrid at the base
exceptional efficiency of the process for disposal of fine mining (section S2) and at the top (Section S3) of the layer. Section S6
tailings and its feasibility as safe mining structures. was constructed with the alternative subballast material, but
without reinforcement, for comparison purposes.
Figure 8.45. Tailings solids concentration versus time in the geotextile
tubes (Silva, 2017).
Figure 8.48. Extraction of samples from the geotextile tubes (Morais 20
20).
Figure 8.46. Images of the experiment in the field and of the numerical
modelling.
Figure 8.49. Geosynthetic reinforced wall for a crusher.
The alternative subballast material consisted of a mixture of
50% of a local soil (silty sand), 25% of mining residues (sandy
silt) and 25% of the conventional subballast material. The
alternative subballast was considerably finer (D85 = 0.9 mm, D50
Figure 8.47 Bird-eye view of the drained stacking of geotextile tubes = 0.076 mm, D10 = 0.0012 mm) than the conventional ballast
that remained in the field for 24 months. (D85 = 23.0 mm, D50 = 5.1 mm, D10 = 0.008 mm).
The reinforcement used was a polyester geogrid, with 30 x 30
8.18 Soil reinforcement in mining mm apertures, tensile strength of 65 kN/m, and tensile stiffness
Soil reinforcement using geosynthetics has been employed in at 5% strain of 520 kN/m. Figure 8.51(c) shows one of the
mining plants and infrastructures (roads and railways) associated experimental sections during a Benkelman beam test. Figure
with ore production and transportation. Figure 8.49 shows the 8.52 presents results of Benkelman beam tests from which it can
use of geosynthetic reinforcement in a retaining structure. be seen that the geogrid reinforced sections behaved in a similar
Becker (2006) investigated the behaviour of a 5 m high and 1700 way to the reference section, constructed with the conventional
m long geogrid reinforced wall to increase the height of an subballast material. As expected, the poorest performance was
existing dyke constructed to contain bauxite residues at a mining obtained in the section with the alternative subballast without
plant owned by Alcoa Aluminum, in Poços de Caldas, MG, reinforcement. The presence of the reinforcement layer also
Brazil (Figure 8.50). The wall was constructed with a silty-clay reduced the horizontal strains in the subballast (Figure 8.53). An
soil and geogrids with tensile strengths of 35 and 55 kN/m and additional benefit brought by the reinforcement was less
tensile stiffness values of 700 and 1100 kN/m. Instrumented breakage of the ballast grains, as shown in Figure 8.54. Thus,
sections showed that the structure behaved very well. geosynthetic reinforcement may allow the use of an alternative
Fernandes et al. (2008) studied the use of mining residues as and cheaper subballast material built with mining residues,
an alternative subballast material for the Victoria-Minas railway, providing a better and more environmentally friendly usage of
which is subjected to intense traffic of freight trains transporting mining residues.
ore from mines owned by Vale Mining Company. Figure 8.51(a)
8.19 Current
shows a cross section of the instrumented experimental sections,
which were subjected to the same traffic of trains as the railway. Geosynthetics are versatile and cost-effective materials that can
Figure 8.51(b) shows four instrumented sections, where section serve many functions in a mine footprint. However, due to the
S1 is the reference section, constructed with the traditional characteristics of the mining activity, appropriate selection and
subballast material. Sections S2 and S3 were constructed with specification of geosynthetic products is of utmost importance.
258
1.5 or 2 mm-thick LLDPE and HDPE have found extensive use
in heap leach pads and more recently in tailings storage facilities.
410 410
In both cases it is important to design to minimize 800 500 1080 500 800
holes/punctures in the geomembrane in the short-term. tubo rail
Geotextile protection layers may be needed in large heap leach drainage ditch
ballast sub ballast
pads to achieve that objective. 650 dorment
200
Geomembranes may also be used as “Raincoat” covers over
heap leach pads in rainy regions to minimize dilution of leach 3%
instrumentation box
3%
solutions and reduce water treatment cost. In addition, raincoats
may be used in dry regions to minimize evaporation and reduce Dimensions in mm.
the demand for water.
(a) Cross-section of the experimental sections.
450 ballast
200 C A A A subballast
subgrade
S1 S2 S3 S6
geogrid
A - Alternative subballast
C - Conventional subballast
Dimensions in mm.
(b) Characteristics of the experimental sections.
(a) General view of the bauxite residues reservoir.
lining
1
5
compacted fill 5.0 m
reinforcement
4.2 m
existing dyke
(b) Cross-section of the reinforced structure.
(c) Benkelman beam tests in one of the sections.
Figure 8.51. Use of alternative subballast material from mining residues
in a railway (Fernandes et al. 2008).
Distance(m)
0 1 2 3 4 5 6 7 8 9 10
0
S1 S2
1 S3 S6
Deflection (mm)
3
C A A A
4
S1 S2 S3 S6
(c) Load test on one of the instrumented sections.
Figure 8.52. Results of Benkelman beam tests on the experimental
Figure 8.50. Geogrid reinforced wall to increase the storage capacity of sections (modified from Fernandes et al. 2008).
a mining reservoir (Becker 2006). Geosynthetic reinforcement can improve the performance of
Geotextile filters may be used in critical and severe conditions. soil structures such as walls and unpaved roads in the mining
Performance tests are required under conditions as close as footprint, as well as infrastructures associated with the mining
possible to those expected in the field to assess their suitability business. Geosynthetics can also be used as reinforcement in
for a given application. Comprehensive filter criteria (including future tailings dams and for the increase of height and storage
retention, permeability, anti-clogging, and endurance/durability capacity of existing dams.
criteria) should be used, and due care exercised in geotextile It has been common in Brazil to find thin geomembranes
filter installation. being used as barriers for waterproofing of contaminated tailings
A significant increase in the use of geotextile tubes to dewater dams (gold, vanadium etc.). Thicknesses equal to or smaller than
mining tailings is foreseen in the coming years. Investigations on 1.5 mm for HDPE geomembranes and even less than 1.0 mm for
this type of geotextile application have shown that geotextile PVC geomembranes have been found. These are very thin
tubes can provide significant reductions in moisture content in membranes, and such thicknesses are unacceptable in many
addition to a safer accumulation of tailings in the mine footprint. countries. Therefore, there is an urgent need to improve the
259
specification of the type and thickness of polymeric barriers in PART B: REMEDIAL ENVIRONMENTAL GEOTECHNICS
mining. Specific standards on the use of geosynthetics in mining
activities should be developed.
9 VAPOUR BARRIERS FOR CONSTRUCTION OVER
16000 BROWNFIELD SITES
S6
14000
12000
Brownfield sites are defined here as locations that had been
deserted because of historical contamination prior to remediation.
S1
)
10000
Despite remediation, many of these sites have persistent low
Strain (
strain gauge
8000 levels of residual contamination. Re-development of these sites
6000 S3 can allow substantial and beneficial use of these prime locations.
4000 Engineered polymeric vapour barriers are used in the
2000 C A A construction of buildings over remediated contaminated sites
S1 S3 S6 (brownfield sites, landfills) to minimize moisture, radon, and
0
1.00E+05 6.00E+05 1.10E+06 1.60E+06 2.10E+06 2.60E+06 volatile organic compound (VOC) migration into the building.
Number of axles Vapour barriers have a specified minimum thickness of 0.15 mm
Figure 8.53. Horizontal strains in the subballast layer. (NYSDOH 2006) although the US EPA recommends a higher
0.75 mm minimum thickness to minimize puncture risk (EPA
No. of axles ( x 10 6 ) 2008).
0.3 0.7 1.4 2.1 In many countries, there are limits on the maximum
60 concentrations of residual contaminants in remediated sites.
Benzene, toluene, ethylbenzene, and xylenes are often used as
Particle breakage, Bg (%)
50
C A A A indicator chemicals because they are some of the most common
40 and volatile. The National Institute for Occupational Safety and
S1 S2 S3 S6
30
Health (NIOSH) recommended exposure limits (REL; e.g., in a
building) are: benzene 3.2 mg/m3; toluene 375 mg/m3;
20 ethylbenzene 435 mg/m3; xylenes 435 mg/m3 (CDC
10
2015a,b,c,d,e,f,). An effective barrier for VOC intrusion must
prevent the peak indoor air quality of a building from exceeding
0 the guidelines and minimize the intrusion.
0 100 200 300 400 500 600 700
The mechanisms of migration of VOCs into a building are
Time (days)
advection and diffusion. Advection is driven by differential
S1 S2 S3 S6 pressures between inside and out and will occur through holes in
Figure 8.54. Breakage of ballast grains. the vapour barrier or geomembrane. Diffusion is driven by a
difference in concentration and will be governed by the diffusion,
Some regions in Brazil are subjected to strong winds. To partitioning, and permeation coefficients of the particular VOC,
avoid air pollution by contaminated dust, mining companies and thickness and polymer used for the vapour barrier or
have used nonwoven geotextiles as covers for high ore piles and geomembrane. Many researchers have examined VOC diffusion
fine tailings. However, this solution has not been efficient, since through polyethylene (PE) based materials (Aminabhavi and
the nonwoven blanket is very light, it is commonly displaced or Naik 1998a,b,c, 1999a,b,c,d,e,f; Sangam and Rowe 2001, 2005;
removed by the wind. Thus, new geosynthetic products and Edil 2003; Joo et al. 2004, 2005; McWatters and Rowe 2007,
installation procedures should be investigated and developed for 2009, 2014, 2015; Park et al 2012; Touze-Foltz et al. 2013; Eun
a better performance of geosynthetic covers for ore and tailings et al. 2014). Jones and Rowe (2016) examined the diffusion of
piles. BTEX for a range of common vapour barriers. McWatters and
Regarding geosynthetic installation, qualified personnel and Rowe (2010, 2018) examined the diffusion of BTEX through a
high-quality installation and construction control are required. coextruded LLDPE/EVOH/LLDPE vapour barrier while
The lack of knowledge on the properties and characteristics of DiBattista and Rowe (2020) examined PCE and TCE diffusion
different geosynthetic products may lead to problems such as through a coextruded LLDPE/EVOH/LLDPE vapour barrier.
those caused by the lack of subgrade preparation and high Under steady-state conditions for the mass flux, f, (the mass
temperatures. GCLs and geomembranes on irregular surfaces of contaminant per unit area per unit time) passing through the
and steep slopes must be protected from puncture using geomembrane or vapour barrier of thickness, H, is given by
geocomposites, geotextiles, or other systems. In addition, a Fick’s first law:
suitable criterion must be used to select suitable needle-punched
GCLs for use on slopes greater than 3H:1V. f = c Pg/ H (9.1)
Mining activity is characterized by the existence of high where c is the difference in concentration between the two sides
loads and the presence of chemically and mechanically of the membrane and Pg is the permeation coefficient. The
aggressive materials and liquids. In addition, the consequences relative resistance of the different products to permeation by
of failures can be extremely serious in terms of loss of human VOC’s is expressed in terms of the permittivity ratio, Pg/H. For
life and environmental damage. In this context, researchers and example, the permittivity of 11 different vapour barriers or
the geosynthetics industry must continue to study and develop geomembranes are expressed relative to the permittivity of a 1.5
products capable of meeting the growing challenges imposed by mm thick HDPE geomembrane in Table 9.1.
mining activities and by increasingly restrictive environmental
legislation. The vapour barrier shown at the top of Table 9.1 allows a
benzene flux 15-fold greater than through HDPE while the
coextruded LLDPE/EVOH/LLDPE geomembrane would only
allow 1/5 of the benzene that would permeate through HDPE.
Thus, there is a 75-fold difference between the best and the worst
diffusion barrier given in Table 9.1
260
Table 9.1 Permittivity [Pg/H] of 11 commercially available vapour These are all mature technologies. Cement-bentonite and
barriers/geomembranes relative to that of a 1.5 mm thick HDPE soil-bentonite cut-off walls were widely used for containment of
geomembrane contamination in the last century but currently, they tend to be
Benzene diffusion at 22oC Thickness [Pg/H]
used for contaminated land management only when other
Material H (mm) [Pg/H]HDPE
remediation technologies do not allow destruction of the
contaminants to regulatory acceptable levels, where the
Perminator Vapour Barrier 0.37 15
contaminants cannot be removed due to the presence of buildings,
Viper Vapour Barrier 0.28 14 infrastructure, etc., or as an interim control measure.
Husky Yellow Guard Vapour Barrier 0.36 12 Containment may, in effect, create a landfill which requires long-
Stegowrap Vapour Barrier 0.35 9.3 term management and regulatory supervision.
PVC 1.02 8 Cut-off walls are currently, and have been for decades, very
XR5 8130 (coated polyester) 0.84 3.7
widely used for hydraulic structures, for example, on the major
programme of post Hurricane Katrina upgrading of dams and
Petrogard VI (coated polyester) 0.81 3.6
levees in the USA.
LLDPE 0.79 2.9 The review presented below will focus processes (a) to (d)
CSPE (Hypalon) 1.14 2.2 above though the discussion on cement-bentonite systems is
LLDPE 1.5 1.5 relevant to all cut-offs using these materials. In the review key
HDPE 1.5 1 issues will be presented but in the interest of brevity, they will
not be discussed in detail, rather readers are referred to
LLDPE/EVOH/LLDPE 0.53 0.2
significant references. For example, discussion, analysis and
review of the application of vertical barriers for pollution
Techniques are now readily available that allow the use of
containment is given in Dominijiani et al. (2021); Evans et al.
both the partitioning and diffusion coefficient in modelling the
(2021); Manassero (1994, 1997); Manassero et al. (1995, 1996,
impact of a particular brownfield site (e.g., McWatters et al.
2000); Manassero and Viola (1992). For an overview of cement-
2019). A 1.5 mm thick HDPE/EVOH/HDPE has been effectively
bentonite systems see Jefferis (2012) and for grouts and slurries
used in biopiles to remediate hydrocarbon contaminated soil
more generally see Jefferis (2014).
which was then reused together with a LLDPE/EVOH/LLDPE
vapour barrier and the McWatters et al. (2019) method of
10.2 Early cut-off walls
analysis to allow safe reuse of this remediated material below
tightly sealed buildings constructed at Casey Station in The concept of excavation under a bentonite slurry to form a
Antarctica. continuous structural wall was advanced by Professor Christian
Veder in 1938 (Xanthakos 1997) though the use of fluids to carry
cut soil and support excavations for water and oil well drilling is
much older.
10 VERTICAL BARRIERS – CUT-OFF WALLS The first field trials of a soil-bentonite slurry trench cut-off
wall began in September 1945 under the supervision of Major
10.1 Cut-off wall construction General M.C. Tyler, United States Army (Retired) ‘as the
originator of the basic idea’ (Kramer 1946). A cut-off 20 ft deep
Vertical barriers (cut-off walls) can be formed by many was formed to protect levees on the Mississippi river from
techniques including: erosion and sand boils.
a) excavation under a self-hardening cement-bentonite slurry Cement-bentonite walls formed by what is now known as the
using a backhoe, clam-shell or cutter-miller to create a vibrated beam method were used for the construction of
cement-bentonite cut-off wall (the choice is of excavation watertight banks on the Upper Danube and Moselle in the late
technology will be mainly dependent on the depth of the 1950s (Maillard and Serota, 1963).
excavation and the materials to be excavated for example, a Ryan et al. (2022) gives a history of cut-off walls, mainly soil-
toe-in to rock can require use of a cutter miller). bentonite and from a broadly US perspective. Jefferis (1997)
b) as above with insertion of a geomembrane prior to the set of gives a review of UK practice. Evans and Dawson (1999)
the cement-bentonite slurry. consider the differences between US and UK practice.
c) excavation under a bentonite slurry followed by backfilling There is a very substantial body of research on cut-off walls.
with a soil-bentonite mix to create a soil-bentonite cut-off For cement-bentonite walls, the work of Caron (for example,
wall. Caron 1973) provides some of the earliest comprehensive studies.
d) excavation under a bentonite slurry followed by backfilling The doctoral thesis of Guner (1978) on slurry mixing, the fluid
with a soil-cement-bentonite mix to create a soil-cement- properties, and the stress-strain behaviour blastfurnace slag-
bentonite cut-off wall. cement-bentonite slurries at different slag proportions by weight
e) trench cutting with a trench cutter/trench saw with of total cementitious material is a seminal study (see also Jefferis
simultaneous injection of a cement-bentonite slurry to 1981). Card (1981) provides data on the hydraulic conductivity
produce an in-situ soil-cement-bentonite wall. of these systems (see also Jefferis 1997 but note §11 below
f) deep soil mixing using a vertically mounted mix head fitted regarding the continuing availability of ground granulated
with cutters rotated about horizontal axes with simultaneous blastfurnace slag, ggbs). Other major works on cut-off walls
injection of a cement-bentonite slurry to create a deep soil include those by Evans, Fratalocchi and Manassero and their co-
mixed (DSM) . workers as cited in the references.
g) deep soil mixing using single or multiple augers rotated about
vertical axe(s) with simultaneous injection of a cement- 10.3 Specifications for vertical barriers
bentonite slurry. This approach is more often used for area
treatment rather than wall formation. The first cut-off walls were principally for the control of seepage
h) jet grouting. through and below hydraulic structures such as dams, dykes and
i) vertical barriers formed by concrete diaphragm walls or levees and this continues to be a major application. For such
secant piles. barriers, a typical requirement was a hydraulic conductivity of
j) any of the above barrier systems coupled with treatment cells <1x10-8 m/s with a minimum specified wall thickness typically
to produce funnel and gate systems. of 600 to 750 mm, though the actual thickness was often
261
controlled by the width of the tool required for the trench bentonite mixes). The exceedingly low hydraulic conductivity of
excavation. The use of cut-off walls for contamination migration Impermix demonstrates that although the particle size of the
control ushered in tighter hydraulic conductivity specifications, reacting materials, at the time of reaction, can exert a strong
typically requiring <1x10-9 m/s. Yet lower hydraulic conductivity influence, it is the pore-structure of the final product and not the
requirements were regarded as unnecessary as below 1x10-9 m/s mineral structure of the raw materials that dominates the
diffusion could dominate and there might be little net benefit hydraulic conductivity of the final product. It can be expected
(Manassero and Shackelford 1994). that the attapulgite reacts with and is removed from the mix just
Hydraulic conductivities at and below 1x10-9 m/s are now as bentonite is from slag-cement-bentonite mixes.
routine and the specification of <1x10-8 m/s for hydraulic
structures can be reduced to <1x10-9 m/s. In principle, this should 10.5 Strength and stiffness of the cut-off wall materials
improve the durability of a cut-off, especially for cement-
bentonite systems as permeating water (even of benign chemical There appear to be no generally established criteria for setting
content), over extended periods, can leach the cementitious the required minimum or maximum strengths of a cut-off wall
phases, for example, removing calcium from the principal material (though there can be for specialized applications such
cementitious components, the calcium silicate hydrates. as for cement-bentonite walls used as shear walls for
Further information on specifications can be found in downstream stability of hydraulic structures in the event of
Jessberger et al. (1991); Schluchter et al. (1991, 1993); Millet et liquefaction).
al. (1992); Jefferis et al. (1999). Specifications can be more focused on ensuring that a
coherent low hydraulic conductivity material is achieved than a
10.4 Hydraulic conductivity of cut-off wall materials requirement for a particular strength and as strength is
traditionally used as an index to assess the quality of engineered
As discussed above a typical specification requirement for a cut- materials, minimum unconfined compressive strengths are
off wall may be a hydraulic conductivity <1x10-9 m/s. This frequently required (e.g., 100 kPa at 28 days, Jefferis et al. 1999).
should be readily achievable with soil-bentonites and also for On occasion, it has been suggested that the stiffness of the
cement-bentonite materials at economic total cementitious cut-off material should match to that of the adjacent ground but
contents provided the cementitious material has a high slag this leads to problems in that (a) it is difficult to design cement-
proportion. This hydraulic conductivity may not be achieved at bentonite mixes to achieve particular stress-strain behaviour, and
28 days but is usually achieved by 90 days. Fratalocchi et al. (b) the strength and stiffness of cement-bentonite mixes will
(1996), Fratalocchi and Pasqualini (2007) and Jefferis (2012) increase with time. At what age are stiffnesses to be matched? A
show that a plot of log hydraulic conductivity – log time gives a year might seem appropriate in terms of property development
straight line until at least 1 year. This provides an excellent but can sign-off of the work be delayed so long and what if the
indicator of hydraulic conductivity development which is useful match is not then achieved? Also, under what drainage
in the design phase of a cut-off to avoid long delays before test conditions should stiffness be assessed?
results are available, to avoid tests of long duration, to select the
optimum mix formulation taking into account long term 10.6 Strain at failure
performance and to investigate hydraulic conductivity trends
when undertaking groundwater compatibility testing. Jefferis Some specifications, for example, those typically in the UK used
(2012) finds that this log-log plot tends to very slightly over- to require the set material to have a hydraulic conductivity
predict hydraulic conductivity reduction with time (i.e., it gives <1x10-9 m/s coupled with a strain at failure of greater than 5%
very slightly low values). under low or zero effective confining stress. However, low
Hydraulic conductivity reduction cannot continue without hydraulic conductivity requires either a high Portland cement
limit. It appears to be the result of both continuing hydration of content or a high slag proportion both of which lead to higher
the cementitious phases (Fratalocchi and Pasqualini 2007) and strength. Strain at failure under low confining stress to strength
also reaction with species such as dissolved inorganic carbon ratios tends to be low and indeed poorly predictable under all
(e.g., bicarbonate ion) if present in the permeating water or from conditions. The 5% strain a failure criterion was debated in the
air, if this is used to pressurise the water in a triaxial permeameter consultations held when developing the UK National
cell without an intermediate gas-proof membrane (carbon Specification for pollution migration control (Jefferis et al. 1999).
dioxide coming from the air). After careful consideration, it was not included in the final
In relation to the hydration of the cement phases, it should be version of the specification on the basis that a cement-bentonite
recognized that clays are not stable in the high pH of material with such properties did not exist under undrained
cementitious systems and will react. Jefferis (2008) found that conditions and/or at low confining stress to strength ratios (and
bentonite could not be identified by x-ray diffraction in samples indeed it still does not exist) or if it did/does exist it cannot be
of slag-cement-bentonite systems. Evans et al. (2020) provide a reliably produced over the full extent of a major wall (Jefferis et
detailed analysis of the topic. As the bentonite reacts, it should al. 1999). From this time, for certainty of hydraulic conductivity
be expected that modest increases in the bentonite content of a for cement-bentonite systems, slag proportions of 70 to 80% of
cement-bentonite mix will increase its strength. This sometimes total cementitious material are regularly adopted (previously the
comes as a surprise to those unfamiliar with the underlying slag proportion was lower to give some reduction in hydraulic
reaction processes. It also should be noted that the bentonite conductivity without substantial increase in strength).
particles may act as nucleation sites for the deposition of Evans et al. (2021) give detailed advice for soil-bentonite
hydrates as the cement reacts with water. If the bentonite is not materials.
pre-hydrated or is poorly hydrated so that it is not well dispersed
to colloidal dimensions the resulting mix will be of higher 10.7 Quality control / quality assurance vertical barriers
(poorer) hydraulic conductivity. For all vertical barrier systems quality assurance of raw materials
Tallard (1984a,b, 1992, 1995, 1997a,b, 1997a,b) developed is fundamental. This will require certificates from suppliers and,
a self-hardening slurry based on ggbs and attapulgite. This if appropriate, control tests on materials as they arrive on site.
material (Trade name Impermix) can achieve hydraulic Accurate batching of the solids into the fluids is also
conductivities as low as 10-12 m/s. It is a fascinating material as essential. For bentonite slurries, this is often done by checking
attapulgite has a lath-like particle shape and not the plate-like density using the industry-popular mud balance. However, the
structure of bentonite (which has, on occasion, been wrongly mud balance has a resolution of only about 15 kg/m3 (Jefferis
assumed to provide low hydraulic conductivity in slag-cement- and Lam 2018) which is insufficient as a bentonite slurry is likely
262
to be batched at a bentonite concentration of perhaps 30 to Table 10.1 Tests on fresh bentonite slurry and trench slurry at a
60 kg/m3. If the bentonite concentration in a bentonite slurry is frequency of 2 per shift
to be checked it should be by moisture content determination Subject Standard Type of Test Target Value
with allowance for the moisture in the bentonite powder as Viscosity >36 s
delivered which may be 10% or more. Poor bentonite hydration Fresh
Filtrate Loss <25
prior to Portland cement addition may result in excessive bleed Bentonite
Density >1.02 g/cm3
of the fresh slurry and high k of the set material. Slurry
API 13B-1 pH 6 to 10
Similarly for cement-bentonite slurries, it is important that the Viscosity >36 s
total cementitious content is controlled preferably by moisture Trench
Density >1.02 g/cm3
content determination. Slurry
pH 6 to 10
If slag is being used it is essential that the individual masses
of slag and cement are accurately batched as the slag/total
Table 10.2 Commonly specified targets for soil-bentonite
cementitious ratio will have substantial impact on the properties backfill at frequency of 2 per shift:
of the final product. Note that addition of more than the design
Subject Standard Type of Test Target Value
quantity of Portland cement will reduce the slag/cement ratio and
ASTM C- 100 mm ±
so may impair the properties of the set material. Just as small Slump
additions of Portland cement to ‘accelerate hardening’ will fail – 143 25
high slag mixes have greater strengths than low slag mixes at all SB Backfill API 13B-1 Density >1.6 g/cm3
times (see, for example, Guner 1978). These are messages that ASTM D- Fines
>15%
can be difficult to get accepted. 1140 Content
For soil-cement-bentonite mixes, the cement-bentonite mix
must be batched as discussed above with soil batched to Table 10.3 Laboratory tests on soil-bentonite backfill at a
appropriate accuracy and considering its moisture content. frequency of 1 per shift
Quality control/assurance testing of the final mix for both Subject Standard Type of Test Target Value
cement-bentonite and soil-bentonite is usually by taking samples ASTM D- Hydraulic
<1x10-9 m/s
of the fluid/mix as delivered to the excavation and at various 5084 conductivity
SB Backfill
depths within it. Tests on-site will include tests for the fluid ASTM D- Fines
>15%
properties. Samples also will be taken for off-site laboratory 1140 Content
testing. These should be stored in sealed tubes under appropriate
temperature conditions and protected from shock, vibration etc. The backfill is essentially 100% saturated and therefore the
(taking particular care during transport of samples from site to measured density of the backfill depends at least partially on how
the test laboratory). For fuller advice see the Notes for Guidance the material is packed into the mud balance used for density
in Jefferis 1999. measurements. The placement of the SB includes enough effort
to ensure there are no air bubbles.
10.7.1 Soil-bentonite walls Hydraulic conductivity testing is done on grab samples sent
By way of example of the considerable level of testing necessary for off-site hydraulic conductivity testing. The grab samples are
for quality control of wall construction, Ruffing and Evans (2022) collected from the backfill, prior to placement into the trench.
suggest the following schedule of tests (Tables 10.1, 10.2 and Hydraulic conductivity testing should be performed at a
10.3) for soil-bentonite walls. minimum frequency of 1 per shift, 1 sample per 30 linear metres
Filtrate loss testing and sand content are specified for fresh along the wall, or 1 sample per 500 cubic metres of backfill
bentonite slurry but not for in-trench slurry. Filtrate loss testing placed, whichever results in the highest number of samples. An
of the in-trench slurry and sand content (fresh or in-trench) can example QC table for offsite soil-bentonite backfill testing is
be performed, but the results should not be compared to shown below with common specified targets.
acceptance criteria (Ruffing et al. 2022) but rather, if measured, As discussed below, the hydraulic conductivity of soil-
they should be collected for information and used as needed to bentonite materials will be affected by the confining pressure
influence decisions in the context of the other results. used in triaxial tests.
The field testing for fines content is used to provide a
preliminary indication of the anticipated hydraulic conductivity. 10.7.2 Cement-bentonite walls
If the fines content of the base soil used to prepare the backfill For cement-bentonite walls, test schedules similar to those
mixtures is below 15%, additional dry bentonite powder or shown in Tables 10.1 to 10.3 will be necessary with tests on both
native fines may need to be added to the SB mix depending upon the bentonite slurry and the cement-bentonite slurry (additionally
the target hydraulic conductivity of the backfill. Pre-construction bleed tests will be required for the latter). Samples must be taken
bench scale studies are used, along with the recognition of the for laboratory testing of set properties (hydraulic conductivity,
differences between field and laboratory mixing, to assess the density, strength). Typical procedures and sampling regimes are
expected hydraulic conductivity for the site-specific base soils set out in Jefferis et al. (1999).
including their fines content. Test regimes such as those set out above will be followed in
The moisture content of the base soils does not significantly many countries. For example, Fratalocchi (2022, pers. comm.)
impact the final properties of the SB backfill. The base soils are confirms that quality control of properties of cement-bentonite
prepared with bentonite-water slurry to meet the required slump. cut-off walls in Italy is by sampling from the trench though for
For base soils with low moisture content, more slurry may be major projects a test panel can be built where piezometric
necessary to achieve the desired slump and workability. For base instruments can be installed whilst the slurry is still fluid for
soils with high moisture content, less slurry may be necessary to periodic hydraulic conductivity tests once the mix has
achieve the desired slump and workability. Since on most sufficiently hardened.
projects, the bentonite in the final SB backfill comes from the Piezocone testing has been successfully applied (Manassero
bentonite in the slurry, as the quantity of slurry added increases, 1994) at least to check the order of magnitude of the hydraulic
the bentonite content of the mixture also increases. Generally, conductivity and the wall continuity (defect checking).
these minor adjustments have no measurable effect on the
10.7.3 Soil-cement-bentonite walls
performance of the backfill. As always, the expected variability
Although not regularly used for cement-bentonite or soil-
of the slurry addition rate should be assessed in the pre-
bentonite walls, coring followed by in-situ hydraulic
construction bench scale study.
263
conductivity testing is regularly used for soil-cement-bentonite volume of the sample (consideration of pore volumes, though
walls in the USA. For these tests, the wall is cored with core often used, is not appropriate when considering leaching and/or
recovery for separate viewing and testing, if required. The reaction processes). It should be noted that the hydraulic
borehole is then subjected to cleaning to remove smeared conductivity tests took several years. Had a material been used
material from the walls, tele-viewer inspection and in-situ falling with a hydraulic conductivity typical of current cut-off walls the
or rising head tests (though note there are concerns about the tests would have required many decades.
analysis of falling head tests conducted in the higher levels of a For contaminated land, there is considerable potential for
wall if above the adjacent groundwater level at the time of the interactions between the contained chemicals and any barrier
test). Coring of the wall requires that the cut-off mix has material. Much research has been done on the effects of
achieved a sufficient strength and this can become an important chemicals on cut-off materials. Fratalocchi and her co-workers
consideration when specifying the strength of the material but have carried out detailed investigations of the effects of
see also the discussion on inducing cracks in the wall at §10.9. aggressive chemicals on cement-bentonite materials (Fratalocchi
et al. 2005, 2006, 2013, 2016, 2021; Fratalocchi and Pasqualini
10.7.4 Test boxes 2007; Brianzoni et al. 2013). Jefferis (1996, 2001) found that
On occasion, perhaps too rarely, the hydraulic conductivity of hydraulic conductivity is a dynamic property of cement-
walls has been investigated by forming a box-out in the wall. For bentonite materials (the cement-bentonite material can react with
this, a parallel wall is created and connected to the main wall by chemicals in an earlier permeant so as to change its hydraulic
closing sections perpendicular to the main wall. A ‘box’ was conductivity to later permeants). Tedd et al. (1997) and Joshi et
created during the first application of the TRD trench cutter for al. (2010) report investigations on two trial containments
soil-mix cut-off wall construction (Gularte et al. 2007). In a installed through chemically aggressive gas works waste. One
much earlier test on a cement-bentonite barrier in the UK in the had a slag-cement-bentonite cut-off and the other had the same
1980s, a 2 x 10 m box was created. A well was sunk in the centre cut-off material plus and HDPE membrane. The containment
and fitted with a submersible pump. After the initial drawdown with an HDPE membrane performed significantly better over
of water in the well, the flow ceased. The soil was relatively time in this very aggressive, sulfate-rich ground.
coarse and it was suggested that the slurry in the surrounding The inclusion of a substantial proportion of pulverized fuel
trenches had penetrated into the soil in the box and effectively ash (pfa) in cement-bentonite mixes can improve their resistance
grouted it up. A second, a 6 x 6 m box, was formed but again to acid and sulfate (Jefferis and Fernandez 2000). However, it
there was effectively zero flow. A hole was then excavated at the should not be assumed that the use of high proportion of
centre of the box and used as a well. The wall was demonstrated blastfurnace slag will improve sulfate resistance as it does in
to have a hydraulic conductivity of the order of 10-9 m/s. Key concrete (Card, 1981).
lessons from these test are that the box needs to be of appropriate For soil-bentonite cut-offs, see, for example, Ruffing et al.
dimensions and pumping continued for sufficient time to achieve 2018.
steady state so that water movement in the unsaturated soil above Jefferis (1992, 2003, 2008 and 2012) demonstrated that
the pumped level has stabilized. Hydraulic conductivity may be understanding the effects of contaminants moving through cut-
more readily assessed by monitoring the water levels in the box off materials require consideration of reaction fronts which can
and in the surrounding ground whilst the box is subject to be addressed via retardation factors which are necessarily
continuous water inflow at an appropriately low rate. The test concentration dependent (the concentration of each of the
duration necessary to prove a hydraulic conductivity of 10-9 m/s. reactive species in the contaminant and their interactions),
is to be likely of the order of months. A further complication is Cut-off walls may provide not only a hydraulic barrier but
that the groundwater regimes on either side of the cut-off wall also retard contaminants migration through the wall (Evans et al.
(and hence the box) may be distinct. 1997 is perhaps the first paper on this for soil-bentonite walls).
10.7.5 Depth of a cut-off wall For research on migration parameters for different
The depth termination criterion for a cut-off wall will be contaminant species through cement-bentonite materials, see for
specified by the designer but achievement of the required depth example, Fratalocchi et al. (2005, 2006, 2013, 2016 and 2021).
must be confirmed in the field throughout the works. This
requires continuous and preferably independent observation and 10.9 Volume change of cement-bentonite materials
confirmatory measurements, for example, with a weighted tape As discussed in Jefferis (2012) and Evans and Jefferis (2014)
at defined intervals along the wall. This will be particularly post-cement-bentonite materials can undergo shrinkage post-
important if the depth criterion is a specified penetration into an placement in the trench as a result of bleed and filtration from
aquiclude layer rather than an overall depth – there can be local the fluid material and for the setting and set material, cooling
deep pockets of high hydraulic conductivity soils. On site from in trench temperature to that of the adjacent ground,
checking of depth is fundamental to the formation of effective continued hydration of the cementitious material and drying for
cut-off walls (as is confirmation that the design overlap is material above the adjacent groundwater level.
achieved for walls formed in panels). As for liners you get what Shrinkage is unlikely to be of any significance for walls
you inspect, not what you expect. constructed in panels and not exposed to significant drying
(slight movements, which can self-heal, may be accommodated
10.8 Durability of cut-off walls at the regular inter-panel joints) However, for soil-cement
Durability of cut-off walls has been a concern from their first use bentonite walls constructed in long reaches coring the wall for
(see for example, Caron 1972). However, they are very few, if in-situ hydraulic conductivity testing has led to the initiation of
any, reports of failure in service. Indeed, it would be useful and cracks across the wall – it could be confirmed that these cracks
instructive to have documentation on any in-service failures. were not present prior to the coring as the recovered cores were
As a general principle, it should be noted that very long-term not themselves cracked, Evans and Jefferis (2014).
leaching by water can be expected to lead to a change in the
properties of cement-bentonite slurries, typically a reduction in 10.10 Repair of vertical barriers
hydraulic conductivity but accompanied by a reduction in Whilst durability of any structure will be a concern, a useful
strength and if subject to confining stress, a reduction in overall aspect of vertical barriers is that they are repairable provided that
volume (Jefferis 2012). For his tests, Jefferis used samples of the surface trace of the barrier remains available – though as for
higher than usual hydraulic conductivity to be able to achieve all barrier systems in-service, identifying zones of leakage is the
substantial volumes of water permeated – over 300 times the
264
greater challenge. In practice reports of failure of slurry trench methods considering a two-dimensional wedge with an assumed
cut-offs for hydraulic structures in service if they exist are planar failure surface under plane strain conditions are used to
extremely rare. analyse trench stability. While soil stratification is often observed
Conversely, vertical barriers formed by the slurry trench in natural environments due to geological sedimentation and
technique are regularly used for the installation, repair or weathering, stability analysis of slurry trenches considering such
deepening of low permeability cores in dams, levees and stratification has not been widely reported. A horizontal slice
hydraulic structures generally. Little (1975) gives examples of method (Li et al. 2013) and a Rankine Theory-based approach
slurry trenches backfilled with plastic concrete for repairs to the (Li et al. 2018) for stability analysis of slurry trenches with
cores of four embankment dams. Currently, cut-off walls are stratification has been developed.
being used for the upgrading of dams and levees in the USA.
10.14 Stress in soil-bentonite walls and backfill consolidation
10.11 Composite vertical barriers
The in-situ stress state of soil-bentonite backfills has a
Just as for liners, composite vertical barriers have been considerable effect on the hydraulic barrier performance of cut-
developed. Jefferis and Fernandez (2000) discuss the inclusion off walls. The effective stress state in these walls is much lower
of an HDPE geomembrane in a cement-bentonite wall to contain than that predicted by a geostatic approach as first demonstrated
acid sulfate waters. The HDPE panels were joined together with by Evans et al. (1985) and further investigated in the field and by
hydrophilic rubber seals in the joints (it was confirmed, by on- back-analysis (Evans et al. 2021). If the confining stress used in
site testing, that these swelled in the strongly alkaline the laboratory tests is based on the vertical geostatic stress
environment of the cement-bentonite slurry). The acid sulphate distribution the hydraulic conductivity of soil-bentonite backfill
soils were underlain by uncontaminated low hydraulic may be significantly underestimated, leading to a non-
conductivity clay. A key cost issue was the depth of toe-in of the conservative design (Filz et al. 2003).
membrane into this clay. A penetration of 1.5 m was adopted and A model considering both the arching mechanism (Evans et
this was confirmed as necessary by advection diffusion, reaction al. 1995) and lateral squeezing effect (Filz 1996, Filz et al. 2004)
modelling with data from laboratory tests extending over many was proposed by Li et al. (2015) to predict the steady-state
months whilst cut-off construction itself was underway. effective stresses in the backfill of soil-bentonite slurry trench
Barr et al. (2015) give an example of the application of a cut-off walls. The arching effect is considered via force
composite vertical barrier for embankment dams and Zhan (see equilibrium, and the lateral squeezing effect of inward
§10.16) an example of a composite wall in China employing a displacement of the trench sidewalls is considered by assuming
GCL as the membrane. the cut-off wall is surrounded by soil which is represented by a
Winkler idealization, as shown in Figure 10.1 This model
10.12 Research on cut-off walls relating to their use in China successfully simulated the stress distribution in a soil-bentonite
cut-off wall at Mayfield, New South Wales, Australia, offering a
Cut-off walls for contamination containment are currently much significant improvement in the prediction of stress in walls.
used in China. Professor Zhan from Zhejiang University and his Another model considering the interaction between backfill and
colleague have provided the following overview of the situation surrounding soil by assuming a sliding wedge under limit
in China. equilibrium conditions based on Coulomb’s theory was proposed
Cut-off walls are commonly constructed for containment of by Ke et al. (2018).
contaminated sites. Typically, a trench, filled with slurry to After placement in a trench, the soil-bentonite backfill
maintain the trench stability, is excavated in the ground, and then consolidates under the vertical load from the weight of the
backfill is placed into the trench displacing the slurry to form a backfill above and the horizontal force from adjacent soils. Full-
cut-off wall. The hydraulic conductivity of the backfill is scale tests have been performed in the past 5 years to monitor the
required to be low, typically less than 1×10 -9 m/s below this changes of earth pressures and pore water pressures in backfills
diffusion may become significant in contaminant transport with time to understand consolidation behaviour of soil-
mechanism. Many laboratory tests have been performed to bentonite cut-off walls (Evans et al. 2017; Barlow et al. 2018;
measure the hydraulic conductivity of soil-bentonite (SB) Tong et al. 2020). The wall constructed at Bucknell University
backfills (Evans 1994; Yeo et al. 2005) and for cement-bentonite was about 194 m long, 0.9 m wide and had an average depth of
(CB) (Jefferis 2003, 2008 and 2012; Opdyke and Evans 2005; 7 m, while two walls constructed by Zhejiang University, were
Joshi et al. 2010) and to study the factors influencing the both 0.6 m wide, 15 m long, 10 m deep (W1) and 10 m long, 6 m
hydraulic conductivity. deep (W2). The earth pressures, pore water pressures in the wall
The recent advances in the research and practice of cut-off and the lateral deformations of adjacent soils have been
walls include stability analyses for slurry trenches in layered monitored. The monitoring data show that the effective stress
soils, determination of stresses in soil-bentonite walls and states of the backfills were nearly isotropic and low (<20 kPa)
backfill consolidation, breakthrough time-based design of cut- and that the primary consolidation required 3~8 months. The
off walls and the use of composite cut-off walls with an HDPE Zhejiang University test also revealed that the interfacial friction
geomembrane or a geosynthetic clay liner (GCL). A brief review on the trench sidewall played an important role in reducing the
of the above advances follows below. total stress and causing a translational soil movement (Tong et al.
2020).
10.13 Stability analysis of slurry trenches in layered soils
The stability of a slurry trench during excavation and prior to 10.15 Breakthrough time-based design of cut-off walls
backfilling is a major concern with the slurry providing a lateral In the design of a cut-off wall, the time for the contaminant to
supporting force to the trench walls. In practice, variably break through the wall should be not less than the designed
distributed surcharges from landfill berms, adjacent solid waste, service life. Transient contaminant transport through slurry walls
excavation machines and nearby buildings may induce large is typically regarded as a one-dimensional advective-dispersive
deformation in the surrounding soils and even trench collapses process. The analytical solutions available to calculate the
during excavation (and of course the converse, if an SB wall is contaminant transport in cut-off walls contain non-elementary
constructed too close to a structure, consolidation of the trench functions or require the solution of Eigen equations. The
backfill over time may cause damage to the structure). Several evaluation of these analytical solutions is nontrivial and
cases of collapse were observed for the trenches excavated in soft generally requires the use of a computer. Accordingly, the wall
clay in coastal areas of China. Commonly force equilibrium
265
thickness corresponding to a designed service life has to be They can prevent further spread of contamination and provide
searched in a trial-and-error manner. This often leads to the time for remediation processes. Cut-off walls are sometimes
determination of the wall thickness by practical experience and used as a long-term solution for contaminated lands with careful
practicality of construction instead of contaminant transport design and high-quality construction in China. Currently, the
analysis. A simplified method for determining the wall thickness plastic concrete cut-off wall is the most popular wall type as
of slurry walls is proposed by Chen et al. (2018) via an much practical experience is available from hydraulic
assumption of decoupled advection-dispersion in the analysis of engineering. Soil-bentonite cut-off walls (using sodium activated
contaminant migration (Figure 10.2). If the wall thickness is Ca-bentonite) are becoming accepted by more and more
given, the breakthrough time of the cut-off wall can be estimated designers due to their cost-effective features. Cement-bentonite
by this method. The relative error for column Peclet numbers PL (CB) cut-off walls are rarely used as the backfill has relatively
over the range 0.01 to 1000 is not greater than 4% when the high hydraulic conductivity unless high cement contents are used
breakthrough criterion of c*/c0 is less than 0.1. This covers (or at least 65% of the total cementitious material is ground
common practical situations for cut-off wall design. granulated blastfurnace slag, Card 1981; Jefferis 1997. The use
When diffusion is dominant in contaminant transport of slag is now standard practice for CB walls in many parts of
through a cut-off wall, the breakthrough time is significantly the world but it is not economically available at all locations).
extended by increasing the retardation factor of the backfill Furthermore, the experience with backfill composition,
materials. Wang et al. (2016) found that the loess amendment of especially with respect to additives, is relatively slight in China.
an SB backfill increased its adsorption capacity with respect to Some technical guidelines for cut-off walls are now available
heavy metals. The Pb(II) adsorption capacity for the SB backfill in China and are being improved by the increasing number of
containing 20% loess by mass was found to be 2 times greater practical applications.
than that of the unamended SB backfill. With this finding, the Composite cut-off walls consisting of an HDPE
breakthrough time for Pb(II) through the loess amended SB wall geomembrane have been applied to about 20 containment and
was evaluated to be about twice that for the unamended SB wall remediation projects of polluted sites in the past 10 years in
when the hydraulic gradient was less than 3. The retardation China. Both the steel frame method and the mechanical roller
effect of the loess amendment was verified by a centrifuge method are used to install the geomembrane in the slurry
modelling test performed in Zhejiang University. The centrifuge supported trench. The maximum geomembrane insertion depth
modelling test demonstrated that the Pb(II) transport distance in to date, 40 m was reached on a site in Changsha city (Figure
loess amended SB wall corresponding to a duration of 50 years 10.3).
in the prototype was 44.3% shorter than that in the unamended
SB wall (Zhan et al. 2020). The research finding shone a light on
the use of loess which is widely deposited across the world as an
adsorptive amendment in the barrier materials for the
containment of heavy metal contaminated sites.
Figure 10.3 HDPE GM cut-off wall (40-m depth) in Changsha, China.
Installation of a GCL in a slurry trench was first implemented
in China in 2017 in Tianjing (Figure 10.4). The maximum depth
of installation reached 15 m. An overlap of 50 cm was made at
the joint of the adjacent GCL panels. After the GCL was installed,
the trench was backfilled with cement-bentonite materials. The
GCL/CB cut-off wall was designed to contain a simple dump of
Figure 10.1. Schematic of the model to estimate the stresses in soil- municipal solid wastes. The potential for leakage at the GCL
bentonite cut-off walls (not to scale) (Modified from Li et al. 2015). panel joints and its countermeasures are under investigation in
the laboratory of Zhejiang University (readers may be interested
to note that the overlap approach was adopted for the first ever
composite geomembrane cut-off wall in the late 1970s. This
employed an HDPE geomembrane in conjunction with an
attapulgite-cement-Dead Sea Water self-hardening slurry, Brice
and Woodward 1984).
Figure 10.2 Configuration of contaminant concentration profile in slurry
wall at breakthrough time for the proposed method (Modified from Chen
et al. 2018)
10.16 Applications of cut-off walls in China Figure 10.4 GCL/CB cut-off wall implemented in China.
For the control of subsurface contamination, containment by cut-
off walls is almost the first choice in China due to its low cost.
266
PART C: LOOKING FORWARD contact with crushed concrete and more briefly on sulfate attack
on crushed concrete.
pH is important as regulatory authorities may place limits on
11 RECYCLED MATERIALS the pH of water discharged from a site. A typical range of
permissible pHs being 6 to 9. It may be noted with regard to the
The construction industry is a major mover and user of materials lower limit that water in equilibrium with atmospheric carbon
and so offers very many opportunities for energy use reduction dioxide will have a pH of about 5.6 without any contribution
and for the re-use of materials especially as it generates so much from sulphur and nitrogen oxides that can be significant in areas
potentially re-usable material. Indeed, today there is almost an affected by industrialization. For the higher pH limit, the sodium
implied duty on all involved in the construction matrix to achieve and potassium oxides and free lime in crushed concrete can result
minimum materials and energy use – trajectory net zero carbon in it having a pH at or over 12.6. Even if/when these species have
and to maximise the re-use, re-purposing and recycling of been leached the pH of leachate in equilibrium with crushed
materials. In this respect one may note that many of the key concrete will be over 9 as a result of dissolution calcium
decisions with regard to the re-use of materials are made at high hydroxide from the calcium silicate hydrates which make the
level – the project scoping stages of works. There is a real need principal contribution to the structural strength of concrete.
for advice at the muddy-boots level for the engineers and other Whilst the pH of leachate from crushed concrete can be expected
on-site operatives. to reduce as leaching progresses it is likely to remain >9 and thus
In the coming years, we can expect major changes in the above typical regulatory limits for very substantial times.
materials we use. Some will cease to be available or be the If the crushed concrete is used as a structural fill it is
subject of major price rises due to limited production and/or important to recognize that the rate of attack by any aggressive
demand from new users. Thus, for example, we can expect chemicals including leaching by water will be much higher than
pulverised fuel ash (pfa) derived from coal burning to cease to for intact structural concrete because of its much higher surface
be available although if existing ash dumps can be mined to area to volume ratio and because it will have sustained some
produce useful material availability may continue for some years. internal cracking in the crushing process. It is also unlikely that
Ground granulated blastfurnace slag (ggbs) has been used for any information will be available on the original mix design
very many years as a replacement for Portland cement in including its water/cement ratio – indeed much crushed concrete
concrete. For example, in the UK sulphate resisting Portland type will have been derived from crushing concretes with a variety of
cement is scarcely available as ggbs-Portland cement blends at mix designs. For all these reasons guidance on the effects of
appropriately high ggbs levels can provide resistance to sulphate chemicals on structural concrete (for example, sulphates) should
attack. Ggbs is a by-product of the reduction of iron ore to iron not be assumed to apply to crushed concrete.
in a blast furnace. However, the iron and steel industry can be
expected to move away from the use of carbon as the reducing 11.1 Some other risks associated with recycled materials
agent (in the process the carbon is oxidised to carbon dioxide).
Thus, ggbs may well be phased out (Tallard 2021, pers. comm.). Risks from other materials offered for re-use in construction
Currently ggbs is rated as having a lower carbon footprint include not only high pH leachates but also deleterious
than Portland cement as its carbon burden is shared with the expansion and emission of gases. These materials include
primary product of the blastfurnace, iron. Ggbs is a major secondary materials, such as: steel slag, old blast furnace slag
constituent of some low carbon cements – a growing market and incinerator bottom ash aggregate (IBAA) derived from the
competing for ggbs supplies. incineration of municipal solid waste. Problems associated with
Slag-cement-bentonite cut-off walls mixes may contain the ferrous slags are discussed in BRE IP/05 (Dunster et al. 2005).
perhaps 65 to 80% slag (by weight of total cementitious Calcium containing leachates from crushed concrete will
components). A phasing out of slag or major price rises due to precipitate calcium carbonate on contact with air or with waters
competition for limited/dwindling supplies will require new cut- containing inorganic carbon species. There is therefore a risk of
off wall mix formulations or, more radically, re-thinking how to blockage of drainage systems etc.
create low hydraulic conductivity vertical barriers in the ground. As crushed concrete will come from the demolition of
The extensive geoenvironmental literature on recycling buildings and other structures, there is a risk that it may contain
materials is often focused on the sustainability benefits and on unwanted and potentially undesirable materials, such as asbestos
research to demonstrate that environmental impacts such as from asbestos insulation, asbestos cement sheeting and pipes.
heavy metal concentrations in aqueous leachates are acceptable. Asbestos fibres may be released by the crushing processes. The
In contrast, much less has been published about the long-term potential for harmful materials should be included in risk
behaviour of recycled and secondary materials. Are they assessments for all recycled materials; risks must be identified
chemically and mechanically stable? Do they have the potential and appropriately managed to protect all those involved from the
for shrinkage, expansion or to damage structures formed from production process, through storage, transport, re-use, future
them or built upon them? Will they emit toxic, explosive or demolition and re-re-use and all associated laboratory testing.
asphyxiant gases? Are the vendors able to provide guarantees
regarding suitability for engineering use or will they choose to 11.2 Residues from incineration of municipal solid waste
limit their liability to the value of the materials supplied? Who
will bear the risks relating to possible in-service failure? The Incineration can be used to reduce the mass and volume of
owner may specify recycled materials and secondary but expect municipal solid waste (MSW) and to recover energy. Crudely
the engineer to bear the risks and the financial consequences of perhaps two thirds of the mass of the waste will become stack
failure. gases. The remaining one third will be mainly incinerator bottom
Furthermore, the environmental significance of more ash (IBA) though there will be a modest amount of flue dust
mundane chemical parameters such as pH tends to receive which may have to go to hazardous waste landfill. It will be
insufficient attention and there is too often inadequate desirable to re-use the incinerator bottom ash. After suitable and
precautionary advice. sufficient processing the ash may be used as incinerator bottom
Jefferis (2019) explored some problems that could flow from ash aggregate (IBAA). The composition of this ash will depend
the use of recycled crushed concrete with the aim of providing a on the composition of the MSW and can be expected to vary
baseline from which to explore some of the problems that can across and between countries and indeed to change over time.
arise from the re-use of construction materials more generally. Expansive reactions can occur in IBAA, notably reaction of
The paper focused on the pH of waters which have been in metallic aluminium in the IBAA with water in the alkaline
267
environment of the ash, and then further reactions of the use, if materials react with chemicals in adjacent materials (e.g.,
aluminium reaction products (Eden 2017). Ideally, metals such natural and anthropogenic chemicals in soils) and so become
as iron and aluminium should be removed from the MSW prior unfit for further re-use. The problems associated with multiple
to incineration. However, they may remain. If an alkali such as re-uses invites consideration of an entropy based approach to
sodium hydroxide or sodium carbonate is added to a sample of recycled materials
an aluminium containing IBA then some fizzing and hydrogen
release can be expected. As many ashes are alkaline there can be
some production of hydrogen without alkali addition. There have 13 OVERARCHING ISSUES
been claims/litigations in respect of hydrogen release.
The paragraphs below set out a few examples of overarching
issues – issues that will affect all of us but on which
geoenvironmental engineers should have particular insights.
12 WASTE MINIMISATION
13.1 Interconnection and Interdependence
As is set out in this paper it is clear that there is a great deal of
current research on landfill engineering and in particular landfill We live in an interconnected world with ever increasing
lining systems. However, looking to the future we must work interdependencies not only at the personal level but also in
towards minimising landfills and seek every opportunity to relation to key areas such as infrastructure and energy supply.
maintain materials within the cycle of utility. Waste minimisation The drive to sustainability and more recently net zero carbon is
is now a well-established concept through considerable likely to increase interdependency unless this is recognised and
opportunities remain in respect of its actual implementation in managed. As a very simple example consider the move towards
construction and more generally in society. electric vehicles. These must be re-charged and will stop if a
We should also expect landfill mining technology to advance, failure of supply persists beyond their charge time. A localised
returning the landfilled (stored) materials to the cycle of utility storm which knocks out part of a distribution system will knock
and the land ideally to multifunctional use. Currently, landfill out all that depends solely on it: homes, industry, infrastructure,
mining tends to be driven by land value. Thus, landfill mining is telecommunications and now electric vehicles. Depots providing
is limited to landfills in areas where there are currently fuel for vehicles running on liquid fuels or gas (lpg) provide a
significant opportunities for development and particularly hard distributed system (provided there is power to dispense the fuel).
developments such as residential units. The net effect is a loss of system redundancy. Very localised
Landfill mining will generate significant amounts of material battery stations would be/will be required to provide the same
for re-use, including metals, building rubble, crushed concrete level of redundancy. Geoenvironmental engineers must add
etc. which will need to be re-purposed. There also may be interdependence to the key touchstones of sustainability, net zero
incinerator bottom ash and flue dust from the incineration of the and resilience.
mixed organic debris for energy recovery. These materials and Interdependence means that the economic cost of events
other recycled and secondary materials will be offered and such as floods, storms, earthquakes, volcanic eruptions are many
indeed perhaps imposed on the geotechnical community. They times greater than the costs of the physical damage caused by the
may come with acceptable environmental credentials (leach rates event (see papers from the Oxford Environmental Change
etc.) but with minimal information on their in-service Institute, for example, Adshead et al. 2021).
geotechnical behaviour. For example, are they mechanical stable? The production of batteries and the magnets for electric
Will they swell, shrink etc. in the medium and long-term? Similar motors is of course not without impacts from the mining and
problems may arise from their use as have become apparent with processing necessary to produce the essential metals.
crushed concrete and incinerator bottom ash aggregate as
discussed above. The potential for such problems should act as a 13.2 Covid 19 and future pandemics
caution to those who choose or are required by specifications to
use them. It is hoped it will also stimulate the geoenvironmental We must expect future construction to be influenced by the
community to undertake research and performance assessments. revised thinking induced by the Covid pandemic and improving
It is trite to say that long-term testing takes a long time. It should preparedness for and resilience to future pandemics. For example,
be put in train now. it may reverse the current trend for population movements into
cities from more rural areas.
12.1 Mixing of materials The trend towards home-based working also may have major
impacts on the type and distribution of homes, offices, and other
The fundamental aim of re-using materials is to keep them in the commercial premises. Indeed, we could be at the beginning of a
cycle of utility. The first uses or re-uses should not compromise major re-visioning of the built environment and associated
future re-uses. Use and re-use of construction materials therefore transport and service infrastructures.
should be undertaken recognising that within a generation or two Many questions hover in our thoughts: How long will any re-
a structure may be demolished and that, at that time, its materials visioning endure? Will we return to pre-Covid 19 thinking? What
must be fit for return to the cycle of utility. will happen/should happen in regions where the state of
A use or re-use that ultimately dooms a material to landfill development is too limited for pandemics to be the issue of the
rather than further use has bought time but potentially at the cost greatest concern?
of very expensive landfills in later generations. Such action will
be particularly unfortunate if the mass and volume of the material 13.3 Climate change
have been increased by blending it with other materials (virgin
or recycled) so that the combined material later fails criteria for Climate change is a matter of immediate concern and engineers
re-use or worse still, as has happened, results in the combined are able to drive change not only with regard to reduction in
material being designated as not-inert and requiring expensive emissions of greenhouse gases but also the design of new works
landfill, indeed possibly toxic waste landfill – a mixture of and the adaptation of the old for climate resilience. Here
crushed concrete and soil will have high pH from the cement resilience includes impacts of sea level rise, more violent storms
possibly making the combined materials toxic waste as it is no and sea states, more intense rainfall, drought, and more extreme
longer simply crushed concrete for which regulatory exemptions climates. Net zero has become the by-word but of course this
may be available. Similar problems can occur during use or re- must include minimisation of all greenhouse gas emissions –
268
noting that methane is a powerful greenhouse gas and that ground. With low viscosity systems which do not seal, loss of
herbivores such as cows are significant producers. It is not grout via preferential pathways or simple gravity driven flow of
simply a matter of changing our approach to hard construction dense solutions becomes a real problem. Is the bio-grout where
but also to agricultural and personal practices. you think it is? Furthermore, grouting at the necessary close
Climate change is currently associated with more extreme centres is expensive. It is not sufficient to research
weather, periods of intense rain, high winds, and sustained biomodification materials alone. Delivery and treatment-
periods of drought. Massive geotechnical works will be required verification systems also require research.
to build resilience and there may be wholesale movements of What is the durability of the bio-grout in service? If there is
populations from environments which become unsustainable – potential for groundwater flow/the ground is permeable there
managed retreat – always provided that it is managed. will be leaching. Indeed for grouted barriers leaching of grout
Whilst climate change deniers may still seek to debate components (e.g., calcium from Portland cement grouts) should
whether anthropogenic carbon dioxide emissions are the cause be recognised as one of the major mechanisms of barrier
of climate change there can be no doubt that carbon dioxide degradation. Many aquifers and all rainwaters are under-
levels in the atmosphere have increased substantially over the saturated with respect to calcite. Will bio-precipitated calcite
last two centuries (from 350 to the current 412 ppm). Amongst remain in-situ? How can design life be estimated?
so many other impacts, this will lead a reduction in the pH of Conversely, if leaching is not significant, will growth of
oceans. larger crystals with depletion of finer crystals (a natural process
to reduce the overall surface energy) affect the mechanical
13.3 Energy production performance of the bio-grout? Can this and other aspects of
In terms of carbon dioxide production, the fossil raw materials Ostwald ripening be an issue?
for energy generation in order of decreasing carbon dioxide No doubt these and many other issues will be addressed in
production are coal (worst), oil and then gas (least emission). the conference.
However, we can and should expect all fossil use for energy
production to be phased out with green(er) energy derived from
hydro, wind, solar and nuclear. The need to store wind and solar
energy is prompting interest in the construction of pump storage 14 ENVIRONMENTAL GEOTECHNICS:
schemes including the use of old mines. Countries, like Canada, The discipline of Environmental Geotechnics has evolved since
the development of vast resources for Hydro production once its beginnings in the late 1970s to early 1980s through the
uneconomic now becomes economic. Liner systems are playing adoption and adaptation of geotechnical engineering knowledge
an increasingly important role in these developments. However, to produce practical solutions for problematic environmental
many of these developments will take considerable time before situations (Sarsby 2019). The design of low-permeability
they are built and oil and gas may remain as chemical feedstocks pollutant barriers, such as waste containment liners and cut-off
until replaced by the green alternatives. These undertakings will walls, may be mentioned as an example of the application of
displace land that otherwise could be used for forestry or food geotechnical engineering expertise, which was mainly gained
production. Major changes can be expected from the drive to from the design of earth dams and embankments and then refined
reduce meat consumption which will not only increase the for environmental purposes (Mitchell et al. 1965; Mitchell 1991;
efficiency of food production (calories per unit area of land) but Daniel 1993; Rowe et al. 1993; Rowe et al. 2004; Shackelford
also reduce methane emissions from those flatulent domesticated 2014; Rowe 2018). When facing these new challenges,
animals. geotechnical engineers understood that knowledge of traditional
soil mechanics was a necessary, but not sufficient condition to
13.4 Mobile energy obtain satisfactory results, and started to broaden their formal
education to include chemistry, geochemistry, and groundwater
Alongside the move to green energy production there will be a hydrology (Goodall and Quigley 1977; Quigley et al. 1983,
change in the energy carrier for vehicles, construction plant etc. 1988; Quigley and Rowe 1986; Shackelford 2005). The
Electric cars soon may be the norm as governments set performance of a containment barrier is in fact not only related
regulations for the phase-out of petrol/diesel cars. Battery to its mechanical and hydraulic features, but also to its ability to
operation also is being found practical for quasi-static limit the mass transport of a pollutant to the groundwater. The
construction plant such as piling rigs. For heavy vehicles and need to integrate hydrogeology and civil engineering became
plant hydrogen currently appears the favoured energy carrier. apparent (Rowe 1988, 1992). This was followed by a need to
integrate polymer engineering in civil engineering and the
13.5 Biocementation increase in knowledge needed to effectively apply geosynthetics
to waste containment applications (Rowe 1998, 2018, 2020).
Across the world there is a substantial corpus of research in
Nowadays, Environmental Geotechnics issues include,
progress on the use of biological techniques for ground
among others, waste containment, waste mechanics, the stability
modification. A significant part of this work is on calcite of landfill liners and covers, the stability of tailing dams and heap
precipitation. In comparison, rather little work seems to be leach pads, the surface containment of lagoons and ponds, soil
addressing the precipitation of other minerals. Work also remediation, contaminated and damaged land reclamation,
continues on the effects of vegetation including plant roots. biologic processes for soil improvement, underground energy
The amount of work in progress is so extensive that it is not exploitation, carbon dioxide sequestration, and nuclear waste
possible to provide a concise overview. There will be papers on disposal. As a result, the professional identity of an expert in the
bio-manipulation in the conference and readers are directed to field of Environmental Geotechnics is characterized not only by
them. However, it is perhaps appropriate to raise a number of a traditional background in soil mechanics and geotechnical
issues. engineering, in addition to structural mechanics and hydraulics,
How is biocementation/biomodification to be applied? All but also by a robust basic training in mathematics, and in
those familiar with geotechnical grouting will be aware of the physical, chemical and biological sciences, which enables them
problems associated with achieving sufficiently complete ground to handle the manifold and complex phenomena that occur in
treatment even with grouts that seal the soil so that treatment can environmental applications. The distinctive traits of the
be at reducing centres working against already partly sealed discipline are the propensity to face new problems through an
269
open-minded approach and the ability to interact with other monitoring related to the modelling of transport and the fate of
disciplines, including social and economic sciences and law. The contaminants, to validate theoretical assumptions, and to verify
pressing need to find practical solutions for the wide range of the reduction in time of contaminant concentrations.
problems that involve the environment has favoured the
tendency to explore new fields of knowledge and expand the
boundaries of the discipline, in contrast to the mainstream
tendency to develop a hyper-specialization focused on particular
subjects.
A possible answer to the question “where are we going?” is
that our discipline will continue along the path of searching for
practical solutions for the emerging, and frequently unexpected,
environmental problems that we will have to face in the future,
relying on an approach which combines geotechnical
engineering expertise with a multidisciplinary cultural
background. Even though we cannot foresee what the future will
bring, we may envisage how we will take on the future
challenges, as the practice in our discipline has built a forma
mentis to deal with problems posed by human progress and the
future evolution of nature (e.g., climate change). In comparison
to other disciplines in the field of geotechnical engineering,
Environmental Geotechnics is not characterized by a specialistic
list of topics, but by a context (i.e., the ground), a goal (i.e., to Figure 14.1. Fundamental interactions of the discipline of Environmental
find practical solutions for environmental problems) and a Geotechnics.
method (i.e., broadening knowledge through a multidisciplinary
approach). Technical and technological advances should make the
We should not be surprised in the future to find individuals collection and management of large sets of data possible,
who have a professional identity in the discipline of including chemical and biological information, which is
Environmental Geotechnics who are engaged in modelling the necessary to support modelling activities.
coupled mechanical, chemical and thermal behaviour of geo- The development of Environmental Geotechnics is driven by
structures interacting with the environment (e.g., through the evolution of social needs, such as the minimisation of
emissions) with the aim of developing a sustainability and materials and energy consumption, the preservation of natural
resilience assessment to select the best design option among the resources, the minimisation of waste generation, and the
available technical and technological alternatives. A systematic maximisation of future reuse options for polluted sites. As a
analysis of the expected evolution of Environmental Geotechnics result, a fundamental role is played by the interaction with
may be attempted based on the current perception of our society, which takes place through the development of new
discipline. It can be postulated that three fundamental regulations and design guidelines. From this point of view, we
interactions will take place in the future in the field of can already perceive a move toward a holistic approach, which
Environmental Geotechnics (Figure 14.1): evaluates environmental, economic, and social benefits of any
an interaction with technical/technological development, project as well as the potential impacts and unintended
an interaction with the society, and consequences. Design choices should be based not only on the
an interaction with new environmental issues. best technical and technological options that are available, but
The previous sections of this state-of-the-art paper have also on a sustainability assessment (e.g., through a life cycle
focused on the description of the most advanced technical and analysis) for the quantification of the direct costs (e.g.,
technological solutions that are available for the containment of construction cost), indirect costs (e.g., social cost of greenhouse
contaminants. The future application of these solutions will gas emissions), and benefits (e.g., employment opportunities). In
increase the efficiency of containment and clean up systems and this way, the key indicators of the environmental, economic and
will determine a relevant environmental benefit. This benefit, in social impacts across the entire project life cycle can be
turn, will stimulate new scientific research to further improve the identified and quantified to make decisions that are fully
outcome of the implemented projects. sustainable (Reddy et al. 2019). Moreover, with the global
Technological developments will significantly enhance the climate change and the consequent extreme events that are
potentiality of the monitoring systems that are used to assess and occurring, a necessary feature of projects is their ability to be
control the performance of operational geo-structures, so that the resilient and maintainable/repairable (i.e., capable of responding
inherent risk to a site can be reduced. Among the emerging to a perturbation or disturbance by resisting damage and
technologies that may change the way geotechnical monitoring recovering/being recovered quickly). Resilience with respect to
is conducted in the future, Soga et al. (2019) mentioned satellite extreme events, such as earthquakes, storms, floods, permafrost
Imaging, distributed fibre optic sensing, LiDAR (Laser Imaging melting, and desertification, is dependent on the mechanical
Detection and Ranging) and computer vision techniques. resistance of geo-structures, which must be able to protect the
Monitoring is expected to play a fundamental role in the environment from the consequences of catastrophic failures.
environmental field, due to the increased acceptance of While the paradigm of sustainability drives the design toward a
Monitored Natural Attenuation (MNA) and Monitored Natural rational risk-based approach, which considers the natural
Recovery (MNR) as tools for the remediation of contaminated attenuation mechanisms that contribute to preserving or restoring
sites (Yong and Mulligan 2019). Monitoring typically involves the environment, the paradigm of resilience highlights the
collecting soil (or sediment) and water samples to analyse them relevance of the mechanical behaviour of geo-structures in
for the presence of contaminants. Laboratory research includes protecting the environment against extreme events.
microcosm studies on the evaluation of the degradation and Although the future will inevitably be characterized by a
transformation of the target organic chemical contaminants, and certain degree of unpredictability, we can expect that future
partitioning tests for both inorganic and organic chemical generations of engineers will have to face emerging issues
contaminants, such as tests for the determination of adsorption related to the impact of human activity on the environment.
isotherms, leaching column tests, and desorption tests. Field These emerging issues may be unexpected, but future engineers
must be ready to face them by adapting their knowledge and by
270
finding suitable solutions. A terrible example of an emerging contaminating lifespan of a landfill in the context of the
issue is that of the COVID-19 epidemic. Soon, the need to face local hydrogeology and how construction of the facility
the impact of this epidemic on the environment will represent an will affect the hydrogeologic conditions,
extraordinary challenge, and this challenge will also involve our service life of the components of the system and how
discipline and will require engineers to study and learn from interactions between system components will change with
reactions to the many previous pandemics that that plagued the time and ageing,
World, see for example, Karlen (1995) and Kilbourne (2008). potential implications of climate change on the
Two specific topics are analysed in the following sub- performance of the facility,
sections: the evolution towards performance-based design in potential failure modes and effects,
place of the prescriptive-based design and the possible potential unintended consequences,
contribution of Environmental Geotechnics to facing the nature and likely effectiveness of monitoring, both in the
emerging issue of the COVID-19 pandemic and future short and long-term,
pandemics. The aim of the first topic is to give an example of the when impacts are likely to be detected by the different
interaction between the discipline of Environmental Geotechnics components of the monitoring system and establishing
and society. Similarly, the second topic is an example of the trigger levels for various types of pre-emptive or remedial
interaction with an emerging issue. action was taken, and
the need for perpetual (i.e., multigenerational) care and
14.1 Performance-based design of landfill lining systems maintenance.
A common performance criterion is that a liner must ensure
The design of pollutant barriers, such as waste containment that the concentrations of pollutants in the groundwater remain
liners, is aimed at minimising both leachate exfiltration and less than prescribed threshold levels at a specified compliance
contaminant transport from the waste to the groundwater system point, which is commonly represented by a monitoring well
(Rowe et al. 2004, Shackelford 2014). The release of located in the underlying aquifer and downstream from the
contaminants that are contained in landfill leachate can be landfill. In fact, the pollutant concentration can be related to a
reduced to environmentally acceptable levels by a barrier system corresponding risk to human health and the environment through
that includes a leachate collection system and a liner system. The a toxicological model that considers the pollutant properties and
leachate collection system typically involves a series of the exposure paths (Dominijanni and Manassero 2021). The
perforated pipes in a granular drainage layer that are aimed at pollutant concentration in the groundwater is obtained from a
minimizing the leachate ponded head. Modern liners may be transport analysis, which considers the migration process from
constituted by compacted clay liners (CCLs) and geosynthetic the waste contained in the landfill to the compliance point.
components, such as geomembranes (GMBs) and geosynthetic The approach to follow is similar to the procedure used for
clay liners (GCLs). GMBs can be coupled with CCLs and/or the risk assessment of contaminated sites that was first developed
GCLs to form composite liners that provide an optimal in the US in the late 1980s (US EPA 1989) and subsequently
groundwater protection performance. Moreover, the aquifer standardised by the ASTM (ASTM 1995; 2000). This procedure
located beneath a landfill is typically not only separated from the is structured in three tiers (Sethi and Di Molfetta 2019):
waste by these artificial layers, but also by a natural foundation the first tier essentially involves comparing the site
or attenuation layer (AL), which plays an important role in contamination with screening concentration values.
limiting contaminant migration (Rowe and Brachman 2004). the second tier involves a simplified contaminant transport
Rowe (2009) stated that in order to minimize the analysis based on analytical, semi-analytical, or simple
environmental impacts of a landfill, a systems engineering numerical solutions, in which part of the input data are
approach to the design should be adopted. This approach derived from on-site investigations, while missing
involves decomposing the entire system into subsystems information is obtained from validated and up-to-date
consisting of simpler identifiable components. The performance databases or from the literature.
of the individual components, as well as the interactions between the third tier represents a more detailed appraisal of risk,
the different components of the system, must be assessed before based on sophisticated models that are solved through
the response of the entire system can be quantified to obtain its advanced numerical methods. These models require
overall engineering performance. enough site-specific chemical, physical and biological data
A performance-based design of a lining system implies a to carry out a full experimental system characterization.
profound change with respect to the traditional prescriptive- Although the first tier is not amenable to a risk assessment for
based design, which is still nowadays the most widespread in the design of a landfill lining system, the second and third tiers
worldwide regulations. Though note that performance based can be applied profitably to evaluate the impact of pollutants,
requirements are typical for vertical barrier systems. For released from waste, on groundwater quality.
example, hydraulic conductivity and thickness will be specified The application of such a risk assessment to the performance-
but not the materials to be used (Jefferis et al. 1999). The benefits based design of landfill lining systems can therefore be carried
of a performance-based design include (Estrin and Rowe 1995): out based on a conceptual model that identifies the leachate
allowing engineers to bring state-of-the-art knowledge to a produced by the waste with the source of contamination and a
project, which in turn is expected to encourage both monitoring piezometer (which may be real or virtual), placed
theoretical and practical research investigations and the downstream of the landfill, with the point of compliance, as
application of innovative technology in the field. shown in Figure 14.2.
putting emphasis on pre-approval design examination The limit value that the pollutant concentration in the
rather than on post-construction monitoring. groundwater can assume, in correspondence to the point of
stimulating a more in-depth scrutiny by regulators and compliance, without producing an unacceptable risk for human
concerned members of the public regarding the adequacy health or the ecosystem may be determined based on a
of the proposed design prior to approval. toxicological model (Redfearn and Roberts 2004; Sethi and Di
performing an analysis of the interaction of the proposed Molfetta 2019). As a result, the verification of the effectiveness
engineering system with the particular hydrogeologic of the lining system and the attenuation layer in containing the
conditions in which the landfill would be located. waste contamination is obtained through a contaminant transport
Performance-based design allows an assessment of the: analysis, which provides the theoretical value of the contaminant
concentration at the point of compliance. A value of the
271
calculated concentration that is lower than the limit value given can be compared to a tier 3 analysis for a polluted site, which can
by the toxicological model is indicative of a satisfactory design, be conducted when a more reliable evaluation of the contaminant
as the related risk is acceptable. concentration in the groundwater is considered necessary.
The variation of dominant transport mechanisms that takes
place between the vertical movement across the landfill lining
system and the underlying attenuation layer and the
predominantly horizontal movement that occurs within the
aquifer represents a difficulty in conducting a contaminant
migration analysis. The vertical transport, which can develop
under unsaturated conditions, is in fact characterized by a low
velocity and is therefore mainly controlled by molecular
diffusion (Shackelford 2014). On the contrary, horizontal
groundwater movement is typically characterized by a velocity
that is some orders of magnitude higher, and the related
Figure 14.2. Reference scheme for the application of the risk assessment contaminant transport is dominated by advection. The numerical
procedure to the performance-based design of a landfill lining system. methods that are suitable for vertical transport may be unsuitable
The pollutant released by the waste migrates vertically through the for solving the horizontal transport in the aquifer, and this may
engineered barrier system and the natural attenuation layer to the implicate a loss of accuracy in the calculation within one of the
underlying aquifer, where pollutant transport then becomes horizontal. two components of contaminant transport. Semi-analytical
methods (Row and Booker loc. cit.) provide one solution to this
Manassero et al. (2000), Rowe and Booker (2000), Guyonnet problem. Another possible solution to this problem is to resort to
et al. (2001), Foose (2010) and Dominijanni and Manassero the Domain Decomposition Method (Bellomo and Preziosi
(2021) derived analytical and numerical solutions for the 1995; Tang et al. 2020). In this method, the domain of the
contaminant transport from landfills under the restrictive considered mathematical problem is partitioned into sub-
assumptions of steady-state conditions and a constant source domains, which can be analysed separately by introducing
concentration in the waste leachate. These conditions exclude the additional interface conditions at the boundaries between them.
possibility of modelling time-varying properties and time- One of the main advantages of the method is the possibility of
dependent phenomena and typically result in conservative running a parallel calculation for the different sub-domains. In
predictions of the groundwater contaminant concentration the scenario of contaminant migration from a landfill, the domain
(Shackelford 1990; Rabideau and Khandelwal 1998; Rowe et al. can be divided into a first sub-domain that includes the landfill
2004). As a result, such solutions should be considered suitable lining system and the underlying attenuation layer, and a second
for a conservative estimate of the risk, in a similar way to a tier sub-domain for the aquifer. The interface between the two sub-
2 analysis for a polluted site. The solutions developed by domains is represented by the top surface of the aquifer, in
Dominijanni and Manassero (2021) allow the contaminant correspondence to which the continuity of the hydraulic head,
concentration distribution to be evaluated in the horizontal the volumetric flux, the contaminant concentration, and the
direction within the aquifer beneath the landfill. In such a way, contaminant mass flux must be imposed. By using such a
the progressive increment of the contaminant concentration that strategy, it is possible to optimize the choice of the numerical
occurs along the direction of the groundwater flow can be method, as each sub-domain can be solved separately, while an
appreciated. The analyses contemplate both the case of thin iterative procedure is implemented to check that the interface
aquifers (i.e., aquifers whose thickness is no more than a few conditions are satisfied. For instance, a value of the hydraulic
metres) and the case of thick confined aquifers, in which the head and the contaminant concentrations, at the top of the
vertical distribution of the contaminant concentration needs to be aquifer, can be assumed at the beginning of the iterative
considered. An analytical solution for the calculation of the procedure. The vertical volumetric flux and the vertical
leakage rate through a composite liner, which consists of a contaminant flux exiting from the attenuation layer are
geomembrane overlying a geosynthetic clay liner, is presented in calculated since these values. By using these fluxes to impose the
Guarena et al. (2020) for the case of a hole located in boundary conditions at the top of the aquifer sub-domain, a
correspondence to a wrinkle in the geomembrane. This solution, distribution of the hydraulic head and contaminant concentration
which can be implemented in a contaminant migration analysis, is found for the groundwater, so that a new estimate of the
considers the influence of bentonite swelling on the hydraulic head and the contaminant concentration at the top of
determination of the hydraulic transmissivity at the interface the aquifer can be derived. The iterative procedure is stopped
between the geomembrane and the underlying geosynthetic clay when the values of the hydraulic head and the contaminant
liner, and the chemico-osmotic component of the water flow that concentration at the top of the aquifer do not change appreciably
is generated by the gradient in the solute concentration. between two consecutive calculation steps.
Dominijanni et al. (2021) have extended the available solutions Another interesting development of the contaminant
to scenarios that involve unconfined flow conditions in thick migration analysis is its application in the frame of a probabilistic
aquifers. These solutions offer the possibility of conducting an approach in which the boundary conditions and the model
analysis that involves a limited number of parameters and allows parameters have a random nature. In fact, a significant difficulty
the influence of the liner properties (e.g., hydraulic conductivity, in the analysis arises from the uncertainty that is encountered in
thickness, defects) and the field conditions (e.g., aquifer the evaluation of the representative values that need to be
thickness, groundwater velocity) on the final result to be assigned to various parameters, such as the leachate contaminant
appreciated under conservative assumptions (steady-state concentration, the hydraulic conductivity of the mineral layers
conditions and constant source concentration). and the number, size and location of the geomembrane defects.
Semi-analytical solutions implemented in popular software In a deterministic approach, the designer must trust in his own
products, such as POLLUTE (Rowe and Booker 1985a; Rowe et judgement to make the most opportune choice of the values that
al. 2004; Rowe and Booker 2005) and MIGRATE (Rowe and have to be assigned to the parameters but cannot, however, know
Booker, 1985b), are available to evaluate the groundwater the combined effect of the variance of the various parameters on
contaminant concentration below a landfill, taking into account the final result of the analysis. The adoption of a probabilistic
the time-evolution of the source concentration in the landfill approach instead allows the random nature of the involved
leachate, as well as time-dependent phenomena, such as parameters to be considered explicitly. In such a way, the final
adsorption, ion exchange and bio-degradation. These solutions
272
results may be related not only to the most representative values hydraulic conductivity and sorptive capacity) through their own
of the involved parameters but also to their variance (e.g., Rowe activities. Any assessment of the sanitary risk related to the
and Fraser 1995). Probabilistic analysis is implemented in the migration of viruses in the subsoil requires the development of
well-established contaminant migration models CONSIM and accurate, reproducible, and cost-effective techniques to identify
LANDSIM models developed by Golder Associates for the UK and quantify the presence of pathogens and their infectiousness
Environment Agency (http://www.consim.co.uk and in particulate media, so that these techniques become familiar to
http://www.landsim.co.uk). most geoenvironmental researchers and engineers. Because of
A performance-based design is an ideal tool for the evaluation the ability of enveloped viruses to retain infectivity for days to
of the project choices based on a sustainability approach, as it months in aqueous environments, and because of the observed
allows the most advanced technologies to be used. Moreover, the persistence of SARS-CoV-2 in unfiltered water with a high
site vulnerability and the natural attenuation mechanisms can be content of organic compounds and suspended solids, major
introduced into the analysis, thereby avoiding recourse to concerns have arisen concerning the discharge of waste waters
oversized lining systems, with relevant economic and from hospitals and health care facilities into the municipal
environmental savings. sewerage systems, due to the likelihood of a significant viral load
The resilience of a project can only be guaranteed through the during the pandemic besides the need for in situ disinfection
coupling of a contaminant migration analysis with a mechanical before discharging, the geosynthetics industry may provide
stability analysis. Mechanical stresses and strains can influence novel lining products (e.g., geomembranes and geosynthetic clay
the barrier performances of the components of a lining system. liners) to prevent virus leakage from buried pipelines and
Rowe and Yu (2018), for instance, reviewed the strains that can containers, which are used to temporarily store pathogen-laden
develop in a geomembrane from gravel above and below it and waste water. Furthermore, the massive usage of disinfectants in
from down-drag due to the weight of the waste and from medical centres, households and public spaces represents in itself
subsequent degradation and consolidation of the waste. These an additional risk factor, as the percentage of both liquid and
strains play a critical role in the long-term performance of a solid waste that contains traces of sodium hypochlorite or similar
geomembrane. Stability analyses should consider the effects of substances is expected to increase during the pandemic, and in
seismic actions, extreme rainfall events and large variations of turn causes the formation of a number of water disinfection by-
the groundwater levels. A key feature for a good design is products, such as the four primary trihalomethanes, which pose
represented by the attention paid to the consequences of potential genotoxic and carcinogenic health risks.
mechanical failures on the barrier performance of a lining Given the above challenges introduced by the outbreak of the
system. pandemic, the Environmental Geotechnics community is called
All the above will be subject to the time and financial upon to contribute to mitigating the negative consequences of the
constraints imposed by the owner. Deciding what is essential and current emergency. As such, the COVID-19 pandemic represents
what merely desirable in a design analysis may be challenging; an opportunity for the development of our discipline, to build
at what point does one refuse to proceed? Furthermore, when upon the unique expertise of geoenvironmental researchers and
adopting/being required to adopt advances in technology, engineers in solving environmentally related geotechnical
designers and contractors must first establish who will carry the engineering problems, and upon the wealth of knowledge that
liability for failure should a new advance in technology not meet comes from microbiology, medicine and other related
expectations. An analysis of some problems relating to the use disciplines. With the aim of further illustrating how
of recycled and secondary materials is set out in Jefferis (2019). Environmental Geotechnics can play a prominent role in
Unexpected problems can, and often do, lead to litigation. safeguarding public health during the pandemic emergency, the
remaining part of this section is devoted to analysing a single
14.2 Emerging issues in Environmental Geotechnics: the issue in detail, namely the disposal of pathogen-laden municipal
COVID-19 pandemic solid wastes in sanitary landfills and the expected performance
of mineral bottom liners in preventing the migration of the
The world-wide spread of the Severe Acute Respiratory infectious pathogen to the surrounding geoenvironment.
Syndrome Coronavirus 2 (SARS-CoV-2), a highly infectious
and pathogenic strain of the newly emerged coronaviruses
SARS-CoV-1 and MERS-CoV, has become a major public 14.2.1 The disposal of SARS-CoV-2 contaminated solid waste
health concern in many countries, and has led governments to SARS-CoV-2 contaminated waste streams that do not undergo a
close national borders, limit the movement of citizens on national suitable treatment, aimed at reducing the viral load, such as
territories, implement quarantine measures and shut down incineration, sterilization by means of autoclave or high-
workplaces, retail activities, schools and universities. In addition temperature disinfection, pose severe threats to their safe
to the fatalities caused by the related disease (COVID-19), other disposal, both in the short-term for the employees who are
adverse consequences are currently being faced by countries involved in the management of solid waste and, hence, are
affected by the pandemic, such as financial instability and exposed to the contaminated bioaerosol (Carducci et al. 2013),
decline in the social order, thus evidencing that effective and in the long-term for the preservation of the quality of the
solutions to global-scale threats, like the global climate change groundwater, which may represent an effective vehicle for the
looming over the near future, can only be provided through transmission of the virus in the environment and, as the main
international cooperation, which implies the real-time source of drinking water, to humans. Because of its potentially
transmission of scientific information and research outcomes high viral load, healthcare, and medical waste (H&MW)
across borders (Tang et al. 2021). represents the contaminated waste stream of greatest concern,
Although the mechanisms by which the pathogenic agent is also considering that the number of existing medical
transmitted to human beings and animals, infects the living hosts incineration/disinfection facilities and regular waste-to-energy
and replicates within their cell tissues are primarily investigated plants may not be sufficient to cope with the exceptionally large
by microbiologists, epidemiologists and public health experts, volume of H&MW generated during the pandemic, so that
new questions have emerged during the emergency, whose relying on sanitary landfills becomes the only viable alternative.
answers can only be provided through a highly cross-disciplinary Similarly, waste streams that are generated from domestic and
approach (Paleologos et al. 2021; Tang et al. 2021; Vaverková et workplace garbage may also contain a non-negligible
al. 2021). For instance, the way viruses interact with and migrate concentration of the virus, due to the spread of the pandemic, and
through the soil environment is largely unknown, as well as how their management therefore requires the adoption of analogous
they can affect the physico-chemical properties of soils (e.g., precautions.
273
Modern engineered landfills are provided with bottom lining particles are also characterized by a negative charge under
systems, whose purpose is to facilitate the removal of the neutral conditions, the system remains stable and unfavourable
leachate, which accumulates in the drainage layer placed above to pathogen adsorption, unless an increase in electrolyte
the lining system, and to prevent migration of pollutants and concentration (Sadeghi et al. 2011) and/or the electrochemical
pathogenic microorganisms into the surrounding subsoil. Bottom valence of the cationic species that are dissolved in the pore
lining systems comprise a geomembrane overlying a low- solution (Sadeghi et al. 2013) cause a compression of the diffuse
permeability mineral layer which, in turn, consists of a double layers and a reduction in the electrostatic repulsive forces,
compacted clay liner (CCL) or a geosynthetic clay liner (GCL). thereby enabling the pathogen and mineral surface to approach
The assessment of the risk for human health that results from the each other as a result of the overwhelming van der Walls forces
migration of the virus requires the transport of the pathogen (Syngouna and Chrysikopoulos, 2010; Theng, 2012).
through the engineered barriers and the underlying aquifer to be The adsorption process described above should not be
modelled, and the virus concentration in the groundwater to be regarded as a complete immobilization of the virus, since a
predicted at a specified compliance point, which is usually perturbation in the chemical boundary conditions of the clay
downstream of the landfill (Dominijanni and Manassero 2021; barrier can cause a fast detachment of the adsorbed pathogen,
Guarena et al. 2020). e.g., because of dilution of the permeating solution. Such a
As an alternative approach, the “P-trap” has been used to likelihood of pathogen release from the soil particles poses the
manipulate redox conditions in drainage lines of mine storage issue of its survival time when attached to geologic media (i.e.,
heaps, so that temporary storage with the ability to alternate whether its inactivation rate is decreased or increased with
redox conditions (aerobic/anaerobic) may help to kill bacterial respect to the free-solution conditions). According to Harvey and
hosts for some viruses. Creation of anaerobic conditions by Ryan (2004), inactivation is slowed down by its attachment to
temporary storage in geomembrane bags has been used to kill the porous media with a high content of clay minerals, whereas an
cholera bacteria in the sewage from refugee camps. acceleration occurs when the virus is adsorbed on coarse-grained
soils with a high content of iron and aluminium oxides, which
14.2.2 Virus adsorption, inactivation, and transport through are characterized by an alkaline isoelectric point. In the former
landfill composite liners case, both the adsorbate and adsorbent carry a net negative
The migration of a virus through the geomembrane liner of a electrical charge and, thus, the interaction forces are relatively
composite liner is governed by the same mechanisms as those of weak: in such a case, the pleomorphic structure of
inorganic chemicals, as the relatively wide dimensions of viruses betacoronaviruses allows a relevant deformation of the protein
cause an extremely slow diffusion, if any at all, through the capsid to take place, so that it can adapt to the clay particle
polymer-based barrier and a preferential transport pathway surface, and at the same time provides high mechanical stability
through geomembrane defects (Shackelford 2014). Viruses may that preserves the integrity and infectiousness of the virus for a
undergo inactivation even before interacting with a clay barrier, longer period of time than in free-solution conditions (Block et
as specific viral components that are required for infection, such al. 2016). In the latter case, the net electrical charges that are
as the nucleic acid core and the outer protein capsid on which the carried by the adsorbate and adsorbent have opposite signs under
host-recognition sites are located, are naturally subjected to near-neutral conditions, so that the virus and mineral surface are
degradation. Moreover, the leachate itself may pose a harsh held by strong electrostatic attractive forces, which in turn lead
environment for the survival of viruses, as the mechanisms of to the disruption of the virus structure and its inactivation.
viral inactivation are accelerated upon an increase in temperature However, an inversion in the sign of the net electrical charge of
and a deviation from neutral pH conditions. In fact, the the virus can take place upon minor pH perturbations on the
microbially mediated consumption of organic matter determines acidic side, which might be encountered in landfills during the
the release of energy in the form of heat and the generation of anaerobic digestion of organic matter. Under such conditions, as
volatile organic acids, which are responsible for the relatively the intersurface potential energy becomes negative (attractive),
low pH during the early degradation stages of solid waste (Chen an enhanced virus uptake and an increased inactivation rate are
et al. 2019). On the basis of the results of an experimental study expected to occur when in contact with clay minerals (McLaren
aimed at measuring the survival of the H6N2 avian influenza and Peterson 1965). These favourable conditions for virus
virus, Graiver et al. (2009) calculated persistence times within adsorption and degradation on the clay particle surface are
methanogenic landfill leachate ranging from approximately 30 further enhanced when GCLs are used in place of CCLs in
days to more than 600 days and observed that the decay in the composite liners, since montmorillonite is able to maintain an
concentration of the infectious virus could be modelled as a first- approximately stable negative net electrical charge at low pH
order reaction, (i.e., the decline is exponential with time. values and is characterized by a higher cation exchange capacity
When the mass balance equation between the liquid and solid 2/g)
phases must be established, the ability of viruses to be reversibly 2/g)
adsorbed by the clay particles should be considered. Such (Lipson and Stotzky 1983; Theng 2012). Finally, the design of
attachment and detachment phenomena are governed by CCLs could be optimized to improve their containment
electrostatic forces, as the protein capsid of viruses carries a net performances against SARS-CoV-2 through the addition of
electrical charge, due to the ionization of carboxyl, amino and sandy soils with a high isoelectric point, such as magnetite sands,
other functional groups (Harvey and Ryan 2004). The net which might result in a larger virus adsorption and inactivation
electrical charge of the virus depends to a great extent on the pH capacity (Moore et al. 1981).
of the pore solution, and is therefore characterized by an The development of a theoretical model that can simulate the
isoelectric point (pI; i.e., the pH value at which the net electrical movement of viruses through clay-based barriers should account
charge is null). The number of deprotonated carboxyl groups for the typical dimensions of the considered pathogen. Indeed, in
above the pI exceeds the number of protonated amino groups, the case of SARS-CoV-2, the diameter of the virus is
and the sign of the virus surface charge is negative, whereas the approximately equal to 120 nm and, as such, is of the same order
opposite situation is encountered below the pI. On the basis of of magnitude as the width of the conductive pores (i.e., the void
the typical values that are assumed for pI, which usually fall space wherein the transport of solvent and solutes takes place) of
within the 3.5 to 7 range (Michen and Graule 2010; Scheller et the GCL bentonite layer, which usually ranges from a few tenths
al. 2020), viruses are negatively charged under neutral to several hundreds of nanometres (Manassero 2020).
conditions. This result has been confirmed for SARS-CoV-2 as Furthermore, several experimental studies have shown that the
a relatively high value of pI, equal to about 6.2, has been transport of inorganic electrolytes through montmorillonite-rich
measured, as reported by Calligari et al. (2020). Because clay
274
clay soils is partially restricted, due to the repulsive electrical Appropriate changes made today can accommodate the currently
forces that arise between the anionic species and the negatively predictable maximum probable flood but may need to be revised
charged clay particles (Musso et al. 2017; Dominijanni et al. in future periodic reviews if, for example, climate change is
2018). Thus, GCLs are expected to behave as semipermeable locally more significant than was anticipated.
membranes, with respect to negatively charged viruses, because The contaminating lifespan of even large MSW landfills,
of two different concurrent mechanisms, namely steric hazardous waste landfills, low-level radioactive waste, and many
hindrance, whereby the narrower bentonite pores are not tailings storage facilities and heap leach facilities could have
accessible to the large-sized pathogen, and electrical hindrance, been estimated using the known contaminants of concern last
whereby the wider bentonite pores can also impose a certain century (a known known based on the contaminants of known
degree of restriction on the migration of the pathogen because of concern). However, it was not practical to estimate a service life
the overlapping of the diffuse-ion swarms. of the geosynthetics used in these facilities in the last century. It
is now possible to make service life predictions and develop a
14.3 Pre-emptive consideration of performance and possible design that, based on what we know today, will have a service
consequences life greater than the contaminating lifespan. However, in the
brief 40-year history of environmental geotechnics it has been
Current environmental geotechnics is focused on design, shown that what we might have considered a reasonable
construction, and operation to ensure safety and environmental contaminating lifespan for the contaminants of concern in the
protection in the context of the known knowns. This is necessary, last century is no longer reasonable for the contaminants of
but not sufficient. Occasionally, consideration is given to the emerging concern discussed in §6.
known unknowns. No consideration is given to the unknown Many landfills produce annual monitoring reports
unknowns because they are unknown at the time a facility is documenting leachate collection, certain aspects of leachate
designed and approved. Unfortunately, in most jurisdictions chemistry, groundwater monitoring data from discrete generally
once a landfill has been approved despite it then being developed widely spaced monitoring wells around the facility, and if there
over many years (often decades), there is no reconsideration of is a leak detection/secondary leachate collection/groundwater
the design when new knowledge (e.g.., better insight into what control layer monitoring data from that layer. These reports
were previously known unknowns or the revelation of previously generally do not consider the big picture questions or what
unknown unknowns) affects the likely performance of what has factors are likely to affect performance over the next decade.
been approved. There is generally no systematic periodic re- Also, except for the data from a leak detection/ secondary
evaluation of the approved facility. However, the first 20 years leachate collection/groundwater control layer, the data collected
of the 21st century have shown that this approach is generally provides very little insight as to the actual performance
unsatisfactory (e.g., see Sections 5 and 6 of this paper). There is of the facility. At best they may detect an unexpected leakage but
a need for a systematic approach to managing over time the even this is doubtful for a single lined landfill monitored by
known unknowns and the unknown unknowns when the facility discrete widely distant monitoring wells. This is particularly
was approved. Perhaps the branch of civil engineering that likely for closed landfills, tailings storage, and mine waste rock
comes closest to meeting this objective the engineering of large disposal facilities. Thus, environmental geotechnics needs to
dams where monitoring and maintenance are embedded in dam face the challenge of finding better approaches to the monitoring
systems engineering. and periodic review of the performance of these large potential
A large dam may have a conceptual design life of 100 years sources of contamination with a very long contaminating
but is unlikely to be decommissioned, if it is playing an essential lifespan.
role in flood control and/or generation of green electricity.
Regulators commonly expect periodic dam safety reviews that
implicitly recognize that dams require continuing care and
maintenance and during their service life may have to deal with
15 CONCLUDING COMMENTS
known unknowns and unknown unknowns at the time they were
constructed. One of the known unknowns when most dams were In the coming decades, geoenvironmental engineers can expect
designed may have been the service life of engineered to have to design, construct, manage, and live with some of the
components, (e.g., spillway gates, spillway, the concrete dam most radical changes in every aspect of life including energy and
wall, an embankment dam subject to predicted but at the time raw materials production, waste minimisation, water shortages,
otherwise unknown settlement). An unknown unknown at the infrastructure resilience, transport systems, land use, and more
time most existing dams were constructed was the extent to frequent flooding and failures due to climate change and systems
which human use of fossil fuels over the last century would interdependence and interconnectedness.
impact the global climate resulting, for example, in more One of the many consequences of climate change will be the
frequent extreme weather events. However, a practice of need for geoenvironmental and geotechnical engineers to follow
conducting periodic reviews of large dams allows a periodic the lead of dam engineers in performing a potential failure modes
evaluation or re-evaluation of potential failure modes and effects and effects analysis for all significant geotechnical and
including but not limited to the possible: geoenvironmental engineering projects. In the third decade of the
(a) consequences due to components of the system 21st century it is no longer sufficient to ensure an adequate factor
approaching the end of their service life, and of safety for slopes and bearing capacity. While this is still
(b) changes to environmental conditions such as climate necessary, we also need to consider the potential unintended
change. consequences of the proposed facility, including environmental
This allows initiation of pre-emptive repair, or replacement, impacts, and identify all knowable/evolving unknowns relevant
or whatever other action is required to maintain minimal danger to proposed undertaking. The next step must be to develop
to human life, property, and the environment. The service life appropriate mitigative strategies for all identified failure modes
may have been a known unknown at the time of construction but and address the known unknowns either directly or by
will become evident from monitoring over time. The rapid monitoring and contingency plans. Given that there will always
impact of the human use of fossil fuels on changing climate and be unknown and unknowns, once constructed, the project should
hence the impact of climate change on the maximum probable be periodically reviewed to identify any deviations from the
flood was an unknown unknown at the time many dams were intended behaviour and identification of new issues that were
designed and built but it can be addressed by appropriate previously unknown. This should be followed by the
upgrades, changes to operating rules, and or infrastructure. implementation of appropriate action to address those changes.
275
It is recognised that all this will require funding streams to be bentonite, the subgrade pore water and permeant chemistry,
available throughout a project’s life and beyond, until it is applied stress, and temperature. k can also be substantially
decommissioned. The cost of this post-closure monitoring and affected by cation-exchange associated with hydration
ongoing maintenance and repairs must be included in the initial followed by wet-dry and freeze-thaw cycles.
business case. It is also recognized that geoenvironmental 9. A useful way of screening GCLs with bentonites from
engineers must be appropriately trained to meet the coming different sources and selecting the most appropriate
challenges. candidates for a given project is to perform a modified swell
This paper has touched on current advances and challenges index test replacing the usual distilled water by either the
on all continents and highlighted the new issues and materials actual or simulated fluid to be contained by the GCL. The
that have emerged in the first two decades of the century. A bentonite(s) with the largest swell index in this modified test
summary of key observations and findings is provided below. are the most likely to be a suitable for the testing indicated
1. Geosynthetics play a major role in many geoenvironmental below. This modified swell index test can be part of a
projects. Because they are easy to use, there is a great specification and can be more useful in CQA than the
tendency to select the cheapest material and have it laid down standard test.
by unskilled labour. There is a prevalent lack of 10. Most designers pay little attention to the geotextiles, the type
understanding of appropriate design and construction of bentonite, the amount of needle-punching (especially
evident in many applications. Geosynthetics need more bundle size), or the mass of bentonite in a GCL in the
attention in undergraduate education by integration, where selection of an appropriate GCL. This is because they have
appropriate, into existing courses and in professional no effect on the GCL hydraulic conductivity when the GCL
development of graduated engineers. is fully hydrated and permeated with distilled water at 35 kPa
2. There are many different resins, antioxidants/stabilizers, and and it is this hydraulic conductivity that is placed on the
carbon blacks that can be used in polyethylene and manufacturer’s GCL product sheets. However, this test bears
polypropylene geosynthetics. The time to nominal failure no similarity to most real field situations and the
and service life of these geosynthetics will be highly corresponding hydraulic conductivity given on the
dependent on compatibility of their constituents with its manufacturer’s GCL product sheets is not a design
chemical, thermal, and UV exposure as well as tensile parameter.
stresses. This can only be assessed by performance-based The in-field hydraulic conductivity is highly dependent (by
testing. Index tests are useful for confirming that the material up to 4 to 5 orders of magnitude) on the aforementioned
delivered to site is consistent with specified properties (i.e., parameters. Therefore, the design hydraulic conductivity for
for quality control and CQA), but they do not provide design a GCL must be established by performing appropriate tests
parameters. on the candidate GCL(s) hydrated with actual or simulated
3. HDPE, LLDPE, or PVC geomembranes are extensively used subgrade pore fluid or, if that is unknown, 10 mmol CaCl2
in fluid containment applications. All three materials will solution, and permeated with either the actual or simulated
ultimately become brittle and fail to maintain their function fluid to be contained. The test should be conducted at
as a hydraulic barrier when they crack. One important representative stress levels. Thus, if the GCL is to be used in
element in the life of polyethylene is the loss of the efficacy a composite liner, a test will be required at both a very low
of antioxidants whereas for PVC it is the leaching of stress (3 to 5 kPa) to simulate conditions below a
plasticizers. This needs to be better understood and geomembrane wrinkle and at the anticipated normal stress
considered in design. where the geomembrane is in direct contact with the GCL. If
4. When designing with HDPE geomembranes, the most wet-dry or freeze-thaw cycles are anticipated, then
important design parameter is the stress crack resistance. appropriate tests should be conducted on specimens with the
However, initial values that are measured, SCRo, are expected cation-exchange to assess the effect of wet dry or
generally not representative of the true material stress crack freeze-thaw cycling on hydraulic conductivity and
resistance since it is influenced by residual stresses geomembrane-clay liner interface transmissivity.
developed during manufacture. With time, these stresses will 11. The short-term hydraulic performance of GCLs permeated
relax and the stress crack resistance will decrease to a by aggressive solutions (e.g., very high or low pH, or high
representative value SCRm. Available data indicate that concentrations of cations, organic solutions) can be
SCRm has an average value of about 37% of the initial value substantially improved by adding polymers to the bentonite
with most values falling between 27% and 47% of the initial to create an EB-GCL. However, insufficient consideration
off-roll value. The actual value can only be estimated based has been given to the potential consequences of polymer
on testing that will take at least 3 months. biodegradation, polymer leaching, and polymer elution. If
5. Depending on the constituents other than the base resin, PVC polymer elution occurs, then open pores will remain within
may uptake moisture which results in significant change in the bentonite and result in a substantially higher long-term
its properties with time. Again, this can only be assessed by hydraulic conductivity of EB-GCLs. Even if elution does not
testing over a period. occur, water soluble polymers will be leached over time
6. New resins and products are coming onto the market and the under field conditions, and they may be biodegraded.
popularity of multi-layered geomembranes is increasing. Polymer elution also reduces interface shear strength. Until
Multi-layered geomembranes have notable advantages but these issues are resolved, extreme caution is advised in the
the potential impact of those layers on long-term use of EB-GCLs, and such use should only be considered
performance requires much more consideration than after a careful failure modes and effects analysis and
received to date. implementation of a plan for monitoring longer-term field
7. Bituminous geomembranes have proven to be very effective performance.
in canals and in the retention of potable water. There is a 12. Applying a polypropylene or polyethylene coating or
growing use of them as liners in mining applications and laminate to a traditional sodium bentonite GCL can also
covers. Care is needed to evaluate their suitability for any substantially improve the short-term hydraulic performance
given project. Problems have been observed at both high and of the GCL. Very little is known about the long-term
low temperatures and with seams. performance of these coatings.
8. The hydraulic conductivity, k, of a GCL after several years 13. Combining a geomembrane and GCL (or CCL) to form a
in a field application is highly dependent on the nature of the composite liner introduces an important parameter: the
276
GMB-GCL or GMB-CCL interface transmissivity, which geomembrane is an excellent barrier to all these compounds.
defines the ability of fluid to spread laterally at the interface However, with typical construction quality assurance, there
contact between the two materials. is high probability that the leakage of many of these
14. The design objective for a barrier system should be to ensure contaminants through a single composite liner will exceed
optimal performance of the entire system, not the allowable levels in the environment unless one or more of
components in isolation. More generally, the potential the following apply.
interaction between the soils or other materials above or (a) there was an appropriate electrical leak location survey
below a component in a liner system must be considered in before and after the geomembrane was covered,
any design. (b) there is a high-level of hydrogeologic protection and
15. Unfortunately, it is not uncommon for a designer to specify attenuation, or
an appropriate GCL considering all the factors above and (c) there is an extra level of engineering in the form of a
then, during construction, the contractor substitutes an secondary leachate collection/ leak detection and liner
alternative GCL to the one that was specified. The alternative system.
is usually proposed because it is cheaper (i.e., it has less mass 20. The impacts associated with the contaminants of emerging
per unit area of geotextile, and/or GCL coating, and/or concern can be kept to an acceptable level for new facilities
bentonite mass per unit area and/or an inferior bentonite) and that satisfy (a) and either (b) or (c) in §15.19 above.
has the same hydraulic conductivity based on the 21. The design life of ponds and lagoons may be a few decades.
manufacturer’s product information sheet. This can lead to The design life of a liner for a dam is commonly about a
failure because the cheaper product is not equivalent in terms century. The contaminating lifespan of a waste disposal
of field performance and the actual mobilized k in the field facility depends on its size (predominantly the thickness of
can be orders of magnitude higher than was given on the waste) and the nature of the waste. It typically ranges from
manufacturer’s specification sheet, with the failure often about 100 to 300 years for municipal solid waste to over 500
occurring a few years after construction and initial years for low-level radioactive waste, to millennia for some
operations for ponds/lagoons, 5 to 10 years for a cover mine waste and long-lived low-level radioactive waste.
system, and several to decades for landfill bottom liners. Thus, the service life of the barrier system components is a
Most of the liability for any such failure will likely land with critical consideration since the system requires a service life
whomever approved the substitution of the alternative greater than the contaminating lifespan or design life as
material for the originally specified material. appropriate.
A similar situation arises for geomembranes to be used in 22. Geosynthetics are versatile and cost-effective materials that
either severe conditions (e.g., aggressive chemical contact or can serve many functions at a mine-site. However, due to the
elevated temperatures) or requiring a long service life. characteristics of the mining activity, appropriate selection,
No substitution of a material that does not fully meet the and specification of geosynthetic products is of utmost
specifications should be permitted without the express importance. 1.5 mm- or 2 mm-thick LLDPE and HDPE have
written approval of the designer of record. found extensive use as liners for heap leach pads and more
16. Care is needed in handling and storing geosynthetics prior to recently for liners and/or covers in tailings storage facilities.
installation. Left exposed to the sun, material rolls can heat In both cases, it is important to design to minimize
up to 60°C in Canada and 75°C in South Africa. A few holes/punctures in the geomembrane in the short-term. Since
months exposure at these temperatures can substantially many of these facilities will require containment for many
reduce the time to nominal failure and ultimately the service decades, if not centuries, it is as important to select the right
life of the material, particularly geotextiles and geonets that materials for the given exposure conditions and design to
are generally not as well stabilized as geomembranes. minimize tensile strains in the geomembrane.
17. On projects where the barrier system plays a critical role, the 23. Geotextile used as filters in landfills, dams and mining
risks of problems during construction can be minimized by applications need to be designed with careful consideration
the construction of trial sections during the early phase of of the potential for particulate, chemical or biologically
construction. This will allow identification of unsuitable induced clogging
materials and construction methods, making all parties A significant increase in the use of geotextile tubes to
familiar with the issues that are going to require careful dewater mining tailings is foreseen in the coming years.
attention during the full-scale construction. The work should 24. Subject to appropriate consideration of the issues raised in
be undertaken under supervision of the CQA consultant. this paper including the effects of exposure conditions and
Once the test section has been constructed, it should then be the selection of compatible materials, the service life of a
exhumed and carefully inspected. Lessons learned need to be composite liner will be equal to the shorter of the service life
documented and implemented in the construction of the of the geomembrane or the GCL. Thus, with appropriate
facility. In barrier system construction: ‘you get what you design (including the selection of materials) and
inspect, not what you expect’. The CQA should be conducted construction, a 1.5 mm- or 2 mm-thick HDPE underlain by a
by an appropriately qualified expert independent of the GCL can provide very long service life and is suitable for use
contractor who should have a budget and enough staff to in dams, ponds, or lagoons, or in covers for closure of
visually monitor all stages of the construction with backup landfills or mine waste facilities, or as a bottom liner for
staff to allow time to do the paperwork. landfills or mine waste facilities. Depending on the exposure
18. Based on available data from monitoring leakage through a conditions, the likely achievable service life may range from
primary composite liner in 122 double-lined landfill cells, decades to millennia.
with good quality assurance but no leak location survey, the 25. Slurry trenching and allied techniques to form vertical cut-
most likely leakage is 70 litres per hectare per day (lphd) off walls are mature technologies with long and successful
with a 66% probability that the leakage will lie between 40 track records.
and 100 lphd and a 95% probability that the leakage will 26. The current generation of these cut-off walls is based on
exceed 5 lphd but be less than 430 lphd. cementitious systems and their properties such as strength
19. Most landfill regulations were developed to contain the and hydraulic conductivity are well established. If a project
contaminants of concern in the last century. Many new requires material properties such as special stress-strain
contaminants of concern have arisen this century. Examples behaviour or unusual chemical resistance beyond current
include PBDEs, BPA, and PFAS. An intact 1.5 mm HDPE experience, then it may be wise to look to other
277
systems.Developing and proving new systems can take years §14: Andrea Dominijanni and R .Kerry Rowe
and certainly will not be possible within a typical contract §15: R. Kerry Rowe and Stephan Jefferis
tender period. As far as possible, the lead authors have endeavoured to revise,
27. New systems will evolve because of innovation but also may and integrate the various contributions to provide a consistent
be driven by changes in material availability (pulverised fuel narrative. Opinions presented within the paper are those the lead
ash and ground granulated blastfurnace slag) and the drive to authors and not necessarily shared by all those who contributed to
reduce carbon emissions (low carbon cements). the paper.
28. The use of sorbents within cut-off walls can delay The authors are grateful for the careful reviews of portions of
contaminant breakthrough. However, unless the sorbed this paper conducted by Anna Norris of Colorado State University
species are degraded within the barrier they are merely stored and Nicolò Guarena of the Politecnico di Torino, and for the
and can be expected to be eluted in due time. Sorption is a contributions from Michela De Camillis and Khizar Khan at Ghent
competitive process and in most cases a concentration University.
dependent process. Contaminants may be desorbed much The proof-reading skills of Drs. F. Abdelaal, A. Bouazza, A. V.
earlier than predicted from studies using just a single DiBattista, Dominijanni, K. Legge, E. Palmeira, J. Scheirs, C.
contaminant or if the chemistry of the leachate changes. Shackelford, and Queen’s University PhD students: Ahmad Al
29. Over time, substantial quantities of contaminants can build Shafie, Antara Arif, Farah Barakat, Mathew Clinton, Yunhui Fan,
within a sorbent system. If contaminants are to be retained YuYan Li, Samuel Makinde, Juliana Reinert, Rodrigo e Silva,
within a barrier, then provision must be made for their Jose e Silva, Michael Somuah, and Mohsen Zafari are greatly
management if/when the barrier is decommissioned. appreciated.
Jefferis (2013) gives an example of arsenic accumulation in The assistance of Dr. Natalie Touze with the French abstract is
iron slime in the ballast of an underground railway track. also very greatly appreciated.
Significant quantities of arsenic accumulated despite the
probably very low concentrations the permeating water.
30. The extensive geoenvironmental literature on recycling
materials is often focused on the sustainability benefits.
While recycling is a notable objective, we need to be vigilant
in avoiding unintended consequences. Consideration must
not only be given to the benefits but to the many questions
that arise. For example, in the proposed application for a
recycled material is it chemically and mechanically stable?
Does it have the potential for shrinkage, expansion or
damage to structures formed from them or built upon them?
Will it emit toxic, explosive or asphyxiant gases? Who will
bear the risks relating to possible in-service failure?
31. As use of fossil is phased out there will be a need to store
wind and solar energy that is prompting interest in the
construction of pump storage schemes including the use of
old mines. It also makes the of new hydro-electric schemes
more economic. Liner systems will likely play an
increasingly important role in these developments.
32. Environmental geotechnics needs to face the challenge of
finding better approaches to the monitoring and periodic
review of the performance of large potential sources of
contamination (e.g., large landfills, tailings storage facilities)
with a very long contaminating lifespan.
16 ACKNOWLEDGEMENTS
The drafts of different sections of this paper were prepared by
groups of authors and edited by R.K. Rowe and S. Jefferis. The
authors of the initial drafts of the discrete sections are as follows:
§1: R. Kerry Rowe and Stephan Jefferis
§2: John Scheirs, Fady B. Abdelaal, and R. Kerry Rowe
§3: Fady B. Abdelaal
§4: Charles D. Shackelford, Joseph Scalia IV, Gemmina Di
Emidio, Mario Manassero, Craig H. Benson, and Takeshi
Katsumi
§5: R. Kerry Rowe
§6: R. Kerry Rowe and Malek Bouazza
§7: Kelvin Legge, Ennio Palmeira, and Tony Zhan
§8: Ennio M. Palmeira, Maria G. Gardoni and Denys Parra
§9: R. Kerry Rowe
§10: Tony Zhan, Stephan Jefferis, Evelina Fratalocchi, and
Jeffrey C. Evans
§11: Stephan Jefferis
§12: Stephan Jefferis
§13: Stephan Jefferis
278
17 APPENDIX: NOTATION
AL Attenuation layer MFI
Melt Flow Index (also known as Melt Flow Rate)
(MFR)
BFE Blown Film Extrusion MSB Multiswellable bentonite
BPA BisPhenolA MSW Municipal solid waste
BPC Bentonite polymer composite MWD Molecular Weight Distribution
BTEX Benzene, Toluene, Ethylbenzene, Xylenes N Total porosity of the bentonite
CaB Calcium bentonite NaB Sodium bentonite
CBSC Capillary Barrier Soil Cover Nl,AV, Average number of layers or lamellae per tactoid
ci Molar concentration of ionic species i NMR Nuclear magnetic resonance
CCL Compacted Clay Liner OIT Oxidative Induction Time (often used for Std-OIT)
COC Cyclic organic carbonate compounds P Probability leakage will exceed the specific Q
cp Peak concentration in the aquifer for the PFOS PBDE PolyBrominated Diphenyl Ether
leakage case examined
DTP Diene terpolymer (additive to increase SCR) PC Propylene carbonate
D* Osmotic effective diffusion coefficient PCB PolyChlorinated Biphenyls
E Void ratio of the bentonite = n/(1 – n) PERT Polyethylene for Raised Temperature
EB Enhanced bentonite PFAS Per- and PolyFluoroAlkyl Substances
EB- Enhanced Bentonite GCL PFOA PerFluoroOctanoic Acid
GCL
FBE Flat Bed (also known as slot die or flat die) extrusion PFOS PerFluoroOctane Sulfonate
Flow Rate Ratio (ratio of MFI at two loads; typically Pg Permeation coefficient through a geomembrane (Pg =
FRR
21.6 and 2.16 kg) Dg x Sgf)
FTIR Fourier-transform infrared POP Persistent Organic Pollutant
GC Glycerol carbonate Ppb Parts per billion
GCD GeoComposite Drain PVF Pore volumes of flow
GCL Geosynthetic clay liner Q Leakage through a hole in a wrinkle (lphd)
GMB Geomembrane QUELTS Queen's University Environmental Liner Test Site
GTX Geotextile RSA Republic of South Africa
HAS Hindered Amine Stabilizers (also known as
(HALS) Hindered Amine Light Stabilizers)
SAP Ssodium polyacrylate
HC HYPER clay SCR Stress Crack Resistance
HDPE High Density Polyethylene SCRo Initial Stress Crack Resistance
HFWC High-Food-Waste-Content SCRm Representative Stress Crack Resistance
OIT test performed at 150oC as per ASTM D5885 to SEM Scanning electron microscopy
HP-OIT
detect retained stabilizer content.
2
I Ionic strength = izi summed over all ionic species Sgf Partitioning coefficient
K Hydraulic conductivity OIT test performed at 200oC as per ASTM D3895 to
Std-OIT
detect retained antioxidants.
kw Hydraulic conductivity with respect to water NCTL Single Point-Notched Constant Load Test
LCS Leachate Collection System UK United Kingdom
LDS Leak detection system USA Unied States of America
LFG LandFill Gas UV Ultra-violet radiation
LFWC Low-Food-Waste-Content XRD X-ray diffraction
LLDPE Linear Low Density Polyethylene GMB-GCL interface transmissivity (m2/s)
lphd Litres Per Hectare per Day m Dimensionless matrix tortuosity factor
Lw Length of a wrinkle with a hole per hectare for the i Membrane efficiency coefficient
PFOS leakage case examined
MDL Method Detection Limit zi Charge of ionic species i
MDPE Medium Density Polyethylene § Section of this paper
279
18 REFERENCES ASTM D5101 2017. Standard test method for measuring the filtration
compatibility of soil-geotextile systems clogging potential, ASTM
Abdelaal, F.B., and Solanki, R., 2021. Shear strength of the International, West Conshohoken, Pennsylvania, USA.
geomembrane–subgrade interface in heap leaching applications. ASTM D5397. Standard test method for evaluation of stress crack
Environmental Geotechnics, 40:1-17. resistance of polyolefin geomembranes using notched constant
Abdelaal, F.B., Rowe, R.K., and Brachman, R. 2014. Brittle rupture of tensile load test, ASTM International, West Conshohoken,
an aged HDPE geomembrane at local gravel indentations under Pennsylvania, USA.
simulated field conditions. Geosynthetics International 21(1), 1-23. ASTM D5514. Standard test method for large-scale hydrostatic puncture
Abdelaal, F. Rowe, R.K., Hsuan, Y., and Awad, R. 2015. Effect of testing of geosynthetics, ASTM International, West Conshohoken,
high temperatures on the physical and mechanical properties of Pennsylvania, USA.
HDPE geomembranes in air. Geosynthetics International 22(3), 207- ASTM D5885. Standard test method for oxidative induction time of
224. polyolefin geosynthetics by high pressure differential scanning
AbdelRazek, A., and Rowe, R.K. 2019. Performance of GCLs in high calorimetry, ASTM International, West Conshohoken, Pennsylvania,
salinity impoundment applications. Geosynthetics International USA.
26(6), 611-628. ASTM D5890 2019. Standard test method for swell index of clay mineral
Acikel, A., Bouazza, A., Gates, W., Singh, R., and Rowe, R.K. 2020. A component of geosynthetic clay liners, ASTM International, West
novel transient gravimetric monitoring technique implemented to Conshohoken, Pennsylvania, USA.
GCL osmotic suction control. Geotextiles and Geomembranes ASTM D6693. Standard test method for determining tensile properties
48(6),755-767. of non-reinforced polyethylene and nonreinforced flexible
Acikel, A., Bouazza, A., Singh, R. M., Gates, W. P. and Rowe, R.K. 2022. polypropylene geomembranes, ASTM International, West
Challenges of the filter paper suction measurements of geosynthetic Conshohoken, Pennsylvania, USA.
clay liners: method, time, capillarity, and hysteresis. Geotechnical ASTM D6766 2020. Standard test method for evaluation of hydraulic
Testing Journal (in press). properties of geosynthetic clay liners permeated with potentially
Acikel, A., Gates, W. , Singh, R., Bouazza, A., Fredlund, D., and Rowe, incompatible aqueous solutions, ASTM International, West
R.K. 2018b. Time-dependent unsaturated behaviour of geosynthetic Conshohoken, Pennsylvania, USA.
clay liners. Canadian Geotechnical Journal 55(12), 1824-1836. ASTM D8117. Standard test method for oxidative induction time of
Acikel, A., Gates, W., Singh, R., Bouazza, A., and Rowe, R.K. 2018a. polyolefin geosynthetics by differential sanning calorimetry, ASTM
Insufficient initial hydration of GCLs from some subgrades: factors International, West Conshohoken, Pennsylvania, USA.
and causes. Geotextiles and Geomembranes 46, 770-781. ASTM E1739 . Standard guide for risk-based corrective action applied
Acikel, A. , Rowe, R.K., and McWatters, R. 2021. Unsaturated behaviour at petroleum release sites, ASTM International, West Conshohoken,
and cation self-diffusion of geosynthetic clay liners in Antarctica. Pennsylvania, USA.
Geosynthetics International 28(4), 350-367. ASTM 2000. E2081 – 00. Standard guide for risk-based corrective action,
Acikel, A., Singh, R., Bouazza, A., Gates, W., and Rowe, R.K. 2015. Technical report, West Conshohocken, PA.
Applicability and accuracy of the initially dry and initially wet Athanassopoulos, C., Benson, C., Donovan, M., and Chen, J. 2015.
contact filter paper tests for matric suction measurement of Hydraulic conductivity of a polymer-modified GCL permeated with
geosynthetic clay liners. Géotechnique 65, 780-787. high pH solutions. Geosynthetics 2015, IFAI, Roseville, Minnesota,
Addis, P., Andruchow, B., and Wislesky, I. 2016. Bituminous USA, 181-186.
geomembrane failure at a co-disposal tailings storage facility. Proc. Azad, F., El-Zein, A., Rowe, R.K., and Airey, D. 2012. Modelling of
Tailings and Mine Waste, Banff, AB, Canada, 9 p. thermally induced desiccation of geosynthetic clay liners in double
Adshead, D., Román García, O., Thacker, S. and Hall, J.W. 2021. composite liner systems. Geotextiles and Geomembranes 34, 28-28.
Infrastructure strategies for achieving the global development Azad, F., Rowe, R.K., El-Zein, A., and Airey, D. 2011. Laboratory
agenda in small islands. Earth's Future, 9(2). investigation of thermally induced desiccation of GCLs in double
Ali, M., Bouazza, A. Singh, R., Gates W., and Rowe, R.K. 2016. composite liner systems.Geotextiles and Geomembranes 29(6),534-
Thermal conductivity of geosynthetic clay liners. Canadian 543
Geotechnical Journal 53(9), 1510-1521. Bahram, M., Mohseni, N., and Moghtader, M. 2016. An introduction to
Alphabet Inc. 2019. Google Earth image, https://bitly.com/ hydrogels and some recent applications. Emerging Concepts in
Al-Tarawneh, K., Buzzi, O., Krabbenhøft, K., Lyamin, A., Sloan, S. 2009. Analysis and Applications of Hydrogels, Majee, Ed., IntechOpen.
An indirect approach for correlation of permeability and diffusion DOI: 10.5772/64301.
coefficients. Defect and Diffusion Forum 283-286, 504-514. Ball, J. 2002. Milestones towards minimum requirements. Proc.
Anderson, R., Rayhani, M., and Rowe, R.K. 2012. Laboratory International Waste Congress IWMSA, August, Durban, South
investigation of GCL hydration from clayey sand subsoil. Africa.
Geotextiles and Geomembranes 31, 31-38. BAM 2015. Guidelines for the certification of protection layers for
Ashe, L. Rowe, R.K., Brachman, R.W.I. and Take, W.A. 2014. geomembranes in landfill sealing systems, Issued by Division 4.3,
Laboratory simulation of bentonite erosion by downslope flow of a Contaminant Transfer and Environmental Technologies, Germany.
GCL, ASCE Journal of Geotechnical and Geoenvironmental Bannour, H., Barral, C., and Touze-Foltz, N. 2013. Flow rate in
Engineering, 140 (8): 04014044-1 to 04014044-9 DOI: composite liners including GCLs and bituminous geomembranes.
10.1061/(ASCE)GT.1943-5606.000114 3rd International Conference on Geotechnical Engineering
Ashe, L. Rowe, R.K., Brachman, R., and Take, W.A. 2015. Laboratory (ICGE'13), Feb. 21-23, Hammamet, Tunisia, 21-23.
study of downslope erosion for ten different GCLs. Journal of Barr, K.M.H., Berry, C.W. and Parker P.J. 2015. The use of a composite
Geotechnical and Geoenvironmental Engineering 141(1), 04014079. HDPE membrane/ bentonite-cement slurry trench cut-off at
Ashmawy, A., El-Hajji, D., Sotelo, N., and Muhammad, N. 2002. Broadwood Loch, Cumbernauld, In The prospect for reservoirs in
Hydraulic performance of untreated and polymer-treated bentonite the 21st Century, Tedd, P. (ed), ICE Publishing.
in inorganic landfill leachates. Clays and Clay Minerals 50(5), 546- Barlow L, Malusis, M. 2018. Assessment of backfill hydraulic
552. conductivity in an instrumented soil- bentonite cut-off wall. Faculty
ASTM D3895. Standard test method for oxidative induction time of Conf. Papers and Presentations. 46.
polyolefinsby differential scanning calorimetry, ASTM International, Barroso, M., Touze-Foltz, N., and Saisi, F. 2006. Validation of the use of
West Conshohocken, Pennsylvania, USA. filter paper suction measurements for the determination of GCL
ASTM D4318 2017. Standard test method for liquid limit, plastic limit, water retention curves. Proc., 8th International Conference on
and plasticity index of soils, ASTM International, West Geosynthetics, Millpress, Rotterdam, The Netherlands, 171-174.
Conshohoken, Pennsylvania, USA. Beck, A. 2015. Available technologies to approach zero leaks.
ASTM D4703. Standard practice for compression molding thermoplastic Geosynthetics 2015, Feb. 15-18, 2015, Portland, Oregon, IFAI,
materials into test specimens, plaques, or sheets, ASTM Roseville, Minnesota, USA.
International, West Conshohoken, Pennsylvania, USA. Beddoe, R., Take, W., and Rowe, R.K. 2010. Development of suction
ASTM D4839 2017. Standard test method for total carbon and organic measurement techniques to quantify the water retention behaviour
carbon in water by ultraviolet, or persulfate oxidation, or both, and of GCLs. Geosynthetics International 17(5), 301-312.
infrared detection, ASTM International, West Conshohoken,
Pennsylvania, USA.
280
Beddoe, R., Take, W., and Rowe, R.K. 2011. Water retention behaviour Bouazza, A., Ali, M., Gates W., and Rowe, R.K. 2017a. New insight on
of geosynthetic clay liners. Journal of Geotechnical and geosynthetic clay liner hydration: the key role of subsoils
Geoenvironmental Engineering 137(11), 1028-1038. mineralogy. Geosynthetics International 24(2),139-150.
Beirigo, E. 2005. Filter behaviour of geotextiles in tailings dams, MSc. Bouazza, A., and Bowders, J., 2010. Geosynthetic clay liners for waste
thesis, Graduate Programme of Geotechnics, University of Brasilia, containment facilities, CRC Press Balkema, Leiden, The
Brasilia, Brazil (in Portuguese). Netherlands.
Bellomo, N., and Preziosi, L. 1995. Modelling Mathematical Methods Bouazza, A., and Gates, W. 2014. Overview of performance
and Scientific Computation, CRC Press, USA. compatibility issues of GCLs with respect to leachates of extreme
Benson, C., Chen, J., and Edil, T. 2014. Engineering properties of chemistry. Geosynthetics International 21(2), 151-167.
geosynthetic clay liners permeated with coal combustion product Bouazza, A., Jefferis, S., and Vangpaisal, T. 2007. Investigation of the
leachates, Report No. 3002003770, Electric Power Research effects and degree of calcium exchange on the Atterberg limits and
Institute (EPRI), Palo Alto, California, USA. swelling of geosynthetic clay liners when subjected to wet–dry
Benson, C., Jo, H., and Musso, T. 2015. Hydraulic conductivity of cycles. Geotextiles and Geomembranes 25(3), 170-185.
organoclay and organoclay-sand mixtures to fuels and organic Bouazza, A., Rouf, M., Singh, R., Rowe, R., and Gates W. 2017b. Gas
liquids. Journal of Geotechnical and Geoenvironmental Engineering diffusion and permeability for geosynthetic clay liners with powder
141(2), 04014094. and granular bentonites. Geosynthetics International 24(6), 604-614.
Benson, C., Kucukkirca, I., and Scalia, J. 2010a. Properties of Bouazza, A., Singh, R., Rowe, R., and Gassner. F. 2014. Heat and
geosynthetics exhumed from a final cover at a solid waste landfill. moisture migration in a geomembrane-GCL composite liner
Geotextiles and Geomembranes 28(6), 536-546. subjected to high temperatures and low vertical stresses. Geotextiles
Benson, C., and Meer, S. 2009. Relative abundance of monovalent and and Geomembranes 42(5), 555-563.
divalent cations and the impact of desiccation on geosynthetic clay Bouazza, A., Vangpaisal, T., and Jefferis, S. 2006. Effect of wet-dry
liners. Journal of Geotechnical and Geoenvironmental Engineering cycles and cation exchange on gas permeability of geosynthetic clay
135(3), 349-358. liners. Journal of Geotechnical and Geoenvironmental Engineering
Benson, C., Oren, A., and Gates, W. 2010b. Hydraulic conductivity of 132(8), 1011-1018.
two geosynthetic clay liners permeated with a hyperalkaline solution. Brachman, R., and Gudina, S. 2008. Geomembrane strains and wrinkle
Geotextiles and Geomembranes 28(2), 206-218. deformations in a GM/GCL composite liner. Geotextiles and
Berube, D., Diebe, P., Rollin, A., and Stark, T. 2007. Massive mining Geomembranes 26(6), 488-497.
evapouration ponds. Geosynthetics, IFAI, Roseville, Minnesota, Brachman, R., Rentz, A., Rowe, R.K., and Take W.A. 2015.
USA, 27-33. Classification and quantification of downslope erosion from a GCL
Block, K.A., Katz, A., Alimova, A., Trusiak, A., Morales, J., Wei, H., when covered only by a black geomembrane. Canadian
Bucher, D., and Gottlieb, P. 2016. Montmorillonite-mediated Geotechnical Journal 52(4), 395-412.
aggregation induces deformation of influenza virus particles. Brachman, R.W.I., Rowe, R.K. and Irfan 2014. Short-Term Local Tensile
Applied Clay Science 124-125, 211-218. Strains in HDPE Heap Leach Geomembranes from Coarse Overliner
Blond, E. and Breul, B. 2014. Behaviour of bituminous geomembranes Materials. ASCE J. of Geotech. and Geoenvir. Engineering,
under elevated-compression stress. Proc. Geosynthetics Mining 140(5): 04014011-1 to 04014011-8.
Solutions, Vancouver, British Columbia, Canada, 325-337. Brachman, R.W.I, Joshi, P., and Rowe, R.K. 2016. “Hydraulic
Bohnhoff, G. 2012. Membrane behaviour, diffusion, and compatibility performance of overlapped geosynthetic clay liner seams requiring
of a polymerized bentonite for containment barrier applications, field-applied supplemental bentonite” ASCE Journal of
Ph.D. dissertation, Colorado State University, Fort Collins, Geotechnical and Geoenvironmental Engineering, 142 (12):
Colorado, USA. 4016067-1 to 11.
Bohnhoff, G., and Shackelford, C. 2013. Improving membrane Brachman, R., Rowe, R.K., and Take, W.A. 2018. Reductions in GCL
performance via bentonite polymer nanocomposite. Applied Clay overlap beneath an exposed geomembrane. Journal of Geotechnical
Science 86, 83-98. and Geoenvironmental Engineering 144(12), 04018094.
Bohnhoff, G., and Shackelford, C. 2014. Hydraulic conductivity of Bradshaw, S., Benson, C., and Scalia, J. 2013. Hydration and cation
chemically modified bentonites for containment barriers. Proc., 7th exchange during subgrade hydration and effect on hydraulic
International Conference on Environmental Geotechnics, A. conductivity of geosynthetic clay liners. Journal of Geotechnical and
Bouazza, S. Yuen, and B. Brown, Eds., Engineers Australia, Geoenvironmental Engineering 139(4), 526-538.
Melbourne, Australia, 440-447 (ISBN 978-1-922107-23-7). Breen, C. 1999. The characterization and use of polycation-exchanged
Bohnhoff, G., and Shackelford, C. 2015. Salt diffusion through a bentonites. Applied Clay Science 15, 187-219.
bentonite-polymer composite. Clays and Clay Minerals 3, 145-162. Breitenbach, A., and Smith, M. 2007. Geomembrane raincoat liners in
Bohnhoff, G., Shackelford, C., Malusis, M., Scalia, J., Benson, C., Edil, the mining heap leach industry. Geosynthetics, IFAI, Roseville,
T., Di Emiddio, G., Katsumi, T., and Mazzieri, F. 2013. Novel Minnesota, USA, 32-39.
bentonites for containment barrier applications. Proc., 18th Breul, B., Reinson, J., Eldridge, T., Stenson G., and Harmon, A. 2006.
International Conference on Soil Mechanics and Geotechnical Bituminous geomembrane in extremely cold conditions. Proc.,
Engineering-Challenges and Innovations in Geotechnics, P. Delage, 8ICG, September 18-22, Yokohama, Japan, Millpress, Rotterdam,
J. Desrues, R. Frank, A. Puech, F. Schlosser, Eds., Presses des Ponts, Netherlands, 395-398.
Paris, Vol. 4, 2997-3000. Brice, G.J. and Woodward, J.C. 1984. Arab Potash Solar Evaporation
Bohnhoff, G., Shackelford, C., and Sample-Lord, K. 2014. Calcium System: Design and development of a novel membrane cut-off wall.
resistant membrane behaviour of a polymerized bentonite. Journal Procedure. Instn, Civil Engrs, Part 1, 1984, 76., Feb., 185-205.
of Geotechnical and Geoenvironmental Engineering 140(3), Brink, D., Shamrock, J., Nortje, R., Johns, J., and Msiza, J. 2009. Some
04013029. pitfalls in the design of lining and capping systems for landfills.
Bokobza, L. 2018. Spectroscopic techniques for the characterization of Landfill Interest Group Conference, Western Cape, South Africa.
polymer nanocomposites: a review. Polymers 10, 7. Brianzoni, V., Fratalocchi, E., Pasqualini, E. 2013. Long-term
Bostwick, L., Rowe, R.K., Take, W.A., and Brachman, R. 2010. performance of cement-bentonite cut-offs in saline and acidic
Anisotropy and directional shrinkage of geosynthetic clay liners. solutions. Proc. 18th ICSMGE: Challenges and Innovations in
Geosynthetics International 17(3),157-170. Geotechnics, Vol. 4, 3001–3004.
Botelho, C. 2019. Behaviour of waterproof barriers and leak detection Brinkmann, F.J.J. 1981. Lekkerkerk, Studies in Environmental Science,
for risk management in vanadium mining tailings dam, Ph.D. Elsevier, Volume 17, 1049-1051.
dissertation, Federal University of Bahia, Bahia, Brazil (in Buchholz, F., and Graham, A. 1998. Modern superabsorbent polymer
Portuguese). technology, Wiley-VCH, New York.
Bouazza, A. 2002. Geosynthetic clay liners. Geotextiles and Building Research Establishment (BRE) 2005. Special Digest 1 Concrete
Geomembranes 20(1), 3-17. in aggressive ground, 3rd edn, Watford.
Bouazza, A. 2021. Interaction between PFASs and geosynthetic liners: Building Research Establishment (BRE). 2011. DG 522. Hardcore for
current status and the way forward. Geosynthetics International, supporting ground floors of buildings: Part 1. Selecting and
28(2), 214-223. specifying materials, Watford.
281
Calligari, P., Bobone, S., Ricci, G., and Bocedi, A. 2020. Molecular Chou, Y-C, Brachman, R.W.I and Rowe, R.K 2021. “Leakage through a
investigation of SARS-CoV-2 proteins and their interactions with hole in a geomembrane beneath a fine-grained tailings”, Canadian
antiviral drugs. Viruses 12(4), 13 pp. Geotechnical Journal, DOI: 10.1139/cgj-2020-0289
Card, G.B. 1981. Properties and performance of bentonite-cement Christian, W., Zainab, B., Tian, K., and Abichou, T. 2020. Effect of
slurries for use as hydraulic cut-offs, University of London PhD specimen preparation on the swell index of bentonite-polymer GCLs.
thesis, 334. Geotextiles and Geomembranes (in press).
Carducci, A., Federigi, I., and Verani, M. 2013. Virus occupational Clinton, M., and Rowe, R.K. 2017. Physical performance of a
exposure in solid waste processing facilities. Annals of Occupational bituminous geomembrane for use as a basal liner in heap leach pads.
Hygiene 57(9), 1115-1127. Proc., 70th Can. Geotech. Conference, Ottawa, Canada.
Carman, P. 1956. Flow of gases through porous media, Butterworths, Cnudde, V., and Boone, M. 2013. High-resolution x-ray computed
London. tomography in geosciences: a review of the current technology and
Carnero-Guzman, G., Bouazza, A., Gates, W., Rowe, R.K. and applications. Earth-Science Reviews 123(8), 1-17.
McWatters, R. 2021. Hydration/dehydration behaviour of Daniel, D.E. 1993. Introduction. In: Daniel, D.E. (Ed.), Geotechnical
geosynthetic clay liners in the Antarctic environment. Geotextiles practice for waste disposal, Chapman and Hall, United Kingdom,
and Geomembranes,49(1),196-209. 33-65.
Caron, C. 1972. Pérennité des systems argile-ciment ou bentonite-ciment Daniel, D., Bowders, J., and Gilbert, R. 1997. Laboratory hydraulic
dans divers types d’application. Construction Tome 27, No. 10, Paris conductivity testing of GCLs in flexible-wall permeameters. Testing
291-296. and Acceptance Criteria for Geosynthetic Clay Liners, L. Well, Ed.,
Caron, C. 1973. Un nouveau style du perforation: la boue autodurcissable. ASTM STP 1308, 208-226.
Annales de l’Institut Technique du Batiment st des Travaux De Camillis, M. 2017. Experimental and numerical study of wet and dry
Publiques, Tome 26, No. 311 Paris, 1-40. cycles on an innovative polymer treated clay for geosynthetic clay
Castillo, J. 2005. Pipe deflection numerical analysis. Personal liners, PhD dissertation, Ghent University, Ghent, Belgium.
communication. De Camillis, M., Di Emidio, G., Bezuijen, A., and Verastegui-Flores, R.
Castillo, J., Hallman, D., Byrne, P., and Parra, D. 2005. Dynamic analysis 2016. Hydraulic conductivity and swelling of a polymer modified
of heap leach pad under high phreatic levels. Proc., XXVII bentonite subjected to wet-dry cycles in seawater. Geotextiles and
Convención Minera, Instituto de Ingenieros de Minas del Perú, Geomembranes 44, 739-747.
Arequipa, Peru. De Camillis, M., Di Emidio, G., Bezuijen, A., Verastegui Flores, D., Van
César, C., Mendoza, J., and Parra, D. 2013. Heap leach pad design in Stappen, J., and Cnudde, V. 2017. Effect of wet-dry cycles on
very aggressive terrains. Proc., Heap Leach Solutions 2013, pp. 291- polymer treated bentonite in seawater: Swelling ability, hydraulic
303, Vancouver, Canada. conductivity and crack analysis. Applied Clay Science 142, 52-59.
Chai, J., and Prongmanee, N. 2020. Barrier properties of a geosynthetic De Camillis, M., Di Emidio, Verastegui-Flores, D., and G., Bezuijen, A.
clay liner using polymerized sodium bentonite. Geotextiles and 2014. Polymer treated clays subject to wet-dry cycling with seawater.
Geomembranes 48, 392-399. Proc., 7th International Conference on Environmental Geotechnics,
Chambers, D., and Bowker, L. 2017. In the dark shadow of the A. Bouazza, S. Yuen, and B. Brown, Eds., Engineers Australia,
supercycle tilings failure risk and public liability reach all time highs. Melbourne, Australia, 1002-1008 (ISBN 978-1-922107-23-7).
Journal of Environments 4, p. 76. Department of Water Affairs and Forestry (DWAF) 1998. Minimum
Chappel, M., Brachman, R., Take, W.A., and Rowe, R.K. 2012a. Large- Requirements for Waste Disposal by Landfill 2nd Edition, Waste
scale quantification of wrinkles in a smooth, black, HDPE Management Series, ISBN 0620-22993-4.
geomembrane. Journal of Geotechnical and Geoenvironmental Di Emidio, G. 2010. Hydraulic and chemico-osmotic performance of
Engineering 138(6), 671-679. polymer treated clays, Ph.D. dissertation, University of Ghent,
Chappel, M., Rowe, R.K., Brachman, R., and Take, W.A. 2012b. A Ghent, Belgium.
comparison of geomembrane wrinkles for nine field cases. Di Emidio, G., Mazzieri, F., Verastegui-Flores, R., Van Impe, W., and
Geosynthetics International 19(6), 453-469. Bezuijen, A. 2015. Polymer-treated bentonite clay for chemical-
Chen, J, Benson, C., Likos, W., and Edil, T. 2017. Interface shear strength resistant geosynthetic clay liners. Geosynthetics International 22(1),
of a bentonite-polymer geosynthetic clay liner and a textured 125-137.
geomembrane. Proc. GeoFrontiers, ASCE, Reston, Virginia, USA, Di Emidio, G., Meeusen, J., Snoeck, D., and Verástegui Flores, R. 2019.
219–226. Enhanced sustainable soils: a review. Proc., 8th International
Chen, J., Benson, C., Edil, T., and Likos, W. 2018. Hydraulic Congress on Environmental Geotechnics, L. Zhan, Y. Chen, and A.
conductivity of geosynthetic clay liners with sodium bentonite to Bouazza, Eds., Springer Nature Singapore Pte Ltd., Vol. 3, 515-522.
coal combustion product leachates. Journal of Geotechnical and Di Emidio, G., Van Impe, W., and Verástegui Flores, R. 2011. Advances
Geoenvironmental Engineering 144(3), 0401800. in geosynthetic clay liners: polymer enhanced clays. Geo-Frontiers
Chen, Y., Li, J., Yang, C., Zhu, B., and Zhan, L. 2017. Centrifuge 2011, ASCE, Reston, Virginia, USA.
modeling of municipal solid waste landfill failures induced by rising Di Emidio, G., Veratsegui-Flores, R., Bezuijen, A. 2013. Beneficial
water levels. Canadian Geotechnical Journal 54(12), 1739-1751. impact of polymer treatment on the swelling and long-term
Chen, Y., Liu, X., Xu, W., et al. 2019. Analysis on stabilization hydraulic efficiency of Ca-bentonites compared to the standard
characteristics and exploitability of landfilled municipal solid waste: sodium activation method. Geosynthetics 2013, IFAI, Roseville,
Case of a typical landfill in China. SCIENCE CHINA Technological Minnesota, USA, 588-594.
Science 49, 199–211, (in Chinese). Di Emidio, G., Verastegui-Flores, R., Mazzieri, F., and Dominijanni, A.
Chen, J., Salihoglu, H., Benson, C., Likos, W., and Edil, T. 2019. 2017. Modified clays for barriers: a review. Innovative Infrastructure
Hydraulic conductivity of bentonite-polymer composite Solutions 2, 47.
geosynthetic clay liners permeated with coal combustion product Doll, P. 1997. Desiccation of mineral liners below landfills with heat
leachates. Journal of Geotechnical and Geoenvironmental generation. Journal of Geotechnical and Geoenvironmental
Engineering 145(9), 04019038. Engineering 123(11), 1001-1009.
Chen Y., Xu, W., Ling, D., Zhan, L., and Gao, W. 2020. A degradation– Dominijanni, A., and Manassero, M. 2021. Steady-state analysis of
consolidation model for the stabilization behaviour of landfilled pollutant transport to assess landfill liner performance.
municipal solid waste. Computers and Geotechnics 118, 103341. Environmental Geotechnics 8(7), 480-494.
Chen Y., Zhan, L., and Gao, W. 2018. Waste mechanics and sustainable Dominijanni, A., Manassero, M., and Puma, S. 2013. Coupled chemical-
landfilling technology: comparison between HFWC and LFWC hydraulic-mechanical behaviour of bentonites. Géotechnique 63(3),
MSWs. Proc., 8th International Congress on Environmental 191-205.
Geotechnics, Vol.1, 3-37. Dominijanni, A., Guarena, N., and Manassero, M. 2018. Laboratory
Chen G-N, Cleall PJ, Li Y-C, Yu Z-X, Ke H, Chen Y-M. 2018. assessment of semi-permeable properties of a natural sodium
Decoupled advection-dispersion method for determining wall bentonite. Canadian Geotechnical Journal 55(11), 1611-1631.
thickness of slurry trench cut-off walls. International Journal of Dominijanni, A., Guarena, N., and Manassero, M. 2021. Risk assessment
Geomechanics-ASCE, 18(5):06018007. procedure for the performance-based design of landfill lining
Chou, Y., Rowe, R.K., and Brachman, W. 2018. Erosion of silty sand systems and cut-off walls, Proc. 3rd International Symposium on
tailings through a geomembrane defect under filter incompatible Coupled Phenomena in Environmental Geotechnics (CPEG), 20-21
conditions. Canadian Geotechnical Journal 55(11), 1564-1575.
282
October 2021, Kyoto, Japan. Japanese Geotechnical Society Special Fan, J-Y and Rowe, R.K. 2022. “Piping of silty sand tailings through a
Publication 9(5), 199-204. circular geomembrane hole”. Geotextiles and Geomembranes (in
Dunster, A.M., Longworth, T.I., Nixon, P.J., Lane, S.J. and Collins, R.J. press).
2005. BRE Information Paper IP/05. Development of sites Fan, J-Y, Rowe, R.K., and Brachman, R.W.I. 2022. “Compressibility and
containing expansive ferrous slags, Watford Permeability of Sand-Silt Tailings Mixtures”, Canadian
Eden, M. 2017. Mechanisms of expansion in incinerator bottom Geotechnical Journal, https://doi.org/10.1139/cgj-2021-0356
aggregate (IBAA) used as aggregate in concrete and as fill in Fan, R., Reddy, K., Yang, Y., and Du, Y. 2020. Index properties, hydraulic
engineering applications. Proceedings of the 16th Euroseminar on conductivity and contaminant-compatibility of CMC-treated sodium
Microscopy Applied to Building Materials, Les Diablerets, activated calcium bentonite. International Journal of Environment
Switzerland. Research and Public Health 17, Article 1863.
Ejezie, J.O., Jefferis, S.A., Lam, C., Sedighi, M. and Syed, M.A. 2020. Fannin, R., Palmeira, E., Atitep Srikongsri, and Gardoni, M. 2008.
Permeation Behaviour of PHPA Polymer Fluids in Sand. Interpretation of the gradient ratio test for geotextile filtration design.
Geotechnique, Proc., 1st Pan American Geosynthetics Conference and Exhibition,
El-Dana, A., and Breul, B. 2010. Coletanche life expectancy. Axter Cancun, Mexico, 1, 95-105.
Coletanche. Fehervari, A., Gates, W., Bouazza, A., and Shackelford, C. 2019.
Eldesouky, H., and Brachman, R. 2018. Calculating local geomembrane Assessment of bentonite compatibility with salinity using
strains from a single gravel particle with thin plate theory. centrifugation-based water retention. Geotechnical Testing Journal
Geotextiles and Geomembranes 46(1), 101-110. 42(2), 275-295.
Elges, H. 1985. Problem soils in South Africa – state of the art dispersive Fehervari, A., Gates, W., Patti, A., Turney, T., Bouazza, A., and Rowe,
soils. The Civil Engineer in South Africa, July 1985. R.K. 2016a. Potential hydraulic barrier performance of cyclic
El-Sebaii, A., Aboul-Enein, S., Ramadan, M., and Khallaf, A. 2011. organic carbonate modified bentonite complexes against hyper-
Thermal performance of an active single basin solar still (ASBS) salinity. Geotextiles and Geomembranes 44, 748-760.
coupled to shallow solar pond (SSP). Desalination 280 (1), 183-190. Fehervari, A., Gates, W., Turney, T., Patti, A., and Bouazza, A. 2016b.
EPA 2001. Geosynthetic clay liners used in municipal waste landfills, Cyclic organic carbonate modification of sodium bentonite for
EPA530-F-97-002 (Revised Issue), US Environmental Protection enhanced containment of hyper saline leachates. Applied Clay
Agency, Washington, D.C. Science 134, 2-12.
(http://www.epa.gov/garbage/landfill/geosyn.pdf). Fernandes, G., Palmeira, E., and Gomes, R., 2008. Performance of
Esford, F., and Janssens, G. 2014. Laboratory test results on a bituminous geosynthetic-reinforced alternative sub-ballast material in a railway
liner exposed to a weak acidic solution. Proc., Geosynthetics Mining track. Geosynthetics International 15(5), 311-321.
Solutions, Vancouver, Canada. 10 pp. Ferreira, J. 2009. Studies on drainage and moisture content reduction of
Estrin, D., and Rowe, R.K. 1995. Landfill design and the regulatory the iron ore of Carajás mine with the use of geosynthetics, MSc.
system. Proc. 5th International Landfill Symposium, Sardinia, Italy, dissertation, Graduate Programme of Geotechnics, University of
Vol. 3, 15-26. Brasília, Brasília, DF, Brazil (in Portuguese).
Evans, J. C. 1994. Hydraulic conductivity of vertical cut-off walls. Filz, G. M. 1996. Consolidation stresses in soil-bentonite backfilled
Hydraulic Conductivity and Waste Containment Transport in Soil, trenches. The 2nd International Congress on Environmental
ASTM, Philadelphia, 79-94. Geotechnics, Balkema, Rotterdam, 497-502.
Evans, J. C., Costa, M. J., and Cooley, B. 1995. The state-of-stress in Filz, G.M., Adams, T. and Davidson, R.R. 2004. Stability of long
soil-bentonite slurry trench cut-off walls. Geoenvironment 2000: trenches in sand supported by bentonite-water slurry. Journal of
Characterization, Containment, Remediation, and Performance in geotechnical and geoenvironmental engineering, 130(9), 915-921.
Environmental Geotechnics, ASCE, Geotechnical Specialty Filz, G.M., Evans, J.C. and Britton, J.P. 2003. Soil-bentonite hydraulic
Publication, New York, NY, 1173-1191. conductivity: measurement and variability. In Proceedings of the
Evans, J.C. and Dawson, A.R., 1999, December. Slurry walls for control 12th Pan American Conf. on Soil Mechanics and Geotechnical
of contaminant migration: A comparison of UK and US practices. In Engineering, Cambridge, Mass. 22-26.
Geo-Engineering for Underground Facilities (pp. 105-120). ASCE. Fleming, I.R., Rowe, R.K. and Cullimore, D.R. 1999. Field observations
Evans, J.C., Fang, H.Y. and Kugelman, I.J., 1985. November. of clogging in a landfill leachate collection system. Canadian
Containment of hazardous materials with soil-bentonite slurry walls. Geotechnical Journal, 36(4):685-707.
In Proceedings of the 6th National Conf. on the Management of Flynn, B., and Carter, G. 1998. Waterproofing material and method of
Uncontrolled Hazardous Waste Sites, 249-252. fabrication thereof, US Patent Number: 6,537,676 B1, March.
Evans, J.C. and Jefferis, S.A. 2014. Volume change characteristics of cut- Folwell, J., Gassner, F., and Phillips, G., 2021. Accelerated ageing testing
off wall materials, Proc. 7th Int. Congress on Environmental of geomembranes for tailings liquor,
Geotechnics, Melbourne. https://www.ausimm.com/globalassets/conferences-and-
Evans, J.C., Larrahondo, J.M. and Yeboah, N. 2020. 2020. Fate of events/mine-waste-and-tailings/abstract-82-accelerated-ageing-
bentonite in slag-cement-bentonite slurry trench cut-off walls. testing-of-geomembranes-for-tailings-liquor.pdf
Proceedings of the Institution of Civil Engineers -Environmental Foose, G.J. 2010. A steady-state approach for evaluating the impact of
Geotechnics, 2020. (ahead of print). solute transport through composite liners on groundwater quality.
Evans J. C., Ororbia M., Gutelius J., Ruffing D., Barlow L., Malusis M. Waste Management 30(8–9), 1577-1586.
2017. Soil-bentonite slurry trench cut-off wall lateral deformations, Foose, G., Benson, C., and Edil, T. 2001. Analytical equations for
consolidation, stress transfer and hydraulic conductivity, predicting concentration and mass flux from composite liners.
Proceedings of the 2nd Symposium on Coupled Phenomena in Geosynthetics International 8(6), 551-575.
Environmental Geotechnics (CPEG2), Leeds, UK 2. Fourie, A., Bouazza, A., Lupo, J., and Abrão, P. 2010. Improving the
Evans, J.C., Prince, M.J. and Adams, T.L. 1997. Metals attenuation in performance of mining infrastructure through the judicious use of
minerally-enhanced slurry walls. No. CONF-970208--PROC. geosynthetics. Proc., 9th International Conference on Geosynthetics,
Evans, J.C., Ruffing, D. and Elton, D. 2021 Fundamentals of Ground Guarujá, Brazil, 193-219.
Improvement Engineering. CRC Press, London, p430. Fowler, J., Larkins, K., and Duke, M. 2005. Dredging aerobis
Ewais, A., Rowe, R.K., Brachman, R., and Arnepalli, D. 2014a. Service- digested biosolids into geotextile tubes for dewatering, New Orleans
life of a high-density polyethylene geomembrane under simulated East Municipal Sewage Treatment Plant, New Orleans, LA. Proc.,
landfill conditions at 85oC. Journal of Geotechnical and 25th Technical Conference of the Western Dredging Association,
Geoenvironmental Engineering 140(11), 04014060. New Orleans, USA, Report No. 507, 247.
Ewais, A., Rowe, R.K., and Scheirs, J. 2014b. Degradation of 2.4 mm Fratalocchi, E., 1996. Hydraulic behaviour of cement-bentonite mixtures.
HDPE geomembrane with high residual HP-OIT. 10th International Ph.D. Thesis, Politecnico di Torino, Turin, Italy (in Italian).
Conference on Geosynthetics, Berlin, Germany. Fratalocchi, E., Brianzoni, V., Di Sante, M., Mazzieri, F., Pasqualini, E.,
Fan, J-Y and Rowe, R.K. 2021. “Seepage through a Circular 2013. Migration of aggressive solutions through CB walls. Proc. 1st
Geomembrane Hole when covered by Fine-Grained Tailings under CPEG, Coupled Phenomena in Environmental Geotechnics, Turin
Filter Incompatible Conditions” Canadian Geotechnical Journal, (in (Italy), Taylor and Francis Group, London, p. 243-252.
press). http://dx.doi.org/10.1139/cgj-2020-0788 Fratalocchi, E., Brianzoni, V., Mazzieri, F., Pasqualini, E., 2016.
Durability of Cement-Bentonite Cut-Off Walls in Sulphate Solutions.
283
ASCE GSP 275, Geo-Chicago 2016. 695-704. DOI: clay liners. Journal of Geotechnical and Geoenvironmental
10.1061/9780784480168.068. Engineering 144(10), 06018011.
Fratalocchi E., Domizi J. and Felici M., 2021. Migration parameters and Gilson-Beck, A. 2019. Controlling leakage through installed
interaction of saline solutions through cement-bentonite cutoffs. geomembranes using electrical leak location. Geotextiles and
Proc. 3rd CPEG, Kyoto, Japan. Geomembranes 47(5), 697-710.
Fratalocchi, E., Domizi, J., Mazzieri, F., 2019. Hydraulic Conductivity Giroud, J., and Bonaparte, R. 1989a. Leakage through liners constructed
and Sorption Capacity of Special Barrier Materials in Inorganic with geomembranes - Part I: Geomembrane liners. Geotextiles and
Solutions. Proc. 8th Int. Congress on Environmental Geotechnics, Geomembranes 8, 27-67.
Hangzhou (China), Vol. 2, p. 461-469. Giroud, J., and Bonaparte, R. 1989b. Leakage through liners constructed
Fratalocchi E., Manassero M., Pasqualini, E, Roccato M. and Spanna C. with geomembranes - Part II: Composite liners. Geotextiles and
1996. Predicting hydraulic conductivity of cement-bentonite slurries. Geomembranes 8, 71-111.
In: Environmental Geotechnics. p. 503-508, A.A. Balkema, Giroud, J., and Fluet Jr., J. 1986. Quality assurance of geosynthetic lining
Rotterdam, ISBN: 9054108487, Osaka (Japan), November, 5-8, systems. Geotextiles and Geomembranes 3(4), 249-287
1996. Giroud, J., and Morel, N. 1992. Analysis of geomembrane wrinkles.
Fratalocchi, E., Pasqualini, E., 2007. Vertical barriers for site Geotextiles and Geomembranes 11(3), 255 276 (Erratum: 12(4),
containment. Proc. XXI CGT, Turin Geotechnical Conf., Torino, 378).
Italy. Giroud, J., Tisseau, B., Soderman, K. and Beech, J. 1995. Analysis of
Fratalocchi, E., Pasqualini, E., Balboni, P., 2006. Performance of a strain concentration next to geomembrane seams. Geosynthetics
cement-bentonite cut-off wall in an acidic sulphate environment. 5th International 2 (6), 1049–1097.
ICEG (Osaka, Japan) Vol.1, 133–139. Gleason, M., Daniel, D., and Eykholt, G. 1997. Calcium and sodium
Fratalocchi, E., Pasqualini, E., Balboni, P., Mozzi, R., 2005. Durability bentonite for hydraulic containment applications. Journal of
assessment of a confinement cut-off wall for a phosphogypsum Geotechnical and Geoenvironmental Engineering 123(5), 438-445.
landfill. Proc. 16th ICSMGE: Geotechnology in Harmony with the Goodall, D.C. and Quigley, R.M. 1977. Pollutant migration from two
Global Environment, Vol. 4, 2257–2260. sanitary landfill sites near Sarnia, Ontario. Canadian Geotechnical
Froneman, J., Moseneke, D., and Madlanga, J. 2015. Shoprite Checkers Journal, 14(2), 223-236.
(Pty) Limited v Member of the Executive Council for Economic Government of Ghana Guidelines 2002. Government of Ghana Ministry
Development, Environmental Affairs and Tourism, Eastern Cape of Local Government and Rural Development Environmental
and Others (CCT 216/14) [2015] ZACC 23; 2015 (6) SA 125 (CC, Protection Agency Landfill Guidelines.
2015 (9) BCLR 1052 (CC) (30 June 2015). Government of Ghana Guidelines 2009. Government of Ghana Ministry
Gao, W., Chen, Y., Zhan L., et al. 2015. Engineering properties for high of Local Government and Rural Development Environmental
kitchen waste content municipal solid waste. Journal of Rock Sanitation Policy.
Mechanics and Geotechnical Engineering 06, 646-658. Graiver, D.A., Topliff, C.L., Kelling, C.L., and Bartelt-Hunt, S.L. 2009.
Gardoni, M., and Morais, F. 2019. Evaluation of the resistance of Survival of the avian influenza virus (H6N2) after land disposal.
geomembranes aged in laboratory and exhumed from gold mining Environmental Science and Technology 43(11), 4063-4067.
tailings dams. Proc., 8th Brazilian Conference of Geosynthetics and Guarena, N., Dominijanni, A., and Manassero, M. 2020. From the design
9th Brazilian Conference of Environmental Geotechnics (Regeo- of bottom landfill liner systems to the impact assessment of
Geossintéticos'2019), São Carlos, SP, Brazil (in Portuguese). contaminants on underlying aquifers. Innovative Infrastructure
Gardoni, M., and Palmeira, E. 2002. Microstructure and pore Solutions 5(1).\
characteristics of synthetic filters under confinement. Geotechnique Guner, A. 1978. The properties and behaviour of bentonite/cement
52(6), 405-418. slurries. University of London PhD thesis, 279.
Gardoni M., Villar L., Ferreira J., and Monteiro M. 2010a. The behaviour Gularte, F., Barneich, J., Burton, J., Fordham, E., Watt, D., Johnson, T.,
of geosynthetics in drainage systems in Carajás mine. Proc., 6th and Weeks, J. 2007. First use of TRD construction technique for soil
International Congress on Environmental Geotechnics, New Delhi, mix cut-off wall construction in the United States. In Designing our
India. Underground Space 1-12.
Gardoni, M., Villar, L., Ferreira, J., and Monteiro, M. 2010b. Laboratory Guo, W., Chu, J., Yan, S., and Nie, W. 2014. Analytical solutions for
study on the performance of geosynthetics drainage systems in geosynthetic tube resting on rigid foundation. Geomechanics and
contact with ore from Carajás mining. Proc., 9th International Engineering 6(1), 65-77.
Conference on Geosynthetics, Guarujá, Brazil Gustitus, S., and Benson, C. 2020. Assessing polymer elution and
Gassner, F. 2009. Field observation of GCL shrinkage at a site in hydraulic conductivity of bentonite-polymer composite geosynthetic
Melbourne Australia. Geotextiles and Geomembranes 27(5), 406- clay liners permeated with aggressive solutions, Proc., GeoAmericas
408. 2020, 4th Pan American Conference on Geosynthetics, Rio de
Gates, W.P., MacLeod, A., Fehervari, A., Bouazza, A., Gibbs, D., Janeiro, Brazil.
Hackney, R., Callahan, D.L. and Watts, M., 2020. Interactions of Gustitus, S., and Benson, C. 2021. Flow-swell index as an indicator of
Per-and Polyfluoralkyl Substances (PFAS) with Landfill Liners. chemical compatibility of bentonite-polymer composite
Advances in Environmental and Engineering Research, 1(4), 40. geosynthetic clay liners. Proc., Geosynthetics 2021, IFAI, Roseville,
https://doi.org/10.21926/aeer.200400 Minnesota, USA.
Gates, W., Shaheen, U., Turney, T., and Patti, A. 2016. Cyclic carbonate- Gustitus, S., Nguyen, D., Chen, J., and Benson, C. 2020. Quantifying
sodium smectite intercalates. Applied Clay Science 124-125, 94-101. polymer loading in bentonite-polymer composites using loss on
Geng, W., Likos, W., and Benson, C. 2016. Viscosity of polymer- ignition and total carbon analyses. Geotechnical Testing Journal
modified bentonite as a hydraulic performance index. Geo-Chicago 44(5), 1448-1466.
2016, GSP 271, ASCE, Reston, Virginia, USA, 498-507. Guyonnet, D., Perrochet, P., Côme, B., Seguin, J.-J., and Parriaux, A.
Ghavam-Nasiri, A., El-Zein, A., Airey, D., Rowe, R.K. 2017. Hydration 2001. On the hydro-dispersive equivalence between multi-layered
and desiccation of geosynthetic clay liners in composite lining mineral barriers. Journal of Contaminant Hydrology 51(3-4), 215-
systems under brine pond conditions: a laboratory investigation. 231.
Proceedings of the 2nd Symposium on Coupled Phenomena in Haase, H., and Schanz, T. 2016. Compressibility and saturated hydraulic
Environmental Geotechnics (CPEG2), Leeds University, Leeds, UK. permeability of clay-polymer composites – experimental and
Ghavam-Nasiri, A., El-Zein, A., Airey, D., Rowe, R., and Bouazza, A. theoretical analysis. Applied Clay Science 130, 62-75.
2019. Numerical simulation of geosynthetic clay liners desiccation Hanson, J., Risken, J., and Yesiller, N. 2013. Moisture-suction
under high thermal gradients and low overburden stress. relationships for geosynthetic clay liners. Proc., 18th International
International Journal of Geomechanics 19(7), 04019069 Conference on Soil Mechanics and Geotechnical Engineering-
Ghavam-Nasiri, A., El-Zein, A., Airey, D., Rowe, R.,K. and Bouazza, A. Challenges and Innovations in Geotechnics, P. Delage, J. Desrues,
2020. Experimental investigation of desiccation of geosynthetic clay R. Frank, A. Puech, F. Schlosser, Eds., Presses des Ponts, Paris,
liners under brine pond conditions. Geosynthetics International 3025-3028
27(6), 593-605. Hao, Z., Sun, M., Ducoste, J. and Barlaz, M. 2017.A model to describe
Ghazizadeh, S., Bareither, C., Scalia, J., and Shackelford, C. 2018. heat generation and accumulation at municipal solid waste landfills.
Synthetic mining solutions for laboratory testing of geosynthetic Proc. GeoFrontiers, ASCE, Reston, Virginia, USA, 281–288.
284
Harvey, R.H., and Ryan, J.N. 2004. Use of PRD1 bacteriophage in Jefferis. S.A., Doe, G. and Tedd, P., assisted by a steering group 1999.
groundwater viral transport, inactivation, and attachment studies. UK National Specification for cut-off walls for pollution migration
FEMS Microbiology Ecology 49(1), 3-16. control, Institution of Civil Engineers.
Haug, M., and Boldt-Leppin. B. 1994. Influence of polymers on the Jefferis, S.A. and Fernandez, A. 2000. Spanish dyke failure leads to
hydraulic conductivity of marginal quality bentonite-sand mixtures. developments in cut-off wall design, Int. Conf. on Geotechnical and
Hydraulic Conductivity and Waste Contaminant Transport in Soil, Geological Engineering, Melbourne, Australia, November.
Daniel, D. and Trautwein, S., Eds., ASTM STP 1142, ASTM, West Jefferis, S.A. and Lam, C. 2018. Using density to determine the solids
Conshohoken, Pennsylvania, USA, 407-421. content of construction slurries, Proceedings of the 19th Int. Conf.
Heller, H., and Keren, R. 2002. Anionic polyacrylamide polymers effect on Soil Mechanics and Geotechnical Engineering, Seoul, 1695-1698.
on rheological behaviour of sodium-montmorillonite suspensions. Jefferis, S.A. 2019a. Recycled crushed concrete: engineering and
Soil Science Society of America Journal 66(1), 19-25. environmental issues. Proc. XVII European Conf on Soil
Hoai, T., and Mukunoki, T. 2020. Combined effects of ammonium Mechanics and Geotechnical Engineering. Reykjavik.
permeation and dry-wet cycles on the hydraulic conductivity and Jefferis, S.A. 2019b. Low Permeability Vertical Barriers: The State of the
internal properties of geosynthetic clay liners. Geotextiles and Art and the Research Needs for the Future: Towards a Sustainable
Geomembranes 48, 912-927. Geoenvironment. Proceedings of the 8th Int. Congress on
Hofford, S. 2015. Survey of fatigue resistance quantification of asphalt Environmental Geotechnics Volume 1 212-214. 10.1007/978-981-
mixture. International Journal of Advanced and Applied Sciences 13-2221-1_16.
2(1) 2015, 42-49 Jessberger H. L., Soyez B., Manassero, M., Street A. 1991. Progress of
Hoor, A., and Rowe, R.K. 2013. Potential for desiccation of geosynthetic the work of the European technical committee no.8 of ISSMFE.
clay liners used in barrier systems. Journal of Geotechnical and Milano: C.I.P.A. s.r.l., S. Margherita di Pula (CA), October, 13-15,
Geoenvironmental Engineering 139(10), 1648-1664. 1993.
Hornsey, W. 2013. Performance of cushion geotextiles for liner Jesseberger H.L., Soyez B., Manassero M., Street A. 1995. Engineering
protection applications. GeoAfrica 2013, 18-20 November, Accra, waste disposal (geotechnics of landfill design and remedial works),
Ghana. Key-note lecture. In: Waste disposal by landfill, Green '93. 21-33,
Hsuan, G., and Koerner, R. 2015. Rationale and background for the GRI A.A. Balkema, Rotterdam, ISBN: 9789054103561, Bolton (UK), 28
GM13 specification for HDPE geomembranes. GSI white paper no. June - 1 July 1993.
32, Geosynthetic Institute, Folsom, Pennsylvania, USA. Jewaskiewitz, S., 2019. Envitech Solutions (Pty) Ltd letter to the
Hu, J., Ke, H., Lan, J., Chen, Y., Meng, M. 2020a. A dual-porosity model Department Environment Affairs, Forestry and Fisheries, ES560-
for coupled leachate and gas flow to vertical wells in municipal solid 2016-L005, 2 October.
waste landfills. Geotechnique 70(5), 406-420. Jia, G., Zhan, T., Chen, Y., and Fredlund, D. 2010. Influence of rainfall
Hu, J., Ke, H., Zhan L., Chen, Z., Lan, J., Powrie, W., Chen, Y. 2020b. pattern on the infiltration into landfill earthen final cover. Advances
Installation and performance of horizontal wells for dewatering at in Environmental Geotechnics, Springer, Berlin, Heidelberg, 641-
municipal solid waste landfills in China. Waste Management 103, 645.
159-168. Jo, H., Benson, C., and Edil, T. 2006. Rate-limited cation exchange in
ICOLD 1986. Geotextiles as filters and transitions in fill dams, Bulletin thin bentonitic barrier layers. Canadian Geotechnical Journal 43,
55, Paris, France. 370-391.
ICOLD 2001. Tailings dams: risk of dangerous occurrences, lessons Jo, H., Katsumi, T., Benson, C., and Edil, T. 2001. Hydraulic
learnt from practical experiences, Bulletin 121, Paris, France, 144 p. conductivity and swelling of nonprehydated GCLs permeated with
Jambeck, J., Geyer, R., Wilcox, C., Siegler, T., Perryman, M., Andrady, single-species salt solutions. Journal of Geoetchnical and
A., Narayan, R., and Law, K. 2015. Plastic waste inputs from land Geoenvrionmental Engineering 127(7), 557-567.
into the ocean. Science 347(6223), 768-771. Joshi K, Kechavarzi C, Sutherland K, Ng M Y A, Soga K, Tedd P. 2010.
James, A., Fullerton, D., and Drake, R. 1997. Field performance of GCL Laboratory and in situ tests for long-term hydraulic conductivity of
under ion exchange conditions. Journal of Geotechnical Engineering a cement-bentonite cut-off wall, Journal of Geotechnical and
123(10), 897-901. Geoenvironmental Engineering, 136(4):562-572.
Jefferis, S.A., 1981. June. Bentonite-cement slurries for hydraulic cut- Joshi, P., Brachman, R.W.I and Rowe, R.K. 2017 Hydraulic
offs. In Proceedings, Tenth Int. Conf. on Soil Mechanics and performance of GCL seams without field-applied supplemental
Foundation Engineering, Stockholm, Sweden, Vol. 1, 435-440. bentonite below a geomembrane wrinkle, ASCE Journal of
Jefferis, S.A. 1992. Contaminant - grout interaction, ASCE Specialty Geotechnical and Geoenvironmental Engineering, 143(9):
Conf., Grouting, Soil Improvement and Geosynthetics, New Orleans, 04017086-1 to 10, doi/10.1061/(ASCE)GT.1943-5606.0001756
February. Joshi, P., Rowe, R.K. and Brachman, R.W.I 2018. Hydraulic
Jefferis, S.A. 1996. Contaminant - barrier interaction: Friend or foe?, performance of GCL seams with factory-applied powered bentonite
Mineralogical Society Conf., Chemical containment of wastes in the below a geomembrane wrinkle. Environmental Geotechnics, 5(5):
Geosphere. 263-270http://dx.doi.org/10.1680/ jenge. 16 .00027
Jefferis, S.A. 1997. The origins of the slurry trench cut-off and a review Joshi, P., Rowe, R.K. and Brachman, R.W.I 2017. “Physical and
of cement-bentonite cut-off walls in the UK. Proc. Intl Containment hydraulic response of geomembrane wrinkles underlying saturated
Technology Conf., St Petersburg, Florida. 9-12 February. fine tailings", Geosynthetics International, 24(1):82-
Jefferis, S.A. 2001. Permeability a dynamic property of barrier materials, 94,http://dx.doi.org/10.1680/jgein.16.00017
Int. Containment and remediation technology Conf. and Exhibition, Jury, W., Gardner, W., and Gardner, W. 1991. Soil physics, 5th edition,
Orlando, Florida, June. John Wiley and Sons, Inc., New York.
Jefferis, S.A. 2003. Long-term performance of grouts and the effects of Kang, J., and Shackelford, C. 2011. Consolidation enhanced membrane
grout by-products, ASCE Specialty Conf., Grouting and Ground behaviour of a geosynthetic clay liner. Geotextiles and
Improvement, New Orleans, February. Geomembranes 29(6), 544-556.
Jefferis, S.A. 2008. Reactive transport in cut-off walls and implications Karlen, A. 1995. Plague's progress, a social history of man and disease.
for wall durability, GeoCongress 2008. Annual Conf. of the Geo- Victor Gollancz, London.
Institute of the American Society of Civil Engineers, New Orleans. Katsumi, T., and Fukagawa, R. 2005. Factors affecting the chemical
Jefferis, S.A. 2012. Cement-bentonite slurry systems, Proc. Fourth compatibility and barrier performance of GCLs. Proc., 16th
International Conference on Grouting and Deep Mixing, Deep International Conference on Soil Mechanics and Geotechnical
Foundations Institute and Geo-Institute of the American Society of Engineering, Osaka, Japan, Sept. 12-16, 2005, Millpress, Rotterdam,
Civil Engineers, pp 1-24, New Orleans. The Netherlands, Vol. 4, 2285-2288.
Jefferis, S.A. 2013. Some unexpected effects of natural and Katsumi, T., Ishimori, H., Onikata, M., and Fukagawa, R. 2008. Long-
anthropogenic chemicals on construction, Geotechnique, March. term barrier performance of modified materials against sodium and
Jefferis, S.A. 2014. Grouts and slurries, in: The Construction Materials calcium permeant solutions. Geotextiles and Geomembranes 26, 14-
Reference Book, second edition, Doran, D.K. and Cather, R., Editors, 30.
Routledge. Kaza, S., Yao, L., Bhada-Tata, P., and Van Woerden, F. 2018. What a
Jefferis, S.A. 2019. Recycled crushed concrete: engineering and waste 2.0: a global snapshot of solid waste management to 2050. The
environmental issues. Proc. XVII European Conf on Soil Mechanics World Bank.
and Geotechnical Engineering. Reykjavik.
285
Ke H, Tong X, Li Y-C, Chen Y-M, Wen Y-D. 2018. Force-equilibrium Legge, K., Legg, P., and Meyer, P. 2007. Coupled solution to heat
based model for predicting stresses in soil-bentonite cut-off walls. induced degradation of containment barriers. Proc., 11th
Journal of Geotechnical and Geoenvironmental Engineering ASCE, International Waste Management and Landfill Symposium, 1-5
144(2):04017112. October, Sardinia, Italy.
Khan, M., Di Emidio, G., and Bezuijen, A. 2021. Matric suction of Li, X., and Zhang, L. 2009. Characterization of dual-structure pore-size
untreated clay and polymerised HYPER clay geosynthetic barriers. distribution of soil. Canadian Geotechnical Journal 46(2), 129-141.
Journal of Environmental Geotechnics, Special Issue: Polymers in Li, T., and Rowe, R.K. 2020. GCL Self-healing: fully penetrating
Geoenvironmental Engineering, under review. hole/slit hydrated with RO water and 10 mM CaCl2 solution.
Khilnani, K, Stark, T., and Bahadori, T. 2017. Comparison of single and Geosynthetics International 27(1),34-47
multi-layer interface strengths for geosynthetic/geosynthetic and Li, T., and Rowe, R.K. 2022. Self-healing of laboratory eroded defects in
soil/geosynthetic interfaces. Proc., GeoFrontiers, ASCE, Reston, a GCL on silty sand. Geosynthetics International (in press).
Virginia, USA, 42–51. Li Y-C, Pan Q, Cleall PJ, Chen Y-M, Ke H. 2013. Stability analysis of
Kilbourne, E.D. 2008. Plagues and pandemics: past, present and future. slurry trenches in similar layered soils. Journal of Geotechnical and
In Global catastrophic risk Bostrom, N and Cirkovik, M.M. Eds Geoenvironmental Engineering ASCE, 139(12):2104-2109.
Global catastrophic risk, pp 287-307, Oxford University Press. Li Y-C, Cleall PJ, Wen YD, Chen YM, Pan Q. 2015. Stresses in soil-
Koerner, R. 2005. Designing with geosynthetics, 5th edition, Pearson- bentonite slurry trench cut-off walls. Geotechnique, 65(10: 843-850.
Prentice Hall, Englewood Cliffs, New Jersey, USA Li Y-C, Wei L, Cleall P J, Lan J-W. 2018. Rankine theory based approach
Koerner, R. 2017. Utilizing PVDs to provide shear strength to saturated for stability analysis of slurry trenches. International Journal of
fine-grained foundation soils, GRI report no. 46, Geosynthetic Geomechanics ASCE, 18(11):06018029.
Institute, Folsom, Pennsylvania, USA Li, Q., Chen, J., Benson, C., and Peng, D. 2021. Hydraulic conductivity
Koerner, R., and Daniel, D. 1995. A suggested methodology for assessing of bentonite-polymer composite geosynthetic clay liners permeated
the technical equivalency of GCLs to CCLs. Geosynthetic Clay with bauxite liquor. Geotextiles and Geomembranes in press.
Liners, R. Koerner, E. Gartung, and H. Zanzinger, Eds., Balkema, Lieske, W., Christ. F., Baille, W., Di Emidio, G., and Wichtmann, T. 2020.
Rotterdam, 73-98. Suction and crack propagation in GCLs subjected to drying and
Koerner, R., and Koerner, J. 2007. GRI’s Second Worldwide Survey of wetting in CaCl2-solutions. Geotextiles and Geomembranes 48(6),
Solid Waste Landfill Liner and Cover Systems, GRI report no. 34, 973-982.
Geosynthetic Institute, Folsom, Pennsylvania, USA Likos, W., and Lu, N. 2006. Pore-scale analysis of bulk volume change
Koerner, G., and Koerner, R. 2017. The durability of exposed from crystalline interlayer swelling in Na+- and Ca2+-smectite.
geomembrane covers. Proc., GeoFrontiers, ASCE, Reston, Virginia, Clays and Clay Minerals 54(4), 516-529.
USA, 139-145. Likos, W., and Wayllace, A. 2010. Porosity evolution of free and
Kolstad, D., Benson, C., Edil, T., and Jo, H. 2004. Hydraulic conductivity confined bentonites during interlayer hydration. Clays and Clay
of a dense prehydrated GCL permeated with aggressive inorganic Minerals 58(3), 399-414.
solutions. Geosynthetics International 11(3), 233-241. Lima, M., Azevedo, M., Zornberg, J., and Palmeira, E. 2018. Capillary
Kozeny, J. 1927. Uber kapillare leitung des wassers im boden. barriers incorporating nonwoven geotextiles. Environmental
Sitzungsber Akad. Wiss. 136(2a), 271-306. Geotechnics 5(3), 168-175.
Kramer, H. 1946. Deep cutoff trench of puddled clay for earth dam and
levee protection, Engineering News Record, 136(26), 76-80. Lin, L., and Benson, C. 2000. Effect of wet-dry cycling on swelling and
Lage, S., Lindner, P., Sinha, P., Kiriy, A., Stamm, M., and Huber, K., 2009. hydraulic conductivity of GCLs. Journal of Geotechnical and
Formation of Ca2+ induced intermediate necklace structures of Geoenvironmental Engineering 126(1), 40-49.
polyacrylate chains. Macromolecules 42, 4288- 4299. Lipson, S.M., and Stotzky, G. 1983. Adsorption of reovirus to clay
Laird, D. 2006. Influence of layer charge on swelling of smectites. minerals: effects of cation-exchange capacity, cation saturation, and
Applied Clay Science 34(1-4), 74-87. surface area. Applied and Environmental Microbiology 46(3), 673-
Lam, C. and Jefferis, S. A. 2021. Effect of sorption on the active 682.
concentration of polymer support fluids, Géotechnique Letters 11, Little, A.L. 1975 in situ diaphragm walls for embankment dams. Proc.
1–6, https//doi.org/10.1680/jgele.20.00062. Conf. on Diaphragm walls and anchorages, Institution of Civil
Lazaro, J. and Breul, B. 2014. Bituminous geomembrane in heap leach Engineers, London. Manassero, M. 1994. Hydraulic conductivity
pads. Proc., Heap Leach Solutions 2014, Info mine, Lima, Peru, 291- assessment of slurry wall using piezocone test. Journal of
303. Geotechnical Engineering, vol. 120(10), 1725-1746, ISSN: 0733-
Lebreton, L., Van Der Zwet, J., Damsteeg, J., Slat, B., Andrady, A., Kaza, 9410.
S., Yao, L., Bhada-Tata, P., and Van Woerden, F. 2018. What a waste Liu, Y., Bouazza, A., Gates, W., and Rowe, R.K. 2015. Hydraulic
2.0: a global snapshot of solid waste management to 2050. The performance of geosynthetic clay liners to sulfuric acid solutions.
World Bank. Geotextiles and Geomembranes 43, 14-23.
Lebreton, L., Van Der Zwet, J., Damsteeg, J., Slat, B., Andrady, A., and Long, X. , Huang, Y., Chi, H., Li, Y., Ahmad, N., and Yao, H. 2018.
Reisser, J. 2017. River plastic emissions to the world’s oceans. Nitrous oxide flux, ammonia oxidizer and denitrifier abundance and
Nature Communications 8, 15611. activity across three different landfill cover soils in Ningbo, China.
Leduc, M., and Smith, M. 2004. Solution collection pipes and Journal of Cleaner Production 170, 288-297.
overstressing of geomembrane liners. The Mining Record, 22-24. Lu, Y., Abuel-Naga, H., Leong, E., Bouazza, A., and Lock, P. 2018.
Lee, J., and Shackelford, C. 2005. Impact of bentonite quality on Effect of water salinity on the water retention curve of geosynthetic
hydraulic conductivity of geosynthetic clay liners. Journal of clay liners. Geotextiles and Geomembranes 46, 707–714.
Geotechnical and Geoenvironmental Engineering 131(1), 64-77. Lupo, J., and Morrison, K. 2005. Innovative geosynthetic liner design
Lee, J., Shackelford, C., Benson, C., Jo, H., and Edil, T. 2005. Correlating approaches and construction in the mining industry. Proc., Geo-
index properties and hydraulic conductivity of geosynthetic clay Frontiers Congress 2005, GSP 140, Reston, Virginia, USA, 1-16.
liners. Journal of Geotechnical and Geoenvironmental Engineering Lupo, J. and Morrison, K. 2007. Geosynthetic design and construction
131(11), 1319-1329. approaches in the mining industry. Geotextiles and Geomembranes
Lee, S., Oren, A., Benson, C., and Dovantzis, K. 2012. Organoclays as 25(2), 96-108
variably permeable reactive barrier (VPRB) media to manage Ma, P., Ke, H., Lan, J., Chen, Y., and He, H. 2019. Field measurement of
NAPLs in ground water. Journal of Geotechnical and pore pressures and liquid-gas distribution using drilling and ERT in
Geoenvironmental Engineering 138(2), 115-127. a high food waste content landfill in Guangzhou, China. Engineering
Legge, K. 2019, An economy and ecosystem symbiosis: barrier systems Geology 250, 21-33.
for water conservation and pollution control. International Paste Malusis, M., and Daniyarov, A. 2016. Membrane efficiency and diffusive
Conference, Cape Town, South Africa. tortuosity of a dense prehydrated geosynthetic clay liner. Geotextiles
Legge, K., Fricker, C., and Mnisi, K. 2015. Pollution prevention and Geomembranes 44, 719-730.
awareness for municipal managers in South Africa. SAICE Civil Malusis, M., Dominijanni, A., Scalia, J., Guarena, N., Sample-Lord, K.,
Engineering Magazine 23(9). Bohnhoff, G., Shackelford, C., and Manassero, M. 2021. Assessing
Legge, K., Fricker, C., Mnisi, K., Nemusibi, K., and Seake, B. 2014. the influence of chemico-osmosis on solute transport in bentonite
Service life considerations of pollution control barrier systems. membranes based on combined phenomenological and physical
SAICE Civil Engineering Magazine 22(7).
286
modeling. 3rd International Symposium on Coupled Phenomena in into a building impacted by a fuel spill in Antarctica. Journal of
Environmental Geotechnics, October 20-22, Kyoto, Japan. Environmental Management, 231 2019): 467-482.
Malusis, M., Kang, J., and Shackelford, C. 2013. Influence of membrane 10.1016/j.jenvman.2018.07.092.
behaviour on solute diffusion through GCLs. Coupled Phenomena Maillard, R. and Serota, S. 1963. Screen grouting of alluvium by the
in Environmental Geotechnics, M. Manassero, A. Dominijanni, S. E.T.F. process, Institution of Civil Engineers Symposium on Grouts
Foti, and G. Musso, Eds., July 1-3, 2013, Torino, Italy, CRC and Drilling Muds in Engineering Practice, Butterworths.
Press/Balkema, Taylor and Francis Group, London, 267-274. Manassero, M. 1994. Slurry wall systems - TC5 report. In: Proceedings
Malusis, M., Kang, J., and Shackelford, C. 2015. Restricted salt diffusion of the First Int. Congress on Environmental Geotechnics: July 10-
in a geosynthetic clay liner. Environmental Geotechnics 2(2), 68-77. 15, 1994. Edmonton, Alberta, Canada. Vancouver:Bi-Tech Publisher
Malusis, M., and Shackelford, C. 2002a. Chemico-osmotic efficiency of Ltd, ISBN: 0921095325, Edmonton (Canada), July, 10-15, 1994.
a geosynthetic clay liner. Journal of Geotechnical and Manassero M. 1997. Pollutants containment via passive barriers.
Geoenvironmental Engineering 128(2), 97-106. Opening Lecture at the Discussion Session 6.1. In: Proceedings of
Malusis, M., and Shackelford, C. 2002b. Coupling effects during steady- the 14th Int. Conf. on Soil Mechanics and Foundation Engineering,
state solute diffusion through a semipermeable clay membrane. Hamburg, 6-12 September 1997. Rotterdam: A.A. Balkema, ISBN:
Environmental Science and Technology 36(6), 1312-1319. 9789054108917, Hamburg (Germany), September, 6-12.
Malusis, M., Shackelford, C., and Olsen, H. 2001. A laboratory Manassero M., Benson C. H., Bouazza A. 2000. Solid waste containment
apparatus to measure chemico-osmotic efficiency coefficients for systems - Key-note Lecture. In: GeoEng 2000: Invited Papers. vol.
clay soils. Geotechnical Testing Journal 24(3), 229-242. 1, 520-642, Lancaster: Technomic Publishing Company Inc., ISBN:
Manassero, M. 2020. Second ISSMGE R. Kerry Rowe Lecture: on the 1587160676, Melbourne, Australia, November, 19-24, 2000.
intrinsic, state, and fabric parameters of active clays for contaminant Manassero M., Fratalocchi E., Pasqualini E., Spanna C., Verga F., 1995.
control. Canadian Geotechnical Journal 57(3), 311-336. Containment with vertical cut-off walls. ASCE Geotechnical Special
Manassero, M., Benson, C.H., and Bouazza, A. 2000. Solid waste Publication, (46 /2), 1142–1172.
containment systems. Proc. GeoEng2000, Technomic, Lancaster, Manassero, M., Pasqualini, E. 1992. Ground pollutant containment
PA, 1, 520-642. barriers. In: Environmental geotechnology: proceedings of the
Manassero, M., and Dominijanni, A. 2003. Modelling the osmosis effect Mediterranean Conf. on Environmental Geotechnology, Cesme,
on solute migration through porous media. Géotechnique 53(5), Turkey, 25-27 May, 1992. A.A. Balkema, Rotterdam. ISBN:
481-492. 9054100559, Cesme (Turkey), May, 25-27.
Manassero, M., Dominijanni, and A., and Guarena, N. 2018. Modelling Manassero M. and Shackelford C. D. 1994. The Role of Diffusion in
hydro-chemo-mechanical behaviour of active clays through the Contaminant Migration through Soil Barriers / Il ruolo della
fabric boundary surface. Proc. China-Europe Conference on diffusione molecolare nella migrazione degli inquinanti attraverso le
Geotechnical Engineering, W. Wu and H. Yu, Eds., Springer barriere minerali. Rivista Italiana di Geotecnica, vol. 1, ISSN: 0557-
International Publishing, Springer Nature Switzerland AG, Vol. 2, 1405.
1618-1626. Manassero M., Van Impe W.F., Bouazza A. 1996. Waste disposal and
Mathur, A., Netravali, N., and O'Rourke, T. 1994. Chemical ageing containment - State of the art lecture. In: Environmental Geotechnics.
effects on the physio-mechanical properties of polyester and A.A. Balkema, Rotterdam. ISBN: 9054108487, Osaka (Japan),
polypropylene geotextiles. Geotextiles and Geomembranes 13(9), November, 5-8, vol. Vol. 3, 193-242.
591-626. Manassero, M., Viola C. 1992. Innovative aspects of leachate
Mazzieri, F., and Di Emidio, G. 2015. Hydraulic conductivity of a dense containment with composite slurry walls; a case history. In: Slurry
prehydrated geosynthetic clay liner. Geosynthetics International Walls: Design, Construction and Quality Control ASTM-STP 1129,
22(1), 138-148. 181-193, Philadelphia: ASTM, ISBN: 9780803114272, Atlantic
Mazzieri, F., Di Emidio, G., and Pasqualini, E. 2016. Effect of wet- City (USA), June, 27-28.
and-dry ageing in seawater on the swelling properties and hydraulic Mazur, A. 1990. A Hazardous Inquiry: The Rashomon Effect at Love
conductivity of two amended bentonites. Applied Clay Science Canal. Harvard University Press.
142(6), 40-51. Meer, S., and Benson, C. 2007. Hydraulic conductivity of geosynthetic
Mazzieri, F., Di Emidio, G., and Van Impe, P. 2010. Diffusion of calcium clay liners exhumed from landfill final covers. Journal of
chloride in a modified bentonite: Impact on osmotic efficiency and Geotechnical and Geoenvironmental Engineering 133(5), 550-563.
hydraulic conductivity. Clays and Clay Minerals 58(3), 351-363. Meyer, W., Meyer, P., and Gundle, C. 2014. Expanding containment
Mazzieri, F., and Pasqualini, E. 2008. Effect of dry/wet cycles and cation barrier boundaries. The 10th International Conference on
exchange on the permeability of a dense prehydrated GCL. Proc., Geosynthetics (10ICG), 21–25 September, Berlin, Germany.
Eurogeo4, Fourth International Conference on Geosynthetic, N. Mitchell, J.K. 1991. Conduction phenomena: from theory to
Dixon, Ed., Edinburgh CD-ROM, paper 125. geotechnical practice, Géotechnique, 41(3):299-340.
McLaren, A.D., and Peterson, G.H. 1965. Physical chemistry and Michell, J., and Soga, K. 2005. Fundamentals of soil behaviour, 3rd
biological chemistry of clay mineral-organic nitrogen complexes. edition, John Wiley and Sons, Inc., New York.
Soil Nitrogen - Monograph 10, American Society of Agronomy, Mitchell, J.K., Hooper, D.R. and Campanella, R.G., 1965. Permeabþ of
Wisconsin, 259-284. compacted clay. Soil Mechanics and Foundations Division, ASCE,
McRory, J., and Ashmawy, A. 2005. Polymer treatment of bentonite clay 91(S4):41-65.
for contaminant resistant barriers. Geo-Fronteirs 2005, E. Rathje, Michen, B., and Graule, T. 2010. Isoelectric points of viruses. Journal of
Ed., GSP 142 Waste Containment and Remediation, ASCE, Reston, Applied Microbiology 109(2), 388-397.
Virginia, USA. Montell Polyolefins Brochure, 1997. Astryn FPA high performance
McWatters, R. and Rowe, R.K. 2010. Diffusive transport of VOCs materials for geomembranes.
through LLDPE and two Co-Extruded Geomembranes. Journal of Millet, R.A., Perez, J.Y. and Davidson, R.R. 1992. USA practice slurry
Geotechnical and Geoenvironmental Engineering, 136(9): 1107- wall specifications 10 years later. In Slurry walls: Design,
1177. construction, and quality control. ASTM STP, Slurry walls design,
McWatters, R.S., Rowe, R.K., Wilkins, D., Spedding, T., Jones, D., Wise, construction and quality control, 1129, 42-66
L., Mets, J., Terry, D., Hince, G., Gates, W.P., Di Battista, V., Shoaib, Moore, R.S., Taylor, D.H., Sturman, L.S., Reddy, M.M., and Fuhs, G.W.
M., Bouazza, A. and Snape, I. 2016b. Geosynthetics in Antarctica: 1981. Poliovirus adsorption by 34 minerals and soils. Applied and
Performance of a composite barrier system to contain hydrocarbon- Environmental Microbiology 42, 963-975.
contaminated soil after 3 years in the field. Geotextiles and Moraes F. and Gardoni, M. 2020. Geotextile tubes to dewater and confine
Geomembranes, 44(5):673-685. gold tailings. 4th Pan American Conference on Geosynthetics, Rio
http://dx.doi.org/10.1016/j.geotexmem.2016.06.001 de Janeiro, Brazil, (in press).
McWatters, R.S., Rutter, A. and Rowe, R.K. 2016a. Geomembrane Morais, F. 2020. Performance of a drained stacking with geotextile tubes
Applications for Controlling Diffusive Migration of Petroleum - New proposal for disposal of mining fine waste, MSc. Thesis.
Hydrocarbons in Cold Region Environments. J. Environ Federal University of Minas Gerais, Dept. of Transport and
Management, 171: 80-94; Geotechnical Engineering, Belo Horizonte, Brazil. (in progress) (in
http://dx.doi.org/10.1016/j.jenvman.2016.05.065 Portuguese).
McWatters,R.S., Rowe, R.K., Wilkins, D., Spedding, T., Hince, G., Moreira, T. 2011. Experimental evaluation of drainage systems with
Richardson, J., and Snape, J. 2019. Modelling of vapour intrusion geosynthetics to decrease iron ore moisture content of the Carajás
287
mining complex, MSc. Thesis, Federal University of Minas Gerais, geosynthetic clay liners used as barriers. Journal of Environmental
Dept. of Transportation and Geotechnical Engineering, Brazil (in Engineering 144(10), 04018100.
Portuguese). Ozhan, H. 2018b. Hydraulic capability of polymer-treated GCLs in
Morgenstern, N., Vick, S. and Van Zyl, D. 2015. Report on Mount Polley saline solutions at elevated temperatures. Applied Clay Science 161,
Tailings Storage Facility Breach, Independent Expert Engineering 364-373.
Investigation and Review Panel, British Columbia, 30 January 2015, Paleologos, E.K., O’Kelly, B.C., Tang, C.-S., Cornell, K., Rodríguez-
unpublished. Chueca, J., Abuel-Naga, H., Koda, E., Farid, A., Vaverková, M.D.,
Morsy, M., and Rowe, R.K. 2017. Performance of blended polyolefin Kostarelos, K., Goli, V.S.N.S., Guerra-Rodríguez, S., Leong, E.C.,
geomembrane in various incubation media based on Std-OIT. Proc., Jayanthi, P., Shashank, B.S., Sharma, S., Shreedhar, S., Mohammad,
GeoFrontiers, ASCE, Reston, Virginia, USA, 1–10. A., Jha, B., Kuntikana, G., Bo, M.W., Mohamed, A.M.O. and Singh,
Msiza, J., 2020. Email Regulatory Standards in Africa for Pollution D.N. 2021. Post COVID-19 water and wastewater management to
Control Barrier Systems confirming little progress with since protect public health and geoenvironment. Environmental
request to IGS Technical Committee on Barriers in 2014, dated 22 Geotechnics 8(3), 193-207.
August 2020. Palmeira, E., Beirigo, E., and Gardoni, M. 2010. Tailings-nonwoven
Mukunoki, T., Nakano, T., Otani, J., and Gourc, J., 2014. Study of geotextile filter compatibility in mining applications. Geotextiles
cracking process of clay cap barrier in landfill using x-ray CT. and Geomembranes 28, 136-148.
Applied Clay Science 101(10), 558–566. Palmeira, E., Fannin, R., and Vaid, Y. 1996. A study on the behaviour of
Musso, G., Cosentini, R., Dominijanni, A., Guarena, N., Manassero, M. soil-geotextile systems in filtration tests. Canadian Geotechnical
2017. Laboratory characterization of the chemo-hydro-mechanical Journal 33(4), 899-912.
behaviour of chemically sensitive clays. Rivista Italiana di Palmeira, E., and Gardoni, M. 2000. The influence of partial clogging
Geotecnica 51(3), 22-47. and pressure on the behaviour of geotextiles in drainage systems.
Musso, G., Cosentini, R.M., Dominijanni, A., Guarena, N., and Geosynthetics International 7(4–6), 403–431.
Manassero, M. 2017. Laboratory characterization of the chemo- Palmeira, E., and Gardoni, M. 2002. Drainage and filtration properties of
hydro-mechanical behaviour of chemically sensitive clays. Rivista nonwoven geotextiles under confinement using different
Italiana di Geotecnica 51(3), 22-47. experimental techniques. Geotextiles and Geomembranes 20, 97-
National Environmental Management Waste Act (NEMWA) 2013, Act 115.
59 of 2008, Regulation 636, (2013. Norms and Standards for Palmeira, E., Gardoni, M., and Bessa da Luz, D. 2005. Soil-geotextile
Disposal of Waste to Landfill, 23 August 2013, South Africa). filter interaction under high stress levels in the gradient ratio test.
Ng, C., Chen, Z., Coo, J., Chen, R., and Zhou, C. 2015d. Gas Geosynthetics International 12(4), 162-175.
breakthrough and emission through unsaturated compacted clay in Palmeira, E., and Trejos Galvis, H. 2017. Opening sizes and filtration
landfill final cover. Waste Management 44, 155-163. behaviour of nonwoven geotextiles under confined and partial
Ng, C., Chen, R., Coo, J., Liu, J., Ni, J., Chen, Y., and Lu, B. 2019. A clogging conditions. Geosynthetics International 24(2), 125-138.
novel vegetated three-layer landfill cover system using recycled Parra D., César C., and Mendoza, J. 2014. Geomembrane and pipe design
construction wastes without geomembrane. Canadian Geotechnical issues in deeper heap leach pads. Proc., Heap Leach Solutions 2014,
Journal 56(12), 1863-1875. Lima, Peru, 167-179.
Ng, C., Coo, J., Chen, Z., and Chen, R. 2016. Water infiltration into a Pavanello, P, Carrubba, P., and Moraci, N. 2018. Dynamic friction and
new three-layer landfill cover system. Journal of Environmental the seismic performance of geosynthetic interfaces. Geotextiles and
Engineering 142(5), 04016007. Geomembranes 46, 715–725.
Ng, C., Liu, J., and Chen, R. 2015a. Numerical investigation on gas Peggs, I. 2008. Prefabricated bituminous geomembrane: a candidate for
emission from three landfill soil covers under dry weather conditions. exposed geomembrane caps for landfill closures. Proc., First Pan
Vadose Zone Journal 14(8). American Geosynthetics Conference and Exhibition, Cancun,
Ng, C., Liu, J., Chen, R., and Coo, J. 2015b. Numerical parametric Mexico, 191-197.
study of an alternative three-layer capillary barrier cover system. Peggs, I., Schmucker, B. and Carey, P. 2005. Assessment of maximum
Environmental Earth Sciences 74(5), 4419-4429. allowable strains in polypropylene geomembranes. Proc.,
Ng, C., Liu, J., Chen, R., and Xu, J. 2015c. Physical and numerical Geofrontiers 2005, ASCE, Reston, Virginia, USA.
modeling of an inclined three-layer (silt/gravelly sand/clay) Petrov, R., Rowe, R.K., and Quigley, R. 1997a. Comparison of laboratory
capillary barrier cover system under extreme rainfall. Waste measured GCL hydraulic conductivity based on three permeameter
Management 38, 210-221. types. Geotechnical Testing Journal 20(1), 49-62.
Norris, A. 2021. Mechanisms of interaction between montmorillonite Petrov, R. Rowe, R.K., and Quigley, R. 1997b. Selected factors
and anionic polymer amendments in geosynthetic clay liners, Ph.D. influencing GCL hydraulic conductivity. Journal of Geotechnical
dissertation, Colorado State University, Fort Collins, Colorado, USA. and Geoenvironmental Engineering 123(8), 683-695.
Norris, A., Aghazamani, N., Conzelmann, J., Scalia, J., and Shackelford, Petrov, R., and Rowe, R.K. 1997. Geosynthetic clay liner (GCL) –
C. 2020. Hydraulic conductivity of geosynthetic clay liners to chemical compatibility by hydraulic conductivity testing and factors
synthetic mine waste leachates. Tailings and Mine Waste 2020, UBC impacting its performance. Canadian Geotechnical Journal 34(6),
Studios, University of British Columbia, 1102329 West Mall, 863-885.
Vancouver, British Columbia V6T 1Z4, Canada. Plaut, R., and Suherman, S. 1998. Two-dimensional analysis of
Norris, A., Scalia, J., and Shackelford, C. 2022. Polymer quantification geosynthetic tube. Acta Mechanica 3(4), 207-218.
methods for geosynthetic clay liners enhanced with anionic Priyanto, D., Whitelaw, B., Rowe, R.K., Barone, F., Abdelaal, F., Zafari,
polymers. Geotechnical Testing Journal 45(2), GTJ20210134.. M., Buckley, J. 2019. Implementation of Research and Development
Norrish, X. 1954. Manner of swelling of montmorillonite. Nature 173, (R&D) Results in the Design of Liner System for the Near Surface
256-257. Disposal Facility (NSDF) – 19089, WM2019 Conference, March 3
Olsen, S., and Kemper, W. 1968. Movement of nutrients to plant roots. – 7, 2019, Phoenix, Arizona, USA.
Advances in Agronomy 20, 91-151. Prongmanee, N., and Chai, J. 2019. Performance of geosynthetic clay
Onikata, M., Kondo, M., Hayashi, N., and Yamanaka, S. 1996. liner with polymerized bentonite in highly acidic or alkaline
Development and characterization of multiswellable bentonite. 2nd solutions. International Journal of Geosynthetics and Ground
International Congress on Environmental Geotechnics, IS-Osaka '96, Engineering 5(3), Article 26.
M. Kamon, Ed., Nov. 5-8, 1996, Osaka, Japan, Balkema, Rotterdam, Prongmanee, N., Chai, J., and Shrestha, S. 2018a. Effects of cations on
Vol. 1, 587-590. consolidation and permeability of polymerized bentonite. Lowland
Onikata, M., Kondo, M., Hayashi, N., and Yamanaka, S. 1999. Complex Technology International 20(3), 297-304.
formation of cation-exchanged montmorillonites with propylene Prongmanee, N., Chai, J., and Shen, S. 2018b. Hydraulic properties of
carbonate: Osmotic swelling in aqueous electrolyte solutions. Clays polymerized bentonites. Journal of Materials in Civil Engineering
and Clay Minerals 47(5), 672-677. 30(10), 04018247.
Opdyke S. M., Evans J. C. 2005. Slag-cement-bentonite slurry walls, Qiu, H., and Yu, J. 2008. Polyacrylate/(carboxymethylcellulose modified
Journal of Geotechnical and Geoenvironmental Engineering, 131(6): montmorillonite) superabsorbent nanocomposite: preparation and
673-681. water absorbency. Journal of Applied Polymer Science 107(1), 118-
Ozhan, H. 2018a. Effects of temperature increase in 0.5 M MgCl2 123.
solution on hydraulic capability of anionic polymer-treated
288
Quigley, R.M. and Rowe, R.K. 1986. Leachate migration through clay in bottom liners and covers for use in
below a domestic waste landfill, Sarnia, Ontario, Canada: Chemical lagoons, secondary containment, landfill and mining applications,
interpretation and modelling philosophies. Hazardous and GIGSA Seminar, 6-7 September, Stone Cradle, Gauteng, South
Industrial Solid Waste Testing and Disposal: Sixth Volume, ASTM Africa.
-term leakage through composite liners
Quigley, R.M., Fernandez, F. and Rowe, R.K. 1988. “Clayey barrier (7th Arthur Casagrande Lecture). Canadian Geotechnical Journal
assessment for impoundment of domestic waste leachate (southern 49(2), 141–169.
Ontario) including clay-leachate compatibility by hydraulic
conductivity testing. Canadian Geotechnical Journal, 25(3):574-
581. 8-
Quigley, R.M., Gwyn, Q.H.J., White, O.L., Rowe, R.K., Haynes, J.E. and ions and mis-
Bohdanowicz, A. 1983. “Leda clay from deep boreholes at conceptions. Geotextiles and Geomembranes 48(2), 137–156.
Hawkesbury, Ontario, Part I, Geology and Geotechnique. - The
53rd Karl Terzaghi Lecture. Journal of Geotechnical and
Rabideau, A., and Khandelwal, A. 1998. Boundary conditions for
modeling transport in vertical barriers. Journal of Environmental Rowe, R.K. -rich soil on the
Engineering 124(11), 1135-1139. performance of an overlying GCL. Journal of Geotechnical and
Ratkievicius, L., Da Cunha Filho, F., De Barros Neto, E., and Santanna, Geoenvironmental Engineering 138(4): 423-431.
ite clay by a cationic surfactant to Rowe, R.K., and Booker, J.R. 1985a. 1-D pollutant migration in soils of
be used as a viscosity enhancer in vegetable-oil-based drilling fluid. finite depth. Journal of Geotechnical Engineering 111(4), 479-499.
-312. Rowe, R.K., and Booker, J.R. 1985b. Two-dimensional pollutant
Razakamanantsoa, A., Barast, G., and Djeran- migration in soils of finite depth. Canadian Geotechnical Journal
performance of activated calcium bentonite treated by polyionic 22(4), 429–436.
charged polymer. Applied Clay Science 59- -114. R solutions for calculating
Reddy, K.R., Kumar, G., and Du, Y.- leakage through composite liner systems. in Developments in
Resiliency Consideration in Polluted Site Remediation, Proc. 8th Theoretical Geomechanics - The John Booker Memorial
International Congress on Environmental Geotechnics, Springer, -
Singapore, Vol. 1, 145-163.
-assessment through a nonhomogeneous soil, © 1983–
methodology: a UK case study for landfills, Water and Environment GAEA Environmental Engineering Ltd, 87 Garden Street, Whitby,
Journal 18(3), 139-143. Ontario, Canada L1N 9E7.
s research
and applications in the mining and mineral processing environment. composite liners, Geosynthetics International 11(4), 273-286.
North American Geosynthetics Society (NAGS) -Geosynthetic
Institute (GSI) Conference, Las Vegas, Nevada. December 14-16. geomembrane with high and low initial high-pressure oxidative
fect of induction time. Geotextile and Geomembranes 42(2), 111–126.
geomembrane colour and cover soil on solar-driven down slope
bentonite erosion from a GCL. Geosynthetics International 23(4), locations with different climatological conditions. Canadian
257– Geotechnical Journal 52 (3), 326-343.
Comparison of wrinkles in white and black HDPE geomembranes. leakage overlain by saturated tailings. Geotextiles and
Journal of Geotechnical and Geoenvironmental Engineering 143(8), -1518.
Rowe, R.K. and Fraser, M.J. 1995. Effect of uncertainty in the
assessment of the potential impact of waste disposal facilities,
challenges and solutions beiynd incremental changes. Proc. Tailings -
and Mine Waste, Vancouver, British Columbia, Canada. 284.
-dry cycles on standard
Water vapour adsorption and desorption in GCLs. Geosynthetics and polymer modified GCLs in covers subjected to flow over the
International 23(2), 86-99 GCL. Geotextiles and Geomembranes 49(4),1155-1165.
long-
gas flow unified measurement system for measuring sequentially term wet-dry cycles under a defect in GMB. Geosynthetics
gas diffusion and gas permeability of partially hydrated geosynthetic International (in press).
-
performance for covering arsenic contaminated mine wastes.
permeability of partially hydrated geosynthetic clay liner under two Geotextiles and Geomembranes 39, 63-77.
stress conditions. Journal of Environmental Geotechnics 3(5), 325- -healing of circular and slit defects in
333 GCLs upon hydration from silty sand under applied stress.
Rowe, R.K. 1987. Pollutant transport through barriers. Proceedings of Geotextiles and Geomembranes 48, 667-683.
ASCE Specialty Conference, Geotechnical Practice for Waste
Di self-healing with hydration from silica sand. Geotextiles and
Rowe, R.K. 1988. Contaminant migration through groundwater: The role Geomembranes 49(4), 871-894.
of modelling in the design of barriers. Canadian Geotechnical Rowe, R.K. and Nadarajah, P. 1996. Estimating leachate drawdown due
Journal, 25(4):778-798. to pumping wells in landfills. Canadian Geotechnical Journal,
Rowe, R.K. 1992. Integration of hydrogeology and engineering in the 33(1):1-
design of waste management sites. Proceedings of the International
Association of Hydrogeologists Conference on “Modern Trends in contaminated water. Geofilters 4th International Conference,
Hydrogeology. Hamilton, pp. 7-21. Stellenbosch, RSA, 1–63.
Rowe, R.K. 1998. Geosynthetics and the minimization of contaminant
migration through barrier systems beneath solid waste. 6th desiccation in low-stress applications Geosynthetics International
International Conference on Geosynthetics, IFAI, Roseville, -315.
Minnesota, USA, Vol. 1, 27- eachate
-term performance of contaminant barrier collection systems in MSW landfills. 6th International Conference
systems. Géotechnique 55(9), 631-678.
3-23.
municipal solid waste landfills to minimize leakage of contaminants
to groundwater. Proc. 3rd Zeng Gou-Xi Lecture, Geosyntethics life of MSW leachate collection systems. Canadian Geotechnical
International 16(6), 391- -178.
289
Rowe, R.K., and Yu, Y. 2018. Tensile strains in geomembrane landfill Ruffing, D.G. and Evans, J.C. 2022,. Slurry trenching, (in prep.).
liners. Keynote Lecture GeoShanghai 2018, Shanghai, China, 1-10. Ruffing, D.G., Evans, J.C. and Coughenour, N. 2018. Soil-Bentonite
Rowe, R.K., and Yu, Y. 2019. Magnitude and significance of tensile Slurry Trench Cut-off Wall Longevity. Int. Foundations Congress
strains in geomembrane landfill liners. Geotextiles and and Equipment Expo, March 5-10, 2018 214-223.
Geomembranes 47(3), 439-458. Ryan,C.R., Ruffing, D, and Evans, J. C. 2022. Soil Bentonite Slurry
Rowe, R.K., Caers, C.J. and Chan, C. 1993. Evaluation of a compacted Trench Cutoff Walls: History, Design, and Construction Practices. In
till liner test pad constructed over a granular subliner contingency Geo-Congress: State of the Art and Practice on Geotechnical
layer, Canadian Geotechnical Journal, 30(4):667-689. Engineering. Reston, VA: ASCE.
Rowe, R.K., Quigley, R., Brachman, R., and Booker, J. 2004. Barrier Sadeghi, G., Behrends, T., Schijven, J.F., and Hassanizadeh, S.M. 2013.
systems for waste disposal facilities, 2nd edition, Taylor and Francis Effect of dissolved calcium on the removal of bacteriophage PRD1
Books Ltd., London. during soil passage: the role of double-layer interactions. Journal of
Rowe, R.K., Bostwick, L., Take, W.A. 2011a. Effect of GCL properties Contaminant Hydrology 144(1), 78-87.
on shrinkage when subjected to Wet-Dry Cycles. Journal of Sadeghi, G., Schijven, J.F., Behrends, T., Hassanizadeh, S.M., Gerritse,
Geotechnical and Geoenvironmental Engineering 137(11), 1019- J., and Kleingeld, P.J. 2011. Systematic study of effects of pH and
1027. ionic strength on attachment of phage PRD1. Ground Water 49(1),
Rowe, R.K., Rayhani, M., Take, W., Siemens, G., and Brachman, R. 12-19.
2011b. GCL Hydration under simulated daily thermal cycles. Salles, F., Douillard, J., Denoyel, R., Bildstein, O., Jullien, M., Beurroies,
Geosynthetics International 18(4), 196-205. I., and Damme, H. 2009. Hydration sequence of swelling clays:
Rowe, R.K., Chappel, M., Brachman, R., and Take, W.A. 2012a. Field evolutions of specific surface area and hydration energy. Journal of
monitoring of geomembrane wrinkles at a composite liner test site. Colloid and Interface Science 333, 510-522.
Canadian Geotechnical Journal 49(10), 1196-1211. Samea, A., and Abdelaal, F. 2019. Chemical durability of bituminous
Rowe, R.K., Yang, P., Chappel, M., Brachman, R., and Take, W.A 2012b. geomembranes in heap leaching applications at 55oC. Geosynthetics
Wrinkling of a geomembrane on a compacted clay liner on a slope. 2019 Conference, Houston, Texas, USA.
Geotechnical Engineering, Journal of the South East Asian Sampaio S., Gardoni M., Palmeira E. 2015. Interface and mechanical
Geotechnical Society 43(3), 11-18. damage tests on geosynthetics used in gold tailings dams in Brazil.
Rowe, R.K., Brachman, R., Irfan, H., Smith, M., and Thiel, R. 2013a. Proc., 2a Conferência Internacional Dam World, LNEC, Lisboa,
Effect of underliner on geomembrane strains in heap leach Portugal, 112-116.
applications. Geotextiles and Geomembranes 40, 37-47. Sampaio S., Gardoni M., Palmeira E., and Park, J. 2014. Stability
Rowe, R.K., Rayhani, M., Take, W.A, Siemens, G. and Brachman, R.W.I analysis of Jacobina’s tailings dam, based on interface parameters of
2013b. Physical modelling of nonwoven-nonwoven GCL shrinkage geomembrane-tailings obtained from laboratory tests, Proc., 10th
under simulated field Conditions”, Geotextiles and Geomembranes, International Conference on Geosynthetics, Berlin, Germany, 1022-
40: 12-19. 1028.
Rowe, R.K., Abdelaal, F., and Brachman, R. 2013c. Antioxidant Sampaio, S., Pimentel K., Halabi, A. and Gardoni, M. 2017. Stability
depletion from an HDPE geomembrane with a sand protection layer. Analysis of a Gold Tailings Dam Lined with HDPE Geomembrane,
Geosynthetics International 20(2), 73-89. Proc., 1st International Conference on Technology and Applications
Rowe, R.K., Ashe, L, Take W.A., and Brachman, R. 2014. Factors of Geosynthetics, Santiago, Chile, 212-220.
affecting the downslope erosion of bentonite in a GCL. Geotextiles Sample-Lord, K., Bohnhoff, G., and Tong, S. 2017. Diffusion of calcium
and Geomembranes 42(5), 445-456 chloride through polymerized bentonite. Proc., GeoFrontiers, ASCE,
Rowe, R.K., Brachman, R., Take, W.A., Rentz, A. and Ashe, L. 2016a. Reston, Virginia, USA, 200–208.
Field and laboratory observations of downslope bentonite migration Santos, L. and Gardoni M. 2014. Study of the durability of
in exposed composite liners. Geotextiles and Geomembranes 44(5), geomembranes for waterproofing of reservoirs of gold tailings dams
686-706 in Brazil. Proc., 10th International Conference on Geosynthetics,
Rowe, R.K., Rentz, A., Brachman, R., and Take W.A. 2016b. Effect of Berlin, Germany, 214-220.
GCL type on downslope bentonite erosion in an exposed liner. Santos, L. and Gardoni, M. 2016. Evaluation of geomembrane
Journal of Geotechnical and Geoenvironmental Engineering degradation aged in laboratory and exposed to weathering. Proc., 3rd
142(12), 4016074. Pan American Conference on Geosynthetics, Miami Beach, USA,
Rowe, R.K., Joshi, P., Brachman, R.W.I and McLeod, H. 2017. Leakage 156-164.
through holes in geomembranes below saturated tailings. ASCE Sarsby, R.W. 2019. Environmental Geotechnics in Practice, Institution of
Journal of Geotechnical and Geoenvironmental Engineering, Civil Engineers (ICE) Publishing, United Kingdom.
143(2):4016099-1 to 10; DOI: 10.1061/(ASCE)GT.1943- Sari, K., and Chai, J. 2013. Self healing capacity of geosynthetic clay
5606.0001606. liners and influencing factors. Geotextiles and Geomembranes 41,
Rowe, R.K., Brachman, R.W.I. and Take, W.A. 2018 Field measurements 64-71.
of overlap reductions for two reinforced fabric-encased GCLs. Scalia, J. 2012. Bentonite-polymer composites for containment
Canadian Geotechnical Journal, 55(5): 631-639. applications, Ph.D. dissertation, University of Wisconsin-Madison,
https://doi.org/10.1139/cgj-2017-0375 Madison, Wisconsin, USA.
Rowe, R.K., Brachman, R., Hosney, M., Take, W.A., and Arnepalli, D. Scalia, J., Bareither, C., and Shackelford, C. 2018b. Advancing the use
2017. Insight into hydraulic conductivity testing of GCLs exhumed of geosynthetic clay liners as barriers. Geotechnical Engineering
after 5 and 7 years in a cover. Canadian Geotechnical Journal 54(8), Journal of the SEAGS and AGSSEA 49(4), 100-114.
1118-1138. Scalia, J and Benson, C. 2011. Hydraulic conductivity of geosynthetic
Rowe, R.K., Morsy, M.S. and Ewais, A.M.R. 2019a. A Representative clay liners exhumed from landfill final covers with composite
Stress Crack Resistance for Polyolefin Geomembranes Used in barriers. Journal of Geotechnical and Geoenvironmental
Waste Management, Waste Management, 100: 18-27. Engineering 137(1), 1-13.
10.1016/j.wasman.2019.08.028 Scalia, J., and Benson, C. 2011. Hydraulic conductivity of geosynthetic
Rowe, R.K., Garcia, J., Brachman, R., and Hosney, M. 2019b. Hydraulic clay liners exhumed from landfill final covers with composite
and chemical performance of geosynthetic clay liners isothermally barriers. Journal of Geotechnical and Geoenvironmental
hydrated from silty sand subgrade. Geotextiles and Geomembranes Engineering 137(1), 1-13.
47(6), 740-754. Scalia, J., and Benson, C. 2016. Evaluation of Na-bentonite-polyacrylate
Rowe, R.K., Priyanto, D. and Poonan, R. 2019c. Factors affecting the mixtures to enhance the chemical resistance of geosynthetic clay
design life of HDPE geomembranes in an LLW disposal facility, liners. Proc., Geo-Chicago 2016, GSP No. 271, ASCE, Reston,
WM2019 Conference, March 3 – 7, 2019, Phoenix, Arizona, USA, Virginia, USA, 388-397.
15p. Scalia, J., and Benson, C. 2017. Polymer fouling and hydraulic
Rowe, R.K., Abdelaal, F., Zafari, M. Morsy, M., and Priyanto, D. 2020. conductivity of mixtures of sodium bentonite and a bentonite-
An approach to geomembrane selection for challenging design polymer composite. Journal of Geotechnical and Geoenvironmental
requirements. Canadian Geotechnical Journal 57(10), 1550-1565. Engineering 143(4), 04016112.
Rowe, R.K., Garcia, J., Brachman, R., and Hosney, M. 2021. Moisture Scalia, J., Benson, C., Bohnhoff, G., Edil, T., and Shackelford, C. 2014.
uptake and loss of GCLs subjected to daily thermal cycles from a Long-term hydraulic conductivity of a bentonite-polymer composite
silty sand subgrade. Geosynthetics International (in press).
290
permeated with aggressive inorganic solutions. Journal of Silva, L. 2017. Geotextile tubes to dewater, confine, and stack finegold
Geotechnical and Geoenvironmental Engineering 140(3), 04013025. ore tailings, M.Sc Thesis, Federal University of Minas Gerais,
Scalia, J., Benson, C., Edil, T., Bohnhoff, G., and Shackelford, C. 2011. Dept. of Transportation and Geotechnical Engineering, Brazil,
Geosynthetic clay liners containing bentonite polymer Department of Transportation and Geotechnical Engineering. Belo
nanocomposite. GeoFrontiers 2011 Advances in Geotechnical Horizonte, Brazil (in Portuguese).
Engineering, J. Han and D. Alazamora, Eds., ASCE, Reston, Silva, V. 2019. Analysis of the behaviour of geotextile pipes for
Virginia, USA, 2001-2009. containment of fine mining waste through numerical modeling, MSc.
Scalia, J., Bohnhoff, G., Shackelford, C., Benson, C., Sample-Lord, K., Thesis, Federal University of Minas Gerais, Dept. of Transportation
Malusis, M., and Likos, W. 2018a. Enhanced bentonites for and Geotechnical Engineering, Brazil, 168 p. (in Portuguese).
containment of inorganic wastes by GCLs. Geosynthetics Silva V., Gardoni, M., and Maia, K. 2020. Numerical modeling of
International 25(4), 392-411. geotextile tubes filling process with gold mine tailings.
Scheirs, J. 2009. A guide to polymeric geomembranes: a practical Geosynthetics International (in press).
approach, John Wiley and Sons, Inc. Simeral, L., and Amey, R. 1970. Dielectric properties of liquid propylene
Scheirs, J., Pospisil, J., O’Connor, M.J., and Bigger, S. W. 1996. carbonate. Journal of Physical Chemistry 74(7), 1443-1446.
Characterization of conversion products formed during degradation Singh, A., 2016. Confirmation letter to the Chamber of Mines SA, Risk
of processing antioxidants. Polymer Durability, 359-374. Based Approach for Assessment Water Use License Applications in
(https://pubs.acs.org/doi/abs/10.1021/ba-1996-0249.ch024) Relation of Facilities for Sction 21 G Water Uses in the Mining
Scheller, C., Krebs, F., Minkner, R., Astner, I., Gil-Moles, M. and Wätzig, Sector, reference WULA/1/2016 dated 29/06/2016.
H. 2020. Physicochemical properties of SARS-CoV-2 for drug Singh, R. and Bouazza, A. 2013. Thermal conductivity of geosynthetics.
targeting, virus inactivation and attenuation, vaccine formulation Geotextiles and Geomembranes 39, 1–8.
and quality control. Electrophoresis 41(13-14), 1137-1151. Singh, M., Kijko, A. and van den Berg, L. 2010. Seismic risk ranking for
Schluchter B., Soyez B., Jessberger H. L., Manassero, M., Puthaar T., large dams in South Africa. Acta Geophysica, Institute of
Street A. 1991. Geotechnics of landfills and contaminated land: Geophysics, Polish Academy of Sciences.
technical recommendations, Berlin: Ernst and Sohn, ISBN: Smith, M. 2004. Drainage pipe deflection for high heaps. The Mining
3433012148, 1-76,. Record, 4th Quarter, 15-19. 2017c). Field measurements of water
Schluchter C., Soyez B. Jessberger H. L., Manassero, M., Ruardi P., storage capacity in a loess–gravel capillary barrier cover using
Street A. 1993. Geotechnics of landfill design and remedial works: rainfall simulation tests. Canadian Geotechnical Journal 54(11),
technical recommendations, Berlin: Ernst and Sohn, ISBN: 1523-1536.
3433012733, 1-158, Soga, K., Ewais, A., Fern, J., and Park, J. 2019. Advances in
Sethi, R., and Di Molfetta, A. 2019. Groundwater Engineering. A Geotechnical Sensors and Monitoring. In: Lu, N., and Mitchell, J.K.
Technical Approach to Hydrogeology, Contaminant Transport and (Eds.), Geotechnical Fundamentals for Addressing New World
Groundwater Remediation, Springer, Switzerland. Challenges, Springer, Switzerland, 29-65.
Shackelford, C. 2008. Selected issues affecting the use and performance Soong, T.-Y., and Koerner, R. 1998. Laboratory study of high density
of GCLs in waste containment applications. Geosynthetics and polyethylene geomembrane waves. Proc., 6th Int. Conf. on
Environment, Proceedings of the Geotechnical Engineering Geosynthetics, Austin, Texas, International Geosynthetics Society, 1,
Conference of Torino (XXI Edition), M. Manassero and A. 301-306.
Dominijanni, Eds., Politecnico di Torino, Torino, Italy, Nov. 27-28, Soule, N., and Burns, S. 2001. Effects of organic cation structure on
2007, Pàtron Editore, Bologna, Italy. behaviour of organobentonites. Journal of Geotechnical and
Shackelford, C.D. 1990. Transit-time design of earthen barriers. Geoenvironmental Engineering 127(4), 363-370.
Engineering Geology 29(1), 79-94. Southen, J., and Rowe, R.K. 2004. Investigation of the behaviour of
Shackelford, C.D. 2005. Environmental issues in geotechnical geosynthetic clay liners subjected to thermal gradients in basal liner
engineering. Proc. 16th International Conference on Soil Mechanics applications. Journal of ASTM International 1(2) ID JAI11470,
and Geotechnical Engineering, Millpress, Rotterdam, The Online; Available: Hwww.astm.orgH; 1 February 2004, 13p. This
Netherlands, Vol. 1, 95-122. was also published by ASTM in hardcopy in Advances in
Shackelford, C.D. 2014. The ISSMGE Kerry Rowe Lecture: The role of Geosynthetic Clay Liner Technology”, Mackey, R.E. and von
diffusion in environmental geotechnics. Canadian Geotechnical Maubeuge, K. (Eds), pp. 121-133.
Journal 51(11), 1219-1242. Southen, J., and Rowe, R.K. 2005. Laboratory investigation of
Shackelford, C., and Daniel, D. 1991. Diffusion in saturated soil: I. geosynthetic clay liner desiccation in a composite liner subjected to
background. Journal of Geotechnical Engineering 117(3), 467-484. thermal gradients. Journal of Geotechnical and Geoenvironmental
Shackelford, C., and Moore, S. 2013. Fickian diffusion of radionuclides Engineering 131(7), 925-935.
for engineered containment barriers: Diffusion coefficients, Southen, J., and Rowe, R.K. 2007. Evaluation of the soil-water
porosities, and complicating issues. Engineering Geology 152(1), characteristic curve for geosynthetic clay liners. Geotextiles and
133-147. Geomembranes 25(1), 2-9.
Shackelford, C., and Scalia IV, J. 2016. Semipermeable membrane Southen, J., and Rowe, R.K. 2011. Numerical modelling of thermally
behaviour in bentonite-based barriers: past, present, and future. induced desiccation of geosynthetic clay liners observed in
GEOVANCOUVER 2016, 69th Canadian Geotechnical Conference, laboratory experiments. Geosynthetics International 18 (5), 289-303.
Oct. 2-5, 2016, Vancouver, Canada, Canadian Geotechnical Society, Stone, J. 1984. Leakage monitoring of the geomembrane for proton
Richmond, British Columbia, Canada, paper 4173. decay experiment. Proc., International Conference on
Shackelford, C., Benson, C., Katsumi, T., Edil, T., and Lin, L. 2000. Geomembranes, IFAI, St. Paul, Minnesota, USA, Vol. 2, 475–480.
Evaluating the hydraulic conductivity of GCLs permeated with non- Syngouna, V.I., and Chrysikopoulos, C.V. 2010. Interaction between
standard liquids. Geotextiles and Geomembranes 18(2-4), 133-161. viruses and clays in static and dynamic batch systems.
Shackelford, C., Meier, A., and Sample-Lord, K. 2016. Limiting Environmental Science and Technology 44(12), 4539-4544.
membrane and diffusion behaviour of a geosynthetic clay liner. Szczerba, M., and Kalinichev, A. 2016. Intercalation of ethylene glycol
Geotextiles and Geomembranes 44(5), 707-718. in smectites: several molecular simulation models verified by x-ray
Shackelford, C., Sevick, G., and Eykholt, G. 2010. Hydraulic diffraction data. Clays and Clay Minerals 64(4), 488-502.
conductivity of geosynthetic clay liners to tailings impoundment Take, W.A., Brachman, R., Rowe, R.K., and Rentz, A. 2014. Temperature
solutions. Geotextiles and Geomembranes 28(2), 149-162. measurements of exposed GMB/GCL composite liners. 10th
Shamrock, J., and Msiza, J. 2015. HDPE geomembrane specification and International Conference on Geosynthetics (10ICG), 21–25
performance implications. Western Cape LIG Seminar, Tulbagh, 15– September, Berlin, Germany.
16 September 2015. Take, W.A., Brachman, R., and Rowe, R.K. 2015. Observations of
Sherard, J., and Dunnigan, L. 1989. Critical filters for impervious soils. bentonite erosion from solar-driven moisture migration in GCLs
Journal of Geotechnical Engineering 115(7), 927-947. covered only by a black geomembrane. Geosynthetics International
Shercliff, D. 1998. Designing with the cylinder test. Proc. of the Polluted 22(1),78–92.
and Marginal Land Conference, Brunel University, London. Tallard, G. 1984a. Slurry Trenches and Hazardous Waste. ASCE Civil
Shukla, S. 2012. Handbook of geosynthetic engineering, 2nd edition, Engineering Magazine, February.
ICE Publishing, London, UK, ISBN 978-0-7277-4175-2. Tallard, G. 1984b. Acuavoir/Enhanced Aquifers, The Military Engineer
Magazine Vol.76#492, March.
291
Tallard, G. 1992. New Trenching Method using synthetic polymers. Trauger, R., and Darlington, J. 2000. Next-generation geosynthetic clay
ASTM publ. STP 1129. liners for improved durability and performance, TR-220, Colloidal
Tallard G. 1992. Hazwaste Barriers, update for the 90’s, Canadian Environmental Technologies Company, Arlington Heights, Illinois,
Geotechnical Society 45th conference, October. USA.
Tallard, G. 1995. From Cement Bentonite to IMPERMIX. D.O.E. US EPA 1989. Risk assessment guidance for superfund (RAGS): part A.
barriers workshop, Denver CO, August. Office of Emergency and Remedial Response, U.S. Environmental
Tallard, G. 1997a. Very Low Conductivity self-hardening slurry for Protection Agency, Washington, D.C., USA.
permanent enclosures. Proc. Intl Containment Technology US EPA 2010. Bisphenol A Action Plan, U.S. Environmental Protection
Conference, St Petersburg, Florida. 9-12 February. Agency
Tallard, G. 1997b. Beyond cement-bentonite: Impermix self-hardening US EPA 2015. Update of Human Health Ambient Water Quality Criteria:
slurry. USCOLD annual conf. 2,4-Dimethylphenol 105-67-9, Office of Science and Technology,
Tang, C., Shi, B., Liu, C., Suo, W., and Gao, L., 2011. Experimental Office of Water, U.S. Environmental Protection Agency Washington,
characterization of shrinkage and desiccation cracking in thin clay DC 20460
layer. Applied Clay Science 52(2), 69-77. US HHS 1999. Toxicological profile for chlorophenols, U.S. Department
Tang, C.-S., Paleologos, E.K., Vitone, C., Du, Y.J., Li, J.S., Jiang, N.J., of Health and Human Services, Public Health Service, Agency for
Deng, Y.F., Chu, J., Shen, Z., Koda, E., Dominijanni, A., Fei, X., Toxic Substances and Disease Registry
Vaverková, M.D., Osi ski, P., Chen, X., Asadi, A., Takeuchi, M.R.H., US HHS 2001. Toxicological profile for pentachlorophenol, U.S.
Bo, M.W., Abuel-Naga, H., Leong, E.C., Farid, A., Baser, T., Department of Health and Human Services, Public Health Service,
O’Kelly, B.C., Jha, B., Goli, V.S.N.S. and Singh, D.N. 2021. Agency for Toxic Substances and Disease Registry
Environmental geotechnics: challenges and opportunities in the post US HHS 2008a. Toxicological profile for cresols, U.S. Department of
COVID-19 world. Environmental Geotechnics 8(3), 172-192. Health and Human Services, Public Health Service, Agency for
Tang, H.S., Haynes, R.D., and Houzeaux, G. 2021. A Review of Domain Toxic Substances and Disease Registry
Decomposition Methods for Simulation of Fluid Flows: Concepts, US HHS 2008b. Toxicological profile for phenols, U.S. Department of
Algorithms, and Applications. Archives of Computational Methods Health and Human Services, Public Health Service, Agency for
in Engineering 28(3), 841-873. Toxic Substances and Disease RegistryVandenberg, L. N., Hauser,
Tedd, P., Holton, I.R., Butcher, A.P., Wallace, S. and Daly, P. 1997. R., Marcus, M., Olea, N., Welshons, W. V. 2007. Human exposure
Investigation of the performance of cement-bentonite cut-off walls to bisphenol A (BPA. Reproductive Toxicology, 24(2), pp. 139–177.
in aggressive ground at a disused gasworks site. Proc. Intl Van Loon, L., Glaus, M., and Müller, W. 2007. Anion exclusion effects
Containment Technology Conf., St Petersburg, Florida. in compacted bentonites: towards a better understanding of anion
Theng, B. 2012. Formation and properties of clay-polymer complexes. diffusion. Applied Geochemistry 22(11), 2536-2552.
Developments in Clay Science 4, 526. Vaverková, M.D., Paleologos, E.K., Dominijanni, A., Koda, E., Tang,
Thiel, R., and Criley, K. 2005. Hydraulic conductivity of partially
prehydrated GCLs under high effective confining stress for three C.S., Manassero, M., O’Kelly, B.C., Xie, Q., Bo, M.W., Adamcová,
real leachates. Waste Containment and Remediation, GSP 142, D., Podlasek, A., Anand, U.M., Mohammad, A., Goli, V.S.N.S.,
GeoFrontiers 2005, ASCE, Reston, Virginia, USA. Kuntikana, G., Palmeira, E.M., Pathak, S. and Singh, D.N. 2021.
Thiel, R., and Richardson, G. 2005. Concern for GCL shrinkage when Municipal solid waste management under COVID-19: challenges
installed on slopes. GRI-18 at GeoFrontiers, Folsom, Pennsylvania, and recommendations. Environmental Geotechnics 8(3), 217-232.
USA. Vryzas, Z., Kelessidis, V., Nalbantian, L., Zaspalis, V., Gerogiorgis, D.,
Thiel, R., and Smith, M., 2004. State of the practice review of heap leach and Wubulikasimu, Y. 2017. Effect of temperature on the rheological
design issues. Geotextiles and Geomembranes 22, 555-568. properties of neat aqueous Wyoming sodium bentonite dispersions.
Thiel, R., Giroud, J., Erickson, R., Criley, K., and Bryk, J. 2006. Applied Clay Science 136, 26-36.
Laboratory measurements of GCL shrinkage under cyclic changes Wang Y, Chen Y, Xie H, Zhang C and Zhan LT. 2016. Lead adsorption
in temperature and hydration conditions. Proc. Eighth International and transport in loess-amended soil-bentonite cut-off wall.
Conference on Geosynthetics, Millpress, Rotterdam, 157-162. Engineering geology, 215: 69-80.
Tian, K., and Benson, C. 2019. Containing bauxite liquor using Wang, Y., Ren, H., Yan, Y., He, S., Wu, S., and Zhao, Q. 2020.
bentonite-polymer composite geosynthetic clay liners. 8th Hindered phenolic antioxidants as heat-oxygen stabilizers for
International Congress on Environmental Geotechnics, L. Zhan, Y. HDPE. Polymers and Polymer Composites 29(9), 1403-1411.
Chen, and A. Bouazza, Eds., Springer Nature Singapore Pte Ltd., Vol. Wichterle, O., and Lím, D. 1960. Hydrophilic gels for biological use.
2, 672-678. Nature 185, 117-118.
Tian, K., Benson, C., and Likos, W. 2016a. Hydraulic conductivity of World Resources Institute 2015. Extreme water shortages are expected
geosynthetic clay liners to low-level radioactive waste leachate. to hit these countries by 2040, article by Jeff Desjardins, August 31,
Journal of Geotechnical and Geoenvironmental Engineering 142(8), 2015, www.visualcapitalist.com.
04016037. Xanthakos, P., 1979. Slurry Walls, McGraw-Hill Book Co.
Tian, K., Benson, C., and Likos, W. 2017. Effect of anion ratio on the Xu, X., Liu, X., Oh, M., and Park, J. 2018. Swelling capacity and
hydraulic conductivity of a bentonite-polymer geosynthetic clay hydraulic conductivity of polymer-modified bentonite under saline
liner. Proc., GeoFrontiers, ASCE, Reston, Virginia, USA,180–189. water conditions. Applied Sciences 8, 1025.
Tian, K., Likos, W., and Benson, C. 2016b. Pore-scale Imaging of Xu, H., Qiu H., Zhu G., Zhan L., Zhang Z., Xu X., Chen Y., and Wang Y.
polymer-modified bentonite in saline solutions. Proc., Geo-Chicago 2019. Comparison of settlement behaviours of high-food-waste-
2016, GSP No. 271, ASCE, Reston, Virginia, USA, 468-477. content (HFWC) and low-food-waste-content (LFWC) MSWs and
Tian, K., Likos, W., and Benson, C. 2019. Polymer elution and hydraulic assessment of their prediction models. SCIENCE CHINA
conductivity of bentonite-polymer composite geosynthetic clay Technological Sciences 62(12), 2271-2292.
liners. Journal of Geotechnical and Geoenvironmental Engineering Xu, H., Wang J., Zhan L., Zhang Z., Xu X., Chen Y., and Yao K. 2020.
145(10), 04019071. Characterization of compression behaviours of high food waste
Tincopa, M., Bouazza, A., Rowe, R.K., and Rahardjo, H. 2020. Back- content (HFWC) MSW and no food waste content (NFWC) MSW
analysis of the water retention curve of a GCL on the wetting path. in China. Waste Management 103, 305-313.
Geosynthetics International 27(5), 523-537. Yamamoto, T. and Yasuhara, A. 1999. Quantities of Bisphenol A leached
Tognon, A., Rowe, R.K., and Moore, I. 2000. Geomembrane strain from plastic waste samples, Chemosphere, 38(11), 2569-2576.
observed in large scale testing of protection layers. Journal of Yeo, S. S., Shackelford, C. D., and Evans, J. C. 2005). Consolidation and
Geotechnical and Geoenvironmental Engineering 126(12), 1194- hydraulic conductivity of nine model soil-bentonite backfills.
1208. Journal of Geotechnical and Geoenvironmental Engineering, ASCE,
Tong X, Li Y-C, Ke H, Li Y, Pan Q. 2020. In situ stress states and lateral 131, No. 10, 1189-1198.
deformations of soil-bentonite cut-off walls during consolidation Yong, R.N., and Mulligan, C.N. 2019. Natural and Enhanced Attenuation
process. Canadian Geotechnical Journal, 57(1):139-148. of Contaminants in Soils. Second Edition. CRC Press, USA.
Touze-Foltz, N., and Farcas, F. 2017. Long-term performance and binder Yu, B., and El-Zein, A. 2019. Experimental investigation of the effect of
chemical structure evolution of elastomeric bituminous airgaps in preventing desiccation of bentonite in geosynthetic clay
geomembranes. Geotextiles and Geomembranes 45(2), 121-130. liners exposed to high temperatures. Geotextiles and Geomembranes
47, 142-153.
292
Yu, B., El-Zein, A., and Rowe, R.K. 2020. Effect of added polymer on Zhang, W., Sun, C., and Qiu, Q. 2016. Characterizing of a capillary
the desiccation and healing of a geosynthetic clay liner subject to barrier evapotranspirative cover under high precipitation conditions.
thermal gradients. Geotextiles and Geomembranes 48(6), 928-939. Environmental Earth Sciences 75(6), 513.
Yu, C., Liao, R., Cai, X., and Yu, X. 2019. Sodium polyacrylate Zhan, L., Xu, X., Chen, Y., Ma, X., and Lan, J. 2015. Dependence of gas
modification method to improve the permeant performance of collection efficiency on leachate level at wet municipal solid waste
bentonite in chemical resistance. Journal of Cleaner Production 213, landfills and its improvement methods in China. Journal of
242-250. Geotechnical and Geoenvironmental Engineering 141(4), 04015002.
Zainab, B., and Tian, K. 2020. Hydraulic conductivity of bentonite- Zhan, L., Li, G., Jiao, W., Wu, T., Lan, J., and Chen, Y. 2017c. Field
polymer geosynthetic clay liners to coal combustion product measurements of water storage capacity in a loess–gravel capillary
leachates. Geo-Congress 2020, GSP 316, ASCE, Reston, Virginia, barrier cover using rainfall simulation tests. Canadian Geotechnical
USA. Journal 54(11), pp.1523-153
Zhan, T., Xu, X., Chen, Y., Ma, X., and Lan, J. 2015. Dependence of gas Zhan, L., Wu, T., Feng, S., Li, G., He, H., Lan, J., and Chen, Y. 2020a.
collection efficiency on leachate level at wet municipal solid waste Full-scale experimental study of methane emission in a loess-gravel
landfills and its improvement methods in China. Journal of capillary barrier cover under different seasons. Waste Management
Geotechnical and Geoenvironmental Engineering 141(4), 04015002. 107, 54-65.
Zhan, L., Qiu, Q., Xu, W., and Chen, Y. 2016. Field measurement of gas Zhan, L., Li, G., Jiao, W., Lan, J., Chen, Y., and Shi, W. 2020b.
permeability of compacted loess used as an earthen final cover for a Performance of a compacted loess/gravel cover as a capillary barrier
municipal solid waste and landfill gas emissions controller in Northwest China. Science of
Zhan L., Xu, H., Chen Y., et al. 2017a. Biochemical, hydrological and The Total Environment 718, 137195.
mechanical behaviours of high food waste content MSW landfill: Zheng, Q-T, Rowe, R.K. and Feng, S-J. 2018. Design of Vertical Landfill
Preliminary findings from a large-scale experiment. Waste Gas Collection Wells Considering Non-homogeneity with depth.
Management 63, 27-40. Waste Management, 82: 26-36.
Zhan L., Xu, H., Chen Y., et al. 2017b. Biochemical, hydrological and /doi.org/10.1016/j.wasman.2018.10.012
mechanical behaviours of high food waste content MSW landfill: Zheng, Q-T, Rowe, R.K. and Feng, S-J. 2019a. Recovery Response of
Liquid-gas interactions observed from a large-scale experiment. Vertical Gas Wells in Non-homogeneous Landfills. Waste
Waste Management 68, 307-318. Management, 83: 33-45. doi.org/10.1016/j.wasman.2018.10.043
Zhan, L., Xu, H., Jiang, X., Lan, J., Chen, Y., and Zhang, Z. 2019. Use Zheng, Q-T, Rowe, R.K. and Feng, S-J. 2019b. Design of Horizontal
of electrical resistivity tomography for detecting the distribution of Landfill Gas Collection Wells in Non-homogeneous Landfills.
leachate and gas in a large-scale MSW landfill cell. Environmental Waste Management, 98: 102-112.
Science and Pollution Research 26(20), 20325-20343. https://doi.org/10.1016/j.wasman.2019.08.017
Zhan LT, You YQ, Zhao R Chen C and Chen YM. 2020. Retarding the Zheng, Q-T, Rowe, R.K. and Feng, S-J. 2020. Control and Estimation of
lead(II) migration in soil-bentonite cut-off wall by loess amendment: Maximum Gas Pressure below Landfill Cover with Horizontal Gas
centrifuge and numerical modeling. Submitted to International wells. Waste Management,114: 33–42,
Journal of Physical Modeling in Geotechnics. https://doi.org/10.1016/j.wasman.2020.06.033
Zhang, W., and Sun, C. 2014. Parametric analyses of evapotranspiration
landfill covers in humid regions. Journal of Rock Mechanics and
Geotechnical Engineering 6(4), 356-365.
293
State of the Art 5: Unsaturated Soils
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
State of the Art Report. Advances in Unsaturated Soil Mechanics: Constitutive
modelling, experimental investigation, and field instrumentation
Nasser Khalili
School of Civil and Environmental Engineering, UNSW Sydney, Australia, [email protected]
Enrique Romero
Department of Civil and Environmental Engineering, UPC, and CIMNE, Spain, [email protected]
Fernando A. M. Marinho
Department of Sedimentary and Environmental Geology - IGc/USP, Brazil, [email protected]
ABSTRACT: This state-of-the-art covers advances in the constitutive modelling, experimental investigation and field measurement
of suction for unsaturated soils. It consists of three distinct parts. Part 1 is devoted to the constitutive modelling and covers major
developments in the field including elasto-plasticity of soil skeleton, encompassing the critical role of unsaturation, and treatment of
cyclic loading, expansive soils, aggregation/cementation/structure, rate dependency of response, and anisotropy. An attempt has also
been made to address the choice of stress variables for constitutive modelling of unsaturated soil including the role of the effective
stress concept principle. Part 2 is devoted to the experimental advances in unsaturated soils, focusing on the hydro-mechanical
phenomena as well as the microstructural effects and the experimental approaches for their examination. Part 3 is dedicated to the
measurement of suction in the field, various probes that can be used for this purpose, their advantages and disadvantages, and the
correct approach for their positioning and installation.
RÉSUMÉ : Cet état de l’art couvre les avancées en matière de modélisation constitutive, d'investigation expérimentale et de mesure sur
le terrain de la succion des sols non saturés. Il se compose de trois parties distinctes. La première partie est consacrée à la modélisation
constitutive et couvre les principaux développements dans le domaine, y compris l'élasto-plasticité du squelette du sol, englobant le rôle
critique de l'insaturation, et le traitement de la charge cyclique, les sols expansifs, l'agrégation/la cémentation/la structure, la dépendance
de la réponse à la vitesse, et l'anisotropie. Une tentative a également été faite pour aborder le choix des variables de contrainte pour la
modélisation constitutive du sol non saturé, y compris le rôle du principe du concept de contrainte effective. La partie 2 est dédiée aux
avancées expérimentales dans les sols non saturés, en se concentrant sur les phénomènes hydromécaniques ainsi que sur les effets
microstructuraux et les approches pour les examiner. La troisième partie est consacrée à la mesure de la succion sur le terrain, aux
différentes sondes qui peuvent être utilisées à cette fin, à leurs avantages et inconvénients, et à l'approche correcte de leur positionnement
et de leur installation.
KEYWORDS: Unsaturated soils, constitutive modelling, experimental investigation, field instrumentation
1 INTRODUCTION as the most appropriate way to tackle the intricate aspects of air
and water interaction within the soil’s pore space and their
Since the pioneering work of Bishop and his co-workers in the cross-coupling with the non-linear deformation of the soil matrix.
1950’s, tremendous advances have been made in the constitutive One the key aims of this state-of-the-art is to address these as
modelling, experimental investigation and field monitoring of well as other points of disagreement and pave the way for the
unsaturated soils. The early work on the mechanics of future developments in the field, as well as highlighting the
unsaturated soils followed the classical approach in soil tremendous advances that have accomplished to date. Topics
mechanics where the soil resistance to shearing and the soil covered will include: stress state variables for multi-phase media,
deformation due to loading were treated separately as two derivation of the effective stress for fluid(s) saturated porous
independent phenomena. The first fully integrated strength- media, hydro-mechanical coupling of water retention response,
deformation model for the behaviour of unsaturated soils, within elasto-plasticity in unsaturated soils, and the treatment of cyclic
a general framework of elasto-plasticity, was introduced by loading, expansive soils, aggregated and structured soils, rate
Alonso and his co-workers in the late 1980’s and early 1990’s. dependency and anisotropy.
This was followed by a rapid succession of seminal contributions In parallel with the constitutive modelling of unsaturated soils,
which have since underpinned perhaps all the advanced significant advances have also been accomplished in the
developments in the constitutive modelling of unsaturated soils. experimental and field investigations of unsaturated soils. These
Nevertheless, the theoretical developments in the mechanics of have clarified the fundamentals of unsaturated soil mechanics
unsaturated has not been without a controversy. There have been and have underpinned the advances in the theoretical
robust discussions and disagreements on the choice of stress developments. Significant progress has been made in the use of
variables and their conjugates for the purposes of constitutive multi-physics and multi-scale techniques and their application to
modelling, the role of the stress state variables as distinct from unsaturated soils. New devices and experimental procedures
stresses that are used for the constitutive modelling, the role of have been developed allowing phase separation, investigation of
suction in unsaturated soil mechanics, the applicability of the coupled processes, and bridging scales for the purpose of
effective stress to unsaturated soils and its true meaning, as well constitutive modelling. These have permitted unparalleled access
297
to small scale microstructural soil features at the pore/grain level The most commonly adopted pair is , the net
with a profound impact on understanding the larger-scale stress tensor, and , the matric suction. Although
phenomenological processes. The advances in the experimental constitutive models using the combination of and
techniques have also permitted investigation of new unsaturated have also been presented (Geiser 1999).
geomaterials (such as rockfill materials, artificially prepared
mixtures, bentonite-based materials) and new multi-physics 2.3 Effective Stress
processes in support of emerging fields in geotechnical
engineering. Of particular interest has been the development of The effective stress principle plays a critical role in the
technologies to improve resiliency of geo-infrastructure, adapt to constitutive modelling of multi-phase media. Often referred to as
climate change, and mitigate the associated risk. the axiom of soil mechanics (Atkinson 2007), it enters the elastic
The section on the experimental and field investigations of as well as elasto-plastic constitutive equations of the solid phase,
this state-of-the-art will describe laboratory techniques and the linking a change in stress to straining or any other relevant
advances made over the last three decades. Particular focus will quantity of the soil skeleton, such as compression, distortion and
be on studying the coupled hydro-mechanical and multi-scale a change of shearing resistance (e.g., see Terzaghi 1936, Biot
phenomena, and the experimental techniques that are advanced 1941, Rice and Cleary 1976, de Boer and Ehlers 1990, Coussy
for the cross-disciplinary applications of unsaturated soil 1995). Without the advent of the effective stress principle, many
mechanics. In particular, the experimental techniques for of the significant achievements of solid mechanics - such as
controlling and measuring suction, hydro-mechanical cells, elasto-plasticity, visco-plasticity, wave propagation, and
microstructural studies, 1-g and centrifuge scaled tests will be numerical methods - could not be extended to soil engineering
discussed. In addition, advances in ground engineering, problems without empiricism.
environmental and energy geotechnics, bioinspired technologies, Based on experimental observations, and for saturated soils,
tailings and industrial processes will be covered. The basic Terzaghi (1936) defined the effective stress, , as “the total
concepts of pore pressure measurement in both the laboratory stress in excess of the neutral stress”
and the field will be presented. A brief historical account of the
development of suction measurements in the field is presented. (1)
The specifics of the most common suction probes are discussed
and their limitations and advantages are highlighted. The correct in which is the effective stress tensor, and is the neutral
positioning of these probes in the field for a reliable measurement stress or the pore water pressure. Terzaghi stated that “all
of suction in the field is also covered. measurable effects of a change of stress, such as compression,
distortion and a change of shearing resistance are exclusively
2 BASIC CONCEPTS due to change in the effective stress”. The first use of this
principle is implicit in Terzaghi’s work on the theory of one-
2.1 Kinematics and Definitions dimensional consolidation, albeit without referring to the term
effective stress (Terzaghi 1923). Bishop and Blight (1963)
Unsaturated soils are three phase porous media consisting of defined the effective stress as that function of the stress state
solid particles and a pore space filled with water and air. One variables that controls the mechanical effects of a change in
fluid, water, is wetting, and the other air, is not. Phases (solid, s, stress, such as volume change and a change in shear strength.
water, w, and air, a) represent the constituents when viewed as a Lade and De Boer (1997) defined the effective stress as “the
part of the mixture. The solid constituent is assumed to be stress that controls the stress-strain, volume change, and
slightly compressible, or incompressible if simplifications are strength behaviour of a given porous medium ….. The pore
desired. Each constituent has a mass and a volume , pressure may be zero or negative, or it may be positive and very
, which make up the total mass and large, but the effective stress must be expressed such that it
the total volume . Intrinsic quantities are produces the same material response for any pore pressure. It is
defined using subscripts and apparent quantities using to be expected that some properties of the given porous medium
superscripts. For example intrinsic mass density of phase is will be part of a more comprehensive effective stress expression,
denoted , whereas the apparent mass density is unlike the simple expression in equation (1), which does not
written as ; hence , where involve any material properties.”.
is the apparent volume fraction of constituent . The apparent In essence, the effective stress principle converts a multi-
volume fractions satisfy the constraint . The phase, multi-stress state porous medium to a mechanically
sign convention of soil mechanics is adopted throughout. equivalent, single-phase, single-stress state continuum, hence
Tensor quantities are identified by boldface letters. permitting the application of the principles of continuum solid
mechanics to fluid-filled deformable porous media (Blight 1967,
2.2 Stress State Variables Khalili et al. 2004, Nuth and Laloui 2008). Specifically, the
effective stress enables capturing “the stress-strain, volume
The smallest number of stress variables controlling the state of a change, and strength behaviour of a porous medium,
system are called the stress state variables of that system. For a independent of the magnitude of the pore pressure”, and solely
multi-phase system of constituents, the number of stress state based on the underlying drained (effective) properties of the soil
variables are equal to , where is the number constraints skeleton (Lade and De Boer 1997).
within the system. For example, the stress state for saturated soils Within this context, the effective stress may be considered as
(a two-constituent material), when no constraints are imposed, is the constitutive stress of the solid skeleton (e.g. Skempton 1960,
captured via two stress state entities; i.e. , the total stress tensor Broja and Koliji 2009, Alonso et al. 2010), with “its seat
and, , the pore water pressure. Imposing the solid constituent exclusively in the solid skeleton of the soil” (Terzaghi 1936).
incompressibility constraint, as a special case, the stress state This distinction is important as it clarifies that effective stress
variables reduce by one, taking the form , in which only pertains to the solid skeleton of the porous medium, and thus
is the identity tensor. only controls the stress state of the solid skeleton rather than the
For unsaturated soils, the stress state variables are , and overall state or equilibrium of the multi-phase system (e.g. see
, where is the pore air pressure. Again, introducing the Coussy 1995, Lade and De Boer 1997, Cheng 2016 and more
incompressibility of the solid grains, the stress state variables recently Guerriero and Mazzoli 2021).
reduce to being one of any pair of the stresses: , Another important, yet less recognised, role of the effective
and (Fredlund and Morgenstern 1977). stress principle is that it furnishes a platform for coupling the
298
deformation of the solid skeleton to the volume change of the straining of the soil skeleton to a change in stress. Similar,
fluid constituents. This is critical for establishing consistent expressions were also made by Hoffman (1928), Lubinski
hydro-mechanical models for multi-phase porous media (Loret (1954), Biot (1955), Schiffman (1970). Terzaghi (1923, 1945)
and Khalili 2000, Khalili et al 2008, Vaunat and Casini 2017). also initially proposed for saturated soils and rocks, but
In its most general form, the effective stress expression for a found experimentally that .
multi-phase porous medium, consisting of fluids, may be Skempton and Bishop (1954) speculated that the effective
expressed as (Khalili 2008) stress is equal to the average intergranular stress acting between
particles and/or clasts. Accordingly, they derived the expression
(2)
(7)
in which is the effective stress tensor, is the total stress
tensor, and fluid pressure. Specialisation of (2) for saturated in which is the contact area between particles per unit area
soils results in of the porous medium. It was also argued that is negligible
for soils, with (7) reducing to (1) for most geotechnical
(3) engineering applications.
Equation (7) was based on satisfying equations of equilibrium
in which is the effective stress parameter for saturated porous for the porous medium. This is frequently presented in soil
media. mechanics textbooks as the proof of Tezaghi’s expression for the
For unsaturated soils, (2) becomes effective stress of saturated soils. However, as noted by Alonso
et al (1987) the form of the equations of equilibrium “is not
(4) necessarily a proof of an effective stress statement for
constitutive modelling” of porous media. The fundamentals of
in which and are the effective stress parameters of the such an approach were also questioned by Bloch (1978). Later,
pore-water pressure and pore-air pressure, respectively. Within based on experimental data using oedometer tests on lead shot,
the practical range of stresses, and for soils with incompressible Skempton (1960) showed that in fact equation (7) was invalid
solids, it can be shown that that the effective stress parameters and that the contact area between particles played no role in the
and satisfy the constraints and that (4) formulation of the effective stress.
reduces to (Bishop 1959) The effective stress expression that was supported by many
investigators (e.g. Biot 1941, Skepmton 1960, Nur and Byerlee,
(5a) 1971, Lade and De Boer 1997, Coussy 2010) involved the ratio
of the bulk modulus of the solid grains, , and the drained bulk
or modulus of the solid skeleton, , as
(5b) (8)
in which is the net stress, is the It was noted that for most soils was significantly larger than
matric or matrix suction and is the effective stress , leading to approaching unity and recovering Terzaghi’s
parameter quantifying the contribution of suction to the effective effective stress expression. However, for rocks and concrete
stress of the solid skeleton. typically attained values less than 1 (e.g. see Brace 1965),
rendering application of Terzaghi’s effective to such media
2.3.1 Effective stress parameter for saturated soils invalid. Proofs of (8) have been presented by Biot (1941) through
The early developments on the role of the effective stress theory of poro-elasticity and invoking existence of elastic
principle in the constitutive modelling of saturated porous media potential, and Nur and Byerlee (1971) through establishing a
were marred with remarkable, yet largely forgotten, strain equivalency between the saturated porous medium and an
controversy, arguments and personal tragedy. Of note were the equivalent continuum with the same underlying mechanical
heated exchanges between Karl von Terzaghi, widely known as properties. Similar proofs using the theory of poro-elasticity have
the father of modern soil mechanics, and Paul Fillunger, the also been present by Geertsma (1957, 1966), Lade and De Boer
father of mixture theory. Paul Fillunger suicided around the time (1997), Coussy (2010) amongst many others. The main premise
of these arguments. A comprehensive account of the exchanges, in all these developments has been the fundamental requirement
and events leading to this tragic outcome, is detailed in the that the effective stress of the solid skeleton is related uniquely,
comprehensive work of de Boer (2000, 2005). and in a one-to-one manner, to the elastic strain of the solid
Numerous investigations have been reported in the literature skeleton through
on the determination of the form of the effective stress for
saturated porous media. Several candidates have been proposed (9)
for the effective stress parameter, .
Fillunger (1930, 1936), on theoretical grounds, and satisfying in which is the underlying drained elastic property tensor of
the equilibrium equations of the entire soil mixture; i.e. using the the solid skeleton, and is the elastic strain tensor. Within this
volume averaged stresses of the solid and water, expressed the context, any stress satisfying (9) is the effective stress of the solid
effective stress as skeleton. Once the effective stress expression is established in the
elastic region, it can then be applied to the elasto-plastic analysis
(6) by invoking an appropriate elastic-plastic model (de Boer and
Ehlers 1990, Loret and Khalili 2000, 2002, Coussy 2010). As
in which is the porosity. Fillunger essentially defined the elaborated by de Boer and Ehlers (1990), the treatment of elastic-
effective stress as the partial/apparent stress of the solid phase, plastic or viscous skeletons would not change the basic statement
expressed as , in which is the intrinsic about the effective stress.
stress of the solids (Bowen 1976). While this expression was Suklje (1969) also used the theory of poro-elasticity but
appropriate for the equilibrium and the balance of momentum of derived the relationship
the solid-water mixture, it was invalid as a constitutive stress for
the mechanical response of the solid skeleton and relating
299
thermodynamic laws to simple volume averaging procedures, or
expressions that are based entirely on intuition.
Earlier expressions of the effective stress parameter for
(10)
unsaturated soils assumed a direct equivalency with the degree
of saturation (Bishop 1959, Bishop and Donald 1961, Bishop and
which is different from the relationship presented in (8) by the Blight 1963). They provided a geometrical interpretation of the
term . In this derivation, Suklje assumed that the effective stress parameter, however no unique relationship could
macroscopic volume change of the solid skeleton due to an all be found between and the degree of saturation. Examples of
around pore pressure is equal to the strain of the solid fraction of the use of degree of saturation as the effective stress parameter,
the soil skeleton. This is clearly not correct as it ignored the , in constitutive modelling of unsaturated soils include
structure of the solid skeleton and change in the porosity as the (Schrefler 1984, Öberg and Sällfors 1995, Bolzon et al. 1996,
solid grains strained. Jommi 2000, Gallipoli et al. 2002, Gallipoli et al., 2003, Wheeler
The following points, from the literature, are key to et al. 2003, Laloui et al 2003, Sheng et al. 2004, Tamagnini 2004,
Sun et al. 2007a,b,c, Nuth and Laloui 2008, Romero and Jommi
understanding the role of the effective stress:
2008). While convenient from a constitutive modelling view
The effective stress is a constitutive stress with the sole role
point, there is overwhelming experimental evidence, gathered
of capturing the mechanical response of the solid skeleton
since 1960’s, against the appropriateness and the use of the
due to a change in stress.
degree of saturation as the effective stress parameter (e.g. see
The effective stress is distinct from stress state variables of
Zerhouni, 1991, Vanapalli and Fredlund 2000, Khalili and
the soil, the partial stress of the solids and the intergranular
Zargarbashi 2010, Alonso et al. 2010, Pereira et al. 2010, Konrad
stress.
and Lebeau 2015, as typical examples). In general,
The effective stress pertains only to the solid skeleton of the
tends to over-estimate the effective stress and the shear strength
soil, and as such it may only be considered as the stress state
of unsaturated clays (Pereira et al. 2010), whereas the reverse
variable of the solid phase. It is not, and it cannot be expected
happens for sands (Konrad and Lebeau 2015). Furthermore,
to be, a stress state variable of the entire soil-water mixture.
implies that, at large suctions, where the degree of saturation
Only for the special case of saturated soils, with the
approaches a limiting residual value, the effective stress and thus
constraint of incompressible solid grains, does the effective
the shear strength are directly proportional to suction. This is not
stress fulfil the role of both the constitutive stress and the
supported by the experimental evidence (Loret and Khalili 2002,
stress state parameter.
Alonso et al. 2010). In addition, on theoretical grounds,
The effective stress is exclusively associated with, and only
incorporating physics at the microscale (Alonso et al. 2010, Lu
with, the elastic component of straining of the solid skeleton
et al. 2010), and based on the thermodynamics of multiphase
(equation 9). This is irrespective of whether the stress-strain
systems (Fuentes and Triantafyllidis 2013, Jian et al. 2017, Einav
response of the soil skeleton is in the elastic, elasto-plastic or
and Liu 2018), it is shown that is only recovered when
elasto-visco-plastic region.
the work of the air-water interface is neglected (Hassanizadeh
The form of the effective stress may be obtained, as is the
and Gray 1990, Houlsby 1997, Hutter et al. 1999, Gray and
convention, by establishing a mechanical equivalency
Schrefler 2001, Borja 2006, Gray and Schrefler 2007, Coussy
between a mutli-phase system and a continuum, with the
2010, Zhao et al. 2010, Nikooee et al. 2013 (to name a few)).
continuum having the same underlying drained mechanical
Consensus is gradually emerging with the view that it is better
properties of the multi-phase system.
to make a function of the effective degree of saturation,
The mechanical equivalency may be established based on a
strain equivalency (e.g. Nur and Byerlee 1971), strength
equivalency (Vanapalli and Fredlund 1996, Khalili and
Khabbaz 1998) or an energy equivalency analysis (de Boer (11)
and Ehlers 1990, Gray and Schrefler 2001, Borja (2006,
Einav and Liu 2018). Irrespective of the way the effective where is the residual degree of saturation.
stress is established, it must be applicable to all mechanical Karube and Kato (1994), Karube et al. (1998), and more
responses of the soil; e.g. deformation and strength. recently Kim et al. (2010), associated the effective stress
An observation in the rock mechanics literature, which is parameter to the contributions of the “bulk water” and “meniscus
contrary to the last point listed above, is that the deformation is water”. Bulk water was defined as the pore water that occupied
controlled by Skempton’s effective stress (8), whereas the shear the pore volume between soil particles, and the meniscus water
strength is correlated with Terzaghi’s effective stress (1). This is was taken to exist at the contact points of the soil particles only.
referred to as a paradox of rock mechanics. However, as Two effective stress parameters were defined: one corresponding
elaborated on by Lade and de Boer (1997), both phenomena are to the bulk water, , and the other to the meniscus water,
controlled by the same form of the effective stress as presented
in (8). They note that “during the process of shearing concrete , (12)
and/or rock, micro fissures develop and open up, and at the time
of peak failure, sufficient deterioration of the solid material has where is the driest degree of saturation obtained from the
occurred such that the bulk modulus of skeleton forming the main wetting path of the soil water retention curve (SWRC) at
shear zone has decreased dramatically. Therefore, the the suction of interest. The sum of and was taken the
expression in equation (8), which is most often employed for the effective stress parameter which reduced to the effective degree
stress-strain calculations, approaches unity near failure, even at of saturation, , (Kim et al. 2010)
high confining pressures. This means that the expression in
equation (8) captures the effective stress for both stress-strain
(13)
and strength behaviour.”
2.3.2 Effective stress parameter for unsaturated soils Similarly, Konrad and Lebeau (2015) divided the degree of
Similar to saturated soils, numerous contributions have been saturation into two components: absorbed, , and capillary
made to the determination of effective stress parameter, , for , and related exclusively to as
unsaturated soils. They range from rigorous analyses based on
300
Lu and Likos (2004, 2006) and Lu et al. (2010) defined the proportional to suction. Again, as stated previously, this is not
effective stress parameter as supported by experimental evidence.
Some investigators have also advocated the use of a ‘suction
(14) stress’ to represent the contribution of suction to the effective
stress of the solid skeleton (e.g. Karube and Kato 1994, Karube
where is the normalised volumetric water content, is the et al. 1998, Lu and Likos 2004, Lu and Likos 2006, Lu et al.
volumetric water content, is the saturated volumetric water 2010).
content and is the residual volumetric water content. For all Karube and Kato (1994), Karube et al. (1998) and Lu and
practical purposes may be approximated as . Likos (2004) defined the suction stress, , as the multiplication
In parallel with the above developments, and by analysing of the effective stress parameter and the matric suction,
strength data from 14 different cases involving glacial tills, silts, . This definition is unproblematic when determining the shear
sandy clays and clays, Khalili and Khabbaz (1996, 1998) strength of unsaturated soils. For the purpose of constitutive
proposed, within a good accuracy, a unique relationship between modelling, however, it is more fundamental to quantify
and the ratio separately from suction, , since not only captures the
contribution of suction to the effective stress, it also acts as the
coupling term between flow and deformation fields in
unsaturated soils (Khalili et al. 2008). When is adopted in a
(15) fully coupled hydro-mechanical analysis, it is necessary to
recover from using .
However, more recently, Lu and Likos (2006) and Lu et al.
where is the suction value marking the transition between (2010) extended the notion of suction stress to include not only
saturated and unsaturated states. For the main wetting path the capillary effects, i.e. , but also physico-chemical effects
, and for the wetting drying path , in which such as van der Waals forces, electrical double-layer forces and
is the air entry value and is the air expulsion value. Khalili cementation in a lumped fashion and relating all these effects to
and Khabbaz (1998) showed that the best-fit value of the matric suction and volumetric water content of the soil. There are
exponent assumes for different soil types. Extension several difficulties with this approach:
of (15) to was given in Russell and Khalili (2006). Physico-chemical effects are controlled by the chemistry of
The effect of suction reversals was in turn examined by Khalili the pore water and the mineralogy of the soil. There are no
and Zargarbashi (2010). direct correlations between the matric suction and/or the
Describing the water retention with the Brooks and Corey water content with the physico-chemical effects in a soil. To
(1964) model, Mašín (2010, 2013) and Khalili (2018) showed capture physico-chemical effects one must introduce
that (15) may alternatively be presented as chemical potential or osmotic suction as an independent state
variable.
(16) Physico-chemical effects are intra-aggregate phenomena and
affect primarily the mechanical properties of the soil rather
in which is the slope of soil water retention curve presented than the effective stress. Therefore, their inclusion in the
in log-log scale. As pointed out by Khalili (2018) equations (15) effective stress statement is inappropriate and cause major
and (16) are equivalent and may be used inter-changeably for the difficulties in the constitutive modelling of soils. For
determination of the effective stress parameter, depending on the example, it is accepted that cementation increases the shear
availability of the relevant input data. strength of soils. However, this is achieved through an
Xu (2004) and Xu et al (2015) adopted fractal mechanics and increase in the strength parameters of the soil rather than an
showed that, using self-similarity of pores in soils, the fractal increase of the effective stress. If one is to introduce the
representation of the effective stress parameter will be of the cementation in the effective stress equation, increasing
form cementation must cause not only an increase in the shear
strength, but a volume contraction of the soil, which is not
the case.
Physico-chemical phenomena are also present in saturated
soils, and through the use of Terzaghi’s effective stress
in which is the fractal dimension of the pore size distribution, principle, their effects are already reflected in the mechanical
with a value that usually lies between 2.4 and 2.6 for natural soils. properties of the soil. Re-including the physico-chemical
Vanapalli et al. (1996) and Fredlund and Vanapalli (2000), effects in the effective stress expression will lead to double
linked the suction contribution to the shear strength of counting of such effects.
unsaturated soil to the aerial fraction of the water constituent, and The usual approach in the constitutive modelling of multi-
proposed phase media is to carefully and systematically identify
physical and chemical processes present in the system and
(17) capture their behaviour through introducing phenomenon
significant variables and parameters. Lumping independent
with being a fitting parameter attaining a value of 1 for non- processes into a single constitutive expression can lead to
plastic granular materials and 3 for highly plastic clays. An intractability of cause and effect within the system.
expression identical to (17) was also presented by Alonso et al.
(2010) based on micro mechanical considerations of soils with 2.4 Soil Water Retention Curve
two dominant pore sizes. In both approaches, , reduced to
for granular materials. This resulted in for non- Another important aspect in the mechanics of unsaturated soils
plastic silts and sands which may not be applicable, as discussed is the soil water retention curve (SWRC), which provides a
earlier. For granular materials is typically less than 0.5. In fundamental relationship between the amount of water held in
addition, both approaches maintained one of the key drawbacks the soil pore space and matric suction. It is the basis for many
of correlating directly with ; that is at large suctions, the empirical relationships for unsaturated soils such as shear
degree of saturation, and hence , in this case, approached a strength, volume change and hydraulic properties (e.g. Biarez et
limiting residual value, making the shear strength linearly al. 1987, Fredlund et al. 1978; Vanappali et al. 1996, Khalili and
301
Khabbaz 1998, Gallipoli et al. 2003; Zhou et al. 2012a,b), and 2007a,b,c, Nuth and Laloui 2008, and Morvan et al. 2011). Some
appears explicitly in the constitutive formulations of unsaturated studies indicate an increase in the value of the pore size
soils, linking pore-phase volume changes to a change in suction distribution index with decreasing void ratio, leading to a steeper
(Dangla et al. 1997, Loret and Khalili 2000, Jommi 2000, SWRC in the unsaturated region for an increase of compaction
Wheeler et al. 2003, Sheng et al. 2004, Khalili et al. 2008, Coussy (e.g. Huang et al. 1998, Karube and Kawai 2001, and Zhou et al.
2010, Khoshghalb and Khalili 2013, Salimzadeh and Khalili 2018). Others suggest an opposite trend with the pore size
2014, Song and Borja, 2014, and Ghaffaripour et al. 2019). distribution index decreasing with decreasing void ratio, leading
Indeed, more research effort has been devoted to the to a flatter SWRC in the unsaturated region (e.g. Gallipoli et al.
quantification and understanding of the water retention 2003, Mašín 2010, Pasha et al. 2017 and 2020). There are also a
behaviour than any other phenomenon relevant to the mechanics number of studies that assume the slope of the SWRC is
of unsaturated porous media. The SWRC is commonly described unaffected by void ratio (e.g. Nuth and Laloui 2008, Gallipoli
as a plot of either gravimetric water content, or volumetric water 2012, Hu 2013, Russell 2014).
content, or degree of saturation, against the logarithm of matric Noting the important role of the void ratio on the water
suction. Due to hydraulic hysteresis, the SWRC typically has two retention characteristics of soils, several SWRC models, with
different branches, one corresponding to wetting and the other to two or more fitting parameters, have been proposed in the
drying. Generally, the water content is higher for drying literature that explicitly take into account the effect of void ratio.
compared to wetting at the same suction value. Examples include the work of Gallipoli et al. (2003), Tarantino
Numerous approaches have been proposed for mathematical (2009), and Sheng and Zhou (2011). Several other investigators
representation of soil water characteristic curve. Summaries have have attempted to capture the void ratio dependency of the
been presented by Fredlund and Xing (1994), Leong and SWRC through the net stress (e.g. see Zhou and Ng 2014).
Rahardjo (1997), Sillers et al (2001) and Lu and Likos (2004). However, such a representation is unlikely to lead to a
The models proposed by Brooks and Corey (1964), van satisfactory outcome, as two samples of the same soil, under the
Genuchten (1980) and Fredlund and Xing (1994), in particular, same net stress, but different loading and unloading histories,
have been popular in the geotechnical engineering community. may possess entirely different void ratios and hence exhibit
The model proposed by Brooks and Corey (1964) is one of the markedly different SWRCs. Also, fitting type models often
earliest and simplest equations for the soil water characteristic require extensive experimental data for the calibration of the
curve. The van Genuchten (1980) model takes into account the model, and tend to perform poorly outside the domain of the
pore size distribution of the soil and provides a continuous soil calibration data.
water characteristic curve. As such, it allows for greater Mašín (2010) was perhaps first to propose a predictive model
flexibility in fitting the degree of saturation over the entire for the evolution of SWRC with void ratio. Adopting equation
suction range. The approach proposed by Fredlund and Xing (15) and invoking the existence of elastic and plastic potential for
(1994) is similar to the van Genuchten (1980) model and is unsaturated soils, he was able to derive an expression that
developed by modifying the pore size distribution function given captured elegantly the evolution with void ratio as
by van Genuchten (1980).
It is generally understood that deformation can markedly (18)
affect the water retention response of soils (Vanapalli et al. 1999,
Ng and Pang 2000, Mašín 2010, Romero et al. 2011, Gallipoli et We recall that is the suction value that separates saturated
al. 2003 and 2012, Salager et al. 2013, Fredlund 2015, Pasha et from unsaturated states of a soil. Dependency of (the slope of
al. 2016 and 2017, Ng et al. 2018, Gao and Sun 2017, Lu and Yi SWRC in a log-log state) on void ratio and suction was in turn
2017, Zhou et al. 2019, Liu et al. 2020, Pasha et al. 2019 and given by
2020). Changes in the soil’s pore structure directly affect water
movement in the soil by altering the size distribution of the pores
as well as the pore connectivity that has direct impact on the (19)
SWRC. Therefore, the SWRC obtained at a specific soil volume
or compacted state cannot be used at other states. In other words, where . and are the values of and
the SWRC must be determined at the void ratio of interest to be at the reference void ratio . was taken at 0.55. Using
of a fundamental and practical value. Ignoring the impact of void (18) and (19) and the form of SWRC at as the reference
ratio on the SWRC can lead to significant error and response, Mašín (2010) was able to predict evolution of a range
misunderstanding in extracting fundamental engineering of soils with a change in void ratio with remarkable accuracy.
properties of unsaturated soils based on the SWRC (Pasha et al. Khoshghalb and Khalili (2013) and Pasha et al. (2017)
2016, Lajmiri at al. 2020). extended Mašín’s work to include hydraulic hysteresis effect,
Many researchers have examined the coupling between and provided an alternative relationship for the evolution of
SWRC and soil volumetric strain. Notable contributions include with as
Wheeler et al. (2003), Gallipoli et al. (2003), Nuth and Laloui
(2008), Mašín (2010), Gallipoli (2012), and Pasha et al. (2016, (20)
2020). Experimental contributions in this area are attributed to
Vanapalli et al. (1999), Romero and Vaunat (2000), Ng and Pang Using the fractal distribution of particle size distribution
(2000), Lee et al. (2005), Miller et al. (2008), Tarantino (2009), Khoshghalb et al (2015) in turn showed that the evolution of
Salager et al. (2013) to name a few. The effect of volume change with can be represented by
on soil microporosity and its link with the SWRC can also be
found in the experimental results of Simms and Yanful (2002), (21)
Koliji et al. (2006), Romero et al. (2011) and Casini et al. (2012).
Despite the important effect of the void ratio and volume
change on the SWRC, a review of the literature shows that in the which is identical to equation (18). In equation (21), is the
majority of the cases the void ratio dependency of SWRC is grain shape factor which assumes a value of one for circular
ignored. This has led to erroneous use and at times contradictory particles and a value greater than one for other shapes. is the
interpretations of SWRC data, particularly in relation to the fractal dimension of the particle size distribution. A similar
evolution of the air entry value and pore size distribution index approach was presented by Russell (2014).
with void ratio (e.g. see Rassam and Williams 1999, Sun et al.
302
3 CONSTITUTIVE MODELLLING OF UNSATURATED Sheng (2015), Bellia et al. (2015), Lei et al. (2016), Mun and
SOILS McCartney (2017), Lloret-Cabot et al. (2013, 2017), Bui et al.
(2017), Zhao et al. (2019), Zhang et al. (2019), Moghaddasi et al.
Many constitutive models have been proposed over the past five (2021), Phan et al. (2021), Ghorbai et al. (2016, 2021a,b),
decades to describe the nonlinear hydro-mechanical behaviour of Komolvilas et al. (2022). It is increasingly recognised that
unsaturated soils. The early work primarily focused on the previous arguments against the use of the effective stress were
applicability of the effective stress equation to describe the false and were invariably based on a linearly equivalent elastic
volume change and shear strength characteristics of unsaturated theoretical framework, which cannot be used to explain non-
soils. Bishop and Blight (1963) obtained strong experimental recoverable volumetric responses such as collapse. As pointed
evidence for the validity of the effective stress in unsaturated out by Loret and Khalili (2000), even in saturated soils, it would
soils by performing a series of “null tests” and demonstrating that be difficult, if not impossible, to explain irrecoverable volumetric
the shear strength and volume change characteristics remain deformations such as dilation and/or collapse (i.e. in metastable
unchanged when the individual components of the effective structures) in terms of the effective stresses alone without
stress are altered but in such a way that the effective stress invoking an appropriate plasticity model. Other stimuli for the
remained constant. However, Jennings and Burland (1962) use of the effective stress have been: i) the mechanical response
questioned the validity of the effective stress in unsaturated soils of soils to changes in total stress, pore water pressure and pore
arguing that it cannot explain the collapse phenomenon upon air pressure can be related to a single stress variable, rather than
wetting. They conducted a series of consolidation tests on several two or three independent stress state variables, ii) both saturated
unsaturated soils, and showed that upon flooding, i.e. a reduction and unsaturated states of the soil can be treated using a single
in suction, all samples collapsed, rather than expanding as is constitutive framework, which is particularly useful when
implied by the effective stress principle. Similar arguments were considering the cyclic response of variably saturated porous
also put forward by Aitchison (1965), Matyas and Radhakrishna media such as embankments and slopes, or when the boundary
(1968), Brackley (1971), Fredlund and Morgenstern (1977) and between saturated and unsaturated states is subject to significant
Gens et al. (1995), among others, leading to many investigators fluctuations.
concluding that the effective stress is not applicable to
unsaturated soils, not fully appreciating what the role of the 3.1 Two Stress State Constitutive Models
effective stress is (explained in section 2.3.1).
Fredlund and Morgenstern (1977) suggested that the The two stress state variables constitutive models are formulated
constitutive behaviour of unsaturated soils can be described by adopting the stress state variables: net stress,
using two independent stress variables, namely net stress (the , and suction, , as stress drivers of the
total stress in excess of pore air pressure) and matric suction mechanical response. With this approach a fully coupled two-
(pore air pressure in excess of pore water pressure), rather than a phase hydro-mechanical elasto-plastic model is required to
single effective stress. Several advanced constitutive models capture the mechanical response of the solid skeleton to
were developed within this framework, including the variations in and . In addition, the approach requires
determination two sets of material parameters: one
q
corresponding to and the other relating to .
Alonso et al (1990) were first to develop a comprehensive
elasto-plastic constitutive model for unsaturated soils within the
two stress state variables framework. Commonly referred to as
s the Barcelona Basic Model (BBM), the model assumed isotropic
plasticity rules, similar to those used in Cam-clay type models.
p
d s The effective stress was replaced by net stress and an additional
pc(0) p SI soil skeleton volumetric strain component was expressed in
d
terms of matric suction. The yield surface was expressed in the
p
mean net stress, , deviatoric stress, , and matric suction,
d
LC
p
, space as
pnet
(22a)
Figure 1. Barcelona Basic Model (BBM).
with
contributions of Alonso et al. (1990), Wheeler and Sivakumar
(22b)
(1995), Cui et al. (1995), Cui and Delage (1996), Vaunat et al.
(2000), Rampino et al. (2000), Georgiadis et al. (2005), Thu et
where is the apparent isotropic preconsolidation pressure
al., (2007), Sheng et al. (2008a,b). Indeed, most of the early
at the suction value of interest, is the cohesion intercept,
advanced constitutive models for unsaturated soils were
is a material constant. A graphical representation of the yield
predominantly based on the two-stress state approach.
surface is presented in Figure 1. At , the cohesion
However, in the past two decades, there has been a shift
intercept, , reduces to zero, and the modified Cam-clay ellipse
towards use of the effective stress approach in constitutive
is recovered in the plane. The yield surface in the
models for unsaturated soils. Examples include Kohgo et al.
plane was defined using two sets of yield curves, referred
(1993), Modaressi and Abou-Bekr (1994), Bolzon et al.
to as the load-collapse (LC) curve and the suction increase (SI)
(1996), Hutter et al. (1999), Loret and Khalili (2000), Karube and
yield locus (Figure 2). The contribution of matric suction to
Kawai (2001), Loret and Khalili (2002), Wheeler et al. (2003),
plastic behaviour was captured through the LC curve, and a
Gallipoli et al. (2003), Laloui et al (2003), Sheng et al. (2003,
suction-dependent apparent cohesion, represented by the
2004), Tamagnini (2004), Santagiuliana and Schrefler (2006), Li
intercept of the yield surface with the deviator stress axis (22b).
(2007a,b), Sun et al. (2007a,b), Samat et al. (2008), Khalili et al.
The LC curve represented the shift in the preconsolidation
(2008), Buscarnera and Nova (2009), Coussy et al. (2010), Sheng
pressure of the soil with unsaturation or suction and, for the first
and Zhou (2011), Zhang and Ikariya (2011), Kikumoto et al.
(2011), Zhou et al. (2012a,b, 2018), Buscarnera and Einav
(2012), Dangla and Pereira (2014), Hu et al. (2015), Zhou and
303
s s
so SI Net stress
LC
LC
Elastic Effective stress
region LC
se=sae
Point 1
pc(0) pnet pc (0), pc (0) pnet , p
Figure 2. Load-Collapse (LC) and suction increase (SI) yield curves. Figure 4. LC curves in effective stress and net stress planes.
time, enabled modelling of collapse upon wetting - one of the key always associated with a pore water that is in suction, not all
characteristics of unsaturated soil behaviour. The LC curve was soils with a suction are unsaturated. A classic example is a
one of the most fundamental and innovative aspect of BBM. It saturated heavily over consolidated soil subjected to
was mathematically expressed as undrained shearing, which generates negative pore pressure
during shearing without being unsaturated.
The form of the LC curve assumed in BBM, see Figure 2,
(23)
violates well established principles of saturated soil
mechanics. To elaborate, we note that below the soil is
in which is the isotropic preconsolidation at zero suction, saturated, and unsaturation only occurs when . We
and is the reference stress at which one reaches the also note that for a saturated soil, suction has no influence on
saturated normal compression line, starting from an unsaturated the preconsolidation pressure, . Hence, for the LC curve
compression line; i.e. though a wetting path that only involves to be valid, must remain constant in the range
elastic swell (see Alonso et al. 1990). and are the slopes ], before increasing with suction at (Loret and
of the normal compression line at the suction of interest and Khalili 2000). This is represented by the dashed line in
reload and unload line in plane (Figure 3). was
assumed to be independent of suction for convenience, even
though there was experimental evidence to the contrary (Alonso pnet
pc(0) pc(s1) pc(s2) pnet
v
s1)
1
v
1
s1 s2
s1)
1 s2
1 s1
s=0
s=0
Figure 3. Evolution of normal compression line with suction - BBM. Figure 5. Convergent evolution of normal compression lines with
suction.
et al. 1990). Figure 4. Now, invoking the validity of effective stress
Undoubtedly, BBM was a major step forward in the principle for saturated soil, the net stress representation of the
constitutive modelling of unsaturated soil, and has formed the LC curve must therefore follow a slope of -1:1 until ,
basis for virtually all subsequent two stress state constitutive before increasing with , as depicted b y the solid line
models proposed for the behaviour of unsaturated soils, e.g. Josa in Figure 4. It is noted that this will result in a non-convex
et al. (1992), Gens and Alonso (1992), Alonso et al. (1999a), yield surface in the BBM.
Wheeler and Sivakumar (1995), Wheeler (1996), Cui and Delage Central to the suction hardening model presented in BBM is
(1996), Vaunat et al. (2000), Wheeler et al. (2002), Chiu and Ng the assumption that the slope of the isotropic compression
(2003), Benatti et al. (2013) among others. Nevertheless, it was line decreases with increasing suction (Figure 3).
formulated based on limited experimental data (Alonso et al. Divergent normal compression lines at different values of
1987), and inevitably included assumptions/simplifications suction are physically inadmissible; i.e. at high stresses, the
which were restrictive. Some aspects are discussed below: normal compression lines for saturated and unsaturated soils
The most obvious restriction in the model is due to the must converge rather than diverge (Figure 5). To explore this
assumption that unsaturation commences from =0, hence further, consider two identical samples of the same soil with
omitting the essential role of in separating saturated from the initial void ratio of . One sample is initialised to
unsaturated states in a soil. While an unsaturated soil is kPa and , and the other to
304
kPa and kPa. Now, both samples are subjected to
a MPa at constant suction. It is evident that the
final void ratio of the samples will be identical. The T
experimental evidence also suggests a
slightly with suction, with a tendency to converge at high
stresses (Figure 6). It is noted from that for ,
equation (23) predicts softening of soil skeleton with suction
rather than hardening. Also, for the case of , the cohesion intercept, , which is assumed to be a linear
applicable to most practical problems, the model predicts no function of suction though the material constant . It is well
suction hardening, i.e. . Both outcomes are understood that the shear strength in unsaturated soils is a
contrary to the physics of unsaturated soils. It is also strongly nonlinear function of suction, and that equation 22b
recognised that the divergent normal compression lines is unlikely to capture the deviatoric response of unsaturated
suggest increasing collapse potential with , which again
1.2
1.1
e
1
0.9
0.8
25 250
pnet (kPa)
Figure 6. Convergent normal compression lines with suction Figure 8. Mechanical response of a normally consolidated soil
(data from Wheeler and Sivakumar 1995) subjected to increasing values of suction – experimental data
(data from Fleureau et al 1993).
is not supported by experimental evidence (Gens et al. 2006).
Another novel feature in the BBM is the introduction of the soils, particularly at the critical state. It is also noted that the
suction increase (SI) locus. This implies that a normally assumption of being constant is akin to assuming that the
consolidated soil, located at Point 1 in Figure 2, subjected to effective stress parameters is independent of
suction loading at constant , will initially undergo suction/degree of saturation.
elastic straining followed by an elasto-plastic response once Finally, for a two stress state model to be complete, it is
the suction exceeds a previously attained maximum, necessary that strain conjugates are defined for each of the
7 . state stresses along with the corresponding elasto-plasticity
8. elements such as the elastic model, the yield surface, the flow
rule, the hardening model, etc. BBM is silent with respect to
- the elasto-plasticity elements associated with the hydraulic
response, and its cross coupling with the mechanical model.
Josa et al. (1992) modified the BBM using non-linear relations
for the variation of void ratio such that collapse strains
diminished with net stress before reducing to zero at high
stresses. Wheeler and Sivakumar (1995) adopted a similar
approach to Alonso et al. (1990) but assumed an associative flow
rule and defined the slope of the CSL in the deviator stress-mean
so ln s net stress plane as a function of matric suction. Wheeler et al.
(2002) attempted to extend the application of (23) to convergent
normal compression lines by requiring that the compression lines
at different values of suction passed a single point at high
stresses. This may not be physically justifiable. This constraint
v s
1 also implied that could be obtained by an elastic cut
through the normal compression lines in the . While
s
such an approach is entirely valid in the plane, it is not
1 correct in the net stress plane as it ignores the stress path taken
by the soil due to the suction increase (for a correct approach, see
Alonso et al. 1990). Cui and Delage (1996) extended BBM to
include anisotropy by using an inclined ellipse as a yield function
and a non-associated flow rule. Wheeler (1996) and Dangla et al.
(1997) included the elasto-plasticity associated with the
Figure 7. Mechanical response of a normally consolidated soil hydraulic model. The coupling of the hydraulic and mechanical
subjected to increasing values of suction – BBM. models was also attempted by Vaunat et al. (2000), Wheeler et
305
al. (2003) and Sheng et al. (2008a,b,c). Sheng et al. (2008a,bc)
adopted the correct form of the LC curve. The resulting non- q
convexity of the yield surface was addressed through an explicit
integration scheme adopting small perturbations of strain and M
suction (Sheng et al 2008c). Despite these attempts, many of the 1
two stress state models proposed in the literature retain all the
original features of BBM without alteration.
3.2 Effective Stress Constitutive Models
In the effective stress approach, a single effective stress, ,
rather than two or three independent stresses, is introduced as the
plasticity driver. The approach is similar to that of saturated soils
except that an additional hardening parameter is introduced to pc pc p
allow for the expansion of the elastic region with unsaturation.
As discussed previously, the key advantage of the effective 2
stress approach is that it permits capturing the mechanical Figure 9. Effective stress model for unsaturated soil -
response of the solid skeleton, both in the elastic as well as elasto- Modified Cam-Clay model.
plastic regions, independent of the hydraulic model, and solely
hydraulic state and the hysteresis on the stiffening of the soil
based on . In this approach, there is no need for the
skeleton (e.g. Gallipoli et al. 2003, Tamagnini 2004, Zhang et al.
introduction of suction, , as an additional independent stress
2007a,b, Gallipoli et al. 2008, Zhou et al. 2012a,b, Komolvilas
variable, as the stress effect of with respect to mechanical
and Kikumoto 2017, Song and Khalili 2019, Moghaddasi et al.
response is a priori accounted for in . Another important
2021).
advantage of the effective stress approach is that the hydraulic
It is instructive to reiterate that even when is taken as , it
model is reduced to the water retention response of the soil.
enters the plasticity model has a hardening parameter rather than
These simplify the constitutive relationships of unsaturated soil,
a stress variable. The role of , in this regard, is formally
and significantly reduce model parameters.
identical to that of temperature and cementation in constitutive
In the effective stress approach, the transition between
modelling of soils, even though the physical effect of
saturated and unsaturated states is seamless and the model
temperature is opposite to that of suction.
parameters are exactly the same of those used in saturated soils
Kohgo et al. (1993) and Modaressi and AbouBekr (1994) were
except for to two entities that can be determined in any soil
amongst first to present an elasto-plastic model for unsaturated
physics laboratories. More importantly, the need for testing in an
based on the effective stress, accounting for the dual effect of
unsaturated state is significantly diminished. Testing soils in an
suction in increasing the effective stress and the preconsolidation
unsaturated state is time consuming, requires sophisticated
pressure. Bolzon et al. (1996) used Bishop’s effective stress for
laboratory equipment and expertise, and has been one of the key
casting the elasto-plastic equations and introduced the suction
inhibitors of the use of unsaturated soil mechanics in practice.
hardening effect directly into the plastic modulus. As a results
Finally, many of the restrictions of BBM, which are related to the
they were unable to predict collapse upon wetting. This was
choice of stress variables, do not apply to the effective stress
rectified by Santagiuliana and Schrefler (2006). Loret and Khalili
models for unsaturated soils.
(2000) proposed a comprehensive framework for constitutive
Using the effective stress framework, the yield surface for an
modelling of unsaturated soils. This framework was formulated
unsaturated soil is simply defined as where is
considering a three-phase porous medium and the effective stress
the parameter controlling the size of the yield surface,
principle. The work presented a fully coupled flow-deformation
determined from the physics of the problem. The plastic potential
model taking into account the elasto-plasticity of the solid
is defined as , where controls the size of the
skeleton as well as the air and water phases within the system.
plastic potential, determined by requiring to pass through the
The key features of this model were: the continuity of behaviour
current stress point, . Similar to saturated soils, the flow rule
at the transition between saturated and unsaturated states;
emanates from the normality rule applied to the stress strain
incorporation of air entry suction directly into the formulation as
conjugates and plastic potential .
the demarcation point between saturated and unsaturated states;
Adopting the modified Cam-clay model as the plasticity
incorporation of suction in the yield surface as the hardening
driver, and invoking the associativity of the flow rule, the yield
parameter; the use of the soil water characteristic curve in
function and the plastic potential for an unsaturated soil in the
determination of coupling between the water and air phases, and
plane is expressed as
capturing the effect of plasticity on the coupling of flow and
deformation fields. Qualitative predictions of the model were
(23)
shown to produce characteristic features of unsaturated soils such
as collapse upon wetting, and plastic followed by elastic response
This is identical to the model for saturated soils (Figure 9), except
of normally consolidated soils during drying. Loret and Khalili
that is quantified using (5) and the preconsolidation pressure,
(2002) developed an elaborate elasto-plastic constitutive model
), is made a function of plastic volumetric strain, ,
for unsaturated soils using as an extension of the Cam-clay
and the hardening parameter, . The hardening parameter, ,
plasticity model. In this work, both the elastic behaviour and the
captures the stiffening effect of suction/unsaturation on the soil’s
failure surface at critical state were defined in terms of the
solid skeleton. Inspired by the work of Alonso et al. (1990), the
effective stress, and the stiffening effect of suction observed in
unsaturation hardening parameter is typically taken as
many experiments (Alonso et al. 1990; Wheeler and Sivakumar
(e.g., Kohgo et al. 1993, Modaressi and AbouBekr 1994, Loret
1995) was captured using a simple hardening model. Suction was
and Khalili 2000, Khalili and Loret 2001, Khalili et al. 2004,
incorporated into the effective stress using the relationship
Sheng et al. 2003, Eberhardsteiner et al. 2003, Borja 2004,
established by Khalili and Khabbaz (1998). The model adopted
Tamagnini and Pastor 2004, Sheng et al. 2004, Ehlers et al. 2004,
an associated flow rule with a split elliptical function for the yield
Russell and Khalili 2006, Mašín and Khalili 2008). However,
surface and plastic potential. The slope of the critical state line in
more recently, it is shown that may be a more suitable
the deviatoric stress-mean effective stress plane was assumed to
hardening parameter, since it naturally captures the effect of
be a material constant. The model required a minimal number of
306
material parameters. Loret and Khalili (2002) used this model to models proposed essentially use the SWRC to obtain the
provide good predictions of the Wheeler and Sivakumar (1995) elastoplastic parameters of the hydraulic model in order to
data. Adopting the same framework Russell and Khalili (2006) reproduce the same SWRC, albeit in an approximate way. A
proposed a unified bounding surface model for sand and clays more prudent and usual approach will be to introduce SWRC as
for monotonic loading of unsaturated soils. Khalili et al. (2008) a constitutive input into the three-phase model. Complete
extended the theoretical developments of Loret and Khalili examples can be found in Loret and Khalili (2000), Khalili et al.
(2000) to include mechanical as well as hydraulic hysteresis (2008), Komolvilas and Kikumoto (2017) and Moghaddasi et al.
effects within the context of the bounding surface plasticity. (2021). In such an approach, the void ratio dependency of SWRC
Other effective stress based models presented in the literature is treated as a material nonlinearity, without invoking an
are due to Lewis and Schrefler (1998), Jommi (2000), Sheng et elaborate elasto-plasticity model.
al. (2003), Wheeler et al. (2003), Gallipoli et al. (2003), Laloui
et al. (2003), Tamagnini (2004), Sun et al. (2003, 2007a, b), constitutive modelling of unsaturated soils through introducing
Mašín and Khalili (2008), Morvan et al. (2010), Wong et al. the concept of equivalent stresses. A normalised yield locus was
(2010), Manzanal et al. (2011 ), Lloret-Cabot et al (2013, 2014, introduced that did not depend on suction; i.e. the change in
2017) , Bellia et al. (2015), Lai et al. (2016), Ghorbani et al. suction did not lead to a change in the yield locus size. They
(2016), Zhang et al. (2019), Ghaffaripour et al. (2019), Bruno stated that the equivalent stress concept may be used to quickly
and Gallipoli (2019), Fabbri et al. (2019), Sitarenios and extend models for saturated soils into unsaturated regime. As
Kavvadas (2020), Moghaddasi et al. (2021).
this approach to describe the complexities associated with the
s
behaviour of unsaturated soils during wetting and drying cycles,
-dS Pr as well as during loading and unloading is not clear and needs
SI further investigation.
3.3 Cyclic Plasticity Models for Unsaturated Soils
p
d p
Conventional elasto-plastic constitutive models fail to describe
SD dS Pr soil behaviour during load reversals. This has, over the years,
prompted the need for enhanced constitutive models capable of
correctly describing the soil response subjected to cyclic loading.
Among the various cyclic plasticity models available in the
p'c p'
literature, kinematic hardening and bounding surface plasticity
Figure 10. Elasto-plastic hydro-mechanical model models have been shown to provide a convenient framework to
(after Wheeler et al. 2003). model a number of observed aspects of cyclic behaviour of soils
such as hysteretic response, memory of past stress history,
As shown by several researchers, irrecoverable deformations smooth degradation of stiffness during loading, small strain
such as collapse can be readily captured within an effective stress stiffness and early onset of plastic strains (Gajo and Muir Wood,
framework by introducing suction or degree of saturation as a 2001). Bounding surface plasticity has attracted a great deal of
hardening parameter in the defining the yield surface (Kohgo et attention due to its simplicity and efficiency in modelling cyclic
al., 1993; Modaressi and AbouBekr, 1994; Loret and Khalili, behaviour. The advantage of the bounding surface formulation is
2000; Khalili and Loret, 2001; Khalili et al., 2004; Russell and that it is geometric in nature with little appeal to the physical
Khalili, 2006; Mašín and Khalili, 2008; Moghaddasi et al., 2021). reasoning of the problem. Yet, it has been shown to accurately
Wheeler et al. (2003) presented an elasto-plastic model based capture the cyclic response of soils and provide a smooth
on three yield surfaces, a yield surface representing mechanical transition from elastic to elasto-plastic behaviour. The method is
behaviour (LC curve) and the other two (SI and SD) representing computationally simple, uses fewer model parameters and results
water retention behaviour of soil (Figure 10). These three of the simulation fit experimental data with reasonable accuracy.
surfaces were coupled through dependency of the LC curve on Khalili et al. (2008) proposed a fully coupled flow deformation
suction and shift in the SWRC with volume change or void ratio model for the cyclic analysis of unsaturated soils including
(see section 2.4). The model was originally developed for 1D hydraulic and mechanical hystereses. The model was developed
isotropic loading, and later extended to the general stress space within the context of the bounding surface plasticity using the
by Lloret-Cabot et al. (2013, 2017). Sheng et al. (2004) extended critical state theory and the effective stress approach. An
Wheeler’s model to deviatoric loading and proposed a more important feature of model was that at any stress point two
general model of hydraulic hysteresis, cast into a classical theory directions of plastic stain increment vector are defined (Khalili et
of elasto-plasticity. In contrast to the previous models, the SI, SD al. 2005). This was shown to be essential to capture volumetric
and LC yield surfaces were allowed to move independently of behaviour of the soil during unloading and re-loading on the dry
each other and non-associative flow rules were defined for SI and side of the critical state line. A void ratio-dependent water
SD surfaces. Variations to Sheng’s model were proposed by Sun retention model was adopted, taking the effect of hydraulic
and his colleagues (i.e. Sun et al., 2007a,b, 2008). Tamagnini hysteresis into account.
(2004) extended the Cam-clay model for unsaturated soils Liu and Muraleetharan (2012a,b) proposed a coupled hydro-
including the hydraulic hysteresis effects. The model was based mechanical constitutive model in the general stress space for
on the effective stress concept using . The constitutive unsaturated sands and silts under both monotonic and cyclic
relationships of Romero and Vaunat (2000) were adopted for the loading conditions. The model was the extension of the isotropic
definition of hysteretic soil water retention curves (SWRCs). In elastoplastic model of Muraleetharan et al. (2009). They
the model, the evolution of the yield surface was governed by considered the hysteretic properties of the SWRC within a
plastic volumetric strain as well as degree of saturation. bounding surface plasticity framework for the mechanical
Common to the above models is that the hydraulic behaviour behaviour. The stress driver of plasticity was assumed to be the
of soils is captured through an elastic-plastic formulation intergranular stress defined as ,
involving definition of yield surfaces (SI and SD), evolution law where is the water volume fraction. They allowed for the
and flow rule. Such a formalism is rarely warranted for one- effect of suction on hardening of the bounding surface through
dimensional phenomena such as the water return behaviour. The the irrecoverable water content calculated from the SWRC
307
model. Pedroso and Farias (2011) as well as Cui and Zhao (2017) Muraleetharan et al. (2009), considering the effect of hydraulic
extended the BBM for the cyclic analysis of unsaturated soils. hysteresis. Bounding surface and subloading surface plasticity
Two yield surfaces were defined in order to provide a smooth frameworks were employed to describe hydraulic and
transition from elastic to plastic behaviour. Oka and Kimoto mechanical behaviour, respectively. The model was later
(2012, 2022), Oka et al. (2019), and Shahbodagh (2011) extended by Ghasemzadeh et al. (2017) for general stress states.
proposed a cyclic elasto-viscoplastic constitutive model for the Kaewsong et al. (2019) proposed a constitutive model for
dynamic analysis of unsaturated soils. The model was developed unsaturated soils within the framework of bounding surface
within the context of the overstress theory and the effective stress plasticity. The model was the extension of the bubble model
concept. The nonlinear kinematic hardening rule, softening due proposed by Al-Tabbaa and Wood (1989). An elliptical elastic
to the structural degradation of soil particles, and suction bubble was defined inside a modified Cam-clay bounding
hardening were taken into account in the model. The suction surface. The size of the elastic bubble was modelled as a function
hardening equation proposed by Cui and Delage (1996) for of suction, degree of saturation, and plastic volumetric strain.
unsaturated silts was adopted in the model. Shahbodagh et al. The model was shown to be capable of capturing the effects of
(2015) proposed a numerical model based on the theory of recent suction history on nonlinear stress–strain relation and
multiphase mixtures for nonlinear large deformation dynamic shear modulus degradation at small strains. Bruno and Gallipoli
analysis of unsaturated porous media including hydraulic (2019) proposed a bounding surface model to describe the
hysteresis. In the model, the coupling between solid and fluid behaviour of unsaturated soils under isotropic loading. The
phases was enforced according to the effective stress principle model was based on the hydraulic law of Gallipoli et al. (2015)
allowing for the suction dependency of the effective stress and the mechanical law of Gallipoli and Bruno (2017). The
parameter. The effect of hydraulic hysteresis on the effective bounding surface theory was used for modelling both hysteretic
stress parameter and soil water characteristic curve were taken water retention and mechanical behaviour of soil. Ghorbani et at.
into account in the model. They showed that the effect of (2021a) also adopted a similar bounding surface approach but
hydraulic hysteresis could markedly alter the response of an included the effect of soil anisotropy.
unsaturated soil during dynamic loading that will invariably
involve complex cycles of strain-induced wetting and drying. 3.4 Constitutive Models for Unsaturated Expansive Soils
Zhou et al (2015) developed a bounding surface model for
unsaturated soils under cyclic loading conditions. The approach The mechanical behaviour of expansive soils, also referred to as
proposed by Gallipoli et al. (2003) was adopted to capture the reactive soils, is influenced by not only the mechanical loading
evolution of bounding surface with suction. Three bounding and matric suction, but also the physico-chemistry of the soil or
surfaces are defined in the model: one describing elastoplastic osmotic suction. A range of constitutive models have been
behaviour during compression; one describing elastoplastic proposed in the past three decades for the behaviour of expansive
behaviour during shearing, and one expressing elastoplastic soils mainly based on an extension of the BBM and the
behaviour as suction changes. Komolvilas and Kikumoto (2017) experimental observations on compacted bentonite in relation to
developed a cyclic elastoplastic model for the analysis of buffer materials in waste disposal facilities. Of particular interest
liquefaction in unsaturated soils. The model was the extension of has been the irrecoverable (plastic) swell observed in expansive
the modified Cam clay model. It was based on the subloading soils due to wetting at low confining pressure.
surface concept and the effective stress approach. The state The most notable contribution on this topic was due to Gens
boundary surface was defined as a function of the degree of and Alonso (1992) who proposed a double structure model for
saturation. The soil water characteristic curve adopted the mechanical behaviour of unsaturated expansive clays, within
considered the effects of specific volume and hydraulic the framework of BBM (Alonso et al., 1990). The key elements
hysteresis. Gholizadeh and Latifi (2018) developed an effective of the model were: i) the mechanical behaviour of the
stress-based hydro-mechanical model for unsaturated soils under macrostructure followed the BBM, ii) the micro-structure was
cyclic loadings. Two separate mechanisms were adopted for the always saturated and elastic, and iii) micro-structure was
mechanical behaviour of the material, i.e. the conventional unaffected by macro-structure, but the deformation of the
plasticity for isotropic loading and a multi-yield surface plasticity microstructure was able to affect the macrostructural level. In
framework for deviatoric loading. Xiong et al. (2019) proposed particular, it was assumed that elastic microstructural “swelling
an elastoplastic constitutive model for unsaturated soil under will affect the soil skeleton through increasing its macro-
monotonic and cyclic loading. The model was developed based structural void ratio. This plastic volume change leads in turn to
on the work of Zhang et al. (2007a,b) and Zhang and Ikariya a movement of the LC to the left (a softening in hardening
(2011). The concept of stress-induced anisotropy proposed by plasticity terms) in response to the new structural arrangement.”,
Zhang et al. (2007a,b) was adopted in the model to capture the and iv) the ratio of macro-structure plastic strain to elastic swell
cyclic response of soil. Bishop’s effective stress and degree of of micro-structure was directly proportional to the over-
saturation were used as state variables in the framework. The consolidation ratio of the soil, attaining a value of zero for a
superloading and subloading concepts were introduced in the normally consolidated compacted clay.
model to consider the influence of overconsolidation and This framework was further refined in the Barcelona
structure on deformation and strength of soils. Cao et al. (2021) Expansive Model (BExM) by Alonso et al. (1999a), where the
proposed a constitutive model for unsaturated soils subject to deformations of two structure levels (micro-structure and
high-cycle traffic loading. The model was developed based on macrostructure) were elaborated and the behaviour of micro-
the BBM and the shakedown concept. Cai et al. (2022) proposed structure was extended into the unsaturated domain,
a fractional-order bounding surface model for unsaturated soils characterised by using the effective stress principle. This
under cyclic loading with constant matric suction. The effective modelling framework was subsequently improved and extended
stress and matric suction are used as the constitutive variables. by Sánchez et al. (2005) and Navarro et al. (2014), and integrated
The effective stress parameter proposed by Khalili and Khabbaz into numerical simulation codes to solve coupled boundary value
(1998) was adopted in the model. The movable mapping centre
rule was used to describe the hysteresis characteristics of the There are several fundamental difficulties associated with the
dynamic stress–strain curve. Ghasemzadeh and Ghoreishian conceptual model underpinning the BExM-type constitutive
Amiri (2013) proposed an elastoplastic constitutive model for models:
unsaturated soils under isotropic loading conditions. The model As shown by Khalili et al. (2010) and Mašín (2013) an
was developed based on the elasto-plastic framework of elastic, isotropic, self-similar expansion of the micro-
308
structure will not lead to a change in the macro-structural unsaturated expansive soils but they adopted an effective stress
configuration of the soil. The volume change of the soil will approach. Lei et al. (2016) used the effective stress approach to
simply be in the form of a magnification of the macro-pore extend the chemo-mechanical model of Loret et al. (2002) using
structure and the associated micro-structural grains. This the general plasticity framework of Loret and Khalili (2002). In
cannot lead to a change in the macro-mechanical behaviour this model, a generalised effective stresses expression was
and softening as was envisaged in the model, introduced incorporating the effects of pore water chemistry.
All volume changes emanating from an underlying elastic They related the mechanical properties to the evolution of molar
deformation (i.e. elastic expansion of the micro-structure) by fraction of absorbed water as well as matric suction as
definition must be reversible. It is unusual to associate a softening/hardening parameters. Guimarães et al. (2013)
plastic straining with an elastic phenomenon without extended BExM to incorporate the chemical effects. They
invoking other processes, adopted BBM for the behaviour of macrostructure and defined a
The function defining the plastic coupling of the macro- modified effective stress expression, including the osmotic
structure to the elastic straining of the micro-structure is Chen et al. (2016) utilised theory
based on limited mechanical reasoning and appears to serve of mixtures to developed a constitutive model for coupled hydro-
only a fitting role, chemo-mechanical analysis of unsaturated highly swelling
The existence of double structure is not specific to materials within the framework of non-equilibrium
compacted bentonite. It is also present in compacted kaolin. thermodynamics. The work included the combined effects of
However, kaolin does not show an expansive behaviour, chemical osmosis and hydration swelling, assuming passive air
The expansive behaviour of clays is heavily influenced by pressure.
their physico-chemical properties of the soil and such effects In summary, a great deal of work has been presented on the
cannot be captured using net stress and suction alone, and mechanics of unsaturated expansive soils in the past three
The macro-structural model is based on BBM without decades. Even so a number of clear gaps remain in the
alteration. knowledge. Some of the questions that require attention include:
Alonso et al. (2011) and Gens et al. (2011) extended the BExM Is the soil structure the only driver of expansive behaviour?
double structure model to include the hydro-mechanical coupling Should physico-chemical effects be considered in the
mechanisms. However, they ignored the dependency of water constitutive modelling of expansive soils? What are the
retention behaviour on the volumetric deformation in their appropriate state parameters with respect to the physico-
models. This issue was later addressed by Della Vecchia et al. chemistry of expansive soils? What are the influences of osmotic
(2013) who developed a fully coupled hydro-mechanical model suction/chemical potential on the strength of reactive soils and
within the double structure framework. This model was an how will the microstructural alterations influence the response?
extension of the model by Romero et al. (2011). Mašín (2013) (The current experimental evidence is inconclusive.) What are
proposed a hypoplastic framework for double structure the combined effects of osmotic and matric suctions on the
hydromechanical modelling of unsaturated expansive clay in an effective stress of the soil, if any, and how can these be
extension of Mašín and Khalili (2008). Fully coupled quantified? What is the appropriate scaling parameter for
hydromechanical models were presented for each structural capturing the contribution of osmotic suction for the effective
level. The effective stress representation by Khalili and Khabbaz stress of the soil skeleton? How does pore water chemistry alter
(1998) was applied at both micro- and macro-structural levels. the soil stiffness and strength and to what extent? What is the
Void ratio dependency of water retention properties at the two impact of repeated cycles of osmotic loading and unloading and
structural levels were also considered. Li et al. (2017a,b) physico-chemical hysteresis on the mechanical response and
presented a constitutive model for unsaturated expansive clays, strength of reactive soils? These, as well as many other
adopting the conceptual model of Baker and Frydman (2009) to unresolved, yet important, fundamental questions will require
extend the work input approach for the constitutive modelling of concerted and systematic research effort.
double structure unsaturated materials. In this model, the
bounding surface concept was employed for modelling the 3.5 Constitutive Models for Unsaturated Aggregated, Cemented,
mechanical behaviour. Fissured Soils
In addition to the double structure approach, some researchers
have adopted a single structure/porosity approach for modelling The hydro-mechanical and constitutive behaviour of unsaturated
the behaviour of expansive unsaturated soils. Sun and Sun (2012) soils can be significantly affected by the presence structure such
presented a model for unsaturated expansive soils, with a focus as aggregation, cementation and fissuring. Several constitutive
on the coupled hydro-mechanical behaviour. This model takes models have been developed to capture the essential features of
into consideration the coupled effects of degree of saturation and the unsaturated soils with structure (Khalili 2008, Yang et al.
void ratio on the mechanical and water-retention behaviour of 2008, Borja and Koliji 2009, Cai et al. 2010, Pereira et al. 2014,
soil. Li and Yang (2017) proposed a constitutive model for the Le Pense et al. 2016, Bruno et al. 2020, Moghaddasi et al. 2021).
hydro-mechanical response of unsaturated expansive soils Khalili (2008) presented a comprehensive hydro-mechanical
through introducing the concept of a macro-structural neutral model for the three-phase analysis flow-deformation analysis of
loading line. The concept was incorporated into the unsaturated fissured porous media. The formulation consisted of three
model of SFG (Sheng et al., 2008a) to derive a volume change separate, yet overlapping models: the deformation model, flow
equation for unsaturated expansive soils. Takayama et al. (2017) of air model and flow of water model. The deformation model
extended the elasto-plastic constitutive model of Ohno et al. was cast in the effective stress space satisfying the equations of
(2007) to unsaturated expansive soils. Cui et al. (2002) developed equilibrium. The flow model was based on the theory of double
a non-linear elastic model to predict the volume change porosity (Khalili and Valliappan 1996). The coupling between
behaviour of swelling clays with dense structure. the air and water flow was established through the water
Physico-chemical effects in the constitutive modelling of retention curve. The coupling between the fluid flows and
expansive soils have been considered by Liu et al. (2005), Zhang deformation was established through the effective stress
and Zhou (2008), Lei et al. (2016), Guimarães et al. (2013), and parameters. Borja and Koliji (2009) presented a double porosity
Chen et al. (2016) to name a few. Liu et al (2005), within the elasto-plastic hydro mechanical model for aggregated
BBM framework, proposed a chemo-mechanical unsaturated unsaturated porous media based on the mixture theory. The work
model following the work of Hueckel (1997) for saturated adopted the effective stress principle, which was derived based
soils extended Hueckel’s work to on the continuum principles of thermodynamics for multi-phase
309
double porous media. Following the classical theory of plasticity, hardening law, albeit restricted to isotropic loading condition.
the yield function was defined in terms of the effective stress but Zou et al. (2013) developed a rheological creep model for
adopting plastic internal variables and local suction in the two unsaturated soils, where time was included explicitly to describe
scales of porosity as the hardening parameters. Yang et al. (2008) the viscoplastic strain. A difficulty with this approach was that
proposed a bounding surface constitutive model incorporating any application of the model requires the definition of an origin
the combined effects of unsaturation and the initial structure. The for the time which cannot be defined in an objective way. De
model was suited for monotonic loading of structured soils with Gennaro and Pereira (2013) developed a viscoplastic constitutive
a fixed projection centre for the mapping rule. Cai et al. (2010) model for time-dependent behaviour of unsaturated geomaterials
presented a binary model for cemented unsaturated soils. The using BBM as the reference elastoplastic framework. The plastic
cementation behaviour was considered elastic where as that due mechanism associated with the SI surface was ignored in their
to unsaturated elasto-plastic. Pereira et al. (2014) presented a model. They used the isotache approach and defined a strain rate
constitutive model for volumetric analysis of cemented/ dependent hardening law to extend the BBM. Bui et al. (2017)
structured unsaturated soils. The model was formulated in the developed a viscoplastic damage model for partially saturated
effective stress space, with suction, plastic volumetric strain and rocks. They obtained the constitutive relations by decomposing
degree of structure as drivers of isotropic hardening/softening. It the inelastic strains into viscoplastic strain and strain due to
was shown that the model could predict the maximum collapse micro-cracking. Separate handling of viscoplastic deformations
due to wetting and loss of structure. Le Pense et al. (2016) from damage deformations is, however, questionable since only
presented a hydro-mechanical constitutive model for clayey soils the combined effects can be measured (Zienkiewicz and
accounting for damage-plasticity couplings as well as hardening Cormeau, 1974).
effects due to suction and plastic volumetric strain. A double
effective stress incorporating both the effect of suction and 3.7 Modelling of Strain Localisation in Unsaturated Soils
damage was defined based on thermodynamical considerations.
Coupling between damage and plasticity was achieved by using Strain localisation, or shear banding, is the concentration of
strain equivalency and the use of the double effective stress into deformation in a narrow zone due to intense shear straining. It is
plasticity equations. Bruno et al. (2020) presented a bounding the precursor to failure and can occur in a range of materials
surface model predicting the combined effects of cementation including alloys, metals, plastics, polymers, granular materials
and partial saturation on the mechanical behaviour of soils and soils. In unsaturated porous media, a breakdown in
subjected to isotropic loading. The loss of cementation caused by interfacial effects; e.g. due to shearing or water infiltration, is
loading, wetting or drying of a normally consolidated soil was shown to reduce strength and cause pronounced shear banding
described through a cementation bonding function which (Cui and Delage 1996, Cunningham et al. 2003, Higo et al.
monotonically decreased with increasing stress. Moghaddasi et 2013).
al. (2021) proposed a fully coupled hydro-mechanical bounding Using the multiphase mixture theory, several approaches have
surface elasto-plastic model for describing the behaviour of been proposed for the analysis of strain localisation in
unsaturated structured soils. The hydraulic characteristics of unsaturated porous media. Schrefler et al. (1996, 2006) and
structured soil were captured through a void ratio-dependent Zhang et al. (1999, 2007) presented a formulation for dynamic
hysteretic water retention model formulated based on the strain localisation in porous media with two fluid phases based
effective stress principle. The effects of structural degradation on an extension of Biot’s formulation assuming passive air
and the degree of saturation on the compressive and tensile pressure. They studied the characteristics of the natural length
strength of the material were considered through controlling the scale which exists in the multiphase models due to the seepage
size of the bounding surface, allowing for a smooth transition of process inducing a rate-dependent behaviour for the soil mixture.
the response from structured to unstructured states. A plastic Borja (2004) developed a mathematical framework, based on
work hardening approach was adopted to consider the effects of Cam-clay plasticity theory, to detect the inception of strain
stress magnitude and accumulated plastic strain on the localisation in partially saturated porous media under plane strain
degradation process. compression at the constitutive level. Ehlers et al. (2004 , 2011),
Shahbodagh (2011), Lazari et al. (2015), Oka et al. (2019) and
3.6 Constitutive Models for Rate-Dependent behaviour of Oka and Kimoto (2012, 2022) developed computational models
Unsaturated Soils for capturing quasi-static and dynamic strain localisation in
multiphase elasto-viscoplastic porous media. The material rate-
Rate dependency is an important element in the mechanical dependency considered was shown to eliminate the numerical
behaviour of geomaterials for capturing creep, creep-induced instability and introduce a natural length-scale into the problem,
failure and the rate of loading on the material response. avoiding the need to perform a diagnostic analysis for the onset
Nevertheless, many of the constitutive models for unsaturated of localisation. Schiava and Etse (2006) investigated the potential
soil have been constructed within the framework of the rate of bifurcation in partially saturated soils under uniaxial and plane
independent theory. strain loading conditions at constant suction. Callari et al. (2010)
Oka et al. (2006) were amongst first to develop elasto- studied, numerically, the response of a perfectly homogeneous
viscoplastic constitutive model for unsaturated soils accounting soil under plane strain compression and showed that the strain
for the effect of unsaturation hardening. A fully coupled three- localisation can be triggered by a heterogeneous effective stress
phase hydro-mechanical model was presented adopting the state induced by fluid flow coupling from the early stages of
advantage of the effective principle within the overstress-type testing. Buscarnera and Nova (2011) investigated the instability
viscoplasticity framework. Oka and Kimoto (2012, 2022), Oka of unsaturated soils under triaxial compression using the
et al. (2019) and Shahbodagh (2011) extended this model to controllability approach proposed by Nova (1994). Peri et al.
allow elasto-viscoplastic dynamic analysis of unsaturated soils. (2014) used the bounding surface model proposed by Khalili et
The nonlinear kinematic hardening rule, softening due to the al. (2008) and derived analytical solutions for the onset of strain
structural degradation of soil particles, and suction hardening localisation in unsaturated soils under undrained, constant water
were taken into account. The suction hardening equation content, and drained loadings. Borja et al. (2013), Song and Borja
proposed by Cui and Delage (1996) for unsaturated silts was (2014), and Song et al. (2017) studied the influence of initial
adopted in the model. De Gennaro et al. (2009) proposed a heterogeneity in unsaturated porous media, with spatial varying
rheological model for partially saturated chalks, including strain density and degree of saturation, on the inception of localisation.
rate effects by means of an extended strain rate-dependent In many of these models, the degree of saturation has been
310
adopted as the effective stress parameter. More recently, Song that this field affords considerable advantages in the following
and Khalili (2019) proposed a peridynamics model, developed aspects:
within the context of the effective stress concept, for strain Provides a closely coupled view of hydraulic and mechanical
localisation analysis of multiphase geomaterials at constant phenomena (multi-physics) where two or more fluids under
suction. different pore filling conditions coexist in a porous medium.
Provides the need to approach phenomena at different scales
3.8 Anisotropic Models for Unsaturated Soils (multi-scale viewpoint at pore/grain level, phenomenological
scale, and upscaling to application).
Unsaturated soils are commonly encountered in civil engineering In recent years, the expansion of geotechnical engineering
practice, formed by compaction, sedimentation or weathering of applications, focusing on resilience to adapt to and mitigate
rock, and invariability exhibit anisotropic behaviour. As pointed climate change effects and the research interest in multi-physics
out Cui and Delage (1996), the microstructure of the soil changes processes is becoming broader (e.g., Gens 2010). There are two
during compaction in order to provide the greatest possible important points to highlight in this context of unsaturated soil
resistance to the applied stresses. This is similar to the effect of mechanics opportunities. On the one hand, the importance of the
sedimentation on natural soft soils and results in the inclination progressive incorporation of reproducible laboratory techniques
of the yield curves indicating the anisotropy of the soil. and well-posed design criteria in codes of practice. On the other
To capture the anisotropic behaviour of unsaturated soils, Cui hand, progress in these new areas requires advanced
and Delage (1996) presented a model for compacted silt and experimental techniques with new technologies. Accordingly,
captured the effects of anisotropy by using an inclined ellipse as besides more traditional geotechnical and mining engineering
a yield function and a non-associated flow rule. Matric suction applications (e.g., Alonso et al. 1987, Fredlund & Rahardjo 1993,
was introduced as an independent variable, contributing to the Fredlund et al. 2012, Fredlund 2017, Tarantino & Di Donna
size of the yield surface. Stropeit et al. (2008) proposed an 2019, Houston 2019, Fredlund 2019), experimental unsaturated
anisotropic elasto-plastic constitutive model for unsaturated soils soil mechanics is of value in different cross-disciplinary topics,
by combining the features of BBM for unsaturated soils with which will be briefly described in the next section, focusing on
anisotropy. This model was further enhanced by D'Onza et al. the testing techniques.
(2010) and Al-Sharrad and Gallipoli (2016) through linking the In this section, we present laboratory techniques and
material anisotropy to both distortion and aspect ratio of the experimental contributions in unsaturated soil mechanics over
constant suction yield surface. More recently, Sutharsan et al. the last two decades relevant to studying the coupled hydro-
(2017) developed a critical state-based constitutive model for mechanical and multi-scale phenomena that the different ground
predicting the unsaturated response of sands. The model was an engineering and cross-disciplinary applications demand.
extension of the framework proposed by Heath et al. (2004). It
used the bounding surface plasticity theory and considered the 4.2 An insight Into Experimental Techniques and Applications
effect of fabric anisotropy. They introduced an anisotropic fabric
parameter enabling the model to capture the effect of sample 4.2.1 Ground engineering
preparation on sand response. More recently, Sitarenios and
Kavvadas (2020) proposed an elasto-plastic constitutive model A significant number of the published experimental research
for unsaturated anisotropic within the effective stress framework. concerns compacted soils, which are affected by changes in
The model was the extension of the modified Cam-clay model water content. Since Proctor’s pioneering work on these
and included the soil anisotropy effects via rotation of the yield materials (Proctor 1933), used in the construction of transport
surface. and hydraulic infrastructures (earth embankments/dams), a
constitutive and experimental framework has been required to
4 EXPERIMENTAL INVESTIGATION study the hydro-mechanical behaviour of these partially
saturated soils at different compaction states (Gens 1995,
4.1 Preamble Tarantino & De Col 2008, Kodikara 2012, Leroueil & Hight
2013, Tatsuoka 2015, Tatsuoka & Gomes Correia 2018). The
The difficulty in understanding the behavioural features of a wide traditional way to approach this hydro-mechanical behaviour has
range of materials found in geological environments, particularly been exploring different initial states in the Proctor plane (dry
under partially saturated conditions, has promoted the use of unit weight vs water content) and mapping the different
experiments and introduced a significant empirical load that has engineering properties to the compaction state, such as
been extended to theoretical and conceptual developments permeability and water retention properties, stiffness, shear
(Alonso 2005). In this regard, the progress of unsaturated soil strength, and collapse/swelling on wetting (e.g., Lawton et al.
mechanics has been based not only on close contact with 1989, Vanapalli et al. 1999, Santucci de Magistris & Tatsuoka
practical engineering applications but also on the development of 2004, Jotisankasa et al. 2009). Nevertheless, advances in
conceptual frameworks, theories, and models, which rely on experimental unsaturated soil mechanics have required
laboratory and in situ experiments. incorporating new constitutive law formulations and numerical
Nevertheless, experimental soil mechanics should not be models to characterise the initial ‘as compacted’ state (but not
solely considered a simple tool for soil model water content and maximum dry unit weight as in the
calibration/validation and parameter estimation. Instead, conventional Proctor plane) (Gens et al. 1995, Sivakumar &
laboratory tests should benefit from crucial advances in Wheeler 2000, Della Vecchia et al. 2013). On the one hand, water
equipment and new techniques for conducting experiments and contents should be transformed into matric suction (considering
making observations. In this regard, experimental soil mechanics void ratio effects) through the adequate branches of water
must reinvent itself and embrace other spaces, such as retention curves, which pose specific experimental difficulties
incorporating new experimental techniques and protocols, and display significant hysteresis during drying and wetting
running multi-scale tests (from microstructural to physical model (Tarantino & Tombolato 2005, Romero et al. 2011). On the other
tests), better controlling multi-physics processes and boundary hand, the attained dry unit weights by compaction should be
conditions, and improving the characterisation at lower structural related to the maximum compaction/preconsolidation stresses
levels (microstructural techniques). (yield stresses) that vary with suction (Alonso et al. 2013).
Furthermore, what does experimental soil mechanics bring to the Indeed, the microstructure set up during compaction is also
table when dealing with partially saturated soils? It can be argued essential information for the initial state, which shapes
311
engineering properties (Gens et al. 1995, Delage et al. 1996, mechanics compared to saturated soils with positive water
Thom et al. 2007, Monroy et al. 2010, Romero 2013, Alonso et pressure (Delage et al. 2008).
al. 2013). Microstructural experimental techniques will be The first classification of experimental techniques for
discussed below (section 4.3.4). controlling the amount of water in soils would be:
In addition, compacted/emplaced materials span a much wider Soaking and air-drying tests.
range of grain size distributions. In this respect, the compacted Constant water content tests.
soil concepts can also be extended to coarser materials such as Controlled degree of saturation experiments and
rockfill material, which is relevant to dams and embankments Controlled-suction tests, either transporting liquid water
(Oldecop & Alonso 2003, Romero et al. 2012a, Oldecop & (control of matric suction) or vapour (control of total
Alonso 2013, Alonso et al. 2016). The similarity in the collapse suction).
response on wetting these materials with contrasting particle The following will briefly outline these experiments. The
sizes and differentiated mechanisms also provides a marked soaking or fast inundation tests are suitable for studying the
multi-scale character to unsaturated soil mechanics. effects of sudden changes observed on rapid flooding, such as
Besides compacted soils, many geotechnical aspects of collapse/swelling (ASTM D4546-21, EN 1997-2:2007) and
partially saturated soils may be observed in swelling and residual shear strength loss on wetting (Nicotera 2000, Alonso et al.
soils in arid regions, which are often unsaturated and structured. 2016). Alternatively, partial soaking can also be carried out to
Geotechnical applications with significant ground-atmosphere achieve a degree of saturation of less than one. Other tests have
interactions can be found on: focused on degradation and swell/shrinkage accumulation on
Plastic clays susceptible to swell/shrink during hydraulic cyclic wetting and drying (Pineda et al. 2014a, Alonso et al.
processes affecting foundations/retaining walls, pavements 2005). All these tests should be complemented with data of water
and railway infrastructures and their remediation (Nelson et retention curves on the corresponding branches (wetting or
al. 2015, Vahedifard et al. 2015, Soltani et al. 2019). drying) and preferably dependent on porosity. As an example of
Residual tropical lateritic soils and saprolites (rock the use of this rapid flooding procedure, Figure 11 shows the
weathering), alluvial/colluvial/aeolian deposits and volcanic collapse response on soaking at different stresses and starting
agglomerates prone to volume change on wetting (Leroueil from hygroscopic conditions (relative humidity around 50%),
& Barbosa 2000, Benatti et al. 2011, El Mountassir et al. and encompassing different materials from silty/clayey soils to
2011, Kholghifard et al. 2014, Ng et al. 2020). coarse granular aggregates. These tests are straightforward to
Soils susceptible to crack during environmental desiccation study the volume changes on soaking at different mean particle
or affected by vegetation, such as flood protection sizes, as well as to discriminate the different phenomena
embankments (Atique et al. 2009). associated with these collapsible phenomena: softening of
Soils prone to salt accumulation under arid climate aggregates and reduction of inter-aggregate pore space in clayey
conditions (De Carteret et al. 2014). soils, softening of silty/clayey bridges between grains, particle
Materials prone to degradation on wetting/drying cycling rearrangements due to loss of suction-induced intergranular
(Fityus et al. 2005, Cardoso et al. 2012, Romero et al. 2014, forces in sands, and particle breakage sensitive to the state
Pineda et al. 2014a, Pineda et al. 2014b, Stirling et al. 2021). (activity) of the fluids and further skeleton rearrangement on
Soils affected by rainfall-induced landslides (Lim et al. 1996, coarser granular aggregates (Nara et al. 2012). In addition,
Alonso et al. 1999b, Alonso et al. 2003, Springman et al.
2003, Askarinejad et al. 2012, Sorbino & Nicotera 2013,
Song et al. 2016, Leong & Ng 2016, Jeong et al. 2017).
Partially saturated soils influenced by cryogenic suction
changes at different ice contents (frost heave and thaw
settlements, effects of cyclic freezing and thawing, soil
freezing water retention curve) (Pelaez et al. 2014, Caicedo
2017, Mao et al. 2018, Teng et al. 2019).
Coupled effects generated by soil-groundwater-biological
systems-atmosphere interactions (e.g., Cui et al. 2013, Tang
et al. 2018) that will be later introduced.
The effect of degree of saturation on electrokinetic
remediation of unsaturated soils (Gabrieli et al. 2008, Yin et
al. 2022).
Settlements of partially saturated loose soils susceptible to
densification caused by earthquake-induced cyclic stresses
(EN 1998-5:2018) and the dynamic response of partially
saturated soils (Pandya & Sachan 2018).
Rammed earth materials (unstabilised or stabilised with
cement/lime or fibrous materials) (Jaquin et al. 2009, Corbin
Figure 11. Collapse on soaking for different materials
& Augarde 2014, Beckett et al. 2018).
and their different mechanisms.
Capillary barrier systems to prevent water infiltration
(Rahardjo et al. 2012, Rahardjo 2015, Scarfone et al. another practical application of these fast-soaking tests is the
2020).Influence of entrapped air bubbles in quasi-saturated possibility of isolating microstructural effects induced on
soils (influence of the compressibility of the pore fluid, water compaction at a constant suction (the one achieved on
retention and permeability at nearly saturated conditions) saturation), since it can be assumed that the microstructure set on
(Sakaguchi et al. 2005, Jommi et al. 2019). compaction is preserved along this rapid process (Santucci de
As shown by these applications, the effects of matric (or total) Magistris & Tatsuoka 2004).
suction, degree of saturation and stress changes are crucial for The constant water content tests are appropriate to study the
correctly understanding the response of these partially saturated compressibility on loading/unloading and the shear strength
materials. Nevertheless, the difficulty of controlling and properties at specific hydraulic conditions (Tarantino & De Col
measuring matric (or total) suction has been one of the reasons 2008, Della Vecchia et al. 2011, Wijaya & Leong 2016). These
for the developmental delay in experimental unsaturated soil tests are performed on conventional laboratory equipment with
312
systems that prevent vapour loss under air-drained conditions. interactions with the mechanical response of the materials. For
Nevertheless, the preferred option is improving the cells and example, the increase in temperature generated by the waste can
measuring suction during these mechanical processes (Tarantino induce cracking due to desiccation in systems open to the
et al. 2011). In case it is not possible to measure the suction atmosphere affecting their functionality (Zornberg et al. 2003),
locally, the tests should be preferably carried out in the low water changes in swelling capabilities on wetting (Romero et al. 2003,
content range, in which water retention curves are not sensitive 2005a), and an increase in the water permeability (Romero et al.
to porosity variations (Romero et al. 2011). Typical (gravimetric) 2001).
water content ranges are usually below 20% in highly plastic Experimental mechanics has been valuable in studying
clays, below 14% in clays and 8% in clayey silts to avoid engineered barriers for radioactive waste disposal. These barrier
noticeable suction changes during compression and shearing materials are emplaced under partially saturated conditions and
(Romero 2013). In this case, the results should be complemented subjected to hydration. Bentonite-based materials have been
with water retention curve data on the corresponding branches extensively studied in this context, due to their low permeability
(wetting or drying) in this high-suction domain. In the light of and their swelling potential upon hydration, which allows closing
the development of techniques for suction measurement, which the engineering gaps. Therefore, studies have concentrated on
are described in section 4.3.2, these tests should progressively be water retention properties, water and air permeability under
included in standards for unsaturated soil testing relevant to partial saturation, compressibility on loading at different
ground properties and geotechnical design. For example, shear suctions, compressibility on suction changes, small-strain shear
strength envelopes with friction angle and total cohesion stiffness and strength, and temperature and chemical effects on
depending on suction and degree of saturation - or water content hydraulic and mechanical properties. Table 1 summarises
- at failure (Vanapalli et al. 1996). selected references for different bentonite types used in
Another useful alternative, although more challenging to engineered barriers.
implement, is the controlled degree of saturation test. These
experiments are suitable for studying the significant role of Table 1. Selected tests performed on different bentonites (engineered
saturation, besides suction, on the compression and shear barriers).
strength response. For example, test results can be helpful to MX-80 type: Swelling
study the evolution of soil compressibility as it approaches Villar 2005 Compressibility
asymptotically the saturated compression line, as they allow Tang & Cui 2005 Water retention curve
evaluating the collapse on soaking at different vertical stresses Karnland et al. 2007 Water permeability
and detecting the maximum collapse zone (collapse strains that Herbert et al. 2008 Temperature effects
increase with confining stress and then decrease). This last issue Tripathy et al. 2015 Chemical effects
has been discussed in models in which the plastic compressibility Molinero-Guerra et al. 2019 Shear stiffness
index depends on the degree of saturation (Zhou et al. 2012a,b, Pintado et al. 2019
Della Vecchia et al. 2013, Alonso et al. 2013). These loading Mesa-Alcantara et al. 2020
tests are performed by automatically controlling the changes of MX-80 type/sand: Water retention curve
pore-water volume and pore volume space , while Wang et al. 2013 Water permeability
monitoring the pore-water pressure beneath a porous interface Tabiatnejad et el. 2016 Microstructure
(Al-Badran 2011, Burton et al. 2016). These volume changes are Compressibility
related through , where is the Chemical effects
constant degree of saturation. Critical issues are matric suction FEBEX: Swelling
equalisation at a constant loading rate since water percolates Lloret et al. 2003 Compressibility
through the low-permeability interface (Burton et al. 2017) and Villar & Lloret 2004 Water retention curve
controlling the evaporation of water in the open-air pressure Villar 2005 Water permeability
system as described below. Castellanos et al. 2008 Temperature effects
Moreover, finally, controlled-suction tests with relative Romero et al. 2005a Chemical effects
humidity regulation (to control total suction) or control of GMZ: Swelling
air/water pressures using a porous interface permeable to the Ye et al. 2013 Water retention curve
wetting fluid to set the capillary mechanism of matric suction. Ye et al. 2014 Water permeability
These techniques are valuable for studying the gradual effects of Chen et al. 2015 Thermal effects
progressive wetting and drying stages and representing stress Chen et al. 2017 Chemical effects
paths for calibrating and validating constitutive models. They He et al. 2019 Microstructure
should be complemented with water content data (or degree of Czech B75: Water retention curve
saturation) throughout the different stages. These controlled- Sun et al. 2019 Thermal effects
suction techniques will be discussed in section 4.3.1. Sun et al. 2020 Microstructure
FoCa clay: Swelling
4.2.2 Environmental and energy geotechnics Imbert & Villar 2006
Experimental unsaturated soil mechanics is also involved in Kunigel V1 and Kunigel V1/sand: Swelling
many geoenvironmental and energy geotechnical applications, Komine 2004 Compressibility
encompassing engineered barriers, sealing materials for Romero et al. 2005b Shear strength
industrial waste dumps, capillary barriers, landfill seal covers to Yamamoto et al. 2019 Gas permeability
protect the environment, energy geo-structures, soil pollution
and remediation, and resilience to adapt to and mitigate climate The already mentioned aspects are also valid when implemented
change effects in geotechnical infrastructures. The common in cover systems for landfills and evapotranspirative ET covers
thread is to understand the coupled thermo-hydro-chemo- with a reduced desiccation and cracking response. The
mechanical processes, in which experimental unsaturated soil quantification and measurement of evapotranspiration, i.e. the
mechanics is essential to broaden the knowledge of these highly flow boundary condition at the soil-atmosphere interface, is of
coupled phenomena. Topics are linked to transporting fluids and interest to these ET cover systems (Cui & Zornberg 2008).
contaminants and water exchanges between the ground and the Experimental unsaturated soil mechanics is also relevant to
atmosphere. At the same time, other phenomena are related to geosynthetic clay liners GCLs. However, it encounters some
heat transfer and chemical effects that present significant experimental challenges, as explained in Bouazza et al. (2013).
313
GCLs and their particular geometric and structural configuration techniques for polluted soils (Alferi 2011). Furthermore, the
have required modifying conventional laboratory techniques for infiltration of DNAPL (Dense NAPLs with higher density than
measuring/controlling suction. Abuel-Naga & Bouazza (2010) water) in saturated soil also concerns capillary interactions
reported a complete review of the existing techniques and their between water and DNAPL (Delage & Romero 2008).
suitability while developing a modified triaxial apparatus that A three-fluid system (air, water and oil) should be considered
controls the water content and measures the suction. Different when dealing with contaminated partially saturated soils. It is
authors have also investigated GCL’s water retention capacity usually assumed that the more wetting fluid (water) fills the
using different techniques (some references are summarised in smallest pores, followed by oil and non-wetting air occupying the
Table 2) and the effects of water salinity (Lu et al. 2018). largest ones (Leverett 1941, Lenhard & Parker 1988). Therefore,
Energy geo-structures, such as energy piles, diaphragm walls cells to apply the axis translation technique should be modified
and tunnels, use the ground, sometimes under partially saturated by treating the high air-entry value ceramic discs, depending on
conditions, for heating/cooling structures and heat storage (e.g., whether water flow is allowed (hydrophilic treatment) or water
McCartney et al. 2016, Vitali et al. 2021). Therefore, the thermal transfer is prevented (hydrophobic treatment with Glassclad®18
properties and the temperature-induced changes in the hydro- used in oil removal at constant water content) (Cui et al. 2003,
mechanical response are key aspects to consider. In such a Manassero et al. 2005, Rabozzi et al. 2006, Alferi 2011). Alferi
context, a challenging issue is the effect of soil saturation on the et al. (2011) used a chilled-mirror psychrometer to determine
heat exchange rate in the long-term response of these systems total suction considering that no significant evaporation of
(Laloui et al. 2014). The degree of saturation influences the Soltrol®170 Isoparaffin Solvent (LNAFL) occurred (Manassero
thermal properties of the ground and, therefore, the heat et al. 2005).
exchange rate, which affects the efficiency of energy piles Another multi-species application is the methane hydrate-
(Akrouch et al. 2016) and can result in an average heat exchange bearing sediments (MHBSs). Natural MHBSs are highly
rate 40% lower than saturated conditions (Choi et al. 2011, susceptible to pressure and temperature changes (hydrate
Aljundi et al. 2020). This fact evidences the importance of dissociation) (Dai & Santamarina 2013, McCartney et al. 2016),
considering partially saturated conditions in selecting thermal and synthetic MHBSs are usually considered to study their
soil parameters for the design (Vieira et al. 2017). Centrifuge thermo-hydro-mechanical behaviour. In addition, the hydrate
tests involving partially saturated states on these thermal content alters the pore geometry, changing the capillary pressure-
applications (foundation’s movements during heating and saturation relationship from the hydrate-free condition
cooling) have been performed by Stewart & McCartney (2014) (Ghezzehei & Kneafsey 2010). Mahabadi et al. (2016) used
and Goode & McCartney (2015). synthetic tetrahydrofuran THF hydrates, which are stable under
atmospheric conditions and miscible in water, to study the
Table 2. Selected experimental techniques in geosynthetic clay liners. significant effects of non-wetting hydrate saturation on the water
Reference Experimental technique retention curves (gas entry pressure increase with increasing
hydrate saturation). The water retention curves were studied with
Southen & Rowe 2004 Pressure plate technique the axis translation technique, using a treated high air-entry value
Non-contact filter paper method, ceramic with a solution of ethylene glycol to avoid THF hydrate
Agus & Schanz 2005 psychrometer (dew point chilled mirror) formation.
technique, relative humidity sensor These different applications have posed significant challenges
Barroso et al. 2006 Filter paper to using experimental techniques for unsaturated soils. Among
Pressure plate (axis translation these developments, the following aspects can be highlighted.
Southen & Rowe 2007 The wetting stage of bentonite-based materials involves
technique)
Abuel-Naga & Thermocouple psychrometer, capacitive operating over a very high suction range (sometimes with an
Bouazza 2010 relative humidity sensor initial total suction of more than 200 MPa) towards
High capacity tensiometers, capacitive saturation. This extended range has entailed a combination
Beddoe et al. 2010
relative humidity sensors of vapour transfer techniques up to a total suction of about 4
Pressure plate, filter paper, relative MPa, followed by liquid transfer using matric suction control
Hanson et al. 2013
humidity (Hoffmann et al. 2005).
Osmotic technique with polyethylene This extended range in bentonites has also required a
Bannour et al. 2014
glycol, vapour equilibrium technique combination of different sensors to measure suction (relative
Rouf et al. 2014 Vapour equilibrium technique humidity probes and dew point psychrometers for total
suction, and high capacity tensiometers for matric suction)
Acikel et al. 2015 Contact filter paper
(Cardoso et al. 2007, Toll et al. 2013, Mendes et al. 2019).
Rouf et el. 2016 Vapour equilibrium technique The vapour transport must be efficient in bentonite-based
Dew point chilled mirror psychrometer materials, so forced convection techniques have been used
Seiphoori et al. 2016 technique, non-contact filter paper by pumping humid air through the material (in case of a high
method air permeability) or laminating along the boundaries of the
Acikel et al. 2018
Dew point chilled mirror psychrometer sample (Dueck 2004, Pintado et al. 2009a,b).
technique In the case of using controlled-suction techniques on
Modified osmotic technique with materials at high temperatures, care must be taken to avoid
Acikel et al. 2020
polyethylene glycol losing vapour through any open system (Romero et al. 2003).
To separate the effects of osmotic and matric suctions in
The determination of the water retention behaviour is even more clayey materials, such as increased pore liquid concentration
complex in multi-species systems, which is the case of soils during drying at controlled relative humidity (Mata et al.
contaminated with liquid organic compounds referred to as 2002) and the osmotic suction changes at constant matric
NAPLs (Non-Aqueous Phase Liquids). LNAPLs (Light NAPLs suction (Mokni et al. 2014).
with lower density than the water) with reduced solubility in the The high air-entry value ceramics used in the axis translation
water remain above the water table in the vadose zone. technique (three-fluid systems) should be treated to avoid
Therefore, experimental techniques on unsaturated soil certain phenomena, such as ice formation of THF or
mechanics can help better understand their flow properties, impeding water passage when controlling oil removal.
which are necessary for the proper performance of remediation
314
Unfortunately, these treatments induce alterations in the air-
entry properties of the ceramics (Alferi 2011).
4.2.3 Bioinspired technologies
Nature-based solutions and soil reinforcement techniques with a
low carbon footprint are attracting increasing interest (UN 2030
Agenda). Recent trends in research in partially saturated soils
involve the use of plants, fungi, and microbial/enzyme/urease
induced calcite precipitation to improve the hydro-mechanical
properties against erosion, landslides, soil collapse, or to favour
soil bearing capacity (Vardon 2014, Ng et al. 2016a, Ng et al.
2019, Tiwari et al. 2021, Assadi-Langroudi et al. 2022).
Biological systems (plants, fungi, bacteria) involved in these
solutions are constantly evolving in time (root growth and decay,
calcium precipitation and eventual dissolution) and space (root Figure 12. Coupled phenomena in soil-vegetation-atmosphere
development), thus continuously generating changes to soil interactions. 1) Phreatic level affected by soil-atmosphere
hydraulic properties (permeability, water retention, Morales et al. interactions (drying/wetting cycles). 2) Soil desiccation
(shrinkage/cracking and effects on infiltration and landslide
2015a, Jotisankasa & Sirirattanachat 2017, Ni et al. 2019a), to initiation). 3) Slope stability (soil cracking and vegetation effects
soil structure (soil fissuring, pores and rock fractures clogging, on soil hydro-mechanical behaviour). 4) Vegetation features
Carminati et al. 2013, Minto et al. 2016, Terzis & Laloui 2018, (root/soil ratios, depth, morpho-physiology) and vegetation
Koebernick et al. 2017), to mechanical behaviour (Yildiz et al. growth/decay (water retention curve, permeability, microstructure,
2018, Terzis & Laloui 2019, Fraccica et al. 2022a) and even to volume change and shear strength properties, carbon sequestration,
soil thermal properties (Venuleo et al. 2016, Oorthuis et al. mucilage chemical effect). 5) Vegetation-induced suction
2021). (‘capillary barrier effect, mechanical strengthening, effect on slope
Soil-vegetation-atmosphere interactions (root water uptake, stability). 6) Soil-vegetation-atmosphere interactions
evapotranspiration, solar radiation, wind, rainfall, runoff, water (evapotranspiration, leaf coverage). 7) Soil erosion and runoff
velocity.
infiltration) affect soil suction (Ng et al. 2016b, Ni et al. 2018),
leading to different coupled phenomena, as depicted, and briefly 2016). Therefore, different ways to formulate this variable have
explained in Figure 12. Laboratory investigations display a been used in literature, sometimes in an empirical way, and
crucial role in these novel applications to provide a according to the specific results observed (Fraccica et al. 2019,
comprehensive understanding of the different processes involved Ng et al. 2016c, Ni et al. 2019b). The choice of a given void ratio
and their future implementation in codes and regulations for formulation has direct consequences on the evaluation of other
nature-based solutions (e.g., Bo et al. 2015). However, variables such as the degree of saturation and the formalisation
experimentally studying partially saturated vegetated soils poses of constitutive laws describing soil water retention,
particular challenges to address. First, a correct laboratory compressibility on loading and suction changes, and volume
sample size must be selected to obtain a representative change upon shearing (Ng et al. 2016c, Ni et al. 2019b, Foresta
elementary volume REV. In this regard, an adequate soil/root et al. 2020, Karimzadeh et al. 2021). Techniques commonly used
ratio should be ensured in laboratory samples (Fraccica et al. in soil mechanics to infer the void ratio, such as paraffin wax
tests (ASTM D7263-21), should be complemented by
consistent with vegetative species under natural conditions to microstructural techniques to explore potential changes
better assess the soil-root hydro-mechanical behaviour. Fissures generated by biological systems on soil pore-size distribution
generated by biological systems and a sound volume of root (Carminati et al. 2013, Scholl et al. 2014, Fraccica et al. 2019,
architecture should be included (Li et al. 2009, Borges et al. Koebernick et al. 2017). Biological systems (roots and bacteria)
2018, Li & Shao 2020). Therefore, a large specimen size, which are often generating suction and hydraulic gradients within the
implies a more complex retrieval or preparation, is often required matrix due to biochemical and physical effects on soil (e.g., root-
to deal with the issues mentioned above. If the soil sample is induced fissures, surfactants released in the matrix, soil water
retrieved, a large and uncommon soil corer is required with consumption by bacteria: Read et al. 2003, Carminati et al. 2010,
consequences on the alteration of the sample. In situ testing Leung et al. 2015, Pham et al. 2016, Liu et al. 2020). These
seems an alternative solution to test large-scale vegetated spatially heterogeneous and time-evolving modifications of the
samples (Comino & Druetta 2010, Cammeraat et al. 2005). hydraulic state should be considered when evaluating soil water
However, attention has to be paid to minimum displacement rates retention properties of bio-remediated soils. Indeed, different
allowed (usually too high), volume change assessment arrangements and response times of the measuring tools within
instrumentation and un-controlled/inhomogeneous hydraulic the soil could lead to discordant/highly scattered observations for
boundary conditions. Suppose the sample is prepared/compacted a given hydraulic state. Therefore, it is essential to carry out
in the laboratory (Pallewattha et al. 2019, Fraccica 2019, Yildiz laboratory experiments at steady-state conditions, ensuring a
et al. 2018). In that case, attention should be paid to ensure water correct equalisation of suction. Given these considerations,
content/dry density homogeneity, facilitate root growth by vegetated soil water retention measurements under darkness or
leaving an adequate pore-space, and control the hydraulic path sun/artificial light conditions have been proposed in the literature
prior to testing (plant irrigation and root water extraction). This to minimise such biological activities (Table 3 summarises
last aspect is fundamental to ensure repeatability, as an excessive different protocols and measuring tools). Matric suction control
matric suction (above that induced on compaction) generated by by axis translation has been used in microbial induced calcium
roots may induce irreversible shrinkage together with soil carbonate precipitation MICP treated soils (Morales et al.
fissuring. A second open issue is the appropriate definition of 2015b). In contrast, it is still unexplored in vegetated soils, as
void ratio and degree of saturation in vegetated soils. The possible damages caused by high air or water pressures on roots
different ways to compute void ratios depend on the stress state, physico-mechanical traits are unknown. Despite providing
either by clogging/stealing void volume at high stresses or reliable total suction values above the permanent wilting point of
opening new fissures at low-stress conditions. They include roots (1.5 MPa), vapour equilibrium technique or psychrometers
attributing the roots to the solid phase, to the gas phase or have been so far rarely used (Fraccica et al. 2019, Morales et al.
identifying them as a fourth (external) phase (Muir Wood et al.
315
Table 3. Unsaturated hydraulic behaviour in vegetated and MICP Table 4. Laboratory tests on vegetated and MICP improved soils.
4
improved soils. hydraulic variable
Reference Laboratory test Sample size
Matric (max. (mm) w s Sr
Other
value allowed Fraccica et al. 2Monotonic d: 200,
Properties Water indirect KM M E
Reference in MPa) or 2022b triaxial comp. h: 400
evaluated content evaluations
total suction 2
Fraccica et al. Uniaxial 2×d: 100
of WRC KM M E
2022a extension h: 50
Vegetated soil 2
Monotonic d: 200
3 Fraccica 2019 KM M E
Ceramic tip triaxial comp. h: 400
2 tensiometer Mahannopkul
Fraccica et Oven- Pore size d: 140
WRC (0.20) & Jotisankasa 2Direct shear KM M -
al. 2019 check 4 distribution h: 150
Dew-point 2019
psychrometer Tan et al. 2 d: 61.8
Direct shear KM - -
WRC and 3 2019 h: 20.0
1 TDR Ceramic tip
Ni et al. Relative Pallewattha et 2
moisture tensiometer - Direct shear 300×300×200 M M E
2019a hydraulic al. 2019
probe (0.10)
conductivity Yildiz et al. 2
Inclinable
WRC 500×500×300 M M E
2018 direct shear
model 2
et al. Multi-stage
depending 250×250×110 M M E
TDR 3
Ceramic tip 2018 direct shear
1
Ni et al. on Void Gonzalez-
WRC moisture tensiometer
2019b ratio and Ollauri & d: 49.0
probe (0.10) 2
Direct shear M M -
root Mickovski h: 23.3
volume 2017
and decay 2
Monotonic d: 61.8
Jotisankasa & WRC and relative 3
Ceramic tip Li et al. 2017 triaxial comp. h: 125.0 M - -
Mass
Sirirattanachat hydraulic tensiometer - Veylon et al. 2
check Direct shear 500×500×300 M E E
2017 ** conductivity (0.10) 2015
WRC / suction Back- Zhang et al. 2Monotonic d: 39.1 KM - -
profile /water analysis Numerical 2010 triaxial comp. h: 80.0
3
1,2 balance during of Ceramic tip modelling Comino &
Leung et al. 1,2
Direct shear 300×300×100 M - -
simulated rainfall infiltrati tensiometer of water Druetta 2010
2015
and on (0.08) and vapour Cammeraat et 1,2 2
60×60×20
evapotranspiratio columns flows Direct shear 1 - - -
al. 2005 600×600×400
n * 3
Unconfined
Water balance / Tiwari et al.
TDR 3
Ceramic tip 2021 comp., split h/d ratio = 2 M - *
2
Ng et al. suction profiles tensile strength
moisture tensiometer -
2014 during simulated Mujah et al. 3
Unconfined
probe (0.10) h/d ratio = 2 - - -
rainfall 2019 comp.
MICP improved soil 3
Vail et al. Desiccation d: 50
WRC M - -
2019 tests h: 6.4
model 3
3
Ceramic tip depending Mahawish et Unconfined d: 100
Chen et al. WRC at different Mass - - -
tensiometer on void al. 2018 comp. h: 200
2021 calcite content check 3
Oedometer,
(0.10) ration and Cardoso et al. Brazilian d: 70
calcite - - *
2018 splitting h: 20
content
WRC after Bahmani et al. 3Unconfined d: 60
Mass 4
Dew-point - - E
Liu et al. 2020 different MICP - 2017 comp. h: 120
check psychrometer Cheng et al. 3Unconfined h/d ratio =
treatment cycles
M - -
WRC at different 4 2017 comp. 1.5÷2.0
Saffari et al. Mass Filter paper
bacterial - Terzis et al. d: 50
2020 check (10) 3
Triaxial comp. M - -
concentrations 2016 h: 100
WRC at different 3
3
Direct shear, (soaking
Saffari et al. Mass Matric paper Morales et al. oedometer, d: 60
bacterial - M M during
2019 check (10) 2015a h: 20
concentrations resonant column test)
3
3
Hanging- Unconfined
Martinez et al. Oven- Morales et al. comp., split d: 38 h: 76
WRC water-column - M M E
2018 check 2015b tensile strength, d: 50 h: 20
(0.004)
4 bender elements
Dew-point 3
Unconfined d: 50
psychrometer Li 2015 comp. h: 100 M - *
Morales et al. Mass 3 Pore size
WRC Axis- d: 100
2015b check distribution
translation Al Qabany & 3Unconfined h: 250
Soga 2013 comp. d: 35.4 - *
technique
1
under light conditions, 2 under darkness conditions h: 100
3
3 Unconfined
matric suction, 4 total suction, WRC: water retention curve Cheng et al. comp., triaxial h/d ratio = M - E
*according to Ng & Leung 2012 2013 comp. 1.5÷2.0
** plant leaves trimmed to minimise transpiration 1
In-situ mechanical investigations, 2 vegetated soils, 3 MICP improved
soils,
4
w: gravimetric water content, s: suction, Sr: degree of saturation, K:
constant water content, M: measurement of water content or suction, E:
determined
*controlled temperature and relative humidity, oven-drying or air-
drying during curing and before testing
316
2015b). Attention must also be paid to dielectric water potential strength and water content/reactive binder ratio of the injected
sensors as they are affected by the organic content of soil (Ni et suspension (Consoli et al. 2016). More recently, direct
al. 2019b), which inevitably increases during vegetation or implications on soil collapsibility have also been observed
bacteria’s activity. Although relative permeability of vegetated (Seiphoori & Zamanian 2022). Although many grouting
soil can be evaluated by indirect methods (Ng & Leung 2012), it applications are above the groundwater level (Le Kouby et al.
should be complemented by in-vivo imaging techniques (i.e. X- 2018), the interpretation of laboratory and in situ tests still relies
ray microtomography, neutron radiography and/or infrared on the assumption of full saturation of the ground. For this
images: Anselmucci et al. 2021, Fraccica 2019, Carminati et al. reason, experimental studies usually do not go beyond the
2010, Parera et al. 2020) to detect preferential paths or suction- estimation of the saturated permeability and the unconfined
induced capillary barriers generated by the roots. compressive strength. This last property is anyhow affected by
In the literature associated with traditional geotechnical suction (Hossain & Yin 2012), usually not considered, generated
applications with bio-treated soils, mechanical tests at controlled by the curing environment or by the chemical process of binder
water content or suction, such as unconfined and triaxial solidification under water depletion conditions. Despite the
compressions and direct shear tests, have been widely used. discussion above, the influence of grouting techniques on soils’
Nevertheless, climate change effects (mainly droughts) will final water retention behaviour (clogging of pores, osmotic
require focusing on shrinkage and cracking phenomena processes, alteration of water absorption) is still poorly explored.
(Lakshmikantha et al. 2012, Ledesma 2016), in which rooted Table 5 summarises selected references in partially saturated
soils are particularly vulnerable due to their high heterogeneity grouted soils.
and interface phenomena with concurrent drying of soil and In addition to grouting, there has been work on chemical
roots. Within this context, direct tensile tests, such as those hydrophobic treatment to soils (e.g., Saulick & Lourenço 2020,
reported by Trabelsi et al. 2018, or wetting/drying paths are still Zhou et al. 2021), where there are theoretical and experimental
poorly investigated in bio-remediated soils. Furthermore, challenges to address at the hydrophobic/hydrophilic interfaces.
interpreting direct tensile tests of rooted soils in a partially
saturated framework is challenging. Therefore, advanced Table 5. Mechanical tests in grouted soils with consideration of variables
techniques should be able to infer principal directions of linked to the soil’s hydraulic state.
deformations (e.g., Particle Image Velocimetry), and tensile 1
variable during curing/testing
strength should be better assessed following a Mohr-Coulomb Reference Laboratory test
criterion with information of the total stress, degree of saturation w s Sr RH
and suction (e.g., Murray et al. 2019, Murray & Tarantino 2019,
Seiphoori & Oedometer test with
Fraccica et al. 2022a). Table 4 summarises selected mechanical
Zamanian inundation KM - E M
studies, focusing on controlling or estimating the samples’ 2022
hydraulic state variables and on the size adequateness of the
specimens. Strongly coupled phenomena occur during Wang et al. Grouting, direct shear at
KM E E C
mechanical tests at different hydraulic states. For example, 2021 different Sr values
matric suction not only enhances soil shear strength but also Fraccica et al. Grouting, Unconfined
affects soil-roots bonding (Fraccica et al. 2022a) and soil-calcite M - E C
2021 compression
contacts (Cheng et al. 2013). Therefore, a more accurate
Hossain & Yin Soil-cement interface direct
measurement/evaluation of suction should be considered in these M C E -
2012 shear at controlled suction
investigations. In MICP improved soils, important suction values
generated by biological activities and curing at given relative Horpibulsuk et
Soil compaction and
humidity are still poorly considered in the interpretation of al. 2010
unconfined compressions at KM - - C
unconfined compression tests (refer to Table 4). constant water contents
Porcino et al. Unconfined compression
4.2.4 Soil chemical improvement - - - C
2011
Several grouting techniques based on the penetration of grout 1
w: gravimetric water content, s: suction, Sr: degree of saturation, RH: relative
into the soil can be used: compaction, jet/mixing, fracture, and humidity, K: constant water content, M: measurement of water content or suction, E:
permeation. Compaction and soil mixing have been extensively determined, C: controlled
studied (e.g., Gallagher & Mitchell 2002, Wong et al. 2018),
while jet/mixing and fracture are hard to implement in laboratory 4.2.5 Tailings and industrial processes
investigations due to the scale involved and the inhomogeneity Wet slurry tailings are conventionally stored above-ground
generated in soil due to excessive binder pressure (fingering, behind earthen dams and are affected by interactions with the
piston effects). Permeation grouting involves imposing low atmosphere. This kind of material can therefore be studied with
pressures on the material to be injected, usually a fluid, not to experimental techniques for unsaturated soils (Fredlund et al.
exceed the soil’s total stress. The process by which non-wetting 2003, Oldecop et al. 2011). In low precipitation areas, these
solutions or resin-type binders are injected and have to impinge materials undergo drying processes, which involve strongly
on other wetting/non-wetting fluids (water/air) present in the soil coupled hydro-chemo-mechanical interactions. Shrinkage of the
can be studied by unsaturated soil’s rheological laws (Wang et surface layer may also initiate cracking due to movement
al. 2021). Interface phenomena such as surface tension, viscosity, restrictions or soil’s heterogeneity above the tensile strength of
and grout pressures should be considered, as well as the relative the soil that depends on suction and the degree of saturation
permeability and the actual stress-state of the soil matrix when (Ledesma 2016, Trabelsi et al. 2018, Tollenaar et al. 2018).
In mining materials subject to drying, desiccation and cracking
optimising injection processes and testing new grouting
may also introduce significant environmental impacts
products. Injection of partially saturated soils with more aqueous
(Rodríguez et al. 2007). Cracking increases the permeability of
suspensions may decrease suction with potential soil volume the tailings and mining waste ponds, forming preferential paths
changes. Imposing high total suctions on curing at relatively low for pollutant generation and infiltration to deeper layers and
relative humidity can damage the solidified binder’s structure yielding oxygen accessibility for oxidation of the metallurgical
(Axelsson 2006, Spagnoli et al. 2021, Fraccica et al. 2021). wastes (Blanco et al. 2013). The long-term stability of tailing
Empirical relationships exist between improved grouted soil dams is also governed by the water content in the shallow
317
unsaturated zone, in which the effects of dissolved salts are also selected surfaces employed in material conveying designs (Fang
significant. Air-drying experiments on metallurgical tailing et al. 2020).
sludges at controlled environmental conditions have been Hygroscopicity of granular materials, among other grain-scale
recently reported by Garino et al. (2021) using a fully effects (Torres-Serra et al. 2021b), affects their flowability and,
instrumented column experiment. Their results evidenced the thus, the performance of conveying processes. Column collapse
formation of a dry tailing crust enclosing a body of mud- experiments have been increasingly used to study a wide range
consistency material. In addition, the authors discussed the role of granular materials with industrial applications (Torres-Serra et
played by the increasing concentrations of dissolved/precipitated al. 2020, 2021a). The effect of particle shape and the morphology
salts in this crust that limit the evaporation rate by increasing of capillary water on granular stability was also studied on wet
osmotic suction. Fine-grained tailings have been studied by sand (Scheel et al. 2008). Particularly, the shear strength of
Wickland et al. (2006), demonstrating that they can keep unsaturated sands in the pendular state (hygroscopic conditions)
saturated conditions for more extended periods thanks to their was characterised by Richefeu et al. (2006), and their increasing
high air-entry value. Oldecop et al. (2011) also observed high in tensile strength with water content was studied by Kim & Sture
situ water contents of old abandoned tailings even for low rainfall (2008) and Lu et al. (2007).
and high evaporation rates. Despite these high saturations, sl ope
4.3 Experimental Techniques in Unsaturated soil Mechanics
Pure water Soil water Soil + soil water
(salinity effects) (curvature of interface +
u w0 = u a0 u w1 = u a1 adsorbed water) 4.3.1 Controlled-suction techniques
This section addresses the research effort during the last two
u w0 u da u v0 u w1 = u da u v1 u a = u da +u v
decades regarding the development of controlled-suction
techniques and the advances in measuring suction (matric,
osmotic, and total). Several seminal state-of-the-art papers have
described these techniques, such as Ridley & Wray (1996),
Delage (2004), Agus & Schanz (2005), Marinho et al. (2008),
Bulut & Leong (2008), Vanapalli et al. (2008), Blatz et al.
flat interface flat interface
(2008), Delage et al. (2008) to cite some of them.
The classification regarding controlled suction techniques can
be based on how water is transferred. For example, water
transport can involve slow vapour phase transfer by regulating
REFERENCE u v1 uv the relative humidity of the air in contact with soil
OSMOTIC: aw MATRIC:
STATE u v0 u v1 , where is the vapour pressure interacting with soil
and the saturation vapour pressure over a flat surface of
w RT u w RT u
pure water at the same temperature. A relationship is therefore
ln v ln a w v
Mw u v0 Mw u v1 required to transform the relative humidity into the different
w RT w RT u components of su ).
ln a w ; s ln v Figure 13 shows a simplified schematic of the suction
Mw Mw u v1
components. The illustration in the middle indicates the osmotic
s
component associated with the activity of the pore water,
, where is the vapour pressure over a flat surface
Figure 13. Suction components (salinity effects and curvature of the of soil water that depends on the ion concentration at the far-
meniscus/adsorbed water). field. The illustration on the right presents the matric component,
, related to the curvature of the meniscus (fluid pressures)
stability problems were not observed even with steep external and the adsorbed water (physico-chemical interactions) (Lu &
slopes (40º-45º) attributed to high matric suctions and Zhang 2019).
cementation effects of precipitated salts. The figure presents the thermodynamic relationship between
Filtered partially sat urated tailings have been studied by the different suction components considering the equality of the
Oldecop & Rodari (2021), who present them as efficiently chemical potential (molar Gibbs function) of water species in the
yielding an unsaturated state, which is preferred as it allows two phases, vapour an ideal gas, water incompressibility and
avoiding the construction of containing dams, it minimises isothermal evolution (Castellan 1983). The psychrometric law
seepage and prevents liquefaction phenomena. Significant long- (first expression at the bottom of the schematics) relates
term effects in filtered tailings include air-drying phenomena to where kg/kmol is the molecular
upon tailing discharge and tailing compaction under the weight mass of water, the molar gas constant, the absolute
of subsequent stack lifts. temperature, and the liquid water density..
Experimental unsaturated soil mechanics is also relevant to From the perspective of experimentally controlling the
powder and bulk solids technology and industrial processes matric suction with measurable variables, only the meniscus
(Mitarai & Nori 2006). The rheological properties of clay curvature (capillarity) is usually considered. Therefore, another
materials for ceramic processing are affected by their water method of handling the matric suction may be via the liquid
content and particle size distribution and the presence of salts, phase transport by controlling the air and water pressures, and
sediments, and expandable minerals (Dondi et al. 2008). , respectively. In this case, matric suction is regulated by the
Moreover, moisture content significantly affects the bulk pressure difference between non-wetting and wetting phases,
handling of the fine fraction of mining ore materials, affecting , acting on both sides of the curved meniscus
their cohesive/adhesive behaviour and handling properties interface. Air behaves as a mixture of several gases (dry air), and
(Cabrejos 2017, Li et al. 2019). The effect on discharge varying amounts of water vapour: , where
trajectories in material conveying lines was investigated using a is the dry air pressure and the vapour pressure.
laboratory-scale model of a belt for various materials (coal, Two different procedures of liquid phase transfer can be used
gravel, magnetite, riversand) at controlled water contents (Ilic & to control the water pressure at constant air pressure. The first
Wheeler 2017). Moreover, wall adhesion tensile force tests based one uses a chemically based control (osmotic technique),
on standard wall friction tests were used to model the contact whereas the second procedure of controlling matric suction
mechanisms between unsaturated bulk iron ore materials and (capillary pressure) uses a hydraulic control with a hanging-
318
water-column method for very low matric suctions (at Most calibrations relating PEG concentration and applied matric
(Berliner et al. 1980, Vanapalli et al. 2008) or axis translation suction have been experimentally determined. A first intuitive
technique for higher matric suctions (at ) (Delage et al. approach is to measure the relative humidity (equivalent to )
2008, Romero et al. 2012b). The axis translation technique will above solutions of PEG using psychrometers (Zur 1966,
be later discussed. Williams & Shaykewich 1969) with calibration curves
independent of PEG’s molecular mass. Dineen & Burland (1995)
and Tang et al. (2010) used a high capacity tensiometer directly
on a sample/thin layer controlled by the osmotic technique
through a selective membrane. However, an alternative approach
has been recently proposed by Lieske et al. (2020) based on a
modified Flory-Huggins thermodynamic polymer solution
equation by knowing the average molar mass, the concentration,
and the PEG-solvent interaction. Different selective membranes
with different molecular weight cut-offs have been used, such as
cellulose acetate (Tripathy et al. 2011) and more resistant
polyethersulfone membranes (Slatter et al. 2000, Yuan et al.
2017). An advantage of the osmotic technique is that there is no
need to apply any air pressure. In addition, high matric suction
can be applied by using high concentration PEG solutions
(maximum matric suction of 8.5 MPa has been used by Cuisinier
& Masrouri 2005). Some of the problems related to this
technique can be summarised as follows:
PEG solution and mainly cellulosic membranes are
susceptible to degradation during long-duration tests.
Membrane effects tend to change the applied matric suction
(concentration polarisation close to the membrane,
membrane fouling and reverse solute draw; Vandoorne et al.
2022).
Calibrations for long-term water volume changes,
particularly with the stability of the weighing system (creep
of the transducers of the electronic balance, vibrations of the
lines connected to the peristaltic pump) and problems
associated with water evaporation and circulation losses
(Abbas et al. 2017).
Axis translation is a widely used technique for regulating
matric suction, which stems from the pressure plate apparatus for
pore water pressure measurement (Gardner 1956, Hilf 1956).
However, since it controls non-wetting and wetting pressures, it
can only regulate the matric suction’s capillary pressure
(meniscus curvature) and not the adsorption mechanism (Lu
2019). Figure 14 1at the top presents a schematic of its
application in an isotropic cell. It involves the increase of the air
pressure in which the soil is immersed , and the
consequent increase of the pore water pressure
, assuming that the curvature of the menisci is not
Figure 14. At the top: application of axis translation technique significantly affected (Hilf 1956, Bocking & Fredlund 1980,
(isotropic stress conditions). The pore size density function of a Delage et al. 2008). In particular, this increase over atmospheric
commercial high air-entry ceramic disk (using mercury intrusion conditions has been criticised for not representing field
porosimetry) is at the middle of the figure. The water retention conditions and strongly affecting (or suppressing) the water
curves of a coarse porous disk, high air-entry ceramic disk and soil
cavitation for capillary water (Baker & Frydman 2009, Lu 2019).
(operative range of axis translation) is presented at the bottom of
the figure. A framework for soil water cavitation has been recently proposed
and experimentally validated using axis translation and
As previously indicated, the osmotic technique uses a hygrometer-based methods by Luo et al. (2021). In addition,
chemically based control at taking advantage of the there are some doubts about how the air pressurisation process
properties of a pumped osmotic solution with high-molecular- affects water held by adsorption mechanisms. A saturated high
mass polyethylene glycol PEG molecules (with activity ) in air-entry value HAEV interface in contact with the soil allows
contact with the soil and separated by a selective semi-permeable independent air and water pressure control (Figure 14 at the
membrane, which allows solvent molecules and ionic species in middle shows the pore size distribution by mercury intrusion
the aqueous soil solution to pass. Description of the technique porosimetry of a commercial HAEV ceramic disk of 1.5 MPa
can be found in Blatz et al. (2008), Delage et al. (2008), Delage bubbling pressure). These interfaces, which limit air transport
& Cui (2008) and Vandoorne et al. (2022). At equilibrium, a pore and are permeable to ionic species in the aqueous soil solution,
pressure deficiency (osmotic pressure), which depends on the are usually made of a) microporous membranes (polyether
PEG concentration (higher matric suctions at increasing sulfone and acrylic copolymer) with air-entry values up to 250
concentrations), is induced on the pore water (soil water with kPa and prepared for improving the time required to reach
activity ) that prevents its flow across the semi- equalisation (Nishimura 2013); b) ceramic-based disks
permeable membrane. (dominant pore sizes typically between 81 and 220 nm and air-
entry values between 0.7 and 1.6 MPa, Romero et al. 2012b); c)
(24) cellulose acetate membranes (MWCO 3500 with largest pore
size about 8 nm, Tripathy et al. 2011); and d) nanoporous glass
319
interfaces (maximum pore sizes about 7 nm and maximum air- (Delage et al. 1998, Tang & Cui 2005, Delage et al. 2008). The
entry value around 7.3 MPa, Mendes et al. 2019). As depicted in main drawback of this experimental technique is that the time to
Figure 14 at the bottom, the soil should preferably present a reach moisture equalisation is exceptionally long because vapour
lower air-entry value than the corresponding one of the ceramic transfer depends on diffusion. Vapour transfer –through the
disk. In addition, to ensure that the top boundary presents no sample or along the boundaries of the sample– can also be forced
water flow (or storage) condition, the top coarse porous disk by a convection circuit driven by an air pump to speed up the
should display a very low air-entry value (this ensures that no process (Blatz & Graham 2000, Lloret et al. 2003, Alonso et al.
water is stored when applying matric suction). 2005, Pineda et al. 2014c). The mass rate transfer of vapour by
A benchmark aimed at comparing different experimental convection (assuming isothermal conditions and constant dry air
techniques for controlling/measuring suction (axis translation, pressure) can be expressed in terms of vapour density or mixing
osmotic technique, high capacity tensiometer and dew-point ratio differences between the reference vessel and the soil
psychrometer) on a mixture of kaolinite, sodium bentonite and (Oldecop & Alonso 2004, Jotisankasa et al. 2007, Romero et al.
sand (reference soil) was presented by Tarantino et al. (2011). 2012b).
Different laboratories tested the two matric suction application One of the difficulties in using the vapour equilibrium
techniques, and similar results were obtained in the range technique is to maintain thermal equilibrium between the
between 20 kPa and 1000 kPa, which gave further confidence in reference system and the sample. A way to minimise this thermal
using the axis translation technique. Nevertheless, a specific effect is achieved by disconnecting the reference system that
discrepancy was observed in the medium range of matric suction regulates RH, and by allowing the equalisation of vapour in the
(between 200 kPa and 400 kPa) between axis translation and a remaining circuit and the soil. Another problem when using the
high capacity tensiometer (for the same water content, the forced convection system is the air pressure differences created
suction measured by the tensiometer was lower). This along the circuit. This phenomenon makes the reference vessel’s
discrepancy has been explained by Marinho et al. (2008), RH not be adequately controlled in the remaining circuit and the
considering an increase in the meniscus curvature, and thus the soil (Pintado et al. 2009a,b). These authors observed that forcing
matric suction, due to compression of entrapped air during water humid air reduced the equalisation time, but the results
pressurisation. Discrepancies between water retention data by highlighted that this flow must be carefully applied to avoid
axis translation and dew point psychrometer for a silty soil have reaching a different total suction than regulated.
also been reported by Bittelli & Flury (2009) for matric suctions A significant issue is to estimate the time required for
> 100 kPa, which were interpreted in terms of the suppression of equilibrating the total suction homogeneously. The conventional
cavitation (Lu 2019). Nevertheless, Boso et al. (2003) observed experimental criterion to define the hydraulic equalisation period
an excellent overlap between high capacity tensiometers is based on measuring volumetric strains and controlling strain
(equilibrated and dynamic curves) and axis translation technique rates below a specified value (typically below 0.1%/day, as
while drying clayey silt. In addition, Hoffmann et al. (2005) suggested by Merchán et al. (2011) based on hydro-mechanical
presented a good continuity between vapour equilibrium and numerical simulations), due to the experimental difficulty in
transistor psychrometer results at elevated total suctions and knowing the precise hydraulic status of the sample during vapour
nearly saturated states controlled with axis translation technique transfer (usually the evolution of vapour transferred to or from
on compacted Febex bentonite. Axis translation has been the sample is not measured).
successfully used by Mokni et al. (2014) to transfer liquids at
different concentrations on a compacted clay and study osmotic 4.3.2 Suction measurement techniques
suction effects on volume change behaviour and shear strength This section will focus on the development of high capacity
at different constant values of matric suction. matric suction tensiometers and the most used psychrometers.
Experimental difficulties concerning its implementation are Significant progress has been made in developing high
associated with the following aspects. capacity tensiometers with fast response time since the probes
The accumulation of diffused air beneath the saturated used by Ridley & Burland (1993) and Tarantino & Mongiovì
HAEV ceramic disk can induce the progressive loss of water (2003). These probes share the same configuration comprising a
continuity, water volume change errors in drained tests and tiny water reservoir, a sufficiently thick high air-entry interface
pore-water pressure measurement errors in undrained tests and a pressure transducer. Therefore, the improvements have
(Airò Farulla & Ferrari 2005, Lawrence et al. 2005, Padilla been focused on using different interfaces, pressure transducers,
et al. 2006). Increasing the water pressure is an efficient way water reservoir volume, and interface saturation. Table 6
to reduce air diffusion rates for specified matric suctions. summarises the improvements made. Within this context, correct
The control of the relative humidity of the air chamber protocols should be followed to pressurise the chamber and
(around 95%) is required to minimise evaporative fluxes ensure the saturation of the high air-entry interface with de-aired
between the soil surface and the air chamber (Romero et al. water (Mendes & Buzzi 2014). In addition, a soil paste/filter
2012b). should be inserted between the stiff high air-entry interface and
Applying air pressure at high degrees of saturation can the soil to improve contact and delay evaporation of the interface.
induce irreversible arrangements in the soil skeleton due to Different prototypes have been recently compared for long-term
pore fluid compression (entrapped air bubbles) (Bocking & matric suction measurements by Mendes & Gallipoli (2020)
Fredlund 1980). If nearly saturated states are expected to be regarding interfaces, pressure transducers, water reservoirs and
reached during the hydraulic paths, it is preferable to increase protective casings. Probes incorporating a small water reservoir
air pressure when the continuity of air is ensured, and then showed a more remarkable ability to sustain suction over a long
maintain the continuous air phase at constant pressure. period without cavitating. Mendes et al. 2019 used a nanoporous
One crucial difficulty when using the axis translation glass interface (96% SiO2 with maximum pore sizes about 7 nm),
technique is estimating the time required to reach matric which allowed reaching matric suctions of 7.3 MPa.
suction equalisation (Oliveira & Marinho 2008, Romero et Nevertheless, despite these different studies, the phenomenon
al. 2012b). of air diffusion through the saturated interface, driven by the
The vapour equilibrium technique, associated with the total gradient of air concentration, remains an open issue when
suction, is implemented by controlling the relative humidity RH considering the long-term stability of these probes and the
of a closed system where the soil is immersed (Figure 13). The possibility of the loss of water continuity. Air diffusion depends
RH of the reference system can be controlled by varying the on the thickness of the interface and the pore water pressure in
chemical potential of different types of aqueous solutions
320
the chamber with an external atmospheric air (Airò Farulla & different. The soil inside the chilled-mirror chamber undergoes
Ferrari 2005, Romero et al. 2012b). some drying before reaching equalisation, and it will follow a
A great effort has also been dedicated to extending main drying path during the measuring period.
psychrometers’ working range. High-range alternatives to the
widely used Peltier thermocouple psychrometry (Campbell 4.3.3 Hydro-mechanical cells and mock-up tests
1979, Fredlund & Rahardjo 1993) were developed in the 90s and Standard cells for saturated soils have been modified to test
2000s: a) transistor psychrometers (Soil Mechanics unsaturated soils using different techniques to determine soil and
Instrumentation SMI type: Woodburn & Lucas 1995) with an water volume changes and apply or measure suction. The initial
upper limit of 70 MPa and long-term measuring time (1 hour), developments of these experimental systems are explained in
and b) chilled-mirror dew-point psychrometers with an upper detail in Delage (2004) and Delage et al. (2008). Table 7
limit of 300 MPa and involving a reduced time of reading (3 to summarises selected developments in different hydro-
10 min). In particular, the latter device has been improved and mechanical cells, in which diverse techniques for controlling
widely used in recent years and has become a reference system total and matric suctions were implemented (vapour equilibrium
to determine total suction (Leong et al. 2003, Cardoso et al. 2007, technique VET, axis translation technique ATT, hanging-water-
column HWC method at atmospheric air pressure, osmotic
Ebrahimi-Birnag & Fredlund 2016). The chilled-mirror dew-
technique OMT, and constant water content CWC tests with
point psychrometer measures the temperature at which
suction measurement).
condensation first appears (dew-point temperature). Then, in
Mock-up tests are usually designed to allow larger-scale
equilibrium with the surrounding air, a soil sample is placed in a investigations of coupled processes before or simultaneously
housing chamber containing a mirror and a photoelectric detector with in situ tests to complement data and demonstrate their
of condensation on the mirror. A thermoelectric (Peltier) cooler feasibility under better-controlled boundary conditions while
precisely controls the mirror’s temperature. The relative validating numerical models to predict the real scenarios better.
humidity is computed from the difference between the dew-point In the field of radioactive waste disposal, several mock-up
temperature and the temperature of the soil sample. experiments have been developed to study the chemo-thermo-
hydro-mechanical behaviour of bentonite-based materials. These
Table 6. High capacity tensiometers (updated information based on materials’ initial phase of saturation generates the most relevant
Delage et al. 2008, Toll et al. 2013). interest from an unsaturated soil mechanics viewpoint since it
Reference Air-entry Water Pressure transducer leads to complex stress/suction paths (Mesa-Alcantara 2021,
value reservoir Bian et al. 2020, Saba et al. 2014, Wang et al. 2013). Some mock-
(kPa) volume ups also consider the coupled thermal response mimicking heat
(mm3) emitting wastes (Chen et al. 2014, Pacovský et al. 2007, Štástka
Ridley & 1500 - Modified/commercial 2014, Huertas et al. 2006, Villar et al. 2012, Martín & Barcala
Burland 1993 (3.5 MPa) 2005). These tests rely on the installation of multiple sensors to
Guan & 1500 Approx. 20 Modified/commercial locally measure temperatures, pore pressures, relative humidity
Fredlund 1997 (1.5 MPa) or total suctions, total stresses, and displacements.
Meilani et al. 500 (1-mm - Modified/commercial Also interesting are the infiltration column tests that are widely
2002 thick ceramic) (1.5 MPa) used for several applications, including capillary barriers (Zhou
Tarantino & 1500 < 4.5 Strain gauged et al. 2021, McCartney & Zornberg 2010, Zhan et al. 2014, Tan
Mongiovi diaphragm (4 MPa) et al. 2018, Yang et al. 2004), geosynthetic clay liners (Bathurst
2003 et al. 2007, Bathurst et al. 2009), road pavements and railway
Take & Bolton 300 - Modified/commercial structures (Duong et al. 2013), contaminated soils (Rahardjo et
2003 (0.7 MPa) al. 2018, Murakami et al. 2008) or landfills (Ng et al. 2016c),
Ridley et al. 1500 Approx. 3 Strain gauged among others. The infiltration columns can be assessed as one-
2003 diaphragm (8 MPa) dimensional boundary condition problems in which several
Lourenço et al 1500 5 Ceramic transducer infiltration regimens can be applied. Sensors such as
2006 (2 MPa) tensiometers, matric potential sensors, pressure transducers or
He et al. 2006 1500 - Modified/commercial time domain reflectrometry TDR probes for volumetric water
(3.5 MPa) content are typically installed at different column heights to
Mahler & 500-1500 5-112 Modified/commercial monitor the saturation front and measure suction and water
Diene 2007 (tensiometer acrylic content.
body) Cracking in desiccating soils due to droughts is receiving more
Jotisankasa et 500 60 Modified/commercial attention since it strongly affects soils’ hydraulic and mechanical
al. 2007 (piezoresistive behaviour and is relevant to many engineering applications
sensor) (shallow foundations, dikes or soil covers for mining, industrial
Oliveira & 1500 - Modified/commercial and municipal waste). However, using small-size laboratory
Marihno 2008 (3.5 MPa) samples has a clear drawback since the mechanical boundary
Mendes & 500-1500 4-800 Modified/commercial conditions imposed by the containers have an impact on the
Buzzi 2014 (3.5 MPa) process of crack formation and propagation (Cuadrado et al.
Mendes et al. 7300 4 Modified/commercial
2019, Lakshmikantha et al. 2018). Therefore, there is increasing
2019 (nanoporous (35 MPa)
use of large-scale specimens in environmental chambers
glass
mimicking environmental variables such as wind velocity, air
interface )
relative humidity or solar radiation (Ledesma 2016, Cui et al.
2013, Yesiller et al. 2000, Lakshmikantha et al. 2018, Rodriguez
et al. 2007, Costa et al. 2013, Peron et al. 2009, Tang et el. 2011,
Cardoso et al. (2007) detected some discrepancies between
Levatti et al. 2019, Tang et al. 2020, Garino et al. 2021). Different
transistor and chilled-mirror dew-point psychrometers in the
instrumentation is employed to understand crack formation and
high suction range (7 to 70 MPa) –systematically larger values
propagation and curling phenomena. Sanchez et al. (2013)
were detected with the dew-point psychrometer–. These authors
presented an automatic 2D profile laser that allows scanning the
suggested that the hydraulic paths undergone by the soil during
overall surface of a drying soil (evolving crack aperture and
the measurement period inside each equipment chamber were
321
depth) and an electronic balance to measure the water loss. A Mendes et al. ATT Double-wall cell with glass to eliminate water
2012 absorption
similar non-contact electro-optical laser-based technique was Muñoz- - Monitoring local changes in water content wit
used by Zielinski et al. (2014) to observe soil curling phenomena Castelblanco et al h an electrical resistivity probe
precisely (evolution of the exposed surface of a natural soil 2012
during controlled drying conditions). Infrared thermal camera Cruz et al. 2012 ATT Biaxial apparatus for plane strain conditions
and PIV technique have also been used by Zeng et al. (2022) to Pineda et al. VET High-pressure apparatus with a forced convect
2014c ion system
track the evolution of desiccation cracking while monitoring the Cai et al. 2014 ATT Temperature and suction control triaxial testin
difference between the temperature of the evaporating surface g
and the atmosphere (soils at higher water contents display more Li et al. 2015 CWC Photogrammetry-based method to measure vol
significant temperature differences). Moreover, the ume changes
Alsherif & VET High suction and elevated temperatures
implementation of ground-penetrating radar technique (Prat et al. McCartney 2015
2013) has allowed detecting the formation of cracks inside the Zhang et al. CWC Photogrammetry-based method to measure vol
soil mass. Electric resistivity sensors (Borsic et al. 2005, Comina 2015 ume changes
et al. 2008, Sentenac & Zielinski 2009, Kong et al. 2012) have Nishimura 2016 VET Creep triaxial cells with cyclic relative humid
ity control system
also allowed monitoring spatial and time evolutions of soil Patil et al. 2016 VET & Control of matric or total suction during shea
fissuring. ATT ring
Mora Ortiz 2016 ATT & C Isotropic cell with the matric suction measure
4.3.4 Microstructural techniques (multi-scale tests) WC ment and local displacement transducer
Banerjee et al. ATT Multi-stage triaxial to reduce testing time
Multi-scale studies associated with the pore network, the 2020
arrangement of grains and their interactions (inter-particle Zhang et al. ATT System to precise inject known amounts of w
contacts/bonding, wettability) are increasingly used to improve 2020 ater into the specimen
the understanding of the hydro-mechanical response and Direct shear apparatus
Caruso & CWC Cell with a high capacity tensiometer
multiphase flow properties. Regarding pore/fissure Tarantino 2004
characterisation, these studies challenge unsaturated soil Ng at el. 2007 ATT & Comparison of ATT and OMT techniques
mechanics due to the wide range of pore sizes (from few nm to OMT
hundreds of m) and their consequences on fluid storativity, Hamid & Miller ATT Direct shear interface tests
2009
permeability, diffusivity, single and multi-phase fluid flow
Jotisankasa & CWC Cell with a high capacity tensiometer
properties. These studies include the following topics (e.g., Juang Mairaing 2010
& Holtz 1986, Delage et al. 1996, Komine & Ogata 1999, Aung Kim et al. 2010 ATT Low confining pressures
et al. 2001, Simms & Yanful 2002, Koliji et al. 2006, Yuan et al. Nam et al. 2011 ATT Multi-stage to reduce testing time
2020, Nguyen et al. 2021, Yuan et al. 2021): Hamidi et al. OMT Higher suction range than with ATT
2013
Pores size, morphology and orientation, Borana et al. ATT Shear strength of unsaturated soil-steel interfa
Pore size distribution changes along thermo-hydro- 2015 ces
mechanical and chemical paths, and gas injection processes, Gallage & Uchim HWC Negative gauge pore-water pressure for a low
ura 2016 suction range
Rosone et al. ATT Adapted conventional cell
Table 7. Hydro-mechanical cells with suction control/measurement. 2016
Reference Technique Improvement Ring shear apparatus
Triaxial cells Infante Sedano et ATT Measurement of water content
Romero et al. ATT Local displacement transducers (laser-based tra al. 2007
1997 nsducer for radial displacements). Suction and Merchán et al. VET Forced convection system to transport vapour
temperature-controlled system 2011
Blatz & Graham VET Suction measurement by thermocouple psychr Oedometer cells
2000 ometry Villar 1999 ATT/VET Cell with a combined system for controlled s
Hoyos & Macari ATT True triaxial apparatus uction
2001 Romero et al. ATT Cell with temperature control. Flushing syste
Toyota et al. ATT Triaxial cell and hollow cylinder apparatus 2003 m and evaporation control.
2001 Oldecop & VET Large cell for rockfill
Meilani et al. ATT Mini-suction probe for matric suction measure Alonso 2004
2002 ment Cuisinier & OMT & Large range of suction by combining two tec
Ng et al. 2002 ATT Differential pressure transducer system to mea Masrouri 2005 VET hniques
sure sample volume change Tarantino & CWC Cell with a high capacity tensiometers
Toyota et al. ATT Hollow cylinder torsional shear apparatus De Col 2008
2004 Airò Farulla et al ATT Wetting/drying cycles
Buenfil et al. ATT Local displacement transducers (laser-based tra . 2010
2005 nsducer for radial displacements) Le et al. 2011 CWC Cell with a high capacity tensiometer
Monroy et al. OMT Miniature tensiometers and active radial stress
Padilla et al. ATT Double-wall cell and diffused-air flushing dev 2014 system
2006 ice Mokni et al. ATT Oedometer and direct shear tests also controll
Jotisankasa et al. HWC Continuous monitoring of water content and s 2014 ing osmotic suction
2007 uction Wijaya & Leong CWC Cell with a high capacity tensiometer
Siemens & Blatz CWC Xeritron sensor to measure suction in the cen 2016
2007 tre of the sample Oldecop & VET Measurement of lateral stress and friction
Alramahi et al. 2 ATT P and S-wave velocity Alonso 2017
008 Cardoso et al. CWC Microfabricated sol-gel relative humidity senso
Rojas et al. 2008 ATT Reduced testing time 2017 r
Romero & ATT Isotropic cell (laser-based transducer for radial Maleksaeedi et al ATT & Continuous measurement of water exchange
Jommi 2008 displacements) 2019 HWC
Uchaipichat & ATT Temperature and suction-controlled cell Multi-functional cells
Khalili 2009 Romero et al. ATT Temperature control system for oedometer and
Perez-Ruiz 2009 ATT True triaxial apparatus for cubical samples un 2001 triaxial cells
der constant-suction control Aversa & ATT System for oedometer or triaxial cells
Chávez et al. VET Large size cell for rockfill with local transdu Nicotera 2002
2009 cers Lourenço et al. VET Tensiometer based suction control system
Biglari et al. ATT Suction-controlled cyclic triaxial tests 2011
2011 Liu et al. 2021 ATT Modular design for pressure plate test, oedom
Hoyos et al. ATT Refined true triaxial apparatus eter test, direct shear test, and triaxial creep t
2011 est
322
number, chemical composition, water content) and the X-ray
Discrimination of the type of the pore space equipment (X-ray energy, intensity) (Hounsfield 1972). During
(micropore/macropore, matrix/fissures, pore body/pore an X-ray tomography scan, a 3D map is reconstructed consisting
throat), of voxels, each representing, through a grey value, the photon
Distribution of the pore volume (micropore volume/ attenuation that the material provides at a specific point in the
macropore volume, volume of fissures, fluid occupancy space. Such a technique has been successfully used in the
effects during multiphase flow displacement processes), investigations of unsaturated soils since voxels constituted by
Connectivity of the pore space, soil grains, water, or air can be easily distinguished due to their
Random distribution of porosity, different attenuation capacities (Salager et al. 2014). Depending
Wettability issues and interaction with grains. on the X-ray image resolution, voxels can include one or multiple
phases: in this last case, the grey value will represent a linear
Among various techniques used to study porous geomaterials at combination of the attenuation values of those phases, weighted
the microstructural scale, mercury intrusion porosimetry MIP by the porosity and the degree of saturation of the material (Luo
and scanning electron microscopy (particularly environmental et al. 2008).
ESEM with digital image analysis) were pioneering techniques Apart from the extensive use of this technique for qualitative
that are still used. MIP is utilised for analysing the pore size purposes in geotechnical engineering, X-ray CT is increasingly
distribution PSD of geomaterials with interconnected porosity being used to observe and quantify features and additional phases
(Delage et al. 1996, Romero & Simms 2008). This technique with a high organic matter, such as bacteria, vegetation and grout
applies absolute pressure to a non-wetting liquid (mercury) to binders, within partially saturated soils (Terzis & Laloui 2019,
enter the empty pores. In addition, Washburn equation is adopted Fraccica 2019, Anselmucci et al. 2021). X-ray CT (i.e., with a
under equilibrated conditions (i.e., null penetration velocity of resolution of the order of microns) has been successfully used to
mercury and constant contact angle that does not vary with the extend and complement the upper limit of soil’s PSD coming
advancing interface) to provide a relationship between the from MIP (Münch & Holzer 2008, Yang et al. 2015, Fraccica et
applied pressure and the entrance size of the intruded pores (e.g., al. 2019). Such result has direct implications for understanding
Juang & Holtz 1986). Sample preparation requires emptying the the hydro-mechanical behaviour of partially saturated soils,
sample of water, which can be dehydrated using controlled whose microstructure is modified by multi-physical phenomena
relative humidity-drying, oven-drying, freeze-drying or critical- (e.g., the evolution of root-induced fissures with suction, fissures
point-drying technique (Delage & Pellerin 1984). For drying generated by gas transport, arrangement and evolution of
sensitive materials, freeze-drying is preferred (Delage & Pellerin solidified grouts or calcite bonds at varying degrees of saturation
1984, Delage et al. 1996). The main limitations of MIP are a) within soil matrix). X-ray CT has also been used to achieve a
enclosed porosity is not measured; b) pores that are accessible deeper understanding of:
only through smaller ones are not detected until the smaller The geometrical and spatial-temporal evolution of liquid
entrance pores are penetrated; c) the apparatus may not have bridges and wetting fronts, fundamental to trace adhesion
enough capacity to enter the smallest pores (non-intruded forces between hydrophilic/hydrophobic grains and
porosity with entrance pore sizes below 7 nm); d) the minimum understand shear strength behaviour of partially saturated
pressure which can be applied limits the maximum detected pore soils (Wildenschild et al. 2005; Berg et al. 2013; Pot et al.
size (non-detected porosity with entrance pore sizes larger than 2015), the impact of grain’s shape and contacts on the
400 m); and e) alteration in the pore volume of compressible phenomenological soil mechanical behaviour (Karatza et al.
materials before mercury penetration starts. MIP results are 2019).
reported as pore size density function PSD, i.e., the log Strain fields and shear bands in unsaturated soils under
differential intrusion curve versus entrance or throat pore size, different stress-paths (e.g., pioneering works of Desrues et al.
which aids visual detection of the dominant pore modes. Data 1996 and Otani et al. 2000 and recent works of Higo et al.
from MIP can be complemented, for pore sizes below 60 nm, 2013 and Takano et al. 2015).
with nitrogen desorption isotherms (although using the Desiccation cracks (Tang et al. 2019), injectability and
adsorption branch is also possible). Data from the latter effects of hydraulic states on grouted/bio-improved soils
technique are interpreted using BJH model, based on which (Minto et al. 2017, Pedrotti et al. 2020).
emptying of pores from condensed adsorptive at decreasing Ice contacts formation and evolution in soil, frost heave
relative nitrogen pressure is re-interpreted by Kelvin equation (Starkloff et al. 2017, Wang et al. 2018, Song et al. 2021).
(e.g., Webb & Orr 1977). Multi-phase flows or gas/solute transport phenomena
The environmental scanning electron microscope ESEM is a (Andrew et al. 2014, Larsbo et al. 2014, Gonzalez-Blanco et
quantitative technique with minimal sample preparation al. 2020).
requirement, and which allows subjecting the sample to Methane hydrate processes (Kneafsey et al. 2007).
hydraulic paths during observation (Komine & Ogata 1999, Ground loss and granular flow for model tunnels or
Montes-H et al. 2003a,b, Romero & Simms 2008). The ESEM is extraction advancement (Takano et al. 2006, Viggiani et al.
a particular type of SEM that works under controlled 2015).
environmental conditions and requires no conductive coating on Estimation of external stress transmission through the soil
the specimen. Montes-H et al. (2003a,b) used ESEM jointly with matrix (Naveed et al. 2016).
a digital image analysis program to estimate bentonite’s Oil-water-air interfaces (Culligan et al. 2006; Gharbi & Blunt
swelling/shrinkage behaviour at different total suctions at 2012).
aggregate scale. Romero & Simms (2008) and Airò Farulla et al. When performing X-ray tomography, the first challenge is
(2010) used the same technique to study the effects of total finding an adequate compromise between REV and resolution,
suction changes on the volumetric behaviour at the as this last characteristic is strictly linked with the specimen’s
microstructural level of different clays. Karatza et al. (2021) used size (Li & Shao 2020). For this reason, partially saturated clays
ESEM to investigate the formation and evolution of meniscus are still poorly investigated at the scale of pore and liquid bridge
structures (capillary bridges) in hydrophobic grain surfaces contact, even if research is pushing towards the development of
during wetting/drying cycles. miniature soil mechanics equipment (Cheng et al. 2020).
Regarding X-ray computer tomography CT, image analysis is Moreover, poor-quality X-ray CT of unsaturated soil with silty
based upon the ability of different materials to attenuate photons or clayey matrix might produce incorrect distinctions of porosity
emitted by an X-ray source in different proportions, depending and grains if excessive noise (i.e., grey value’s scattering) or
on multiple characteristics (i.e., bulk density, porosity, atomic
323
edge/ring artefacts are present in the images. Despite the 8. X- ed soils.
limitations of X-ray images (strictly linked with the experience Reference Soil (USCS) Applications
in using the equipment), many investigations on partially
Li & Shao 2020 CL
saturated soils use this powerful technique. A selection is
presented in Table 8 jointly with the type of soils investigated. Moscariello & Cuomo Artificial sands
2019 HM behaviour:
Given that X-rays are more attenuated by soil grains than water
and that neutron tomography offers a higher contrast in the Tang et al. 2019 CH Soil bulk density
visualization of the water phase, the two techniques have been Pore-size
Karatza et al. 2019 Zeolite granules distribution
recently used in a complementary way (Kim et al. 2013, 2016, (d50 = 1.36 mm) Tortuosity
Stavropoulou et al. 2020). Khaddour et al. 2018 SP WRC
A good combination of microstructural techniques and Strain fields and
phenomenological measuring tools can be successfully used to Karatza et al. 2018 Caicos sand
shear bands
link microstructural features of partially saturated soils such as Mukunoki et al. 2016 SP under mechanical
liquid bridges, through the evolution of their curvature, with stresses.
grain’s adhesion forces (Hueckel et al. 2020). Takano et al. 2015 Yamazuna Sand Evolution of liquid
(d50 = 0.54 mm) bridges/wettability/c
ontact angle
4.3.5 1-g scaled experiments and centrifuge tests Pot et al. 2015 Silt (56%) with sand
Effects of grain size
Table 9 summarises selected experimental 1-g scaled and (27%) and clay (17%)
and shape on soil
centrifuge N-g (where N is the scale factor, and g is the Earth’s Yang et al 2015 Sandstone HM behavior
gravity) set-ups proposed to study the unsaturated mechanical Desiccation cracks
Berg et al. 2013 Sandstone
behaviour of geomaterials under distinct flow and water content
regimes. Among 1-g testing, the well-established granular Higo et al. 2013 SP
column collapse experiment was first proposed for the Bull et al. 2022 Sand (71%) with fines
investigation of dry granular flows by Lajeunesse et al. 2004 and (29%)
Lube et al. 2005. Unsaturated collapse experiments using natural Kemp et al. 2022 Clay/silt/sand/gravel
materials are of interest though scarce (e.g., Fern & Soga 2017). layers Bio-improved soils:
Scale model embankments are built to investigate rainfall Pore size
infiltration into various soil types and the consequences on Anselmucci et al. 2021 SP distribution
stability. The performance of permeable geosynthetics on sandy WRC
Fraccica 2019 SM
soils was assessed by Garcia et al. (2007), instrumenting a soil Root volume and
Terzis & Laloui 2019 SP architecture
box with pore pressure and moisture content sensors. Rainfall
infiltration in gravelly soils was also studied with a 2D seepage Koebernick et al. 2017 Sandy loam Root-induced
fissures
set-up by Dong et al. (2017). Computed tomography was used to Minto et al. 2017 Sandstone
Root-induced grains
identify the main contributing effects to the observed infiltration Keyes et al. 2017 Sand (52.7%) and fines kinematics
rates. Other 1-g experiments were designed to characterise the (47.3%)
unsaturated behaviour of transparent soils (Iskander 2018,
Siemens et al. 2013) and hydrophobic soils (Karatza et al. 2021, Carminati et al. 2013 Sand (92%) with fines
Zhou et al. 2021), as well as the effect of freezing/thawing Kim et al. 2021 Clay
(Caicedo 2017). Gonzalez-Blanco et al. Granular bentonite
The term centrifuge testing is used to describe the 2020 Multiphase flows,
measurement of physical properties of unsaturated soils such as gas/solute transport
suction or permeability using geo-centrifuges (Timms et al. Andrew et al. 2014 Limestone
2014), whereas centrifuge physical modelling refers to the Larsbo et al. 2014 Silty and clayey soils
extrapolation of reduced scale model tests to full-scale
Pedrotti et al. 2020 Sand (d50 = 1.2 mm) Grouted soils
geotechnical applications (Caicedo & Thorel 2014). Another
advantage of centrifuge modelling is observing the hydraulic Takano et al. 2006 Dry sand
Granular flows and
long-term soil response in reduced time frames (Dell’Avanzi et Viggiani et al. 2015 Fine sand ground loss processes
al. 2004). Scaling laws for centrifuge modelling of unsaturated (d50 = 0.2 mm)
soils were recently reviewed, focusing on soil water retention Song et al. 2021 Volcanic ash sand and
scaling (Mirshekari et al. 2018). The capillary rise was studied in terrace deposits Ice
formation/evolution in
sandy soils (d10 Wang et al. 2018 CL partially saturated
the effects of gravity on menisci shape, unlike in coarser soils soils
Starkloff et al. 2017 Sandy and silty soils
(Rezzoug et al. 2004). Scaling laws are generally defined by
dimensional analysis as in saturated soils, albeit showing Culligan et al. 2006 Glass beads Oil-water-air
discrepancies for unsaturated flow problems. Nevertheless, Gharbi & Blunt 2012 Carbonate limestones interfaces
analytic determination of scaling laws provides a consistent Kim et al. 2016 Sand Projectile
framework for scaling steady-state and unidimensional flow in impact/penetration
engineering
2004). Relevant scaling factors for unsaturated soil phenomena,
including heat exchange, infiltration, evaporation, and rainfall
(Askarinejad et al. 2014, Caicedo & Thorel 2014), are
summarised in Table 10. The factors are given a length scaling
factor N=L_prototype/L_model, where L_model is the length of
the small scale physical model that is compared to the quantity
in the full-scale prototype. In equivalent terms, it corresponds to
an acceleration scaling N=a_model/g, where a_model is the
centrifugal acceleration that scales up gravity g.
324
Table 9. Classification of selected experimental set-ups for investigating Geotechnical applications of centrifuge modelling can be found
the unsaturated mechanical behaviour of granular materials. on:
Flow regime Saturation state Characterization of the at-rest coefficient of lateral earth
Reference Funicular/ pressure of unsaturated soils (Li et al. 2014) and expansive
Steady Transient Pendular
Capillary clay swelling (Gaspar et al. 2019).
1-g (granular collapse Slope stability (Higo et al. 2015) including root-induced
tests)
phenomena (Leung et al. 2017, Liang et al. 2017a) and live
Artoni et al. 2013 pole reinforcement (Ng et al. 2017; Kamchoom & Leung
Gabrieli et al. 2013 2018).
Morse et al. 2014 Climatic or environmental chambers (Archer & Ng 2018);.
Santomaso et al. 2018
Rainfall (Bhattacherjee & Viswanadham 2018, Khan et al.
2018) and seepage (Beckett & Fourie 2018).
Torres-Serra et al. 2018 Seismic induced settlement of shallow foundations on
Pinzón & Cabrera 2020 unsaturated soils (Borghei et al. 2020).
Taylor-Noonan et al. 2 The leading experimental technologies applied in centrifuge
020 investigation, also recently reviewed by Take (2018), include:
N-g Image analyses (Beckett & Fourie 2018) and especially the
Ng et al. 2008
Particle Image Velocimetry (PIV) (Iskander 2018).
3D printing (Liang et al. 2017b).
Askarinejad et al. 2014 HR sensors (Takada et al. 2018), tensiometers (Basson et al.
Caicedo et al. 2015 2021, Jacobsz 2018, Kwa & Airey 2018), laser-based
Rotisciani et al. 2016 displacement and fibre optic strain sensors.
Image analysis techniques such as the PIV are widespread in the
Lozada et al. 2018
observation of flow kinematics (Morse et al. 2014, Pinzón &
Lalicata et al. 2020 Cabrera 2020, Torres-Serra et al. 2020). The PIV is also known
Lucas et al. 2020 as Digital Image Correlation (DIC) in geomechanics frameworks
Escobar et al. 2021 (Hall 2012). In 2D-surface DIC analyses the displacement fields
are recovered, starting from an initial reference configuration, by
cross-correlation of the subsequent image pairs of the deforming
material. Resolution of the PIV analysis depends on the size of
Table 10. Scaling factors for centrifuge modelling concerning length
scale N of full-scale prototypes.
the correlation windows, the pixel subsets being traced, and of
Quantity Prototype/model
the search region or region of interest ROI, as well as on the
recording frame rate, defining the time spacing between analysis
Length N
points. Recent advancements on the PIV technique include its
Volume and mass N3
enhancement to large displacement and deformation
Time (diffusion) N2
observations for landslide modelling in both 1-g and N-g (Pinyol
Acceleration, gravity N-1
et al. 2017). PIV is also currently used in the study of local frost
Force N-2
deformation (Wang et al. 2020) and unsaturated soil-structure
Stress, moduli and strength 1
interaction (Vo et al. 2016, Shwan 2019, Speranza et al. 2020).
Rezzoug et al. 2004 Novel imaging applications using infrared techniques are of
Capillary rise N special interest to the study of unsaturated materials. The
Dell’Avanzi et al. 2004a saturation degree of soils can be measured by short wave infrared
SWIR images from the different absorbance of incident light
Discharge velocity N-1
existing between water and the solid particles. SWIR imaging
Suction 1
was thereby successfully used to observe near-surface water
Caicedo & Thorel 2014 infiltration profiles for sandy soil column set-ups (Parera et al.
Rate of water content N-2 2020, Sadeghi et al. 2017). Hydraulic soil properties were also
Rate of volumetric strain N-2 recently derived from SWIR measurements of volumetric water
Heat flux N-1 contents (Bandai et al. 2021).
Rain intensity N-1
Rain duration N2 5 FIELD INSTRUMENTATION
Rain frequency N-2 5.1 Preamble
Rate of evaporation N-1
Wind velocity 1 The quantities required to be measured in the field, to contribute
-1
to the design and monitoring of engineering works, are
Vapour pressure deficit N associated with the type of mechanical demands that the structure
Askarinejad et al. 2014 imposes on the soil as well as how the soil will react. It is the
Seepage velocity (macro/micro) N-1
function of the engineer to understand the phenomena involved
in each situation and thus define the quantities to be determined
Seepage time (macroscale) N2 and/or monitored. When the soil is in a saturated state there are
Seepage time (microscale) N only two phases, namely, the solid phase and the liquid phase. In
Hydraulic gradient (macro/micro) N-1 this situation the determination of the stress state, by means of
a
assuming centrifuge arm length much larger than the model
pore pressure measurements, using piezometers, is quite
length
common and relatively usual. When the problems involve soils
in an unsaturated state, the mere presence of air in the voids,
combined with the water, increases the difficulty of monitoring
for physical reasons. This section will discuss how to choose and
define the installation position of suction sensors in the field. The
325
physical principles that are used in the measurements will be In 1849, Bourbon obtained a patent for what is known to this day
briefly given, in order to allow the reader not only to know the as the Bourbon gauge pressure. But in 1733, an unprecedented
working principle of the sensors, but also to be able to choose, pressure measurement was taken by Stephen Hales. He was the
calibrate and correctly install the sensors. Examples of how soil first to measure blood pressure, although it was on a horse. The
suction is measured are presented, providing examples from the measurement was made with what we know today as a
literature and highlighting for each type the most appropriate way piezometer (Figure 15). It is noteworthy that in this case the
to install the sensors and determine their optimal positioning. response time of the measurement system must be immediate.
5.2 Pore Water Pressure measurement
To understand the differences and similarities between positive
and negative pore pressure measurements, it is essential to recall
some fundamental concepts of pressure measurement in liquids.
The measurement of pore water pressure is of utmost importance
for the vast majority of engineering works. Furthermore, most of
the problems that geotechnical structures face are associated with
variations in the pore-pressure of water. In the early days Galileo
Galilei made important observations for the knowledge of
pressure in liquids. With the development of a hydraulic pump
by Galileo at the end of the 16th century, he noted the limitation
Figure 17. Types of piezometers (a) Casagrande (b) Open
standpipe.
It is important to understand the principle enunciated by the one
who, nowadays, gives the name of the pressure unit we use.
Pascal's principle state s that in a fluid within a closed system,
any change in pressure at any point in the fluid will be transmitted
to all points in the fluid as well as to the fluid-containing system.
This principle is valid for both positive and negative pressures,
keeping within the limitations associated with cavitation,
observed, although not fully understood by Galileo.
Measured pressure values are referenced to absolute zero or
atmospheric pressure. When measuring absolute pressure, the
sensor must have vacuum at the back of the sensor. In the case of
relative pressure measurement, the pressure varies from one
Figure 15. Illustration showing Stephen Hales performing a blood
pressure measurement on a horse.
location to another as the pressure is corrected for level
conditions. Air (atmospheric) pressure decreases with increasing
of his invention in raising water to a maximum height of 10m. altitude. This pressure is the relative pressure or pressure relative
This limitation was never explained by Galileo, who nevertheless to the level where the measurement is made. Figure 16 illustrates
recorded this observation. Studies related to atmospheric and the different ways to interpret the measured values. Pressure
liquid pressure measureme nt, developed by Galileo, Torricelli, values below atmospheric pressure measured in porous materials
Viviani, Perier, Von Gueriche and Hooke, led Boyle to establish such as soil are called suction and can exist even for values below
the law between pressure and air volume at the same temperature.
Figure 16. Pressure definitions and their references. Figure 18. (a) Pressure transducer with a porous tip (b) piezometer
installed in the field.
326
absolute zero. 2021).
Positive pore water pressure measurement requires continuity Measurement of some characteristic (e.g., electrical
between the sensor and the water. Casagrande or standpipe-type conductivity, capacitance, therm al conductivity, etc.)
piezometers record the piezometric pressure or the water level, associated with suction, measured not necessarily in the soil,
respectively. Measurements are made on the surface by but in a porous element with a well-defined characteristic
measuring height, or the pressure at a given point. A schematic that allows establishing a pressure equilibrium with the soil.
drawing of these piezometers can be seen in Figure 17. These sensors use porous elements in which the quantities
Piezometers without connection to the atmosphere measure pore related to suction are measured. As these quantities, in
water pressure using pressure transducers that require an general, require the material to be in an unsaturated condition,
interface between the soil pore water and the water of the measurements for very low and very high suction values are
transducer. compromised.
Figure 18 schematically i llustrates the tip of a piezometer In all situations there will be a flow after installation, between
with a pressure transducer. This system is composed of a porous the soil water and the sensor, or vice versa. It is this transfer
element that will establish contact between the soil pore water process that determines the sensor response time and the
and the water within the transducer (for the case of a tensiometer). applicability of the sensor to specific cases. Ideally a method for
The placement of piezometers for measuring positive pore measuring the negative pressure of soil water (suction) should
water pressure is intuitively straightforward and requires the have the following characteristics:
sensor to be positioned below the phreatic line. Bearing in mind Calibration must be easy to verify and be reliable over time
that in situations where the piezometer is installed in unsaturated and for each sensor
soil, but the expectation is to measure positive pressure, It must guarantee the measurement of the suction type to
equipment that allows re-saturation of the system should be which it is specified
chosen. If the system has the presence of air, inaccurate values It must allow measurements at suction levels suitable for
may be measured and the response time is impaired. each case
The piezometer that allows the measurement of negative It must be easy to use and economically viable
(relative) pore water pressure is named tensiometer. According It should require little or no maintenance.
to Or (2001), the inventor of the tensiometer was the American
plant physiologist Burton E. Livingston. In 1908, Livingston Referring once again to Figure 16, we could separate the sensors
developed a system to control the soil moisture available to that measure pressure below atmospheric pressure (they would
plants, which was perfected in 1918, and which has all the be the tensiometers) from those that measure above (the
elements of a tensiometer, and used nowadays. piezometers). However, this separation would be too simplistic
and would not take into account the advances that are already
observed in the pore pressure gauges that are available. Of course,
sensors that measure pressure below atmospheric pressure play
the role of tensiometers. However, the same sensor can measure
both positive and negative pressures. Tarantino et al. (2008)
present a review of the types of piezometers and the development
of sensors that allow measuring both values above and below
atmospheric pressure and even below absolute zero. Details of
the functioning of tensiometers that measure pressure below
absolute zero can be found in the literature (e.g., Marinho et al.
2008).
5.3 Sensors for Measuring Suction in the Field
In pore pressure measurements, whether positive or negative,
there will be an interaction between the sensor and the pore water.
In both cases, the pore water characteristic can influence the
measured value depending on the characteristics of the sensor
Figure 19. Main sensors for suction measurement and field use, (a) used. However, as pointed out previously, one of the most
Conventional tensiometer (b) High capacity tensiometer (c) important aspects for these sensors is the response time, which
Capacitive sensor (d) Electro-resistive sensor (e) Thermal sensor. depends on the operating principle of the sensor, the hydraulic
Determining negative water pressure does not necessarily require characteristics of the soil and the interface that facilitates the
the sensor to measure the water pressure directly, nor does it interaction between the soil water and the water in the sensor.
necessarily require the measurement to be taken with the sensor Suction measurement for use in warning systems can only be
in physical contact with the soil. Field measurements generally considered if tensiometers are used. This is due to the response
measure matrix suction. The measurement principle can be time of this type of sensor, which is the smallest of them all.
divided into two: Table 11 presents the main sensors used to measure suction in
Direct measurement of water pressure using tensiometers the field, its range and working principle. All sensors listed are
that use a porous element as the interface between the pore commercially available. The HCT can be found with a higher
water and the transducer water, in the same way as measurement capacity (up to 1.5 MPa), but it still requires some
piezometers. These sensors, called tensiometers, generally adjustments to allow its use in the field.
do not allow suction measurements beyond 1 atm. Teher are Figure 19 illustrates the five sensors referred to in Table 11.
some tensiometers which have been designed to allow The tensiometer shown in Figure 19a has the transducer close to
suction measurements greater than 1 atm, as will be the ceramic capsule, however the other models use transducers
mentioned later. Other sensors that use the osmotic principle that are on the surface. The high capacity tensiometer illustrated
are also promising (e.g., Bakker et al., 2007; Rahardjo et al., in Figure 19b allows for larger suction measurements than the
conventional tensiometer. Other models can also be found with a
327
greater measurement capacity. However, the process of range. The granular matrix sensor (Fig. 19d) is very robust and
saturation and maintenance of capacity in the field is still a very interesting for use at great depths. Figure 19e shows the
challenge. Figure 19c shows a capacitive sensor, which has a low thermal conductivity sensor.
maintenance requirement and a very wide suction measurement
Table 11 – Suction measurement sensor for field measurement and its range.
Sensor Suction range Working principle
Tensiometer 0 to -1 atm Water pressure
HCT 0 – 150 kPa Water pressure
GMS 0 to – 200 kPa Electrical resistivity
Teros 21 5 – 100 MPa Capacitance
Thermal Conductivity 1 kPa to 1 MPa Thermal conductivity
All sensors that use an indirect way of measuring suction, need used to measure suction can compromise the monitoring project.
the presence of a porous element that will serve as a reference to Thus, it is essential that the choice of sensor and its positioning
convert the measurements made according to the operating are analysed together.
principle of each one. As such, the porous material incorporated Figure 20 presents, schematically, a soil profile indicating the
into the sensor has an unknown amount of hydraulic hysteresis, pore-water pressure profile and the retention curve, in terms of
causing differences in relationships between suction and the degree of saturation and water content. In Figure 20a, in addition
measured quantity (e.g. dielectric permittivity, resistivity) for to the equilibrium profile (no infiltration and no evaporation), the
wetting and drying. In general, the calibration used by the hypothetical profiles associated with the dry and rainy seasons
manufacturers is associated with the drying path. In this way, it are also indicated. Linked with the profiles, Figure 20b shows the
is essential to check the hysteresis that each sensor of this type relationship between the degree of saturation and depth, while
can have, and in particular the scanning curves. This verification Figure 20c presents the variation of water content with depth. It
must be carried out according to the level of seasonal variation
that is expected in each case according to the position of the
sensor in the active zone (further described). Some examples of
studies and comments related to hysteresis in sensors that use
ceramics or other porous material are presented by several
authors (eg Campbell & Gee, 1986; Feng et al., 2002; Bulu &
Leong, 2008; Yates & Russell, 2022; among others).
Suction is the negative soil pore water pressure; however, it is
not necessary to measure the pressure directly, it is possible, and
often necessary and useful, to determine some variable that is in
equilibrium with the pore water pressure. For a deeper
Figure 20. (a) soil suction profiles at equilibrium showing the Figure 21. Association between TMI and active zone depth.
profiles for dry and wet seasons (b) SWRC in terms of degree of
saturation (b) SWRC in terms of water content.
should be noted that the depth in this case refers to the suction
understanding of this concept, the reader is referred to the work when the system is in equilibrium (suction = - ).
of Edlefsen & Anderson (1943). The importance of understanding the suction profiles and how
they present themselves in each case is fundamental for the
5.4 Fundamental Aspects for Monitoring Suction in Unsaturated choice of the sensor and its positioning. As shown in Figure 20a
Soils below the phreatic surface the soil is saturated and with positive
pressure and that above this line the soil has negative pore water
Selection of an appropriate sensor for the measurement of suction pressure. It should be noted, however, that even above the
depends on the expected range of and location/depth at which phreatic surface the soil can remain saturated and with negative
suction is to be monitored. In the same way piezometers are pressure. This makes pore pressure measurements unfeasible
installed in positions of interest for the project, the suction when using piezometers that are not capable of measuring
sensors must also be positioned depending on the interest of the negative pore pressure. The quantification of this capillary
project, but also on the expected response, so that the values do saturated range can be made using the SWRC (e.g., Nadai et al.
not deviate from the limits of the sensor used. It should always 2022). Figure 20b illustrates how the separation point between
be kept in mind that the systematic maintenance of any sensor the saturated and unsaturated condition is determined using the
SWRC. The height of saturation by capillarity can reach tens of
328
meters and the definition of this height can be fundamental for generate different profiles and may not establish an easily
the instrumentation planning in projects. The capillary rise height defined active zone.
is related to the air entry suction of the material, as indicated by The equilibrium profile of suction, as already mentioned,
Kumar & Malik (1990). As we approach the surface, the suction allows you to easily identify the suction value when there is no
profile is affected by the seasonal weather variation, establishing infiltration or evaporation. In this way, associated with the depth
what is called active zone, shown in Figure 20a. In order to better of the active zone, the suction values to be measured can be
understand the sensor selection process, and its placement in the evaluated. Figure 22a reproduces the suction profile in the
field, the following sections describe typical situations and how equilibrium condition, illustrating the schematic range of pore
one should interpret them. pressure variation and definition of the active zone. In principle,
when placing a sensor below the active zone, the suction value is
5.4.1 Climatic condition and active zone well established and directly related to the distance to the
It is critical to consider the expected flow pattern when phreatic surface.
positioning the suction sensors. Flow is caused by both Some sensors, such as the tensiometer, generally have an
infiltration and evaporation; and in some cases capillary rise. It installation depth limitation. Ordinary tensiometers have at their
is known that sensors very close to the surface can be subjected lower end a ceramic with high air entry pressure (e.g., 1 bar), and
to suctions above those that the sensor can measure (see Figure at the other end a pressure transducer or vacuum gauge. In this
20a), generating the need for early maintenance, which in many system, each meter of depth means a reduction in the capacity to
cases can lead to sensor loss due to difficulties in removing them measure suction of 9.81 kPa. In this way, the depth of installation
for maintenance. This behaviour depends not only on the of suction sensors must also be evaluated when choosing the type
structure to be monitored, but also on local climatic conditions. of instrument. When the tensiometer has the transducer together
The climate and soil, as well as the geometry of the structure, with the ceramic, as in the case of the tensiometer shown in
induce the establishment of the active zone. The active zone (Hs) Figure 19a, this problem does not exist.
is the depth that undergoes seasonal variations, suction and/or As previously described, taking the water level as a reference,
moisture content, due to climate and/or vegetation. It should be it is possible, in a first evaluation, to estimate the suction at the
noted, however, that this active zone can also be influenced by sensor installation point and thus better position the sensors,
the variation of the water level, which in turn affects the height always keeping in mind the fluctuations due to the climate, in the
of capillary rise described below. active zone.
The choice of suction sensor positions must be such that it
allows the chosen sensor to remain within the measurement range
and permits the definition of suction profile over time or other
information that may be relevant for the project such as the
degree of saturation (via SWRC). When the phreatic surface is
close to the surface (tens of meters or less) the suction profile can
be inferred as illustrated in Figure 20a, and the suction can be
estimated at the depth of the active zone. As we will see, the
active zone varies between 2 and 4 meters in most cases.
Therefore, it is possible to evaluate which type of sensor will Figure 22. (a) Concept of equilibrium suction profile (b) Concept of
allow the measurement of this suction. When positioning the equilibrium suction for regions with deep or absent phreatic level
sensor below this depth, little variation is expected with time. .
When the objective is to measure suction in a location where
Some authors present the relationship between the TMI the phreatic surface is very deep, which generally occurs in arid
(Thornthwaite Moisture Index) and the depth of the active zone. or semi-arid regions, the use of field sensors becomes almost
(e.g., Fityus et al, 1998; Mitchell, 2008) and this relationship is impossible. Under these conditions the expected suction below
used in Australian standards, for example, relevant to the active zone is very high. Figure 22b schematically illustrates
foundations in expansive soils (AS2870, 2011). According to the suction distribution under these conditions and presents the
Mitchell (2008) climate changes, which will certainly change the equilibrium suction that is established, below the active zone.
TMI values in different locations, will induce a variation in the This suction is not related to the equilibrium profile described
active zone that can create serious problems. The monitoring of before. The equilibrium suction will still, in conditions where the
these variations in terms of the active zone is fundamental and phreatic surface is deep, have some variation, but as mentioned
the measurement of suction becomes of extreme necessity in the by Vann & Houston (2021) insignificant from a practical point of
short and medium term. view. Neither these variatio ns are significant to affect eventual
Figure 21 presents some relationships found in the literature settlement and shear strength calculations. In many cases where
correlating the depth of the active zone with the TMI. The the equilibrium suction is very high, it may be more appropriate
purpose of presenting this relationship is to help define the ideal to take samples and measure the suction in the laboratory.
positioning for the installation of sensors for measuring suction. Russam and Coleman (1961) were among the first to establish
As previously mentioned, the active zone typically varies a relationship between equilibrium suction and a climate index,
between 2 and 4 m for dry-sub-humid, wet sub-humid and humid in this case the TMI (Thornthwaite Moisture Index). However, as
climates, with a tendency to increase for dry-sub-humid climates. mentioned by Vann & Houston (2021), the relationship between
Positioning the sensor within the active zone will necessarily lead TMI and suction is weak, as it depends on other factors such as
to fluctuations in measurements depending on local climatic soil type, profile heterogeneity and also surface condition.
aspects. It should be noted that vegetation, surface use and other Furthermore, Karunarathne et al. (2012) and Sun (2015) showed
aspects can affect near-surface oscillations. that different TMI can be obtained through various approaches
It is usual to take the pheatic surface as a reference, as and hypotheses. Even considering all these difficulties, it will
illustrated in Figure 20a. For simplification, we will adopt only interesting to examine the relationship between equilibrium
two suction distribution situations with depth for discussion. One suction and IMR. Figure 23 illustrates this relationship as
where the phreatic surface is known and another where it is obtained/estimated by several authors.
absent or very deep. Situations where there may be a continuous The data presented by Russam and Coleman (1961) separate
flow of water from some source, or the existence of vegetation the relationship according to soil type. Within this approach, it is
329
sensors and their measurement ranges are presented, allowing a
direct evaluation of the sensor that can be used as a function of
the expected equilibrium suction.
The equilibrium profile described above is partially formed by
capillary effects. This, often ignored, can be of fundamental
importance in the evaluation of the behaviour of structures.
Depending on the material to be monitored, whether it is a
landfill, a tailings bed, a dry stack or a natural soil, the presence
of water at the bottom can induce a capillary rise that can reach
tens of meters. It should also be noted that the phreatic surface
may be tens of meters below the base of the landfill or natural
slope. The association between saturated zone of the capillary
rise (saturated fringe) and water infiltration can generate what
can be called a pore water pressure bomb, since when the
infiltration meets the capillary zone, the capillary height is
converted into positive pore water pressure (e.g., Vaughan 1985,
Jayatilaka & Gillham 1996). Suction measurement, performed
correctly, using suitable sensors and correctly positioned, is
Figure 23. Relationship between equilibrium suction and TMI. essential to detect saturated regions with negative pore water
pressures. These regions are prone to liquefaction if other
possible to observe that more plastic soils have a higher features are present.
equilibrium suction. The other authors present relationships that The association of water content measurement with suction
are associated wit h plastic soils and, more specifically, measurement can significantly help in the interpretation of data
expansive soils. The data from McKeen & Johnson (1990), and this is done using the SWRC or measuring the water content
interpreted here, indicate suction values for dry sub-humid, wet in the field. However, given the intrinsic differences between the
sub-humid and humid climates, which range from approximately two measurement processes, great care must be taken when
400 to 100 kPa. As in the hypothetical condition of a very dry installing sensors to measure water content. Two materials with
climate and with constant relative humidity (RH), the suction different geotechnical characteristics may have equal suction
value has to come into equilibrium with the RH. The figure values when in equilibrium, but different water content values.
shows the suctions that are related to a 50% and 70% RH, which In this way, the suction measurement is independent of the soil
indicates that the data should tend towards these values, or on which the sensor is installed. Very heterogeneous materials
another constant RH value. The curves presented by Saha et al. can generate difficulties in the interpretation of suction values
(2019) and Vann & Houston (2021) indicate suctions above 500 and moisture content.
kPa in all climates. On the right side of Figure 23, the most used
Table 12. Examples of structures where suction was measured in the field
Type of sensor Application Reference
GMS Railroad track Castro et al. (2021)
GMS Slope Barbosa et al. (2010)
GMS Slope Mendes & Marinho (2010)
MPS-6 Natural ground Tian et al. (2018)
Tensiometer Landfill cover Alam et al. (2019)
Tensiometer Slope Garg et al (2015)a
Tensiometer Slope Sestrem et al. (2018)
Tensiometer Slope Tu et al. (2009)
Tensiometer Slope Vieira & Marinho (2001)
Tensiometer Slope Yang et al. (2019)
Tensiometer - HC Embankment Mendes et al. (2008)
Tensiometer Natural Ground Silva Junior (2011)
Tensiometer - HC Natural ground Cui et al. (2008)
Tensiometer - HC Slope Toll et al. (2011)
Tensiometer - HC Landfill cover Maldaner & Marinho (2012)
Tensiometer and Thermal Sensors Slope with vegetation Garg et al. (2015)b
330
5.4.2 Where to use the sensors and its requirements Agus S.S. & Schanz T. 2005. Comparison of four methods for m
The vast majority of geotechnical structures involve soils with easuring total suction. Vadose Zone J. 4(4), 1087-1095.
negative pore water pressure (regardless of whether it is saturated Airò Farulla C. & Ferrari A. 2005. Controlled suction oedometric
tests: analysis of some experimental aspects. Proc. Int. Sympo
or not). In many of these cases, suction controls safety or at least sium on Advanced Experimental Unsaturated Soil Mechanics,
induces additional safety, when properly designed. There are also Trento, Italy, June 27-29, 2005. In Advanced Experimental Un
situations where the saturation is guaranteed, as is the case with saturated Soil Mechanics EXPERUS 2005. A. Tarantino, E. R
some mining tailings. Even though the suction measurement in omero and Y.J. Cui (eds.). A.A. Balkema Publishers, Leiden:
the field is not common, understanding the importance of suction 43-48.
is a key factor for a proper instrumentation system. Table 12 lists Airò Farulla C., Ferrari A. & Romero E. 2010. Volume change b
ehaviour of a compacted scaly clay during cyclic suction chan
some examples of use of suction measurement sensors in
ges. Can. Geotech. J. 47(6), 688-703.
geotechnical structures. Aitchison G.D. 1965. Soil properties, shear strength and consolidat
ion. In proceeding 6th International Conference Soil Mechanic
5.5 Summary remarks Foundations Engineering 3, 318-321.
Akrouch G.A., Sánchez M. & Briaud J.L. 2016. An experimental,
Monitoring engineering works with a focus on determining analytical and numerical study on the thermal efficiency of en
suction is an unusual procedure. However, more and more the ergy piles in unsaturated soils. Computers and. Geotechnics 71
climatic aspects have demanded monitoring to establish the , 207-220.
Al Qabany A. & Soga K. 2013. Effect of chemical treatment use
suction profile. In addition, many situations require an d in MICP on engineering properties of cemented soils. Géote
understanding of the water pressure situation. This section chnique, 63, 331-339.
presented the necessary aspects for the choice and positioning of Al-Badran Y. M. H. 2011. Volumetric yielding behavior of unsatu
suction sensors that can be applied to any type of engineering rated fine-grained soils. PhD thesis, Ruhr-Universität Bochum,
structure. The main aspects to be highlighted are: Bochum, Germany.
The available options of sensor for suction measurement Alferi M. 2011. Theoretical and experimental study of soil remedi
ation systems based on the mobilization of fluid phases. PhD
allow measurement in a range well suited for engineering Thesis, Politecnico di Torino, Italy.
use. Alferi M., Romero E., Dominijanni A. & Manassero M. 2011. L
The position of the sensor is of fundamental importance not NAPL retention in partially saturated silty sand. Unsaturated S
only for the proper functioning of the equipment but also oils, Two Volume Set (1st ed.), 1445-1450. CRC Press.
for its correct interpretation. Aljundi K., Vieira A., Maranha J., Lapa J. et al. 2020. Effects of
Understanding the possible suction profiles is one of the temperature, test duration and heat flux in thermal conductivit
y measurements under transient conditions in dry and fully sat
fundamental aspects for choosing and positioning the urated states. E3S Web of Conferences 195, 04007.
sensors. Alam M.J.B., Hossain M.S., Sarkar L. & Rahman N. 2019. Evalu
The local climate plays a key role in the suction limits to ation of Field Scale Unsaturated Soil Behavior of Landfill Co
be measured. The preliminary assessment based on the ver through Geophysical Testing and Instrumentation. In Geo-
possible profiles helps in the correct definition of the Congress 2019: Geoenvironmental Engineering and Sustainabili
monitoring system. ty (pp. 1-11). Reston, VA: American Society of Civil Enginee
rs.
Of the sensors presented here, the tensiometer is the most Alonso E.E, Pinyol N. & Gens A. 2013. Compacted soil behaviou
accurate and precise. However, it is the one that requires r: initial state, structure and constitutive modelling. Géotecniqu
the most maintenance when used in the field. e, 63(6), 463-478.
Alonso E.E. 2005. Las catástrofes y el progreso de la geotecnia. I
n Spanish. Real Academia de Ingeniería, Madrid: 1-80.
6 ACKNOWLEDGEMENTS Alonso E.E., Gens A. & Delahaye C.H. 2003. Influence of rainfal
l on the deformation and stability of a slope in overconsolidat
The authors wish to acknowledge Dr Alessandro Fraccica, Dr ed clays: A case study. Hydrogeology Journal 11(1): 174-192.
Laura Gonzalez-Blanco, Dr Joel Torres, Dr Juliana Knobelsdorf, Alonso E.E., Gens A. & Hight D. 1987. Special problem soils. G
Dr Michael Habte, Dr Ahmad Jafari, Dr Babak Shahbodagh, Dr eneral Report. Proc. 9th Eur. Conf. on Soil Mech. and Found.
Soodeh Samimi and Dr Mahnoush Gharehdaghikhajehghiasi for Engng, (Dublin, Ireland) 3: 1087-1146.
Alonso E.E., Gens A. & Josa A. 1990. A constitutive model for
their kind help while writing this state-of-the-art. The authors
partially saturated soils. Géotechnique 40(3), 405-430.
thank Prof. Adrian Russell and Prof. Anthony Leung for critically Alonso E.E., Vaunat J. & Gens A. 1999a. Modelling the mechani
reading the manuscript and making several useful remarks. cal behaviour of expansive clays. Engineering geology 54(1-2),
173-183.
Alonso E.E., Lloret A. & Romero E. 1999b. Rainfall induced defr
7 REFERENCES ormations of road embankments. Rivista Italiana di Geotecnica
33(1): 8-15.
Abbas M.F., Elkady T.Y. & Al-Shamrani M.A. 2017. Calibrations Alonso E.E., Pereira J. Vaunat J. & Olivella S. 2010. A microstr
for volume change measurements using osmotic suction contro ucturally based effective stress for unsaturated soils. Geotechni
l technique. HBRC Journal, 13(1), 39-46. que, 60, 12, 913–925.
Abuel-Naga H. & Bouazza A. 2010. A novel laboratory technique Alonso E.E., Romero E. & Hoffmann C. 2011. Hydromechanical
to determine the water retention curve of geosynthetic clay li behaviour of compacted granular expansive mixtures: experime
ners. Geosynthetics International 17(5), 313-322. ntal and constitutive study. Géotechnique 61(4), 329-344.
Acikel A.S., Bouazza A., Gates W.P., Singh R.M. & Rowe R.K. Alonso E.E., Romero E. & Ortega E. 2016. Yielding of rockfill i
2020. A novel transient gravimetric monitoring technique impl n relative humidity-controlled triaxial experiments. Acta Geotec
emented to GCL osmotic suction control. Geotextiles and Geo hnica 11(3), 455-477.
membranes 48(6), 755-767. Alonso E.E., Romero E., Hoffmann C. & García-Escudero E. 200
Acikel A.S., Gates W.P., Singh R.M., Bouazza A., Fredlund D.G. 5. Expansive bentonite-sand mixtures in cyclic controlled-sucti
& Rowe R.K. 2018. Time-dependent unsaturated behaviour of on drying and wetting. Engineering Geology 81(3), 213-226.
geosynthetic clay liners. Can. Geotech. J. 55(12), 1824-1836. Alramahi B., Alshibli K.A., Fratta D. & Trautwein S. 2008. A su
Acikel A.S., Singh R.M., Bouazza A., Gates W.P. & Rowe R.K. ction-control apparatus for the measurement of P and S-wave
2015. Applicability and accuracy of the initially dry and initia velocity in soils. Geotechnical Testing Journal 31(1), 12-23.
lly wet contact filter paper tests for matric suction measureme Al-Sharrad M., Wheeler S.J. & Gallipoli D. 2012. Influence of an
nt of geosynthetic clay liners. Géotechnique 65(9), 780-787. isotropy on yielding and critical states of an unsaturated soil.
331
In Unsaturated Soils: Research and Applications 129-136. Spri Barroso M., Touze-Foltz N. & Saidi F.K. 2006. Validation of the use of
nger, Berlin, Heidelberg. filter paper suction measurements for the determination of GCL
Alsherif N.A. & McCartney J.S. 2015. Thermal behaviour of unsaturated water retention curves. Proc. 8th Int. Conf. on Geosynthetics,
silt at high suction magnitudes. Géotechnique 65(9), 703-716. Yokohama, Japan. Goesynthetics, J. Kuwano and J. Koseki (eds.)
Al-Tabbaa, A., & Wood, D.M. 1989. An experimentally based bubble Rotterdam, 171-174. ISBN 90-5966-044-7
model for clay. In International Symposium on Numerical Models in Basson J.A., Broekman A. & Jacobsz S.W. 2021. TD-DAQ: A low-cost
Geomechanics. Elsevier Applied Science Publishers, 91-99. data acquisition system monitoring the unsaturated pore pressure
Andrew M., Bijeljic B. & Blunt, J. M. 2014 Pore-by-pore capillary regime in tailings dams. Hardware X10, e00221.
pressure measurements using X-ray microtomography at reservoir Bathurst R.J., Ho A.F. & Siemens G. 2007. A column apparatus for
conditions: Curvature, snap-off, and remobilization of residual CO2, investigation of 1-D unsaturated-saturated response of sand-
Water Resour. Res., 50, 8760–8774. geotextile systems. Geotechnical Testing Journal 30(6), 433-441.
Anselmucci F., Andò E., Viggiani G., Lenoir N. et al. 2021. Use of x-ray Bathurst R.J., Siemens G. & Ho A.F. 2009. Experimental investigation of
tomography to investigate soil deformation around growing roots. infiltration ponding in one-dimensional sand-geotextile columns.
Géotechnique Letters 11(1), 96-102. Geosynthetics International 16(3), 158–172.
Archer A. & Ng C.W.W. 2018. A new environmental chamber for the Beckett C.T.S. & Fourie A.B. 2018. Flow visualisation in a geotechnical
HKUST centrifuge facility. Physical Modelling in Geotechnics 1(71), centrifuge under controlled seepage conditions. Physical Modelling
489-494. A. McNamara, S. Divall, R. Goodey, N. Taylor, S. in Geotechnics 2, 823-828.
Stallebrass and J. Panchal (eds.). CRC Press, London. A. McNamara, S. Divall, R. Goodey, N. Taylor, S. Stallebrass and J.
Artoni R., Santomaso A.C., Gabrieli F., Tono D. et al. 2013. Collapse of Panchal (eds.). CRC Press, London. Beckett C.T.S., Augarde C.E.,
quasi-two-dimensional wet granular columns. Physical Review E Easton D. & Easton T. 2018. Strength characterisation of soil-based
87(3), 032205. construction materials. Géotechnique 68(56), 400-409.
AS2870, 2011. Australian Standard: Residential Slabs and Footings. Beddoe R.A., Take W.A. & Rowe R.K. 2010. Development of suction
Standards Australia. measurement techniques to quantify the water retention behaviour of
Askarinejad A., Beck A. & Springman S.M. 2014. Scaling law of static GCLs. Geosynthetics International 17(5), 301-312.
liquefaction mechanism in geocentrifuge and corresponding Bellia Z., Ghembaza M.S. and Belal T. 2015. A thermo-hydro-
hydromechanical characterization of an unsaturated silty sand mechanical model of unsaturated soils based on bounding surface
having a viscous pore fluid. Can. Geotech. J. 52(6), 708–720. plasticity. Computers and Geotechnics 69, 58-69.
Askarinejad A., Casini F., Bischof P., Beck A. et al. 2012. Rainfall Benatti J.C.B., Miguel M.G., Rodrigues R.A. & Vilar O.M. 2011.
induced instabilities: a field experiment on a silty sand slope in Collapsibility study for tropical soil profile using oedometric test
northern Switzerland. Rivista Italiana di Geotecnica RIG (3), 50-71. with controlled suction. In Unsaturated soils (eds E.E. Alonso and A.
Assadi-Langroudi A., O’Kelly B.C., Barreto D., Cotecchia F. et al. 2022 Gens), CRC Press/Balkema 1, 193–198.
Recent advances in nature-inspired solutions for ground engineering Benatti J.C.B., Rodrigues R.A. & Miguel M.G. 2013. Aspects of
(NiSE). International Journal of Geosynthetics and Ground mechanical behavior and modeling of a tropical unsaturated soil.
Engineering 8(1), 1-36. Geotechnical and Geological Engineering 31(5), 1569-1585.
ASTM D4546-21. Standard test methods for one-dimensional swell or Berg S., Ott H., Klapp S.A., Schwing A. et al. 2013. Realtime 3D Imaging
collapse of soils. ASTM International. of Haines Jumps in Porous Media Flow. Proc. Natl. Acad. Sci. U.S.A.
ASTM D7263-21. Standard test method of laboratory determination of 110(10), 3755–3759.
density and unit weight of soil specimens. ASTM International. Bhattacherjee D. & Viswanadham B.V.S. 2018. Development of a rainfall
Atkinson J. 2007. The mechanics of soils and foundations, CRC Press, p. simulator in centrifuge using modified Mariotte’s principle. Physical
480. Modelling in Geotechnics 1(45), 337-342. CRC Press.
Atique A., Sánchez M. & Romero E. 2009. Investigation of crack Bian X., Cui Y.J., Zeng L.L. & Li X.Z. 2020. State of compacted
desiccation in soil from a flood protection embankment. Proc. 4th bentonite inside a fractured granite cylinder after infiltration.
Asia-Pacific Conference on Unsaturated Soils, Newcastle, Australia, Applied Clay Science, 186, 105438.
Taylor & Francis, London, 1, 51–56. Biarez J., Fleureau J.M., Zerhouni M.I., & Soepandji B.S. 1987.
Aung K.K., Rahardjo H., Leong E.C. & Toll D.G. 2001. Relationship Variations de volume des sols argileux lors de cycles de
between porosimetry measurement and soil-water characteristic drainagehumidification. Revue Francaise de Geotech. 41, 63–72.
curve for an unsaturated residual soil. Geotechnical and Geological Biglari M., Jafari M.K., Shafiee A., Mancuso C. & D’Onofrio A. 2011.
Engineering 19, 401-416. Shear modulus and damping ratio of unsaturated kaolin measured by
Aversa S. & Nicotera M.V. 2002. A triaxial and oedometer apparatus for new suction-controlled cyclic triaxial device. Geotechnical Testing
testing unsaturated soils. Geotechnical Testing Journal 25(1), 3-15. Journal 34(5), 525-536.
Axelsson M. 2006. Mechanical tests on a new non-cementitious grout, Bishop A.W. 1959. The principle of effective stress, Teknisk Ukeblad 106,
silica sol: A laboratory study of the material characteristics. 859–863.
Tunnelling and Underground Space Technology 21(5), 554-560. Bishop A.W. & Blight G.E. 1963. Some aspects of effective stress in
Bahmani M., Noorzad A., Hamedi J. & Salimi F. 2017. Improving sand saturated and partly saturated soils. Géotechnique 13(3), 177-197.
properties using microbial-induced calcite precipitation method. Bishop A.W. and Donald, I.B. 1961. The experimental study of partly
Journal Clean WAS 1(2), 1-5. saturated soil in triaxial apparatus. Proc., 5th Int. Conf. SMFE,
Baker R. & Frydman S. 2009. Unsaturated soil mechanics. Critical Dunod, Paris, 13–21.
review of physical foundations. Engineering Geology 106(1–2): 26– Bittelli M & Flury M. 2009. Errors in water retention curves determined
39. with pressure plates. Soil Sci. Sc. Am. J. 73(5), 1453-1460.
Bakker G., van der Ploeg M.J., de Rooij G.H., Hoogendam C.W., Gooren Biot M. A. 1941. General theory of three-dimensional consolidation. J.
H.P., Huiskes C., Koopal L.K. & Kruidhof H. 2007. New polymer Appl. Phys. 12, 155–164.
tensiometers: Measuring matric pressures down to the wilting point. Biot M. A. 1955. Theory of elasticity and consolidation for a porous
Vadose zone journal 6(1), 196-202. anisotropic solid. J. Appl. Phys. 26, 182-185.
Bandai T., Sadeghi M., Babaeian E., Tuller M. et al. 2021. Blanco A., Lloret A., Carrera J. & Olivella S. 2013. Thermo-hydraulic
Characterization of Unsaturated Water Flow in Soils Using Short- behaviour of the vadose zone in sulphide tailings at Iberian Pyrite
Wave Infrared Imaging through Inverse Modeling [Abstract]. ASA, Belt: Waste characterization, monitoring and modelling.
CSSA, SSSA International Annual Meeting, Salt Lake City, UT. Engineering Geology 165, 154-170.
Banerjee A., Puppala A.J. & Hoyos L.R. 2020. Suction-controlled Blatz J. & Graham J. 2000. A system for controlled suction in triaxial
multistage triaxial testing on clayey silty soil. Engineering Geology tests. Géotechnique 50(4), 465-469.
265, 105409. Blatz J., Cui Y.J. & Oldecop L. 2008. Vapour equilibrium and osmotic
Bannour H., Stoltz G., Delage P. & Touze-Foltz N. 2014. Effect of stress technique for suction control. Geotechnical and Geological
on water retention of needlepunched geosynthetic clay liners. Engineering 26 (6): 661–673.
Geotextiles and Geomembranes, 42(6), 629-640. Blight, G.E. 1967. Effective stress evaluation for unsaturated soils. J. Soil
Barbosa E.E.M., Prado R.L., Mendes, R.M. & Marinho, F.A.M. 2010. Mech. Found. Div., 932, 125–148. Journal of Geotechnical
Estimation of Moisture Content Using the GPR Method: A Engineering, ASCE, 104 GT2, 303-304.
Comparative Evaluation in Laboratory and Field Experiments. Bloch P. 1978. Discussion to: Stress state variables in unsaturated soils.
Brazilian Journal of Geophysics, 28, 691-701. (In Portuguese) Bo M.W., Fabius M., Arulrajah A. & Horpibulsuk S. 2015.
332
Environmentally friendly slope stabilization using a soil nail and root Dordrecht.
system in Canada. In Ground improvement case histories: chemical, Burton G.J., Pineda J.A., Sheng D., Airey D.W. & Zhang F. 2016.
electrokinetic, thermal and bioengineering methods. B. Indraratna, J. Exploring one-dimensional compression of compacted clay under
Chu and C. Rujikiatkamjorn (Eds.). Butterworth-Heinemann 629– constant degree of saturation paths. Géotechnique 66(5), 435–440.
654. Burton G.J., Pineda J.A., Sheng D., Airey D.W. et al. 2017. Reply:
Bocking K.A. & Fredlund D.G. 1980. Limitations of the axis translation Exploring one-dimensional compression of compacted clay under
technique. Proc. 4th Int. Conf. on Expansive Soils, Denver, Colorado, constant degree of saturation paths. Géotechnique 67(1), 86-90.
117-135. Buscarnera G. & Einav I. 2012. The yielding of brittle unsaturated
Bolzon G., Schrefler B.A. & Zienkiewicz O.C. 1996. Elastoplastic soil granular soils. Géotechnique 62(2): 147–160.
constitutive laws generalized to partially saturated states. Buscarnera G. & Nova R. 2009. An elastoplastic strainhardening model
Géotechnique 46(2), 279-289. for soil allowing for hydraulic bonding–debonding effects.
Borana L., Yin J.H., Singh D.N. & Shukla S.K. 2015. A modified suction- International journal for numerical and analytical methods in
controlled direct shear device for testing unsaturated soil and steel geomechanics 33(8), 1055-1086.
plate interface. Marine Georesources & Geotechnology 33(4), 289- Buscarnera G. & Nova R. 2011. Modelling instabilities in triaxial testing
298. on unsaturated soil specimens. International journal for numerical
Borges J.A., Pires L.F., Cassaro F.A., Roque W.L. et al. 2018. X-ray and analytical methods in geomechanics 35(2), 179-200.
microtomography analysis of representative elementary volume Cabrejos F. 2017. Effect of moisture content on the flowability of crushed
(REV) of soil morphological and geometrical properties. Soil and ores. EPJ Web of Conferences 140, 08011.
Tillage Research, 182, 112-122. Cai G., Han B., Zhou A., Li J. & Zhao C. 2022. Fractional-order
Borghei A., Ghayoomi M. & Turner M. 2020. Centrifuge tests to evaluate bounding surface model for unsaturated soils under cyclic loading.
seismic settlement of shallow foundations on unsaturated silty sand. Computers and Geotechnics 141, 104529.
Geo-Congress 2020: Geotechnical Earthquake Engineering and Cai G.Q., Zhao C.G., Li J. & Liu Y. 2014. A new triaxial apparatus for
Special Topics, Minnesota, 318, 198-207. testing soil water retention curves of unsaturated soils under different
Borja R.I. 2004. Cam-Clay plasticity. Part V: A mathematical framework temperatures. Journal of Zhejiang University SCIENCE A 15(5),
for three-phase deformation and strain localization analyses of 364-373.
partially saturated porous media. Computer methods in applied Cai, G.Q., Zhao, C.G., & Qin, X.M. 2010. Structural bonding-breakage
mechanics and engineering 193(48-51), 5301-5338. constitutive model for natural unsaturated clayey soils. Acta
Borja R.I. 2006. On the mechanical energy and effective stress in Mechanica Sinica 26(6), 931-939.
saturated and unsaturated porous continua. International Journal of Caicedo B. & Thorel L. 2014. Centrifuge modelling of unsaturated soils.
Solids and Structures 43, 1764–1786. Journal of Geo-Engineering Sciences 2, 83-103.
Borja R.I. & Koliji A. 2009. On the effective stress in unsaturated porous Caicedo B. 2017. Physical modelling of freezing and thawing of
continua with double porosity. Journal of the Mechanics and Physics unsaturated soils. Géotechnique 67(2), 106-126.
of Solids, 57, 1182–1193. Caicedo B., Tristancho J. & Thorel L. 2015. Mathematical and physical
Borja R.I., Song, X., & Wu, W. 2013. Critical state plasticity. Part VII: modelling of rainfall in centrifuge. Int. Journal of Physical
Triggering a shear band in variably saturated porous media. Modelling in Geotechnics. A. McNamara, S. Divall, R. Goodey, N.
Computer Methods in Applied Mechanics and Engineering 261, 66- Taylor, S. Stallebrass and J. Panchal (eds.). CRC Press, London
82. 15(3), 150-164.
Borsic A., Comina C., Foti S., Lancellotta R. & Musso G. 2005. Imaging Callari, C., Armero, F., & Abati, A. 2010. Strong discontinuities in
heterogeneities with electrical impedance tomography: laboratory partially saturated poroplastic solids. Computer Methods in Applied
results. Géotechnique 55(7), 539-547. Mechanics and Engineering 199(23-24), 1513-1535.
Boso M., Romero E. & Tarantino A. 2003. The use of different Cammeraat E., van Beek R. & Kooijman A. 2005. Vegetation succession
measurement techniques to determine water retention curves. Proc. and its consequences for slope stability in SE Spain. Plant and Soil
Int. Conf. Mechanics of Unsaturated Soils, Weimar, Germany, 278, 135–147.
September 18-19, 2003. 93 Springer Proceedings in Physics Campbell G.S. 1979. Improved thermocouple psychrometers for
(Volume 1). Unsaturated Soils: Experimental Studies. T. Schanz measurement of soil water potential in a temperature gradient. J.
(ed.). Springer-Verlag, Berlin, 1, 169-181. Phys. E: Sci. Instrum. 12, 739-743.
Bouazza A., Zornberg J., McCartney J.S. & Singh R.M. 2013. Cao Z., Chen J., Alonso E.E., Tarragona A.R., Cai Y., Gu C. & Zhang Q.
Unsaturated geotechnics applied to geoenvironmental engineering 2021. A constitutive model for the accumulated strain of unsaturated
problems involving geosynthetics. Engineering Geology 165, 143-
153. Numerical and Analytical Methods in Geomechanics 45(7), 990-
Bowen R.M. 1976. Theory of mixtures. In Continuum Physics, vol. III, 1004.
Eringen, AC (ed.). Academic Press: New York, 1-127. Cardoso R., Maranha das Neves E. & Alonso E.E. 2012. Experimental
Brackley I.J. 1971. Partial collapse in unsaturated expansive clay. CSIR. behaviour of compacted marls. Géotechnique 62(11), 999–1012.
Brooks R., & Corey A. 1964. Hydraulic properties of porous media. Cardoso R., Pires I., Duarte S.O.D & Monteiro G.A. 2018. Effects of
Hydrology Paper No.3, Colorado State University, Fort Collins, CO. clay’s chemical interactions on biocementation, Appl. Clay Sci. 156,
Bruno A.W. & Gallipoli D. 2019. A coupled hydromechanical bounding 96-103.
surface model predicting the hysteretic behaviour of unsaturated Cardoso R., Romero E., Lima A. & Ferrari A. 2007. A comparative study
soils. Computers and Geotechnics 110, 287-295. of soil suction measurement using two different high-range
Brace W.F. 1965. Some new measurements of linear compressibility of psychrometers. In Experimental Unsaturated Soil Mechanics 79-93.
rocks, J. Geophys. Res., 70, 391-398. Springer-Verlag, Berlin.
Bruno A.W., Gallipoli D., Rouainia M. & Lloret-Cabot M. 2020. A Cardoso R., Sarapajevaite G. Korsun O., Cardoso S. & Ilharco L. 2017.
bounding surface mechanical model for unsaturated cemented soils Microfabricated sol-gel relative humidity sensors for soil suction
under isotropic stresses. Computers and Geotechnics 125, 103673. measurement during laboratory tests. Can. Geotech. J. 54(8), 1176-
Buenfil C., Romero E., Lloret A. & Gens A. 2005. Experimental study 1183.
on the hydro-mechanical behaviour of a silty clay. In Unsaturated Carminati A., Moradi A.B., Vetterlein D., Vontobel P. et al. 2010.
Soils - Advances in Testing, Modelling and Engineering Dynamics of soil water content in the rhizosphere. Plant and Soil,
Applications. CRC Press, 15-28. 332, 163–176.
Bui T.A., Wong H., Deleruyelle F., Xie L.Z. & Tran D.T. 2017. A Carminati A., Vetterlein D., Koebernick N., Blaser S. et al. 2013. Do roots
thermodynamically consistent model accounting for viscoplastic mind the gap?. Plant and Soil 367, 651–661.
creep and anisotropic damage in unsaturated rocks. International Caruso M. & Tarantino A. 2004. A shearbox for testing unsaturated soils
Journal of Solids and Structures 117, 26-38. at medium to high degrees of saturation. Géotechnique 54(4), 281-
Bull D.J., Smethurst J.A., Meijer G.J., Sinclair I. et al. 2022. Modelling 284.
of stress transfer in root-reinforced soils informed by four- Casini F., Vaunat J., Romero E., Desideri A. 2012. Consequences on
dimensional X-ray computed tomography and digital volume water retention properties of double-porosity features in a compacted
correlation data. Proc. R. Soc. A 478: 20210210. silt. Acta Geotech 7(2):139–150.
Bulut R. & Leong E.C. 2008. Indirect measurement of suction. In Castellan G.W. 1983. Physical chemistry. Addison-Wesley Publishing
Laboratory and field testing of unsaturated soils 21-32. Springer, Company, Reading, Massachusetts, 3rd ed.
333
Castellanos E., Villar M.V., Romero E., Lloret A. et al. 2008. Chemical on unsaturated soils (unsat'95), Paris, France. volume 2.
impact on the hydro-mechanical behaviour of high-density FEBEX Cui Y.J., Ta A.N., Hemmati S., Tang A.M. et al. 2013. Experimental and
bentonite. Physics and Chemistry of the Earth, Parts A/B/C, 33, numerical investigation of soil-atmosphere interaction. Engineering
S516-S526. Geology 165, 20-28.
Castro G., Pires J., Motta R., Bernucci L., Marinho F. & Merheb A. 2021. Cui Y.J., Tang A.M., Mantho A.T. & De Laure E. 2008. Monitoring field
Unsaturated numerical analysis of a railroad track substructure soil suction using a miniature tensiometer. Geotechnical Testing
considering climate data. Transportation Geotechnics 31, 100662. Journal 31(1), 95–100.
Chávez C., Romero E. & Alonso E.E. 2009. A rockfill triaxial cell with Cui Y.J., Yahia-Aissa M. & Delage P. 2002. A model for the volume
suction control. Geotechnical Testing Journal 32(3), 1–13. change behavior of heavily compacted swelling clays. Engineering
Chen L., Liu Y.M., Wang J., Cao S.F. et al. 2014. Investigation of the geology 64(2-3), 233-250.
thermal-hydro-mechanical (THM) behavior of GMZ bentonite in the Cui Y.J. & Zornberg J.G. 2008. Water balance and evapotranspiration
China-Mock-up test. Engineering geology 172, 57-68. monitoring in geotechnical and geoenvironmental engineering.
Chen X., Pao W., Thornton S. & Small J. 2016. Unsaturated hydro- Geotech. Geol. Eng. 26(6), 783-798.
mechanical–chemical constitutive coupled model based on mixture Cuisinier O. & Masrouri F. 2005. Hydromechanical behaviour of a
coupling theory: Hydration swelling and chemical osmosis. compacted swelling soil over a wide suction range. Engineering
International Journal of Engineering Science 104, 97-109. Geology 81(3), 204-212.
Chen Y., Gao Y., Ng C.W.W. & Guo H. 2021. Bio-improved hydraulic Culligan, K. A., Wildenschild, D., Christensen, B. S. B., Gray, W. G., &
properties of sand treated by soybean urease induced carbonate Rivers, M. L. 2006. Pore-scale characteristics of multiphase flow in
precipitation and its application Part 1: Water retention ability. porous media: A comparison of air–water and oil–water experiments.
Transportation Geotechnics, 27, 100489. Advances in Water Resources, 29(2), 227-238.
Chen Y.G., Zhu C.M., Ye W.M., Cui Y.J. et al. 2015. Swelling pressure Cunningham, M.R., Ridley, A.M., Dineen, K., & Burland, J. B. 2003. The
and hydraulic conductivity of compacted GMZ01 bentonite under mechanical behaviour of a reconstituted unsaturated silty clay.
salinization–desalinization cycle conditions. Appl. Clay Sci. 114, Géotechnique 53(2), 183-194.
454-460. Dai S. & Santamarina J.C. 2013.
Chen Z.G., Tang C.S., Shen Z., Liu Y.M. et al. 2017. The geotechnical bearing sediments. Geophysical Research Letters 40(21), 5637-5641.
properties of GMZ buffer/backfill material used in high-level Dangla P., Malinsky L., & Coussy O. 1997. Plasticity and imbibition-
radioactive nuclear waste geological repository: a review. drainage curves for unsaturated soils: A unified approach. Proc. 6th
Environmental Earth Sciences 76(7), 270. Int. Symp. on Numerical Models in Geomechanics, Montreal: 141–
Cheng A.H.D. 2016. Poroelasticity. Springer. P. 877. 146.
Cheng L., Cord-Ruwisch R. & Shahin M.A. 2013. Cementation of sand Dangla, P., & Pereira, J. M. 2014. A thermodynamic approach to effective
soil by microbially induced calcite precipitation at various degrees stresses in unsaturated soils incorporating the concept of partial pore
of saturation. Can. Geotech. J. 50, 81-90. deformations. Vadose Zone Journal 13(5).
Cheng L., Shahin M.A. & Mujah D. 2017. Influence of key De Boer R. 2000. Theory of porous media, highlights in the historical
environmental conditions on microbially induced cementation for development and current state. Springer, Berlin/New York, 618pp.
soil stabilization. J. Geotech. Geoenviron. Eng. 143(1), 04016083. De Boer R. 2005. The engineer and the scandal: a piece of science history.
Cheng, Z., Wang, J., Coop, M. R., & Ye, G. 2020. A miniature triaxial Springer, Berlin/Heidelberg, 293pp.
apparatus for investigating the micromechanics of granular soils De Boer R. & Ehlers W. 1990. The development of the concept of
with in situ X-ray micro-tomography scanning. Frontiers of effective stresses. Acta Mech 83(1–2), 77–92.
Structural and Civil Engineering. 14(2): 357–373 3. De Carteret R.S., Buzzi O., Fityus S. & Liu X. 2014. Effect of Naturally
Chiu C.F. & Ng C.W.W. 2003. A state-dependent elasto-plastic model for Occurring Salts on the Tensile and Shear Strength of Bitumen-Sealed
saturated and unsaturated soils. Géotechnique 53(9), 809-829. Unbound Granular Road Pavements. Journal of Materials in Civil
Comina C., Foti S., Musso G. & Romero E. 2008. An advanced cell to Engineering, ASCE, USA, 26(6), 1-13.
monitor spatial and time variability in soil with electrical and seismic De Gennaro V. & Pereira J.M. 2013. A viscoplastic constitutive model
measurements. Geotechnical Testing Journal 31, 404-412. for unsaturated geomaterials. Computers and Geotechnics 54, 143-
Comino E. & Druetta A. 2010. The effect of Poaceae roots on the shear 151.
strength of soils in the Italian alpine environment. Soil & Tillage De Gennaro V., Pereira J.M., Gutierrez M. & Hickman R.J. 2009.
Research 106, 194–201 Viscoplastic modeling of fluids filled porous chalks. Rivista Italiana
Consoli N.C., Ferreira P.M.V., Tang C.S., Marques S.F.V. et al. 2016. A Di Geotechnica 1, 44-64.
unique relationship determining strength of silty/clayey soils – Delage P. & Pellerin M. 1984. Influence de la lyophilization sur la
Portland cement mixes. Soils and Foundations 56(6), 1082–1088. structure d’une argile sensible du Quebec. Clay Minerals 19, 151-
Corbin A & Augarde C. 2014. Fracture energy of stabilized rammed earth. 160.
Procedia Materials Science 3, 1675-1680. Delage P. & Romero E. 2008. Geoenvironmental testing. Geotechnical
Costa S., Kodikara J.K. & Shannon B. 2013. Salient factors controlling and Geological Engineering 26(6), 729-749.
desiccation cracking of clay in laboratory experiments. Delage P. & Cui Y. J. 2008. An evaluation of the osmotic method of
Géotechnique 63(1), 18-29. controlling suction. Geomechanics and Geoengineering 3 (1), 1.
Coussy O. 1995. Mechanics of porous continua, John Wiley, New York. Delage P. 2004. Experimental unsaturated soil mechanics. Proc. 3rd Int.
Coussy O. 2010. Mechanics of porous solids, John Wiley, New York. Conference on Unsaturated Soils 3, 973-996.
Coussy O., Pereira J.M. & Vaunat, J. 2010. Revisiting the Delage P., Audiger M., Cui Y.J. & Howat M. 1996. Microstructure of a
thermodynamics of hardening plasticity for unsaturated soils. compacted silt. Can. Geotech. J. 33(1): 150-158.
Computers and Geotechnics 37(1-2), 207-215. Delage P., Howat M.D. & Cui Y.J. 1998. The relationship between
Cruz J.A., Hoyos L.R. & Lizcano A. 2012. Unsaturated soil response suction and swelling properties in a heavily compacted unsaturated
under plane strain conditions using a servo/suction-controlled clay. Engineering Geology 50, 31-48.
biaxial apparatus. In: Mancuso C., Jommi C., D’Onza F. (eds) Delage P., Romero E. & Tarantino A. 2008. Recent developments in the
Unsaturated Soils: Research and Applications. Springer, Berlin, techniques of controlling and measuring suction in unsaturated soils.
Heidelberg, 31-38. Proc. 1st Int. Conference on Unsaturated Soils, E-UNSAT 2008, 33-
Cuadrado A., Encalada D., Ledesma A. & Prat P.C. 2019. Soil surface 52, Durham, United Kingdom. hal-00331841
boundary condition in desiccating soils. Proc. of the XVII ECSMGE Dell’Avanzi E., Zornberg J.G. & Cabral A.R. 2004. Suction profiles and
Reykjavik, Iceland, 1-6 September 2019. Geotechnical Engineering, scale factors for unsaturated flow under increased gravitational field.
foundation of the future. Soils and Foundations 44(3), 79–89.
Cui K. & Zhao T. 2017. Unsaturated dynamic constitutive model under Della Vecchia G., Jommi C. & Romero E. 2013. A fully coupled elastic-
cyclic loading. Cluster Computing 20(4), 2869-2879. plastic hydromechanical model for compacted soils accounting for
Cui Y.J. & Delage P. 1996. Yielding and plastic behaviour of an clay activity. Int. J. Numer. Anal. Meth. Geomech. 37: 503-535.
unsaturated compacted silt. Géotechnique 46(2), 291-311. Della Vecchia G., Joomi C. & Romero E. 2011. Radial stress paths in
Cui Y.J., Delage P. & Alzoghbi P. 2003. Retention and transport of a triaxial tests on compacted Boom clay: stiffness and anisotropy. Proc.
hydrocarbon in a silt. Géotechnique 53(1), 83-91. 5th Asian-Pacific Conference on Unsaturated Soils, Pattaya,
Cui, Y.J., Delage, P., & Sultan, N. 1995. An elasto-plastic model for Thailand, 273-279.
compacted soils. In proceedings of the first international conference Desrues J., Chambon R., Mokni M. & Mazerolle F. 1996. Void ratio
334
evolution inside shear bands in triaxial sand specimens studied by Hunter Valley (Newcastle, NSW 11-13 July) p. 251-265.
computed tomography, Geotéchnique 46(3), 529-546. Fleureau, J.M., Kheirbek-Saoud, S., Soemitro, R., & Taibi, S. 1993.
Dineen K. & Burland J.B. 1995. A new approach to osmotically Behavior of clayey soils on drying–wetting paths. Canadian
controlled oedometer testing. In E.E. Alonso & P. Delage (eds.), Geotechnical Journal 30(2), 287-296.
Unsaturated Soils, Proc. 1st Int. Conf. on Unsaturated Soils, Paris, 2, Foresta V., Capobianco V. & Cascini L. 2020. Influence of grass roots on
459-465. Rotterdam: Balkema. shear strength of pyroclastic soils. Can. Geotech. J. 57(9), 1320–
Dondi M., Iglesias C., Dominguez E., Guarini G. & Raimondo M. 2008. 1334.
The effect of kaolin properties on their behaviour in ceramic Fraccica A. 2019. Experimental Study and Numerical Modelling of Soil-
processing as illustrated by a range of kaolins from the Santa Cruz Roots Hydro-Mechanical Interactions. Ph.D. thesis Universitat
and Chubut Provinces, Patagonia (Argentina). Appl. Clay Sci. 40(1), Politècnica de Catalunya, Université de Montpellier.
143-158. Fraccica A., Romero E. & Fourcaud T. 2019. Multi-scale effects on the
Dong H., Huang R. & Gao Q.F. 2017. Rainfall infiltration performance hydraulic behaviour of a root-permeated and compacted soil. In IS-
and its relation to mesoscopic structural properties of a gravelly soil Glasgow 2019 Proc. 7th International Symposium on Deformation
slope. Engineering Geology 230, 1–10. Characteristics of Geomaterials, 1-5. EDP Sciences.
D’Onza F., Gallipoli D. & Wheeler S.J. 2010. Effect of anisotropy on the Fraccica A., Romero E. & Fourcaud T. 2022a. Tensile strength of a
prediction of unsaturated soil response under triaxial and oedometric compacted vegetated soil: Laboratory results and reinforcement
conditions. Unsaturated soils 2, 787-794. interpretation. Geomechanics for Energy and the Environment,
D'Onza F., Gallipoli D., Wheeler S., Casini F., Vaunat J., Khalili N., 100303.
Laloui L., Mancuso C., Mašín D., Nuth M. & Pereira J.M. 2011. Fraccica A., Romero E. & Fourcaud T. 2022b. Large-scale triaxial tests
Benchmark of constitutive models for unsaturated soils. of vegetated soil at low confining stresses. In: EGU 2022 General
Géotechnique 61(4), 283-302. Assembly, May 2022, Wien (Austria).
Dueck A. 2004. Hydro-mechanical properties of a water unsaturated Fraccica A., Spagnoli G., Romero E., Arroyo M. & Gómez R. 2021.
sodium bentonite. Laboratory study and theoretical interpretation. Exploring the mechanical response of low-carbon soil improvement
PhD Thesis, Lund University, Sweden. mixtures. Can. Geotech. J. Just-IN https://doi.org/10.1139/cgj-2021-
Dumont M., Taibi S., Fleureau J.M., Abou Bekr N. & Saouab A. 2010. 0087.
Modelling the effect of temperature on unsaturated soil behaviour. Fredlund D.G. 2015. Relationship between the laboratory SWCCs and
Comptes Rendus Geoscience 342(12), 892-900. field stress state. Raton, FL: CRC Press.
Duong T.V., Trinh V.N., Cui Y.J., Tang A.M. & Calon N. 2013. Fredlund D.G. & Rahardjo H. 1993. Soil mechanics for unsaturated soils.
Development of a Large-Scale Infiltration Column for Studying the John Wiley & Sons, Inc. New York.
Hydraulic Conductivity of Unsaturated Fouled Ballast. Geotechnical Fredlund D.G. 2017. Role of SWCC in unsaturated soil mechanics.
Testing Journal 36(1), 1–10. Geoffrey Blight Lecture. Proc. 19th Int. Conf. on Soil Mechanics and
Ebrahimi-Birang N. & Fredlund D. 2016. Assessment of the WP4-T Geotechnical Engineering, Seoul: 57-80.
Device for Measuring Total Suction. Geotechnical Testing Journal, Fredlund D.G. 2019. Determination of unsaturated soil property
39(3), 20150118. functions for engineering practice. Jerry Jennings Lecture, Proc. 17th
Edlefsen N.E. & Anderson A.B.C. 1943. The thermodynamics of soil African Regional Conference on Soil Mechanics and Geotechnical
moisture. Hilgardia 16, 31–299. Engineering, Cape Town South Africa: 3-19.
Ehlers W., Avci O., & Markert B. 2011. Computation of slope movements Fredlund D.G., Rahardjo H. & Fredlund M.D. 2012. Unsaturated soil
initiated by rain-induced shear bands in small-scale tests and in situ. mechanics in engineering practice. John Wiley & Sons, Inc.
Vadose Zone Journal 10(2), 512–525. Fredlund M.D., Fredlund D.G., Houston S.L. & Houston W.N. 2003.
Ehlers W., Graf T. & Ammann M. 2004. Deformation and localization Assessment of unsaturated soil properties for seepage modeling
analysis of partially saturated soil. Computer methods in applied through tailings and mine wastes. Proc. 10th Int. Conf. Tailings Mine
mechanics and engineering 193(27-29), 2885-2910. Waste, Vail, Colorado, US, 149-157.
Einav . & Liu M. 2018. Hydrodynamic derivation of the work input to Fredlund, D.G., & Morgenstern, N.R. 1977. Stress state variables for
fully and partially saturated soils, Journal of the Mechanics and unsaturated soils. Journal of the geotechnical engineering division
Physics of Solids 110, 205-217. 103(5), 447-466.
El Mountassir G., Sánchez M., Romero E. & Soemitro R.A.A. 2011. Fredlund D., Morgenstern N. & Widger R. 1978. The shear strength of
Behaviour of compacted silt used to construct flood embankment. unsaturated soils. Can. Geotech. J., 15(3), 313–321.
Geotechnical Engineering 164(3), 195-210. Fredlund D.G., Xing A. 1994. Equations for soil-water characteristic
EN 1997-2:2007 Eurocode 7: Geotechnical design Part 2: Ground curve. Canadian Geotechnical Journal 31, 521-532.
investigation and testing. CEN. Fuentes W. & Triantafyllidis T. 2013. On the effective stress for
EN 1998-5:2018 Eurocode 8: Design of structures for earthquake unsaturated soils with residual water. Géotechnique 63(16): 1451-
resistance Part 5: Foundations, retaining structures and geotechnical 1455.
aspects. CEN. Gabrieli F., Artoni R., Santomaso A. & Cola S. 2013. Discrete particle
Eberhardsteiner J., Hofstetter G., Meschke G. & Mackenzie-Helnwein P. simulations and experiments on the collapse of wet granular columns.
2003. Coupled material modelling and multifield structural analysis Physics of Fluids 25(10), 103303.
in civil engineering, Eng. Comput. 20, 524–558. Gabrieli L., Jommi C., Musso G. & Romero E. 2008. Influence of
Escobar A., Caicedo B. & Cabrera M. 2021. Interaction between a electroosmotic treatment on the hydro-mechanical behaviour of
cylinder and a partially saturated soil for compaction analysis. clayey silts: preliminary experimental results. Journal of Applied
Transp. Geotech. 30, 100600. Electrochemistry 38(7): 1043-1051.
Fabbri A., Wong H.K.K., Lai B., Bui T.A. & Branque D. 2019. Gajo, A., & Muir Wood, D. 2001. A new approach to anisotropic,
Constitutive elasto-plastic model for fine soils in the unsaturated to bounding surface plasticity: general formulation and simulations of
saturated saturation zone. Computers and Geotechnics 111, 10-21. natural and reconstituted clay behaviour. International journal for
Fang T., Chen W., Plinke J., Wheeler C. & Roberts A. 2020. Study of the numerical and analytical methods in geomechanics 25(3), 207-241.
wall adhesive tensile contact of moist iron ore bulk solids. Gallage C. & Uchimura T. 2016. Direct shear testing on unsaturated silty
Particuology 50, 67-75. soils to investigate the effects of drying and wetting on shear strength
Fern E.J. & Soga K. 2017. Granular column collapse of wet sand. parameters at low suction. Journal of Geotechnical and
Procedia Engineering 175, 14–20. Geoenvironmental Engineering 142(3), 04015081.
Fillunger P. 1936. Erdbaumechanik? Selbstverlag des Verfassers, Wien. Gallagher P.M. & Mitchell J.K., 2002. Influence of colloidal silica grout
Fillunger P. 1930. Auftrieb und Unterdruck in Staumauern, Transactions on liquefaction potential and cyclic undrained behavior of loose sand.
second World Power Conference 9, VDI-Verlag, Berlin, 323-329. Soil Dynamics and Earthquake Engineering 22, 1017–1026.
Fityus S.G., Cameron D.A. & Walsh P.F. 2005. The shrink swell test. Gallipoli D. & Bruno A.W. 2017. A bounding surface compression model
Geotechnical Testing Journal 28(1): 1-10. with a unified virgin line for saturated and unsaturated soils.
Fityus S., Walsh P., Kleeman P. 1998. The influence of climate as Géotechnique 67(8), 703-712.
expressed by the Thornthwaite Index on the design depth of moisture Gallipoli D., Bruno A.W., D'onza F. & Mancuso C. 2015. A bounding
change of clay soils in the Hunter Valley. Conference on surface hysteretic water retention model for deformable soils.
Geotechnical Engineering and Engineering Geology in the Hunter Géotechnique 65(10), 793-804.
Valley. Geotechnical Engineering and Engineering Geology in the Gallipoli D., Gens A., Chen G. & D’Onza F. 2008. Modelling unsaturated
335
soil behaviour during normal consolidation and at critical state. USA, OTC-20550-MS.
Computers and Geotechnics 35(6), 825-834. Gholizadeh E. & Latifi M. 2018. A coupled hydro-mechanical
Gallipoli D., Gens A., Sharma R. & Vaunat J. 2003. An elasto-plastic constitutive model for unsaturated frictional and cohesive soil.
model for unsaturated soil incorporating the effects of suction and Computers and Geotechnics 98, 69-81.
degree of saturation on mechanical behaviour. Géotechnique 53(1), Ghorbani J., Airey D.W. & El-Zein A. 2018. Numerical framework for
123-135. considering the dependency of SWCCs on volume changes and their
Gallipoli D., Gens A., Vaunat J. & Romero E. 2002. Role of degree of hysteretic responses in modelling elasto-plastic response of
saturation on the normally consolidated behaviour of soils. In unsaturated soils. Computer Methods in Applied Mechanics and
Proceedings 3rd International Symposium on Unsaturated Soil Engineering 336, 80-110.
(Unsaturated Soils): 115-120. Recife, Brazil. Ghorbani J., Nazem M., & Carter J.P. 2016. Numerical modelling of
Gallipoli D. 2012. A hysteretic soil-water retention model accounting for multiphase flow in unsaturated deforming porous media. Computers
cyclic variations of suction and void ratio. Geotechnique 62(7): 605– and Geotechnics 71, 195-206.
616. Ghorbani J., Airey D.W., Carter J.P. & Nazem M. 2021a. Unsaturated soil
Gao, Y. & Sun D.A. (2017). Soil-water retention behavior of compacted dynamics: Finite element solution including stress-induced
soil with different densities over a wide suction range and its anisotropy. Computers and Geotechnics 133, 104062.
prediction. Comput Geotech 91: 17-26. Ghorbani I. & Airey D.W. 2021b. Modelling stress-induced anisotropy in
Garcia E.F., Gallage C.P.K. & Uchimura T. 2007. Function of permeable multi-phase granular soils. Computational Mechanics 67 (2), 497-
geosynthetics in unsaturated embankments subjected to rainfall 521.
infiltration. Geosynthetics International 14(2), 89–99. Gonzalez-Blanco L., Romero E.E. & Marschall P. 2020. Gas transport in
Gardner W.R. 1956. Calculation of capillary conductivity from pressure granular compacted bentonite: coupled hydro-mechanical
plate outflow data. Soil Sci. Soc. Am. Proc. 20: 317-320. interactions and microstructural features. In Proc. EUNSAT 2020,
Garg A., Coo J.L. & Ng C.W.W. 2015a. Field study on influence of root Oct. 19-21, Lisbon, Portugal E3S Web of Conferences, 195, 04008-
characteristics on soil suction distribution in slopes vegetated with 1. EDP Sciences.
Cynodon dactylon and Schefflera heptaphylla. Earth surface Gonzalez-Ollauri A. & Mickovski S.B. 2017. Plant-soil reinforcement
processes and landforms 40(12), 1631-1643. response under different soil hydrological regimes. Geoderma 285,
Garg A., Leung A.K. & Ng C.W.W. 2015b. Comparisons of soil suction 141-150.
induced by evapotranspiration and transpiration of S. Heptaphylla. Goode J.C. & McCartney J.S. 2015. Centrifuge modeling of end-restraint
Canadian Geotechnical Journal 52(12), 2149–2155. effects in energy foundations. J. Geotech. Geoenviron. Eng. 141(8),
Garino L.M., Oldecop L.A., Romero E. & Rodriguez R.L. 2021. Tailings 04015034.
desiccation process studied in environmental chamber experiment. Gray W.G. & Schrefler B.A. 2001. Thermodynamic approach to effective
Proc. of the Institution of Civil Engineers – Geotechnical stress in partially saturated porous media. Eur. J. Mech. A-solid, 20,
Engineering 1-32, Ahead of Print. 4, 521-538
Gaspar T.A.V., Jacobsz S.W., Smit G. & Osman A.S. 2019. An expansive Gray W.G. & Schrefler B.A. 2007. Analysis of the solid phase stress
clay for centrifuge modelling. Proc. 17th Eur. Conf. Soil Mech. tensor in multiphase porous media. International Journal for
Geotech. Eng. ECSMGE 2019, 421-428. Numerical and Analytical Methods in Geomechanics 31(4), 541-581.
Geiser F. 1999. ‘Comportement me´canique d’un limon non sature:Etude Geertsma J. 1957. The effect of fluid pressure decline on volumetric
experimentale et modelisation constitutive.’’ PhD thesis, Swiss changes of porous rocks, Trans. AIME, 210, 331-340.
Federal Institute of Technology, Lausanne, Switzerland. Geertsma, J. 1966. Problems of rock mechanics in petroleum production
Gens A. 1995. Constitutive modelling. Application to compacted soils. engineering, Proc. 1st Int. Congr. Rock Mechanics, Lisbon, I, R. 359,
Proc. 1st Int. Conf. on Unsaturated Soils (Paris, France) 3, 1179- 585.
1200. Guan Y. & Fredlund D.G. 1997. Use of tensile strength of water for the
Gens A. 2010. Soil-environment interactions in geotechnical engineering. direct measurement of high soil suction. Can. Geotech. J. 34, 604-
Géotechnique 60(1), 3-74. 614.
Gens A. & Alonso E.E. 1992. A framework for the behaviour of Guimarães L.D.N., Gens A., Sánchez M. & Olivella S. 2013. A chemo-
unsaturated expansive clays. Canadian Geotechnical Journal 29(6), mechanical constitutive model accounting for cation exchange in
1013-1032. expansive clays. Géotechnique 63(3), 221-234.
Gens A., Alonso E.E., Suriol J. & Lloret, A. 1995. Effect of structure on Guerriero M. & Mazzoli S. 2021. Theory of effective stress in soil and
the volumetric behaviour of a compacted soil. Proc. 1st Int. Conf. on rock and implications for fracturing processes: A review.
Unsaturated Soils, Paris 1: 83-88. Geosciences 11, 119.
Gens A., Jouanna P. & Schrefler B.A. 1995. Modern issues in non- Hall S.A. 2012. Digital Image Correlation in Experimental
saturated soils. CISM course and Lectures No. 357. Geomechanics. ALERT Doctoral School 2012 Advanced
Gens A., Sánchez M. & Sheng D. 2006. On constitutive modelling of Experimental Techniques in Geomechanics, 69-102.
unsaturated soils. Acta Geotechnica 1(3), 137-147. Hamid T.B. & Miller G.A. 2009. Shear strength of unsaturated soil
Gens A., Vállejan B., Sánchez M., Imbert C., Villar M.V. & Van Geet M. interfaces. Can. Geotech. J. 46(5), 595-606.
2011. Hydromechanical behaviour of a heterogeneous compacted Hamidi A., Habibagahi G. & Ajdari M. 2013. A modified osmotic direct
soil: experimental observations and modelling. Géotechnique 61(5), shear apparatus for testing unsaturated soils. Geotechnical Testing
367-386. Journal 36(1), 20-29.
Georgiadis K., Potts D.M. & Zdravkovic L. 2005. Three-dimensional Hanson J.L., Risken J.L. & Yesiller N. 2013. Moisture-suction
constitutive model for partially and fully saturated soils. relationships for geosynthetic clay liners. Proc. of the 18th Int. Conf.
International Journal of Geomechanics 5(3), 244-255. on Soil Mechanics and Geotechnical Engineering, Paris, 3025-3028.
Ghaffaripour O., Esgandani G.A., Khoshghalb A. and Shahbodagh B. Hassanizadeh S.M., & Gray W.G. 1990. Mechanics and thermodynamics
2019. Fully coupled elastoplastic of multiphase flow in porous-media including interphase boundaries.
unsaturated porous media using a meshfree method. International Advances in Water Resources 13(4):169–186.
Journal for Numerical and Analytical Methods in Geomechanics He L., Leong E.C. & Elgamal A. 2006. A miniature tensiometer for
43(11), 1919-1955. measurement of high matric suction. American Society of Civil
Gharbi, O. & Blunt, J. M., 2012 The impact of wettability and Engineers Reston Va, 897–1907.
connectivity on relative permeability in carbonates: A pore network He Y., Zhang K.N. & Wu D.Y. 2019. Experimental and modeling study
modeling analysis, Water Resour. Res., 48, W12513. of soil water retention curves of compacted bentonite considering
Ghasemzadeh H. & Amiri S.G. 2013. A hydro-mechanical elastoplastic salt solution effects. Geofluids 2019, 4508603.
model for unsaturated soils under isotropic loading conditions. Heath, A.C., Pestana, J.M., Harvey, J.T., & Bejerano, M.O. 2004.
Computers and Geotechnics 51, 91-100. Normalizing behavior of unsaturated granular pavement materials.
Ghasemzadeh H., Sojoudi M.H., Amiri S.G. & Karami M.H. 2017. Journal of Geotechnical and Geoenvironmental Engineering 130(9),
Elastoplastic model for hydro-mechanical behavior of unsaturated 896-904.
soils. Soils and Foundations 57(3), 371-383. Herbert H.J., Kasbohm J., Sprenger H., Fernández A.M. & Reichelt C.
Ghezzehei T.A. & Kneafsey T.J. 2010. Measurements of the capillary 2008. Swelling pressures of MX-80 bentonite in solutions of
pressure-saturation relationship of methane hydrate bearing different ionic strength. Physics and Chemistry of the Earth, Parts
sediments. In Offshore Technology Conference. Houston, Texas, A/B/C 33, S327-S342.
336
Higo Y., Lee C.W., Doi T., Kinugawa T. et al. 2015. Study of dynamic CRC Press, London.
stability of unsaturated embankments with different water contents Jaquin P. A., Augarde C.E., Gallipoli D. & Toll D.G. 2009. The strength
by centrifugal model tests. Soils and Foundations 55(1), 112-126. of unstabilised rammed earth materials. Géotechnique 59(5), 487–
Higo, Y., Oka, F., Sato, T., Matsushima, Y., & Kimoto, S. 2013. 490.
Investigation of localized deformation in partially saturated sand Jayatilaka C.J. & Gillham R.W. 1996. A deterministic-empirical model
under triaxial compression using microfocus X-ray CT with digital of the effect of the capillary fringe on near-stream area runoff 1.
image correlation. Soils and Foundations, 53(2), 181-198. Description of the model. Journal of Hydrology 184(3-4), 299-315.
Hilf J.W. 1956. An investigation of pore-wa¬ter pressure in compacted Jennings J.E.B. & Burland J.B. 1962. Limitations to the use of effective
cohesive soils. PhD Thesis. Technical Memo No.654, United States stresses in partly saturated soils. Géotechnique 12(2), 125-144.
Bureau of Reclamation, Denver. Jeong S., Lee K., Kim J. & Kim Y. 2017. Analysis of rainfall-induced
Hoffman O. 1928. Permeazoni d'Acqua e loro Effeti nei Muri di Ritenuta, landslide on unsaturated soil slopes. Sustainability 9(1280), 1-20.
Hoepli, Milan. Jiang Y., Einav I., & Liu M. 2017. A thermodynamic treatment of
Hoffmann C., Romero E. & Alonso E.E. 2005. Combining different partially saturated soils revealing the structure of effective stress. J.
controlled-suction techniques to study expansive clays. Proc. Int. Mech. Phys. Solids 100 (2017) 131–146.
Symposium on Advanced Experimental Unsaturated Soil Mechanics, Jommi C. 2000. Remarks on the constitutive modelling of unsaturated
Trento, Italy, June 27-29, 2005. Advanced Experimental Unsaturated soils. In: Tarantino A, Mancuso C, editors. Experimental evidence &
Soil Mechanics EXPERUS 2005. A. Tarantino, E. Romero & Y.J. theoretical approaches in unsaturated soils. Rotterdam: Balkema;
Cui (eds.). A.A. Balkema Publishers, Leiden: 61-67. 139–53.
Horpibulsuk S., Rachan R., Chinkulkijniwat A., Raksachon Y. & Jommi C., Muraro S., Trivellato E. & Zwanenburg C. 2019. Experimental
Suddeepong A. 2010. Analysis of strength development in cement- results on the influence of gas on the mechanical response of peats.
stabilized silty clay from microstructural considerations. Géotechnique 69(9), 753-766.
Construction and Building Materials 24, 2011–2021. Josa, A., Balmaceda, A., Gens, A., & Alonso, E. E. 1992. An elastoplastic
Hossain M.A. & Yin J.H. 2012. Influence of grouting pressure on the model for partially saturated soils exhibiting a maximum of collapse.
behavior of an unsaturated soil-cement interface. J. Geotechnical and In 3rd international conference on computational plasticity,
Geoenvironmental Engineering 138(2), 193-202. Barcelona 1, 815-826.
Houlsby G.T. 1997. The work input to an unsaturated granular material, Jotisankasa A. & Mairaing W. 2010. Suction-monitored direct shear
Geotechnique 47(1), 193-196. testing of residual soils from landslide-prone areas. J. Geotechnical
Hounsfield G.N. 1972. A method of and apparatus for examination of a and Geoenvironmental Engineering 136(3), 533-537.
body by radiation such as X- or gamma-radiation. British Patent No Jotisankasa A. & Sirirattanachat T. 2017. Effects of grass roots on soil-
1.283.915, London. water retention curve and permeability function. Can. Geotech. J. 54,
Houston, S.L. 2019. It is time to use unsaturated soil mechanics in routine 1612–1622
geotechnical engineering practice. Journal of Geotechnical and Jotisankasa A., Coop M. & Ridley A. 2007. The development of a suction
Geoenvironmental Engineering 145(5), 5. control system for a triaxial apparatus. Geotechnical Testing Journal
Hoyos L.R. & Macari E.J. 2001. Development of a stress/suction- 30(1), 69-75.
controlled true triaxial testing device for unsaturated soils. Jotisankasa A., Coop M. & Ridley A. 2009. The mechanical behaviour of
Geotechnical Testing Journal 24(1), 5-13. an unsaturated compacted silty clay. Géotechnique 59(5): 415- 428.
Hoyos L.R., Pérez-Ruiz D.D. & Puppala A.J. 2011. Refined true triaxial Juang C.H. & Holtz R.D. 1986. A probabilistic permeability model and
apparatus for testing unsaturated soils under suction-controlled the pore size density function. Int. J. Numer. Anal. Meth. Geomech.
stress paths. Int. Journal of Geomechanics 12(3), 281–291. 10, 543-553.
Hu R. 2013. A water retention curve and unsaturated hydraulic Kaewsong R., Zhou C. & Ng C.W.W. 2019. Modelling effects of recent
conductivity model for deformable soils: consideration of the change suction history on small-strain stiffness of unsaturated soil. Canadian
in pore-size distribution. Geotechnique 16:1389–1405. Geotechnical Journal 56(4), 600-610.
Hu R., Chen Y.F., Liu H.H. & Zhou C.B. 2015. A coupled stress–strain Kamchoom V. & Leung A.K. 2018. Hydro-mechanical reinforcements of
and hydraulic hysteresis model for unsaturated soils: live poles to slope stability. Soils and Foundations 58(6), 1423-1434,
Thermodynamic analysis and model evaluation. Computers and ISSN 0038-0806.
Geotechnics 63, 159-170. Karatza Z., Ando E., Papanicolopulos S.A., Ooi J.Y. & Viggiani G. 2018.
Huang S., Barbour S.L. & Fredlund D.G. 1998. Development and Evolution of deformation and breakage in sand studied using X-ray
verification of a coefficient of permeability function for a tomography. Géotechnique 68(2), 107-117.
deformable unsaturated soil. Can. Geotech. J. 35(3):411–25. Karatza Z., Andò E., Papanicolopulos S.A., Viggiani G. & Ooi J.Y. 2019.
Effect of particle morphology and contacts on particle breakage in a
of a single contaminant. International journal for numerical and granular assembly studied using X-ray tomography. Granular Matter
analytical methods in geomechanics 21(1), 43-72. 21(3), 1-13.
Hueckel, T., Mielniczuk, B., & El Youssoufi, M. S. 2020. Adhesion-force Karatza Z., Buckman J., Medero G.M. & Beckett C.T.S. 2021. Evolution
micro-scale study of desiccating granular material. Géotechnique, of meniscus structures in hydrophobic granular systems. Journal of
70(12), 1133-1144. Hydrology, 603(Part C), 126954.
Huertas F., Fariña P., Farias J., García-Siñériz J.L et al. 2006. Full-scale Karimzadeh A. A., Leung A. K., Hosseinpour S., Wu, Z. et al. 2021.
engineered barrier experiment: Updated final report. Technical Monotonic and cyclic behaviour of root-reinforced sand. Canadian
Publication 05-0/2006. Madrid: Enresa. Geotechnical Journal 58: 1915–1927.
Hutter K., Laloui L. & Vulliet L. 1999. Thermodynamically based Karnland O., Olsson S., Nilsson U. & Sellin P. 2007. Experimentally
mixture models of saturated and unsaturated soils. Mechanics of determined swelling pressures and geochemical interactions of
-338. compacted Wyoming bentonite with highly alkaline solutions.
Ilic D. & Wheeler C.A. 2017. Transverse bulk solid behaviour during Physics and Chemistry of the Earth, Parts A/B/C 32(1-7), 275-286.
discharge from troughed belt conveyors. Advanced Powder Kraube, D. & Kato, S. 1994. An ideal unsaturated soil and the Bishop’s
Technology 28(9), 2410-2430. soil. Proc. 13th Int. Conf. Soil Mechanics and Foundation Eng., 1,
Imbert C. & Villar M.V. 2006. Hydro-mechanical response of a bentonite 43-46.
pellets/powder mixture upon infiltration. Appl. Clay Sci. 32(3-4), Karube, D., Kato, S., Honda, M. and Kawai, K. 1998. A constitutive
197-209. model for unsaturated soil evaluating effects of soil moisture
Infante Sedano J., Vanapalli S.K. & Garga V.K. 2007. Modified ring distribution. Proc. 2nd Int. Conf. on Unsaturated
shear apparatus for unsaturated soils testing. Geotechnical Testing Soils, Beijing, 1, 485-490.
Journal 30(1), 39-47. Karube D. & Kawai K. 2001. The role of pore water in the mechanical
Iskander M. 2018. Transparent soils turn 25: Past, present, and future. behavior of unsaturated soils. Geotechnical & Geological
Physical Modelling in Geotechnics 1, 389-394. A. McNamara, S. Engineering 19(3), 211-241.
Divall, R. Goodey, N. Taylor, S. Stallebrass and J. Panchal (eds.). Kemp, N., Angelidakis, V., Luli, S., & Nadimi, S. 2022. How Do Roots
CRC Press, London. Interact with Layered Soils?. Journal of Imaging 8(1), 5.
Jacobsz S.W. 2018. Low cost tensiometers for geotechnical applications. Keyes S.D., Cooper L., Duncan S., Koebernick N. et al. 2017.
Physical Modelling in Geotechnics 1, 305-310. A. McNamara, S. Measurement of micro-scale soil deformation around roots using
Divall, R. Goodey, N. Taylor, S. Stallebrass and J. Panchal (eds.). four-dimensional synchrotron tomography and image correlation. J.
337
R. Soc. Interface 14, 20170560. Kodikara J. 2012. New framework for volumetric constitutive behaviour
Khaddour G., Riedel I., Andò E., Charrier P.et al. 2018. Grain-scale of compacted unsaturated soils. Can. Geotech. J. 49(11), 1227–1243.
characterization of water retention behaviour of sand using X-ray CT. Koebernick N., Daly K.R., Keyes S.D., George T.S. et al. 2017. High-
Acta Geotechnica, 13(3), 497-512. resolution synchrotron imaging shows that root hairs influence
Khalili N. 2008. Two-phase fluid flow through fractured porous media rhizosphere soil structure formation. New Phytologist 216(1), 124–
with deformable matrix. Water Resources Research 44, W00C04. 135.
Khalili N. 2018. Guidelines for the application of effective stress Kohgo Y., Nakano M. & Miyazaki T. 1993. Theoretical aspects of
principle for shear strength and volume change determination in constitutive modelling for unsaturated soils. Soils and foundations
unsaturated soils. Australian Geomechanics Journal 53 (1), 37-47. 33(4), 49-63.
Khalili N. & Khabbaz M.H. 1996. The concept of effective stress in Koliji A., Laloui L., Cuisinier O. & Vulliet L. 2006. Suction induced
unsaturated soils. UNICIV Report R-360, The University of New effects on the fabric of a structured soil. Transport in Porous Media
South Wales, Sydney, Australia. 64, 261-278.
Khalili N. & Khabbaz M.H. 1998. A unique relationship for for the Komine H. & Ogata N. 1999. Experimental study on swelling
determination of the shear strength of unsaturated soils. characteristics of sand–bentonite mixture for nuclear waste disposal.
Geotechnique 48(5), 681-687. Soils and Foundations 39(2), 83–97.
Khalili, N., & Loret, B. 2001. An elasto-plastic model for non-isothermal Komine H. 2004. Simplified evaluation for swelling characteristics of
analysis of flow and deformation in unsaturated porous media: bentonites. Engineering Geology 71(3-4), 265-279.
formulation. International Journal of Solids and Structures 38(46-47), Komolvilas V. & Kikumoto M. 2017. Simulation of liquefaction of
8305-8330. unsaturated soil using critical state soil model. International Journal
Khalili N., Geiser F. & Blight G.E. 2004. Effective stress in unsaturated for Numerical and Analytical Methods in Geomechanics 41(10),
soils: Review with new evidence. International journal of 1217-1246.
Geomechanics 4(2), 115-126. Komolvilas, V. Kikumoto M. & Kyokawa H. 2022. Mechanism of
Khalili N., Habte M.A. & Zargarbashi S. 2008. A fully coupled flow wetting induced deformation and failure of unsaturated soils. Int. J.
deformation model for cyclic analysis of unsaturated soils including Numer. Anal. Methods. Geomech. DOI: 10.1002/nag.3336
hydraulic and mechanical hystereses. Computers and Geotechnics Kong L.W., Bai W. & Guo A.G. 2012. Effects of cracks on the electrical
35(6), 872-889. conductivity of a fissured laterite: A combined experimental and
Khalili N., Habte M. A. & Valliappan S. 2005. A bounding surface statistical study. Geotechnical Testing Journal 35(6), 870-878.
plasticity model for cyclic loading of granular soils. International Konrad J.M. & Lebeau M. 2015. A capillary-based effective stress
Journal for Numerical Methods in Engineering, 63(14), 1939-1960. formulation for predicting the shear strength of unsaturated soils.
Khalili N., Uchaipichat A. & Javadi A.A. 2010. Skeletal thermal Canadian Geotechnical Journal 52(12):150603144058007.
expansion coefficient and thermo-hydro-mechanical constitutive Karunarathne A.M.A.N., Gad E.F., Sivanerupan S., & Wilson J.L. 2013.
relations for saturated homogeneous porous media. Mechanics of Review of residential footing design on expansive soil in Australia.
Materials 42(6): 593-598. Editor: Bijan Samali, Mario M. Attard and Chongmin Song. 22nd
Khalili N. & Valliappan S. 1996. Unified theory of flow and deformation Australasian Conference on the Mechanics of Structures and
in double porous media. European Journal of Mechanics. A, Solids Materials (ACMSM22). Sydney, Australia. Taylor & Francis. pp.
15(2): 321-336. 575-579.
Khalili N. & Zargarbashi S. 2010. Influence of hydraulic hysteresis on Kwa K.A. & Airey D.W. 2018. The development of a small centrifuge
effective stress in unsaturated soils. Géotechnique 60(9), 729-734. for testing unsaturated soils. Physical Modelling in Geotechnics 1,
Khan I.U., Al-Fergani M. & Black J.A. 2018. Development of a rainfall 513-518. A. McNamara, S. Divall, R. Goodey, N. Taylor, S.
simulator for climate modelling. Physical Modelling in Geotechnics Stallebrass and J. Panchal (eds.). CRC Press, London.
1, 507-512. A. McNamara, S. Divall, R. Goodey, N. Taylor, S. Lade P.V. & De Boer R. 1997. The concept of effective stress for soil,
Stallebrass and J. Panchal (eds.). CRC Press, London. concrete and rock. Géotechnique 47(1), 61–78.
Kholghifard M., Ahmad K., Ali N., Kassim A. & Kalatehjari R. 2014. Lai B.T., Wong H., Fabbri A. & Branque D. 2016. A new constitutive
Collapse/swell potential of residual laterite soil due to wetting and model of unsaturated soils using bounding surface plasticity (BSP)
drying-wetting cycles. Natl. Acad. Sci. Lett., 37(2), 147–153. and a non-associative flow rule. Innovative Infrastructure Solutions
Khoshghalb A. & Khalili N. 2013. A meshfree method for fully coupled 1(1), 1-8.
analysis of flow and deformation in unsaturated porous media. Int. J. Lajeunesse E., Mangeney-Castelnau A. & Vilotte J.P. 2004. Spreading of
Numer. Anal. Methods Geomech. 37(7), 716–743. a granular mass on a horizontal plane. Physics of Fluids 16(7), 2371–
Khoshghalb, A., Pasha, A.Y., & Khalili, N. 2015. A fractal modelfor 2381.
volume change dependency of the water retention curve. Lajmiri, A., Bagherieh, A. R., & Azizi, F. 2020. The simultaneous effect
Géotechnique, 65(2), 141–146. of void ratio and hydraulic hysteresis on effective stress parameter
Kikumoto, M., Kyokawa, H., Nakai, T., & Shahin, H. M. 2011. A simple in unsaturated soils. European Journal of Environmental and Civil
elasto-plastic model for unsaturated soils and interpretations of Engineering https://doi.org/10.1080/19648189.2020.1715263.
collapse and compaction behaviours. Unsaturated Soils. Alonso & Lakshmikantha M.R., Prat P.C. & Ledesma A. 2012. Experimental
Gens. 849-855. evidence of size effect in soil cracking. Can Geotech J. 49(3), 264–
Kim B.S., Shibuya S., Park S.W. & Kato, S. 2010. Application of suction 284.
stress for estimating unsaturated shear strength of soils using direct Lakshmikantha R.M., Prat P.C. & Ledesma A. 2018. Boundary effects in
shear testing under low confining pressure. Can. Geotech. J. 47(9), the desiccation of soil layers with controlled environmental
955-970. conditions. Geotechnical Testing Journal 41(4), 675-697.
Kim F.H., Penumadu D., Gregor J., Kardjilov N. & Manke, I. 2013. High- Lalicata L.M., Rotisciani G.M., Desideri A., Casini F. & Thorel L. 2020.
resolution neutron and X-ray imaging of granular materials. Journal Physical modelling of piles under lateral loading in unsaturated soils.
of Geotechnical and Geoenvironmental Engineering 139(5), 715- E3S Web Conference 195, 01021.
723. Laloui L., Klubertanz G. & Vulliet L. 2003. Solid–liquid–air coupling in
Kim F.H., Penumadu D., Kardjilov N. & Manke, I. 2016. High-resolution multiphase porous media. International Journal for Numerical and
X-ray and neutron computed tomography of partially saturated Analytical Methods in Geomechanics 27(3), 183-206.
granular materials subjected to projectile penetration. International Laloui L., Olgun C.G., Sutman M., McCartney J.S. et al. 2014. Issues
Journal of Impact Engineering 89, 72-82. involved with thermo-active geotechnical systems: characterization
Kim K., Rutqvist J., Harrington J.F., Tamayo-Mas E. & Birkholzer J.T. of thermo-mechanical soil behavior and soil-structure interface
2021. Discrete dilatant pathway modeling of gas migration through behavior. J. Deep Found. Inst. 8(2), 108–120.
compacted bentonite clay. International Journal of Rock Mechanics Larsbo M., Koestel J. & Jarvis N. 2014. Relations between macropore
and Mining Sciences 137, 104569. network characteristics and the degree of preferential solute
Kim T.H. & Sture S. 2008. Capillary-induced tensile strength in transport. Hydrology and Earth System Sciences 18(12), 5255-5269.
unsaturated sands. Can. Geotech. J. 45(5), 726-737. Lawrence C.A., Houston W.N., Houston S.L. & Harraz A.M. 2005.
Kneafsey T.J., Tomutsa L., Moridis G.J., Seol Y. et al. 2007. Methane Pressure pulse technique for measuring diffused air volume. Proc.
hydrate formation and dissociation in a partially saturated core-scale Int. Symp. Adv. Exp. Unsaturated Soil Mechanics, Trento, Italy, June
sand sample. Journal of Petroleum Science and Engineering 56(1-3), 27-29, 2005. Advanced Experimental Unsaturated Soil Mechanics
108-126. EXPERUS 2005. A. Tarantino, E. Romero and Y.J. Cui (eds.). A.A.
338
Balkema Publishers, Leiden: 9-13. deformations during triaxial testing using a photogrammetry-based
Lawton E.C., Fragaszy R.J. & Hardcastle J.H. 1989. Collapse of method. Can. Geotech. J. 53(3), 472–489.
compacted clayey sand. J. Geotech. Engng. 115(9): 1252-1267. d tomography (CT) be used to
viscoplasticity in strain localization analysis of multiphase Earth Sciences 79, 29.
geomaterials. International Journal for Numerical and Analytical Li W. & Yang Q. 2017. A macro-structural constitutive model for
methods in Geomechanics 39(14), 1570-1592. partially saturated expansive soils. Bulletin of Engineering Geology
Le Kouby A., Guimond-Barrett A., Reiffsteck P. & Pantet A. 2018. and the Environment 76(3), 1075–1084.
Influence of drying on the stiffness and strength of cement-stabilized Li X.S. 2007a. Thermodynamics-based constitutive framework for
soils. Geotech. Geol. Eng. 36(3), 1463–1474. unsaturated soils. 1: Theory. Géotechnique 57(5), 411-422.
Le Pense S., Arson C. & Pouya A. 2016. A fully coupled damage- Li X.S. 2007b. Thermodynamics-based constitutive framework for
plasticity model for unsaturated geomaterials accounting for the unsaturated soils. 2: A basic triaxial model. Géotechnique 57(5),
ductile–brittle transition in drying clayey soils. International Journal 423-435.
of solids and structures 91, 102-114. Li Y., Wang Y., Wang Y. & Ma C. 2017. Effects of root spatial distribution
Le T.T., Cui Y.J., Munoz J.J., Delage P. et al. 2011. Studying the hydric on the elastic–plastic properties of soil-root blocks. Sci Rep. 7, 800.
and mechanical coupling in Boom clay using an oedometer equipped Liang T., Bengough A.G., Knappett J.A., Wood D.M. et al. 2017a.
with a high capacity tensiometer. Frontiers of Architecture and Civil Scaling of the reinforcement of soil slopes by living plants in a
Engineering in China 5(2), 160-170. geotechnical centrifuge. Ecological Engineering 109, 207–227.
Ledesma A. 2016. Cracking in desiccating soils. E3S Web of Liang T., Knappett J.A., Bengough A.G. & Ke Y.X. 2017b. Small-scale
Conferences 9, 03005. modelling of plant root systems using 3D printing, with applications
Lee I.M., Sung S.G., Cho G.C. 2005. Effect of stress state on the to investigate the role of vegetation on earthquake-induced
unsaturated shear strength of a weathered granite. Can. Geotech. J. landslides. Landslides 14(5), 1747-1765.
42(2):624–631. Lieske W., Tripathy S., Baille W. & Schanz T. 2020. An alternative
Lei X., Wong H., Fabbri A., Limam A. & Cheng Y.M. 2016. A chemo- approach for determining suction of polyethylene glycols for soil
elastic–plastic model for unsaturated expansive clays. International testing. Géotechnique Letters 10(1), 45-49.
Journal of Solids and Structures 88, 354-378. Lim T.T., Rahardjo H., Chang M.F. & Fredlund D.G. 1996. Effect of
Lenhard R.J. & Parker J.C. 1988. Experimental validation of the theory rainfall on matric suctions in a residual soil slope. Can. Geotech. J.
33, 618-628.
phase systems for monotonic drainage paths. Water Resources Liu B., Zhu C., Tang C.S., Xie Y.H. et al. 2020. Bio-remediation of
Research 24(3), 373-380. desiccation cracking in clayey soils through microbially induced
Leong E.C., Tripathy S. & Rahardjo H. 2003. Total suction measurement calcite precipitation (MICP). Engineering geology 264, 105389.
of unsaturated soils with a device using the chilled-mirror dew-point Liu C. & Muraleetharan K.K. 2012a. Coupled hydro-mechanical
technique. Géotechnique 53(2), 173–182. elastoplastic constitutive model for unsaturated sands and silts. I:
Leong E.C., & Rahardjo H. 1997. Review of soil-water characteristic Formulation. International Journal of Geomechanics 12(3), 239-247.
curve equations. Journal of Geotechnical and Geoenvironmental Liu C. & Muraleetharan K.K. 2012b. Coupled hydro-mechanical
Engineering 123(12), 1106-1117. elastoplastic constitutive model for unsaturated sands and silts. II:
Leroueil S. & Barbosa P.S. de A. 2000. Combined effect of fabric, Integration, calibration, and validation. International Journal of
bonding and partial saturation on yielding of soils. Proc. Unsaturated Geomechanics 12(3), 248-259.
Soils for Asia. Singapore. Balkema: 527-532. Liu K., Chen W.B., Yin J.H., Feng W.Q. & Borana L. 2021. A novel
Leroueil S. & Hight D.W. 2013. Compacted soils: From physics to multifunctional apparatus for testing unsaturated soils. Acta
hydraulic and mechanical behaviour. Advances in Unsaturated Soils. Geotechnica, 16(12), 3761-3778.
Taylor & Francis Group, London: 41-59. Liu X., Zhou A., Shen S.L., Li J. & Sheng D. 2020. A micro-mechanical
Leung A.K., Garg A. & Ng C.W.W. 2015. Effects of plant roots on soil- model for unsaturated soils based on DEM. Computer Methods in
water retention and induced suction in vegetated soil. Eng. Geol. 193, Applied Mechanics and Engineering 368, 113183.
183–197. Liu Y., Cai G., Zhou A., Han B., Li J. & Zhao C. 2021. A fully coupled
Leung A. K. & Ng C. W. W. 2016. Field investigation of deformation constitutive model for thermo-hydro-mechanical behaviour of
characteristics and stress mobilisation of a soil slope. Landslides unsaturated soils. Computers and Geotechnics 133, 104032.
13(2): 229 – 240. Liu Z., Boukpeti N., Li X., Collin F., Radu J.P., Hueckel T. & Charlier R.
Leung A.K., Kamchoom V. & Ng C.W.W. 2017. Influences of root-
induced soil suction and root geometry on slope stability: A clays: a feasibility study. International journal for numerical and
centrifuge study. Can. Geotech. J. 54(3), 291-303. analytical methods in geomechanics 29(9), 919-940.
Levatti H.U., Prat P.C. & Ledesma A. 2019. Numerical and experimental Lloret A., Villar M.V., Sánchez M., Gens A., Pintado X. & Alonso E. E.
study of initiation and propagation of desiccation cracks in clayey 2003. Mechanical behaviour of heavily compacted bentonite under
soils. Computers and Geotechnics 105, 155-167. high suction changes. Géotechnique 53(1), 27–40.
Leverett M. 1941. Capillary behavior in porous solids. Transactions of Lloret-Cabot M., Sánchez M. & Wheeler S.J. 2013. Formulation of a
the AIME 142(01), 152-169.
Lewis, R.W., Schrefler, B.A., & Rahman, N. A. 1998. A finite element incorporating mechanical–water retention couplings. International
analysis of multiphase immiscible flow in deforming porous media Journal for Numerical and Analytical Methods in Geomechanics
for subsurface systems. Communications in numerical methods in 37(17), 3008-3035.
engineering 14(2), 135-149. Lloret-Cabot M., Wheeler S.J. & Sánchez M. 2017. A unified mechanical
Li B. 2015. Geotechnical properties of biocement treated sand and clay. and retention model for saturated and unsaturated soil behaviour.
Ph.D. Thesis, Nanyang Technological University, Singapore. Acta Geotechnica 12(1), 1-21.
Li C., Moreno-Atanasio R., O’Dea D. & Honeyands T. 2019. Lloret-Cabot M., Wheeler S.J., Pineda J.A., Sheng D. & Gens A. 2014.
Experimental study on the physical properties of iron ore granules Relative performance of two unsaturated soil models using different
made from Australian iron ores. ISIJ International 59(2), 253-262. constitutive variables. Canadian geotechnical journal 51(12), 1423-
Li J., Yin Z.Y., Cui Y. & Hicher P.Y. 2017. Work input analysis for soils 1437.
with double porosity and application to the hydromechanical Loret B., Hueckel T. & Gajo A. 2002. Chemo-mechanical coupling in
modeling of unsaturated expansive clays. Canadian Geotechnical saturated porous media: elastic–plastic behaviour of homoionic
Journal 54(2), 173-187. expansive clays. International Journal of Solids and Structures
Li J.H., Zhang L.M., Wang Y. & Fredlund D.G. 2009. Permeability tensor 39(10), 2773-2806.
and representative elementary volume of saturated cracked soil. Can.
Geotech. J. 46(8), 928–942. International journal for numerical and analytical methods in
Li J.S., Xing Y.C. & Hou Y.J. 2014. Centrifugal model test on the at-rest geomechanics 24(11), 893-927.
coefficient of lateral earth pressure in unsaturated soils. Physical Loret B. & Khalili N. 2002. An effective stress elastic–plastic model for
Modelling in Geotechnics 2, 931-936. Proc. 8th Int. Conf. on unsaturated porous media. Mechanics of Materials 34(2), 97-116.
Physical Modelling in Geotechnics ICPMG 2014. Lourenço S.D.N., Gallipoli D., Toll D.G., Augarde C.E. & Evans F.D.
Li L., Zhang X., Chen G. & Lytton R. 2015. Measuring unsaturated soil 2011. Towards a tensiometer based suction control system for
339
laboratory testing of unsaturated soils. Geotechnical Testing Journal Martinez A., Huang L. & Gomez M. G. 2018. Thermal conductivity of
34(6), 755–764. MICP-treated sands at varying degrees of satura
Lourenço S.D.N., Gallipoli D., Toll D.G., Evans F.D. 2006. Development Letters, 9(1), 15-21.
of a commercial tensiometer for triaxial testing of unsaturated soils. Mašín, D. 2010. Predicting the dependency of a degree of saturation on
Geotechnical Special Publication 147, ASCE, Reston, (2), 1875- void ratio and suction using effective stress principle for unsaturated
1886. soils. Int. J. Numer. Anal. Methods Geomech., 34(1), 73–90.
Lozada C., Thorel L. & Caicedo B. 2018. Bearing capacity of circular Mašín D. & Khalili N. 2008. A hypoplastic model for mechanical
footings resting on unsaturated desiccated soils. Int. J. Physical response of unsaturated soils. International Journal for Numerical
Modelling in Geotechnics 19(3), 154–166. and Analytical Methods in Geomechanics 32(15), 1903-1926.
Lu N. 2019. Revisiting axis translation for unsaturated soil testing. J. Mašín D. 2013. Double structure hydromechanical coupling formalism
Geotechnical and Geoenvironmental Engineering 145(7): 02819001. and a model for unsaturated expansive clays. Engineering Geology
Lu Y., Abuel-Naga H., Leong E.C., Bouazza A. & Lock P. 2018. Effect 165, 73-88.
of water salinity on the water retention curve of geosynthetic clay Mata C., Romero E. & Ledesma A. 2002. Hydro-chemical effects on
liners. Geotextiles and Geomembranes 46(6), 707-714. water retention in bentonite-sand mixtures. In Proc 3rd Int Conf.
Unsaturated Soils, Jucá de Campos and Marinho (eds), Recife,
stress in unsaturated soil. Water Resources Research 46, W05515. Brazil, 1,283–288.
Lu N., & Likos W.J. 2004, Unsaturated Soil Mechanics, p. 556, John Matyas E.L. & Radhakrishna H.S. 1968. Volume change characteristics
Wiley, New York. -448.
Lu N., & Likos W.J. 2006. Suction stress characteristic curve McCartney J.S. & Zornberg J.G. 2010. Effects of infiltration and
forunsaturated soil, J. Geotech. Geoenviron. Eng., 132(2), 131–142. evaporation on geosynthetic capillary barrier performance. Can.
Lu N., Wu B. & Tan C.P. 2007. Tensile Strength Characteristics of Geotech. J. 47(11), 1201-1213.
Unsaturated Sands. J. Geotechnical and Geoenvironmental McCartney J.S., Sánchez M. & Tomac I. 2016. Energy geotechnics:
Engineering, 133(2), 144-154. Advances in subsurface energy recovery, storage, exchange, and
Lu N. & Zhang C. 2019. Soil sorptive potential: Concept, theory, and waste management. Comput. Geotech. 75, 244-256.
verification. J. Geotechnical and Geoenvironmental Engineering McKeen, R. G., & Johnson, L. D. (1990). Climate-controlled soil
145(4): 04019006. design parameters for mat foundations. Journal of Geotechnical
Lu, N. & Yi D. 2017. Correlation between Soil-Shrinkage Curve and engineering, 116(7), 1073-1094.
water retention characteristics. J Geotech Geoenviron Eng 143(9): Meilani, I., Rahardjo H., Leong E.C. & Fredlund D.G. 2002. Mini suction
04017054. probe for matric suction measurements. Can. Geotech. J. 39(6),
Lubinski A. 1954. The theory of elasticity for porous bodies displaying a 1427-1432.
strong pore structure, Proc. 2nd US Nat. Congr. Appl. Mech., Ann Mendes J. & Buzzi O. 2014. Performance of the University of Newcastle
Arbor, 247-256. high capacity tensiometers. Proc. 6th Int. Conf. on Unsaturated soils,
Lube G., Huppert H.E., Sparks R.S.J. & Freundt A. 2005. Collapses of Sydney, Australia, 1611-1616.
two-dimensional granular columns. Physical Review E 72(4), 41301. Mendes J. & Gallipoli D. 2020. Comparison of high capacity tensiometer
Lucas D., Herzog R., Iten M., Buschor H. et al. 2020. Modelling of designs for long-term suction measurements. Physics and Chemistry
landslides in a scree slope induced by groundwater and rainfall. Int. of the Earth, Parts A/B/C 115,102831.
J. Physical Modelling in Geotechnics 20(4), 177–197. Mendes J., Gallipoli D., Tarantino A. & Toll D. 2019. On the
Luo L., Lin H. & Hallek P. 2008. Quantifying Soil Structure and development of an ultra-high-capacity tensiometer capable of
Preferential Flow in Intact Soil Using X-ray Computed Tomography. -564.
Soil Science Society of America Journal 72(4), 1058-1069. Mendes J., Toll D.G., Augarde C.E. & Gallipoli D. 2008. A system for
Luo S., Likos W.J. & Lu N. 2021. Cavitation of water in soil. Journal of field measurement of suction using high capacity tensiometers.
Geotechnical and Geoenvironmental Engineering 147(8). Unsaturated Soils: Advances in Geo-Engineering - Proceedings of
Mahabadi N., Zheng X & Jang J. 2016. The effect of hydrate saturation the 1st European Conference on Unsaturated Soils, Durham, UK,
on water retention curves in hydrate-bearing sediments. Geophys. July 2008, 219-225.
Res. Lett. 43, 4279–4287. Mendes J., Toll D.G. & Evans F. 2012. A double cell triaxial system for
Mahannopkul K. & Jotisankasa A. 2019. Influences of root concentration unsaturated soils testing. In Mancuso C., Jommi C., D’Onza F. (eds
and suction on Chrysopogon zizanioides reinforcement of soil. Soils ) Unsaturated Soils: Research and Applications. Springer, Berlin, H
and Foundations 59(2), 500–516. eidelberg. 5-10.
Mahawish A., Bouazza A. & Gates W.P. 2018. Effect of particle size Mendes R. M. & Marinho F.A.M. 2010. Monitoring water content and
distribution on the bio-cementation of coarse aggregates. Acta suction profiles in slopes of Serra do Mar. Revista Fundações e
Geotech. 13, 1019-1025. Obras Geotécnicas, Editora Rudder, SP, 80-84 (in Portuguese).
Mahler C.F. & Diene 2007. Tensiometer development for high suction Merchán V., Romero E. & Vaunat J. 2011. An adapted ring shear
analysis in laboratory lysimeters, Experimental unsaturated soil apparatus for testing partly saturated soils in the high suction range.
mechanics, (ed. T. Schanz), Springer,103-115. Geotechnical Testing Journal 34(5), 433-444.
Maldaner L.deS. & Marinho F.A.M. 2012. Field methodology for Mesa-Alcantara A. 2021. Hydro-mechanical behaviour of pellet/powder
mixture of bentonite and impact of gas migration. PhD. Thesis.
Congres, Firenze, IT. pp 1-10. Univeristat Politècnica de Catalunya, Barcelona, Spain.
Maleksaeedi E., Nuth M., Momoh N. & Chekired M. 2019. A modified Mesa-Alcantara A.M., Romero E., Mokni N. & Olivella S. 2020.
oedometer setup for simultaneously measuring hydromechanical Microstructural and hydro-mechanical behaviour of bentonite pellets
stress-strain paths for soils in the unsaturated state. Geotechnical and powder mixtures. E3S Web of Conferences 195, 04003. EDP
Testing Journal, 43(2), 287-309. Sciences.
Manassero M., Musso G., Rabozzi C. & Ribotta L. 2005. Retention Miller G.A., Khoury C.N., Muraleetharan K.K., Liu C., Kibbey T.C.G.
curves for a polluted soil. In: Proc. Int. Symp. Adv. Exp. Unsaturated 2008. Effects of soil skeleton deformations on hysteretic soil water
Soil Mechanics, Trento, Balkema, 459-465. characteristic curves: experiments and simulations. Water Resour.
Manzanal D., Pastor M. & Merodo J.A.F. 2011. Generalized plasticity Res. 44(5).
model for saturated and unsaturated soils. Part Minto J.M., Hingerl F.F., Benson S.M. & Lunn R.J. 2017. X-ray CT and
II: Unsaturated soil modeling. International Journal for Numerical multiphase flow characterization of a ‘bio-grouted’ sandstone core:
and Analytical Methods in Geomechanics 35(18), 1899-1917. The effect of dissolution on seal longevity. International Journal of
Mao Y., Romero E. & Gens A. 2018. Ice formation in unsaturated frozen Greenhouse Gas Control, 64, 152-162.
soils. Proc. 7th Int. Conf. on Unsaturated Soils. Hong Kong: 597- Minto J.M., MacLachlan E., El Mountassir G. & Lunn R.J. 2016. Rock
602. fracture grouting with microbially induced carbonate precipitation,
Marinho F.A.M., Take W.A. & Tarantino A. 2008. Measurement of Water Resour. Res. 52(11), 8827–8844.
Mirshekari M., Ghayoomi M. & Borghei A. 2018. A Review on Soil-
Laboratory and field testing of unsaturated soils 3-19. Springer, Water Retention Scaling in Centrifuge Modeling of Unsaturated
Dordrecht. Sands. Geotechnical Testing Journal 41, 20170120.
Martin P.L. & Barcala J.M. 2005. Large scale buffer material test: Mock- Mitarai N. & Nori F. 2006. Wet granular materials. Advances in Physics
up experiment at CIEMAT. Engineering Geology 81(3), 298-316. 55(1-2), 1-45.
340
Mitchell P.W. 2008. Footing Design for Residential Type Structures in Murray I. & Tarantino A. 2019. Mechanisms of failure in saturated and
Arid Climates. Australian Geomechanics Journal 43, 51-68. unsaturated clayey geomaterials subjected to (total) tensile stress.
Modaressi A. & Abou-Bekr N. 1994. A unified approach to model the Géotechnique 69(8),701–712.
behaviour of saturated and unsaturated soils. Proceedings of the 8th Murray I., Tarantino A. & Francescon F. 2019. A tensile strength
International Conference on Computer Methods and Advances in apparatus with the facility to monitor negative pore-water pressure.
Geomechanics. Balkema, Rotterdam 1507-1513. Geotechnical Testing Journal 42(5): 1384-1401.
Moghaddasi H., Shahbodagh B. & Khalili N. 2021. A bounding surface Nam S., Gutierrez M., Diplas P. & Petrie J. 2011. Determination of the
plasticity model for unsaturated structured soils. Computers and shear strength of unsaturated soils using the multistage direct shear
Geotechnics 138, 104313. test. Engineering Geology 122(3-4), 272-280.
Mokni N., Romero E. & Olivella S. 2014. Chemo-hydro-mechanical Nara Y., Morimot K., Hiroyoshi N., Yoneda T. et al. 2012. Influence of
behaviour of compacted Boom Clay: joint effects of osmotic and relative humidity on fracture toughness of rock: implications for
matric suctions. Géotechnique 64(9), 681-693. subcritical crack growth. Int. J. Solids Struct. 49, 2471–2481.
Molinero-Guerra A., Cui Y.J., He Y., Delage P. et al. 2019. Navarro V., Asensio L., Yustres Á., Pintado X. & Alonso J. 2014. An
Characterization of water retention, compressibility and swelling elastoplastic model of bentonite free swelling. Engineering geology
properties of a pellet/powder bentonite mixture. Engineering 181, 190-201.
Geology 248, 14-21. Naveed M., Schjønning P., Keller T., de Jonge L.W. et al. 2016.
Monroy R., Zdravkovic L. & Ridley A. 2010. Evolution of Quantifying vertical stress transmission and compaction-induced
microstructure in compacted London clay during wetting and soil structure using sensor mat and X-ray computed tomography.
loading. Géotechnique 60(2): 105-119. Soil and Tillage Research, 158, 110-122.
Monroy R., Zdravkovic L. & Ridley A.M. 2014. Evaluation of an active Nelson J.D., Chao K.Ch., Overton D.D. & Nelson E.J. 2015. Foundation
system to measure lateral stresses in unsaturated soils. Geotechnical Engineering for Expansive Soils. John Wiley & Sons, Inc.
Testing Journal 37(1), 71-84. Ng C.W.W. & Leung A.K. 2012. Measurements of drying and wetting
Montes-H G., Duplay J., Martinez L. & Mendoza C. 2003a. Swelling- permeability functions using a new stress-controllable soil column.
shrinkage kinetics of MX80 bentonite. Applied Clay Science 22, J. Geotech. Geoenviron. Eng. ASCE 138 (1), 58–65.
279-293. Ng C.W.W., Akinniyi D.B. & Zhou C. 2020. Experimental study of
Montes-H G., Duplay J., Martinez L., Geraud Y. & Rousset-Tournier B. hydromechanical behaviour of a compacted lateritic sandy lean clay.
2003b. Influence of interlayer cations on the water sorption and Canadian Geotechnical Journal 57 (11): 1695-1703.
swelling-shrinkage of MX80 bentonite. Applied Clay Science 23, Ng C.W.W., Chen R., Coo J.L., Liu J. et al. 2019. A novel vegetated three-
309-321. layer landfill cover system using recycled construction wastes
Mora Ortiz R.S. 2016. Efectos de la microestructura en el without geomembrane. Can. Geotech. J., 56(12), 1863–1875.
comportamiento hidromecánico de suelos compactados. PhD. Thesis. Ng C.W.W., Coo J.L., Chen Z.K. & Chen R. 2016a. Water infiltration
Univeristat Politècnica de Catalunya, Barcelona, Spain. into a new three-layer landfill cover system. J. Environmental
Morales L., Romero E., Jommi C., Garzón E. & Giménez A. 2015a. Engineering, 142(5), 04016007.
Feasibility of a soft biological improvement of natural soils used in Ng C.W.W., Garg A., Leung A.K. & Hau B. 2016b. Relationships
compacted linear earth construction. Acta Geotechnica 10, 157–171. between leaf and root area indices and soil suction induced during
Morales L., Romero E., Jommi C., Garzón E. & Giménez A. 2015b. drying-wetting cycles. Ecol. Eng. 91, 113–118.
Ageing effects on the small-strain stiffness of a bio-treated Ng C.W.W., Ni J.J., Leung A.K. & Wang Z.J. 2016c. A new and simple
compacted soil. Géotechnique Letters 5, 217–223. water retention model for root-permeated soils. Géotechnique Lett.
Morse M.S., Lu N., Wayllace A., Godt J.W. & Take W.A. 2014. 6, 106–111.
Experimental Test of Theory for the Stability of Partially Saturated Ng C.W.W., Leung A.K. & Woon K. 2014. Effects of soil density on
Vertical Cut Slopes. J. Geotechnical and Geoenvironmental grass-induced suction distributions in compacted soil subjected to
Engineering 140(9), 04014050. rainfall. Can. Geotech. J. 51(3), 311–321.
Morvan M., Wong H. & Branque D. 2010. An unsaturated soil model Ng C.W.W., Leung A.K., Yu R. & Kamchoom V. 2017. Hydrological
with minimal number of parameters based on bounding surface Effects of Live Poles on Transient Seepage in an Unsaturated Soil
plasticity. International Journal for Numerical and Analytical Slope: Centrifuge and Numerical Study. J. Geotechnical and
Methods in Geomechanics 34(14), 1512-1537. Geoenvironmental Engineering 143(3), 04016106.
Morvan M., Wong H., & Branque D. 2011. Incorporating Ng C.W.W., Springman S.M. & Alonso E.E. 2008. Monitoring the
porositydependent hysteretic water retention behavior into a new Performance of Unsaturated Soil Slopes. Geotechnical and
constitutive model of unsaturated soils. Can. Geotech. J. Geological Engineering 26(6), 799–816.
48(12):1855–1869. Ng C.W.W., Zhan L.T. & Cui Y.J. 2002. A new simple system for
Moscariello M. & Cuomo S. 2019. Wetting test and X-ray computed measuring volume changes in unsaturated soils. Can. Geotech. J.
tomography of volcanic unsaturated sands. Géotechnique Letters, 39(3), 757-764.
9(4), 314-321. Ng C.W.W., Zhou C. & Chiu C.F. 2020. Constitutive modelling of state-
Muir Wood D., Diambra A. & Ibraim E. 2016. Fibres and soils: a route dependent behaviour of unsaturated soils: an overview. Acta
towards modelling of root-soil systems. Soils Found. 56(5), 765-778. Geotechnica 15(10), 2705-2725.
Mujah D., Cheng L. & Shahin M.A. 2019. Microstructural and Ng C.W.W., Cui Y.J., Chen R. & Delage P. 2007. The axis-translation and
geomechanical study on biocemented sand for optimization of MICP osmotic techniques in shear testing of unsaturated soils: A
process. J. Mater. Civ. Eng. 31, 1-10. comparison. Soils and Foundations 47 (4): 675–684.
Mukunoki T., Miyata Y., Mikami K. & Shiota E., 2016. X-ray CT Ng C.W.W. & Pang Y. 2000. Influence of stress state on soil-water
analysis of pore structure in sand. Solid Earth, 7(3), 929-942. characteristics and slope stability. J Geotech Geoenviron. Eng.
Mun W. & McCartney J.S. 2017. Constitutive model for drained 126(2):157–66.
compression of unsaturated clay to high stresses. Journal of Ng C.W.W., Sadeghi H., Hossen S.K.B., Chiu C.F., Alonso E.E.,
Geotechnical and Geoenvironmental Engineering 143(6), 04017014. Baghbanrezvan S. 2018. Water retention and volumetric
Münch B. & Holzer H. 2008. Contradicting Geometrical Concepts in characteristics of intact and recompacted loess. Canadian Geotech J.
Pore Size Analysis Attained with Electron Microscopy and Mercury Revue Canad. Geotech. 53(8): 1258–1269.
Intrusion. J. Am. Ceram. Soc. 91, 4059–4067. Nguyen V., Pineda J.A., Romero E. & Sheng D. 2021. Influence of soil
Muñoz-Castelblanco J., Delage P., Pereira J.M. & Cui Y.J. 2012. A Local microstructure on air permeability in compacted clay. Géotechnique
Monitoring of Water Content in Unsaturated Soil Triaxial Testing. In: 71(5), 373-391.
Mancuso C., Jommi C., D’Onza F. (eds) Unsaturated Soils: Research Ni J. J., Leung A. K., & Ng C. W. W. 2018. Modelling soil suction
and Applications. Springer, Berlin, Heidelberg 19-24. changes due to mixed species planting. Ecological Engineering, 117:
Murakami M., Sato N., Anegawa A., Nakada N. et al. 2008. Multiple 1 – 17.
evaluations of the removal of pollutants in road runoff by soil Ni J.J., Leung A.K. & Ng C.W.W. 2019a. Unsaturated hydraulic
infiltration. Water Research 42(10-11), 2745-2755. properties of vegetated soil under single and mixed planting
Muraleetharan K.K., Liu C., Wei C., Kibbey T.C. & Chen L. 2009. An conditions. Géotechnique 69(6), 554–559.
elastoplatic framework for coupling hydraulic and mechanical Ni J.J., Leung A.K. & Ng C.W.W. 2019b. Modelling effects of root
behavior of unsaturated soils. International Journal of Plasticity growth and decay on soil water retention and permeability. Can.
25(3), 473-490. Geotech. J. 56, 1049–1055.
341
Nicotera M.V. 2000. Interpretation of shear response upon wetting of Shear strength of a vegetated soil incorporating both root
natural unsaturated pyroclastic soils. Experimental Evidence and reinforcement and suction. Transportation Geotechnics 18, 72-82.
Theoretical Approaches in Unsaturated Soils, 177-192. Balkema. Pandya S. & Sachan A. 2018. Effect of matric suction and initial static
Nikooee E., Habibagahi G., Hassanizadeh S.M. & Ghahramani A. 2013. loading on dynamic behaviour of unsaturated cohesive soil.
Effective stress in unsaturated soils: A thermodynamic International Journal of Geotechnical Engineering 12(5): 438-448.
approach based on the interfacial energy and hydromechanical coupling. Parera F., Pinyol N.. & Alonso E.E. 2020. Massive, continuous, and non-
Transp. Porous Med. 96: 369–396. invasive surface measurement of degree of saturation by shortwave
Nishimura T. 2013. Application of micro-porous membrane technology infrared images. Can. Geotech. J. 58(6), 749-762.
for measurement of soil-water characteristic curve. Proc. 18th Int. Pasha A.Y., Khoshghalb A., Khalili N. 2016. Pitfalls in interpretation of
Conference on Soil Mechanics and Geotechnical Engineering, Paris, gravimetric water content-based soil-water characteristic curve for
1171-1174. deformable porous media. Int J Geomech ASCE 16(6): D4015004.
Nishimura T. 2016. Investigation on creep behavior of geo-materials with Pasha A.Y., Khoshghalb A., Khalili N. 2017. A hysteretic model for the
suction control technique. In E3S Web of Conferences 9, 18002. evolution of water retention curve with void ratio. J. Eng. Mech.
EDP Sciences. ASCE DOI:10.1061/(ASCE)EM.1943-7889.0001238.
Nova, R. 1994. Controllability of the incremental response of soil Pasha A.Y., Khoshghalb A., Khalili N. 2019. Can degree of saturation
specimens subjected to arbitrary loading programmes. Journal of the decrease during constant suction compression of an unsaturated soil?
Mechanical behavior of Materials 5(2), 193-202. Computers and Geotechnics 106, 199-204.
Nur A. & Byerlee, J.D. 1971. ‘An exact effective stress law for elastic Pasha A.Y., Khoshghalb A., Khalili N. 2020. Evolution of isochoric water
deformation of rock with fluids. J. Geophys. Res. 76-26, 6414–6419. retention curve with void ratio. Computers and Geotechnics. 122:
Nuth M. & Laloui L. 2008. Effective stress concept in unsaturated soils: 103536.
Clarification and validation of a unified framework. International Patil U.D., Hoyos L.R. & Puppala A.J. 2016. Characterization of
journal for numerical and analytical methods in Geomechanics 32(7), compacted silty sand using a double-walled triaxial cell with fully
771-801. automated relative-humidity control. Geotechnical Testing Journal,
Oberg A., Sallfors G. 1995. A rational approach of the determination of 39(5), 742-756.
shear strength parameters of unsaturated soils. 1st Int. Conf. on Pedroso D.M. & Farias M.M. 2011. Extended Barcelona basic model for
Unsaturated soils, Paris, 151-158. unsaturated soils under cyclic loadings. Computers and Geotechnics
Ohno, S., Kawai, K., & Tachibana, S. 2007. Elasto-plastic constitutive 38(5), 731-740.
model for unsaturated soil applied effective degree of saturation as a Pedrotti M., Wong C., El Mountassir G., Renshaw J.C. & Lunn R.J. 2020.
parameter expressing stiffness. Journal of JSCE 63(4), 1132-1141. Desiccation behaviour of colloidal silica grouted sand: a new
Oka F., Shahbodagh B. & Kimoto S. 2019. A computational model for material for the creation of near surface hydraulic barriers.
Engineering Geology, 270, 105579.
International Journal for Numerical and Analytical Methods in Pelaez R., Casini F., Romero E., Gens A. & Viggiani G.M.B. 2014.
Geomechanics 43(1), 138-165. Freezing-thawing tests on natural pyroclastic samples. International
Oka, F., & Kimoto, S. 2012. Computational modeling of multiphase Conference on Unsaturated Soils. Unsaturated Soils: Research &
geomaterials. CRC press. Applications. Sydney: CRC Press: 689-1694.
Oka, F., & Kimoto, S. 2022. Computational Multiphase Geomechanics. Pereira J.M., Coussy O., Alonso E.E., Vaunat J. & Olivella S. 2010. Is the
CRC Press. degree of saturation a good candidate for Bishop’s parameter?.
Oka, F., Kodaka, T., Kimoto, S., Kim, Y.S., & Yamasaki, N. 2006. An Unsaturated Soils - Proc. Fifth Int. Conf. on Unsaturated Soils, Sep
elasto-viscoplastic model and multiphase coupled FE analysis for 2010, Barcelona, Spain. 913-919.
unsaturated soil. In Unsaturated Soils, 2039-2050. Pereira J.M., Rouainia M. & Manzanal D. 2014. Combined effects of
Oldecop L, Garino L., Muñoz J.J., Rodriguez R. & García C. 2011. structure and partial saturation in natural soils. Journal of Geo-
Unsaturated behaviour of mine tailings in low precipitation areas. Engineering Sciences 2(1-2), 3-16.
Proc. 5th Int. Conf. Unsaturated Soils, Barcelona, Spain, 1425-1430. Perez-Ruiz D.D. 2009. A refined true triaxial apparatus for testing
Unsaturated Soils 2010. E.E. Alonso and A. Gens (eds.). CRC Press, unsaturated soils under suction-controlled stress paths. PhD Thesis,
London. The University of Texas at Arlington, USA.
Oldecop L. & Alonso E.E. 2003. Suction effects on rockfill Peri D., Zhao G. & Khalili N. 2014. Strain localization in unsaturated
compressibility. Géotechnique 53(2), 289-292. elastic-plastic materials subjected to plane strain compression.
Oldecop L. & Alonso E.E. 2004. Testing rockfill under relative humidity Journal of Engineering Mechanics 140(7), 04014050.
control. Geotechnical Testing Journal 27(3), 269-278. Peron H., Hueckel T., Laloui L. & Hu L. 2009. Fundamentals of
Oldecop L. & Alonso E.E. 2013. Rockfill dams. Advances in Unsaturated desiccation cracking of fine-grained soils: experimental
Soils. Taylor & Francis Group, London, 61-86. characterisation and mechanisms identification. Can. Geotech. J.,
Oldecop L. & Rodari G. 2021. Unsaturated mine tailings disposal. Soils 46(10), 1177-1201.
and Rocks 44(3), e2021067421. Pham V., Nakano A., van der Star W.R.L.., Heimovaara T. & van Paassen
Oldecop, L. & Alonso, E. 2017. Measurement of lateral stress and friction L. 2016. Applying MICP by denitrification in soils: A process
in rockfill oedometer tests enabling the analysis of the experimental analysis. Environmental Geotechnics 5(2), 79-93. 15.00078
results in the '-q space. Geotechnical Testing Journal, 40(5), 1-11 . Phan D.G., Nguyen G.D., Bui H.H. & Bennett T. 2021. Constitutive
Oliveira O.D. & Marinho F.A.M. 2008. Suction equilibration time for a modelling of partially saturated soils: Hydro-mechanical coupling in
high capacity tensiometer. Geotechnical Testing Journal 31(1), 101- a generic thermodynamics-based formulation. International Journal
105. of Plasticity 136, 102821.
Oorthuis R et al. 2021. Slope orientation and vegetation effects on soil Pineda J., Romero E., Alonso E.E. & Pérez T. 2014c. A new high-
thermo-hydraulic behavior. An experimental study. Sustainability pressure triaxial apparatus for inducing and tracking hydro-
13(1), 14. mechanical degradation of clayey rocks. Geotechnical Testing
Or D. 2001. Who invented the tensiometer?. Soil Science Society of Journal 37(6), 933-947.
America Journal 65(1), 1-3. Pineda J.A., Alonso E.E. & Romero E. 2014a. Environmental
Otani, J., Mukunoki, T., & Obara, Y., 2000. Application of X-ray CT degradation of claystones. Géotechnique 64(1), 64-82.
method for characterization of failure in soils. Soils and Foundations Pineda J.A., Romero E., De Gracia M. & Sheng D. 2014b. Shear strength
40(2), 111-118. degradation in claystones due to environmental effects.
Pacovský J., Svoboda J. & Zapletal L. 2007. Saturation development in Géotechnique 64(6), 493–501.
the bentonite barrier of the Mock-Up-CZ geotechnical experiment. Pintado X., Lloret A. & Romero E. 2009a. Assessment of the use of the
Physics and Chemistry of the Earth, Parts A/B/C 32(8-14), 767-779. vapour equilibrium technique in controlled-suction tests. Can.
Padilla J.M., Perera Y.Y., Houston W.N., Perez N. & Fredlund D.G. 2006. Geotech. J. 46(4), 411-423.
Quantification of air diffusion through high air-entry ceramic disks. Pintado X., Lloret A. & Romero E. 2009b. Reply to the discussion by
Proc. 4th Int. Conf. on Unsaturated Soils, Carefree, Arizona, April 2- Leong et al. on “Assessment of the use of the vapour equilibrium
6, 2006. Unsaturated Soils. Geotechnical Special Publication No. technique in controlled-suction tests”. Can. Geotech. J. 46(12),
147. G.A. Miller, C.E. Zapata, S.L. Houston and D.G. Fredlund 1482–1484.
(eds.). ASCE, Reston, Virginia, 2: 1852-1863. Pintado X., Romero E., Suriol J., Lloret A. & Madhusudhan B.N. 2019.
Pallewattha M., Indraratna B., Heitor A. & Rujikiatkamjorn C. 2019. Small strain shear stiffness of compacted bentonites for engineered
342
barrier system. Geomechanics for Energy and the Environment 18, their hydraulic properties. Engineering Geology 165: 3-19.
1-12. Romero E., Alonso E.E., Alvarado C. & Wacker F. 2012a. Effect of
Pinyol N., Alvarado M., Parera F. & Yerro A. 2017. Novel Procedure to loading history on time dependent deformation of rockfill. In
Validate MPM Results by Means of PIV Measurements. Procedia Unsaturated Soils: Research and Applications 419-424. Springer,
Engineering 175, 332–340. Berlin, Heidelberg.
Pinzón G. & Cabrera M.A. 2020. Planar granular column collapse: an Romero E, Musso G & Jommi C. 2012b. Experimental techniques for
experimental proxy for the study of transitional granular flows. Proc. hydromechanical and electro-chemo-hydraulic processes. ALERT
13th Int. Symp. Landslide (ISL 2020). Cartagena. M.A. Cabrera, L. Doctoral School 2012: advanced experimental techniques in
F. Prada-Sarmiento, & J. Montero (Eds.). geomechanics. Grenoble CEDEX: ALERT Geomaterials, 157-202.
Porcino D., Marcianò V. & Granata R. 2011. Undrained cyclic response Romero E. & Jommi C. 2008. An insight into the role of hydraulic history
of a silicate grouted sand for liquefaction mitigation purposes. on the volume changes of anisotropic clayey soils. Water Resour. Res.
Geomechanics and Geoengineering: An International Journal 6(3), 44: W12412.
155–170. Romero E. & Simms P. 2008. Microstructure investigations in
Pot V., Peth S., Monga O., Vogel L.E. et al. 2015. Three-dimensional unsaturated soils. A review with special attention to mercury
distribution of water and air in soil pores: comparison of two-phase intrusion porosimetry and environmental scanning electron
two-relaxation-times lattice-Boltzmann and morphological model microscopy. Geotechnical and Geological Engineering 26(6), 705-
outputs with synchrotron X-ray computed tomography data. 72.
Advances in water resources, 84, 87-102. Romero E., Della Vecchia G. & Jommi C. 2011. An insight into the water
Prat P.C., Ledesma A., Cuadrado A. & Levatti H.U. 2013. Ground retention properties of compacted clayey soils. Géotechnique 61(4):
penetrating radar system for detection of desiccation cracks in soils. 313-328.
Proc. 3rd Int. Symp. Computational Geomechanics (COMGEO III). Romero E., Facio J.A., Lloret A., Gens A. & Alonso E.E. 1997. A new
Krakow, Poland, International Centre for Computational suction and temperature controlled triaxial apparatus. Proc. of the
Engineering, 249-258. 14th International Conference on Soil Mechanics and Foundation
Proctor R.R. 1933. Fundamental principles of soil compaction.
Engineering News Record V11(9), 148-156. Romero E., García I. & Knobelsdorf J. 2005b. Gas permeability
Rabozzi C., Ribotta L. & Gremigni G. 2006. Retention curves and evolution of a sand/bentonite during. Advanced Experimental
hydraulic properties of a soil contaminated by NAPL. Proc. 5th Unsaturated Soil Mechanics: Proc. Int. Symp. Advanced
ICEG Environmental Geotechnics: Opportunities, Challenges and Experimental Unsaturated Soil Mechanics, Trento, 3, 385-390. CRC
Responsibilities for Environmental Geotechnics, Cardiff, 1232-1239. Press.
Thomas Telford Publishing. Romero E., Gens A. & Lloret A. 2003. Suction effects on a compacted
Rahardjo H., Santoso V.A., Leong E.C., Ng, Y.S. et al. 2012. Performance clay under non-isothermal conditions. Géotechnique 53(1), 65-81.
of an instrumented slope covered by a capillary barrier system. Romero E., Gens A. & Lloret, A. 2001. Temperature effects on the
Journal of Geotechnical and Geoenvironmental Engineering, 138(4), hydraulic behaviour of an unsaturated clay. Geotechnical and
481–490. Geological Engineering 19, 311-332.
Rahardjo H. 2015. Capillary barrier as a slope protection. Distinguished Romero E., Vaunat J. & Merchán V. 2014. Suction effects on the residual
Lecture AP-UNSAT2015 Conference, Guilin, China October 23-26, shear strength of clays. J. Geo-Engineering Sciences 2(1–2), 17-37.
2015. Unsaturated Soil Mechanics – From Theory to Practice. Z. Romero E., Villar M.V. & Lloret A. 2005a. Thermo-hydro-mechanical
Chen, C. Wei, D. Sun and X. Xu (eds.). CRC Press. behaviour of two heavily overconsolidated clays. Engineering
Rahardjo H., Gofar N. & Satyanaga A. 2018. Effect of concrete waste Geology 81(3), 255-268.
particles on infiltration characteristics of soil. Environmental Earth Romero, E. & Vaunat, J. 2000. Retention curves of deformable clays.
Sciences 77(9), 1-12. Experimental evidence and theoretical approaches in unsaturated
Rahardjo H., Shen Y., Tsen-Tieng D. L., Ramos-Rivera J., Nong X. & soils, In Proceedings of International Workshop on Unsaturated Soil,
Hamdany A.H. 2021. New Osmotic Tensiometer Development. Trento, Italy, Balkema, Rotterdam, pp. 91–106 91-106.
Geotechnical Testing Journal 44(3), 722-740. Rosone M., Airò Farulla C. & Ferrari A. 2016. Shear strength of a
Rampino C., Mancuso C. & Vinale F. 2000. Experimental behaviour and compacted scaly clay in variable saturation conditions. Acta
modelling of an unsaturated compacted soil. Canadian Geotechnical Geotechnica 11, 37-50.
Journal 37(4), 748-763. Rotisciani G.M., Casini F., Desideri A. & Sciarra G. 2016. Modelling of
Read D.B., Bengough A.G., Gregory P.J., Crawford J.W. et al. 2003 Plant imbibition process in an embankment scale model. E3S Web
roots release phospholipid surfactants that modify the physical and Conference, 9, 16009.
chemical properties of soil. New Phytologist 157, 315–326. Rouf M.A., Bouazza A., Singh R.M., Gates W.P. & Rowe R.K. 2016. Gas
Rezzoug A., König D. & Triantafyllidis T. 2004. Scaling Laws for flow unified measurement system for sequential measurement of gas
Centrifuge Modeling of Capillary Rise in Sandy Soils. J. diffusion and gas permeability of partially hydrated geosynthetic
Geotechnical and Geoenvironmental Engineering 130(6), 615–620. clay liners. Can. Geotech. J., 53(6), 1000-1012.
Rice J.R. & Cleary M.P. 1976. Some basic stress diffusion solutions of Rouf M.A., Singh R.M., Bouazza A., Gates W.P. & Rowe R.K. 2014.
fluid saturated elastic porous media with compressible constituents. Evaluation of a geosynthetic clay liner water retention curve using
Rev. Geophys. Space Phys., 14, 227–291. vapour equilibrium technique. Unsaturated Soils: Research &
Richefeu V., el Youssoufi M.S. & Radjaï F. 2006. Shear strength Applications 1003-1009. Khalili, Russell & Khoshghalb (Eds).
properties of wet granular materials. Physical Review E 73(5), Taylor & Francis Group, London, ISBN 978-1-138-00150-3
0513041. Russell, A.R. 2014. How water retention in fractal soils depends on
Ridley A.M. & Burland J.B.1993. A new instrument for the measurement particle and pore sizes, shapes, volumes and surface areas.
of soil moisture suction. Géotechnique 43(2), 321-324. Géotechnique 64(5), 379-390.
Ridley A.M. & Wray W.K. 1996. Suction measurement: A review of Russell A.R. & Khalili N. 2006. A unified bounding surface plasticity
current theory and practices. Proc. 1st Int. Conf. on Unsaturated model for unsaturated soils. International Journal for Numerical and
Soils, Paris, September 6-8, 1995. Unsaturated Soils. E.E. Alonso Analytical Methods in Geomechanics 30(3), 181-212.
and P. Delage (eds.). A.A. Balkema / Presses des Ponts et Chaussées, Russam K. & Coleman, J.D. 1961. The effect of climatic factors on
Paris, 3: 1293-1322. subgrade moisture conditions. Géotechnique 11(1): 22–28.
Ridley A.M., Dineen K., Burland J.B. & Vaughan P.R. 2003. Soil matrix Saba S., Cui Y.J., Tang A.M. & Barnichon J.D. 2014. Investigation of the
suction: some examples of its measurement and application in swelling behaviour of compacted bentonite–sand mixture by mock-
geotechnical engineering. Géotechnique 53(2), 241–253. up tests. Can. Geotech. J. 51(12), 1399-1412.
Rodríguez R., Sánchez M., Ledesma A. & Lloret A. 2007. Experimental
and numerical analysis of desiccation of a mining waste. Can. Resolution Shortwave Infrared Imaging of Water Infiltration into
Geotech. J. 44(6), 644–658. Dry Soil. Vadose Zone Journal 16(13), 1-10.
Rojas J.C., Mancuso C. & Vinale F. 2008. A modified triaxial apparatus Saffari, R., Nikooee, E., & Habibagahi, G. 2020. The effect of microbial
to reduce testing time: Equipment and preliminary results. In calcite precipitation on the retention properties of unsaturated fine-
Unsaturated Soils. Advances in Geo-Engineering 119-126. CRC grained soils: discussion of the governing factors. E3S Web of
Press. Conferences, 195, 05009. EDP Sciences.
Romero 2013. A microstructural insight into compacted clayey soils and Saffari, R., Nikooee, E., Habibagahi, G., & van Genuchten, M. T. 2019.
343
Effects of biological stabilization on the water retention properties Sestrem L.P., Kormann A.C.M. & Marinho F.A.M. 2018. A field study
of unsaturated soils. J. Geotechnical and Geoenvironmental on the influence of rainfall intensity, suction and load distribution, in
Engineering, 145(7), 04019028. a reinforced unsaturated slope in Brazil. In Landslides and
Sakaguchi A., Nishimura T. & Kato M. 2005. The effect of entrapped air Engineered Slopes. Experience, Theory and Practice (pp. 1829-
on the quasi-saturated soil hydraulic conductivity and comparison 1834). CRC Press.
with the unsaturated hydraulic conductivity. Vadose Zone Journal Shahbodagh, B. 2011. Large deformation dynamic analysis method for
4(1): 139-144. partially saturated elasto-viscoplastic soils. PhD Dissertation, Kyoto
Salager S., Khaddour G., Charrier P. & Desrues J. 2014. An investigation University, Kyoto, Japan.
into unsaturated states of granular media using X-ray computed Shahbodagh, B., Khalili, N., & Esgandani, G.A. 2015. A numerical
tomography. Unsaturated Soils: Research & Applications. CRC model for nonlinear large deformation dynamic analysis of
Press, Boca Raton, 703–709. unsaturated porous media including hydraulic hysteresis. Computers
Salager S., Nuth M., Ferrari A., Laloui L. 2013. Investigation into water and Geotechnics 69, 411-423.
retention behaviour of deformable soils. Can Geotech J 50(2): 200– Sheng D. 2011. Review of fundamental principles in modelling
208. unsaturated soil behaviour. Computers and Geotechnics 38(6), 757-
Salimzadeh S. & Khalili, N. 2014. Consolidation of unsaturated lumpy 776.
clays. J. Geo-eng. Sci., 2(1), 67–82. Sheng D., Fredlund D.G. & Gens A. 2008a. A new modelling approach
Samat, S., Vaunat, J., & Gens, A. 2008. A thermomechanical framework for unsaturated soils using independent stress variables. Canadian
for modelling the response of unsaturated soils. In Proc. 1st Geotechnical Journal 45(4), 511-534.
European Conf. on Unsaturated Soils, Durham, UK, 547-558. Sheng D., Gens A., Fredlund D.G. & Sloan S.W. 2008b. Unsaturated soils:
Sanchez M., Atique A., Kim S., Romero E. & Zielinski M. 2013. from constitutive modelling to numerical algorithms. Computers and
Exploring desiccation cracks in soils using a 2D profile laser device. Geotechnics 35(6), 810-824.
Acta Geotechnica 8, 583–596. Sheng D., Pedroso D.M. & Abbo A.J. 2008c. Non-convexity and stress-
Sánchez M., Gens A., do Nascimento Guimarães L. & Olivella S. 2005. path dependency of unsaturated soil models. Computational
A double structure generalized plasticity model for expansive Mechanics 42(5), 685-694.
materials. International Journal for numerical and analytical methods Sheng D., Sloan S.W. & Gens A. 2004. A constitutive model for
in geomechanics 29(8), 751-787. unsaturated soils: thermomechanical and computational aspects.
Sánchez M., Gens A., Guimarães L. & Olivella S. 2008. Implementation Computational Mechanics 33(6), 453-465.
algorithm of a generalised plasticity model for swelling clays. Sheng D., Sloan S.W., Gens A. & Smith D.W. 2003. Finite element
Computers and Geotechnics 35(6), 860-871. formulation and algorithms for unsaturated soils. Part I: Theory.
Santagiuliana R. & Schrefler B.A. 2006. Enhancing the Bolzon– International journal for numerical and analytical methods in
Schrefler–Zienkiewicz constitutive model for partially saturated soil. geomechanics 27(9), 745-765.
Transport in porous media 65(1), 1-30. Sheng D. & Zhou A.N. 2011. Coupling hydraulic with mechanical
Santomaso A.C., Volpato S. & Gabrieli F. 2018. Collapse and runout of models for unsaturated soils. Canadian Geotechnical Journal 48(5),
granular columns in pendular state. Physics of Fluids 30(6), 063301. 826-840.
Santucci de Magistris F. & Tatsuoka F. 2004. Effects of moulding water Shwan B. 2019. The effect of suction on ground surface settlement for a
content on the stress strain behaviour of a compacted silty sand. Soils tunnel. E3S Web of Conferences 92, 17008.
& Found. 44(2): 85–101. Siemens G. & Blatz J. 2007. Triaxial apparatus for applying liquid
Saulick K. & Lourenço S.D.N. 2020. Hydrophobisation of clays and nano infiltration with controlled boundary conditions and internal suction
silica for ground engineering. E3S Web of Conferences 195, 03039. measurement. Journal of Geotechnical and Geoenvironmental
Scarfone R., Wheeler S.J. & Lloret-Cabot M. 2020. Conceptual hydraulic Engineering 133(6), 748-752.
conductivity model for unsaturated soils at low degree of saturation Siemens G.A., Take W.A. & Peters S.B. 2013. Physical and numerical
and its application to the study of capillary barrier systems. Journal modeling of infiltration including consideration of the pore-air phase.
of Geotechnical and Geoenvironmental Engineering 146 (10). Can. Geotech. J. 51(12), 1475-1487.
Scheel M., Seemann R., Brinkmann M., Di Michiel M. et al. 2008. Sillers W.S., Fredlund D.G. & Zakerzadeh N. 2001. Mathematical
Morphological clues to wet granular pile stability. Nature Materials attribute of some soil –water characteristic curve models.
7(3), 189–193. Geotechnical and Geological Engineering, 19, 243 -283.
Schiffman R.L. 1970. The stress components of a porous medium. Silva Junior A.C. 2011. Predict of suction and moisture of tropical soils
Journal of Geophysical Research 75(20), 4035-4038. using mechanistic model. 2011. 137 f. MSc dissertation. Federal
Schiava R. & Etse G. 2006. Constitutive modeling and discontinuous University of Goias (in Portuguese).
bifurcation assessment in unsaturated soils. Journal of Applied https://repositorio.bc.ufg.br/tede/handle/tde/1336
Mechanics 73(6), DOI:10.1115/1.2202349 Simms P.H. & Yanful E.K. 2002. Predicting soil-water characteristic
Scholl P., Leitner D., Kammerer G., Loiskandl W. et al. 2014. Root curves of compacted plastic soils from measured pore-size
induced changes of effective 1D hydraulic properties in a soil distributions. Géotechnique 52(4), 269-278.
column. Plant Soil 381, 193–213. Sitarenios P. & Kavvadas M. 2020. A plasticity constitutive model for
Schrefler B.A., 1984. The finite element method in soil consolidation unsaturated, anisotropic, nonexpansive soils. International Journal
(with applications to surface subsidence), Ph.D. Thesis. University for Numerical and Analytical Methods in Geomechanics 44(4), 435-
College of Swansea, C/Ph/76/84, Swansea (UK). 454.
Schrefler, B.A., Sanavia, L., & Majorana, C.E. 1996. A multiphase Sivakumar V. & Wheeler S.J. 2000. Influence of compaction procedure
medium model for localisation and postlocalisation simulation in on the mechanical behaviour of an unsaturated compacted clay. Part
geo 1: Wetting and isotropic compression. Géotechnique 50(4): 359-368.
International Journal on Experiments, Modelling and Computation Skempton W.A. 1960. Effective stress in soils, concrete and rocks. Conf.
of Materials and Structures 1(1), 95-114. on Pressure and Suction in Soils, London, 4–16.
Schrefler, B.A., Zhang, H.W., & Sanavia, L. 2006. Interaction between Skempton, A. W., and Bishop, A. W. (1954). Soils, in Building materials,
different internal length scales in strain localization analysis of fully their elasticity and inelasticity, (ed. M. Reiner) North Holland
and partially saturated porous media— Publishing Company, Amsterdam, The Netherlands, 417-482.
journal for numerical and analytical methods in geomechanics 30(1), Suklje, L. (1969). Rheological aspects of soil mechanics, Wiley
45-70. Interscience, New York, P. 123.
Seiphoori A. & Zamanian, M. 2022. Improving mechanical behaviour of Slatter E.E., Jungnickel C.A., Smith D.W. & Allman M.A. 2000.
collapsible soils by grouting clay nanoparticles. Engineering Investigation of suction generation in apparatus employing osmotic
Geology. In Press, Journal Pre-proof. methods. Proc. 1st Asian Conference on Unsaturated Soils, 297–302.
Seiphoori A., Laloui L., Ferrari A., Hassan M. & Khushefati W.H. 2016. Singapore.
Water retention and swelling behaviour of granular bentonites for
application in Geosynthetic Clay Liner (GCL) systems. Soils and modelling unsaturated soils. International Journal for Numerical and
Foundations, 56(3), 449-459. Analytical Methods in Geomechanics 36(14), 1667-1681.
Sentenac P. & Zielinski M. 2009. Clay fine fissuring monitoring using
miniature geo-electrical resistivity arrays. J. Environ. Geol. 59, 205- of elastoplastic constitutive models for unsaturated soils.
214. International Journal of Geomechanics 15(2), 04014041.
344
Soltani A., Deng A., Taheri A., Mirzababaei M. & Vanapalli S. 2019. bentonite for deep geological radioactive waste repositories: High-
Swell–shrink behavior of rubberized expansive clays during temperature experiments and thermodynamic modeling.
alternate wetting and drying. Minerals 9, 224. Engineering Geology 269, 105549.
Song B., Nakamura D., Kawaguchi T. & Kawajiri S. 2021. Development Sun H., Mašín D., Najser J., Ned la V. & Navrátilová E. 2019. Bentonite
of a technique for observing the frost heaving process in soil using microstructure and saturation evolution in wetting–drying cycles
an industrial micro-focus X-ray CT scanner. Geomate Journal 21(84), evaluated using ESEM, MIP and WRC measurements.
112-120. Géotechnique 69(8), 713-726.
Song X, Borja RI. 2014. Mathematical framework for unsaturated flow Sun X. 2015. The impact of climate as expressed by Thornthwaite
in the finite deformation range. Int. J. Numer. Meth. Eng. 97(9): 658– Moisture Index on residential footing design on expansive soil in
792. Australia. Master degree thesis. School of Civil Environmental and
Song Y.S., Cho Y.C. & Hong S. 2016. Analyses on variations in the Chemical Engineering. College of Science Engineering and Health.
unsaturated characteristics of a mine waste-dump slope during RMIT University
rainfall. Environmental Earth Sciences 75(14), 1106. Sun W. & Sun D.A. 201
Song, X., & Khalili, N. 2019. A peridynamics model for strain behaviour of unsaturated compacted expansive soils. International
localization analysis of geomaterials. International Journal for Journal for Numerical and Analytical Methods in Geomechanics
Numerical and Analytical Methods in Geomechanics 43(1), 77-96. 36(8), 1002-1022.
Sutharsan T., Muhunthan B. & Liu Y. 2017. Development and
air system with random heterogeneity via stabilized mixed finite implementation of a constitutive model for unsaturated sands.
elements. International Journal for Numerical Methods in International Journal of Geomechanics 17(11), 04017103.
Engineering 112(13), 1926-1950. A., Wu W. & Springman S.M. 2018.
Sorbino G. & Nicotera M.V. 2013. Unsaturated soil mechanics in rainfall- Experimental validation of a coupled hydro-mechanical model for
induced flow landslides. Engineering Geology 165: 105-132. vegetated soil. Géotechnique 68(5), 375-385.
Southen J.M. & Rowe R.K. 2004. Investigation of the behavior of Tabiatnejad B., Siddiqua S. & Siemens G. 2016. Impact of pore fluid
geosynthetic clay liners subjected to thermal gradients in basal liner salinity on the mechanical behavior of unsaturated bentonite–sand
applications. In Advances in Geosynthetic Clay Liner Technology: mixture. Environmental Earth Sciences 75(22), 1-10.
2nd Symposium. ASTM International. Takada Y., Ueda K., Iai S. & Mikami T. 2018. Liquefaction behaviour
Southen J.M. & Rowe R.K. 2007. Evaluation of the water retention curve focusing on pore water inflow into unsaturated surface layer. In:
for geosynthetic clay liners. Geotextiles and Geomembranes 25(1), Physical Modelling in Geotechnics 2, 1017-1021. A. McNamara, S.
2-9. Divall, R. Goodey, N. Taylor, S. Stallebrass and J. Panchal (eds.).
Spagnoli G., Romero E., Fraccica A., Arroyo M. & Gómez R. 2021. The CRC Press, London.
effect of curing conditions on the hydro-mechanical properties of a Takano D., Lenoir N., Otani J. & Hall S.A. 2015. Localised deformation
metakaolin-based soilcrete. Géotechnique. Ahead of Print. in a wide-grained sand under triaxial compression revealed by X-ray
Speranza G., Ferrari A. & Laloui L. 2020. A physical model for the tomography and digital image correlation. Soils and Foundations
interaction between unsaturated soils and retaining structures. 4th 55(4), 906-915.
European Conference on Unsaturated Soils, E-UNSAT 2020, Takano D., Otani J., Nagatani H. & Mukunoki T. 2006. Application of x-
Cardoso R., Jommi C. & Romero E. (eds), E3S Web of Conferences ray CT on boundary value problems in geotechnical engineering:
195, 03013. EDP Sciences. research on tunnel face failure. In GeoCongress 2006: Geotechnical
Springman S.M., Jommi C. & Teysseire P. 2003. Instabilities on moraine Engineering in the Information Technology Age 1-6.
slopes induced by loss of suction: a case history. Géotechnique 53(1), Takayama Y., Tachibana S., Iizuka A., Kawai K. & Kobayashi I. 2017.
3-10. Constitutive modeling for compacted bentonite buffer materials as
Starkloff T., Larsbo M., Stolte J., Hessel R. & Ritsema C. 2017. unsaturated and saturated porous media. Soils and foundations 57(1),
Quantifying the impact of a succession of freezing-thawing cycles 80-91.
on the pore network of a silty clay loam and a loamy sand topsoil Take W.A. & Bolton M.D. 2003. Tensiometer saturation and the reliable
using X-ray tomography. Catena 156, 365-374. measurement of soil suction. Géotechnique 53(2), 159-172.
Štástka J. 2014. Mock-up Josef Demonstration Experiment. Tunel 23(2), Take W.A. 2018. Current and emerging physical modelling technologies.
65-73. ISSN 1211-0728. Physical Modelling in Geotechnics 1, 101-109. A. McNamara, S.
Stavropoulou E., Andò E., Roubin E., Lenoir N. et al. 2020. Dynamics of Divall, R. Goodey, N. Taylor, S. Stallebrass and J. Panchal (eds.).
water absorption in callovo-oxfordian claystone revealed with CRC Press, London.
multimodal x-ray and neutron tomography. Frontiers in Earth Tamagnini R. 2004. An extended Cam-clay model for unsaturated soils
Science, 8, 6. with hydraulic hysteresis. Géotechnique 54(3), 223-228.
Stewart M.A. & McCartney J.S. 2014. Centrifuge modeling of soil- Tamagnini R. & Pastor M. 2004. A thermodynamically based model for
structure interaction in energy foundations. J. Geotech. Geoenviron. unsaturated soils: a new framework for generalized plasticity. In
Eng.140(4), 04013044. Proc. 2nd Int. Workshop on Unsaturated Soils, Capri.
Stirling R.A., Toll D.G., Glendinning S., Helm P.R. et al. 2021. Weather- Tan H., Chen F., Chen J. & Gao Y. 2019. Direct shear tests of shear
driven deterioration processes affecting the performance of strength of soils reinforced by geomats and plant roots. Geotextiles
embankment slopes. Géotechnique 71(11), 957-969. and Geomembranes, 47(6), 780-791.
Stropeit, K., Wheeler, S., & Cui, Y.J. 2008. An anisotropic elasto-plastic Tan S.H., Wong S.W., Chin D.J., Lee M.L. et al. 2018. Soil column
model for unsaturated soils. In: 1st European Conference on infiltration tests on biomediated capillary barrier systems for
Unsaturated Soils, Durham, England, 625-631. mitigating rainfall-induced landslides. Environmental Earth
Sciences 77(16), 1-13.
Tang A.M. & Cui Y.J. 2005. Controlling suction by the vapour
initial densities. International Journal for Numerical and Analytical equilibrium technique at different temperatures and its application in
Methods in Geomechanics 27(12), 1079-1098. determining the water retention properties of MX80 clay. Can. Geot.
J. 42, 287-296.
Tang A.M., Cui, Y.J., Qian, L.X., Delage, P. & Ye W.M. 2010. Calibration
compacted soils. International journal for numerical and analytical of the osmotic technique of controlling suction with respect to
methods in geomechanics 31(11), 1257-1279. temperature using a miniature tensiometer. Can. Geotech. J. 47(3),
Sun D.A., Sheng D. & Sloan S.W. 2007b. Elastoplastic modelling of 359–365.
hydraulic and stress–strain behaviour of unsaturated soils. Tang A.M., Hughes P.N., Dijkstra T.A., Askarinejad A. et al. 2018.
Mechanics of Materials 39(3), 212-221. Atmosphere-vegetation-soil interactions in a climate change context;
Sun D.A., Cui H.B., Matsuoka H. & Sheng D.C. 2007c. A three- impact of changing conditions on engineered transport infrastructure
dimensional elastoplastic model for unsaturated compacted soils slopes in Europe. Quarterly Journal of Engineering Geology and
with hydraulic hysteresis. Soils and Foundations 47(2), 253-264. Hydrology 51, 156-168.
Sun D.A., Sheng D., Xiang L. & Sloan S.W. 2008. Elastoplastic Tang C.S., Cheng Q., Leng T., Shi B. et al. 2020. Effects of wetting-
prediction of hydro-mechanical behaviour of unsaturated soils under drying cycles and desiccation cracks on mechanical behavior of an
undrained conditions. Computers and Geotechnics 35(6), 845-852. unsaturated soil. CATENA 194, 104721.
Sun H., Mašín D., Najser J. & Scaringi G. 2020. Water retention of a Tang C.S., Shi B., Liu C., Suo W.B. & Gao L. 2011. Experimental
345
characterization of shrinkage and desiccation cracking in thin clay landslides and slopes. Quarterly Journal of Engineering Geology and
layer. Applied Clay Science 52(1-2), 69-77. Hydrogeology 44(1), 23–33.
Tang, C. S., Zhu, C., Leng, T., Shi, B., Cheng, Q., & Zeng, H. 2019. Tollenaar R.N., van Paassen L.A. & Jommi C. 2018. Small-scale
Three-dimensional characterization of desiccation cracking behavior evaporation tests on clay: influence of drying rate on clayey soil layer.
of compacted clayey soil using X-ray computed tomography. Can. Geotech. J. 55, 437-445.
Engineering Geology, 255, 1-10. Torres-Serra J., Rodríguez-Ferran A. & Romero E. 2021a. Classification
Tarantino A. 2009. A water retention model for deformable soils. of granular materials via flowability-based clustering with
Geotechnique 59(9):751–762. application to bulk feeding. Powder Technology 378, 288–302.
Tarantino A. & De Col E. 2008. Compaction behaviour of clay. Torres-Serra J., Rodríguez-Ferran A. & Romero E. 2021b. Study of grain-
Géotechnique 58(3), 199 –213. scale effects in bulk handling using discrete element simulations.
Tarantino A. & Di Donna A. 2019. Mechanics of unsaturated soils: Powder Technology 382, 284–299.
simple approaches for routine engineering practice. Rivista Italiana Torres-Serra J., Romero E. & Rodriguez-Ferran A. 2018. Hygroscopicity
di Geotecnica 4: 5-46. issues in powder and grain technology. Proc. 7th Int. Conf.
Tarantino A., Gallipoli D., Augarde C.E., De Gennaro V. et al. 2011. Unsaturated Soils, Hong Kong, 805–810.
Benchmark of experimental techniques for measuring and Torres-Serra J., Romero E. & Rodríguez-Ferran A. 2020. A new column
controlling suction. Géotechnique 61(4), 303–312. collapse apparatus for the characterisation of the flowability of
Tarantino A. & Mongiovì L. 2003. Calibration of tensiometer for direct granular materials. Powder Technology 362, 559–577.
measurement of matric suction. Géotechnique 53(1), 137-141. Toyota H., Sakai N. & Nishimura, T. 2001. Effects of the stress history
Tarantino A., Ridley A.M. & Toll D.G. 2008. Field measurement of due to unsaturation and drainage condition on shear properties of
suction, water content, and water permeability. Geotechnical and unsaturated cohesive soil. Soils and Foundations 41(1), 13-24.
Geological Engineering 26(6), 751-782. Toyota H., Nakamura K. & Sramoon, W. 2004. Failure criterion of
Tarantino A. & Tombolato S. 2005. Coupling of hydraulic and unsaturated soil considering tensile stress under three-dimensional
mechanical behaviour in unsaturated compacted clay. Géotechnique stress conditions. Soils and Foundations 44(5), 1-13.
55(4), 307-317. Trabelsi H., Romero E. & Jamei M. 2018. Tensile strength during drying
Tatsuoka F. & Gomes Correia A. 2018. Importance of controlling the of remoulded and compacted clay: the role of fabric and water
degree of saturation in soil compaction linked to soil structure design. retention. Appl. Clay Sci. 162, 57–68.
Transportation Geotechnics 17: 3-27. Tripathy S., Bag R. & Thomas H.R. 2015. Enhanced isothermal effect on
Tatsuoka F. 2015. Compaction characteristics and physical properties of swelling pressure of compacted MX80 bentonite. Engineering
compacted soil controlled by the degree of saturation. Proc. 6th Int. Geology for Society and Territory 6, 537-539. Springer, Cham.
Conf. on Deformation Characteristics of Geomaterials, Buenos Aires, Tripathy S., Tadza M.Y.M. & Thomas H.R. 2011. On the intrusion of
40-78. polyethylene glycol during osmotic tests. Géotechnique Letters 1 (3),
Taylor-Noonan A., Arpin N., Cabrera M., Siemens G. & Take W.A. 2020. 47–51.
Mechanism of air entry during collapse of saturated and unsaturated Tu X.B., Kwong A.K.L., Dai F.C., Tham L.G. & Min H. 2009. Field
columns of transparent granular soil. EGU General Assembly 2020. monitoring of rainfall infiltration in a loess slope and analysis of
Teng J., Shan F., He Z., Zhang S. et al. 2019. Experimental study of ice failure mechanism of rainfall-induced landslides. Engineering
accumulation in unsaturated clean sand. Géotechnique 69(3), 251- Geology 105(1-2), 134-150.
259. Uchaipichat A. & Khalili N. 2009. Experimental investigation of thermo-
Terzaghi K. 1923. Die Berechnung der DurchlaÈssigkeitsziffer des Tones hydro-mechanical behaviour of an unsaturated silt. Géotechnique
aus dem Verlauf der hydrodynamischen Spannungserscheinungen, 59(4), 339–353.
Sitzungber. Akad. Wiss. Wien, 132, 125-138. Vahedifard F., Leshchinsky B.A., Mortezaei K. & Lu N. 2015. Active
Terzaghi K. 1936. The shearing resistance of saturated soils. Proc. 1st Int. earth pressures for unsaturated retaining structures. J. of
Conf. Soil Mech., Cambridge, Mass., 1, 54-56. Geotechnical and Geoenvironmental Engineering, 141, 11.
Terzaghi K. 1945. Stress conditions for the failure of saturated concrete Vail M., Zhu C., Tang C.S., Anderson L. et al. 2019. Desiccation cracking
and rock, Proc. Am. Soc. Testing Materials 45, 777-801. behavior of MICP-treated bentonite. Geosciences, 9(9), 385.
Terzis D. & Laloui L. 2018. 3-D micro-architecture and mechanical van Genuchten M.T. 1980. A closed-form equation for predicting the
response of soil cemented via microbial-induced calcite precipitation. hydraulic conductivity of unsaturated soils. Soil Science Society of
Scientific Reports 8, 1416. America Journal, 44(5), 892-898.
Terzis D. & Laloui L. 2019. Cell-free soil bio-cementation with strength, Vanapalli S.K. & Fredlund D.G. 2000. Comparison of different
dilatancy and fabric characterization. Acta Geotechnica 14, 639–656. procedures to predict unsaturated soil shear strength. Geotechnical
Terzis D., Bernier-Latmani R. & Laloui L. 2016. Fabric characteristics Special Publication 287(99), Geo-Denver Conference.
and mechanical response of bio-improved sand to various treatment Vanapalli S.K., Fredlund D.G. & Pufahl D.E. 1999. The influence of soil
conditions. Géotechnique Letters 6, 50–57. structure and stress history on the soil-water characteristics of a
Thom R., Sivakumar R. Sivakumar V. Murray E.J. & Mackinnon P. 2007, compacted fill. Géotechnique 49(2): 143-159.
Pore size distribution of unsaturated compacted kaolin: the initial Vanapalli S.K., Fredlund D.G., Pufahl D.E. & Clifton A.W. 1996. Model
states and final states following saturation. Géotechnique 57(5), 469- for the prediction of shear strength with respect to soil suction. Can.
474. Geotech. J. 33, 379-392.
Thu T.M., Rahardjo H. & Leong E.C. 2007. Elastoplastic model for Vanapalli S.K., Nicotera M.V. & Sharma R.S. 2008. Axis translation and
unsaturated soil with incorporation of the soil-water characteristic negative water column techniques for suction control. Geotechnical
curve. Canadian Geotechnical Journal 44(1), 67-77. and Geological Engineering 26(6): 645–660.
Tian Z., Kool D., Ren T., Horton R. & Heitman J.L. 2018. Determining Vandoorne R., Gräbe P.J. & Heymann G. 2022. Polyethylene glycol and
in-situ unsaturated soil hydraulic conductivity at a fine depth scale membrane processes applied to suction control in geotechnical
with heat pulse and water potential sensors. Journal of Hydrology osmotic testing. International Journal of Geotechnical Engineering
564, 802-810. 16(1): 103-122.
Timms W., Whelan M., Acworth I., McGeeney D. et al. 2014. A novel Vann J.D. & Houston S. 2021. Field Suction Profiles for Expansive Soil.
Centrifuge Permeameter to characterize flow through low Journal of Geotechnical and Geoenvironmental Engineering
permeability strata. Physical Modelling in Geotechnics – Proc. of the 147(9),04021080.
8th Int. Conf. on Physical Modelling in Geotechnics 2014, ICPMG Vardon P.J. 2014. ‘Climatic influence on geotechnical infrastructure: a
2014, 193-199. review’. J. Environ. Geotech. 2, 166–174.
Tiwari N., Satyam N. & Sha Vaughan P.R. 1985. Pore pressures due to infiltration into partly saturated
performance evaluation of expansive soil biotreated with indigenous slopes. First International Conference on Geomechanics in Tropical,
bacteria using MICP method. Scientific Reports 11, 10324. Lateritic and Saprolitic Soils. Brasilia, Vol. 2. ABMS. 61-71.
Toll D.G., Lourenço S.D.N. & Mendes J. 2013. Advances in suction Vaunat J., Cante J.C., Ledesma A. & Gens A. 2000. A stress point
measurements using high suction tensiometers. Engineering algorithm for an elastoplastic model in unsaturated soils.
Geology 165, 29-37. International Journal of Plasticity 16(2), 121-141.
Toll D.G., Lourenço S.D.N., Mendes J., Gallipoli D., Evans F.D., Vaunat J. & Casini 2017. A procedure for the direct determination of
Augarde C.E., Cui Y.J., Tang A.M., Rojas J.C., Pagano L., Mancuso Bishop's parameter from changes in pore size distribution.
C., Zingariello C. & Tarantino A. 2011. Soil suction monitoring for Geotechnique 67(7), 631-636.
346
Venuleo S., Laloui L., Terzis D., Hueckel T. & Hassan M. 2016. unsaturated soil based on bounding surface plasticity. Journal of
Microbially induced calcite precipitation effect on soil thermal Rock Mechanics and Geotechnical Engineering 2(2), 135-142.
conductivity. Géotechnique Letters 6, 39–44. Woodburn J.A. & Lucas B. 1995. New approaches to the laboratory and
Veylon G., Ghestem M., Stokes A. & Bernard A. 2015. Quantification of field measurement of soil suction. Proc. 1st Int. Conf. on Unsaturated
mechanical and hydric components of soil reinforcement by plant Soils, Paris, September 6-8, 1995. Unsaturated Soils. E.E. Alonso
roots. Can. Geotech. J. 52, 1839–1849. and P. Delage (eds.). A.A. Balkema / Presses des Ponts et Chaussées,
Vieira A., Alberdi-Pagola M., Christodoulides P., Javed S. et al. 2017. Paris, 2, 667-671.
Characterisation of Ground Thermal and Thermo-Mechanical Xiong Y.L., Ye G.L., Xie Y., Ye B., Zhang S. & Zhang F. 2019. A unified
Behaviour for Shallow Geothermal Energy Applications. Energies constitutive model for unsaturated soil under monotonic and cyclic
10(12), 2044. loading. Acta Geotechnica 14(2), 313-328.
Vieira, A. M and Marinho, F.A.M. (2011). Seasonal Suction Variation on Xu Y. 2004. Fractal approach to unsaturated shear strength. J. Geotech.
a Residual Soil Slope in São Paulo. III Brazilian Slope Stability Geoenviron. Eng. ASCE, 130: 3, 264–273.
Congress (COBRAE). pp.287-295 (In Portuguese). Xu Y. & Cao L. 2015. Fractal representation for effective stress of
Viggiani G., Andò E., Takano D. & Santamarina J.C. 2015. Laboratory unsaturated soils. Int. J. Geomech.15(6): 04014098.
X-ray tomography: a valuable experimental tool for revealing Yamamoto S., Sato S., Tomoyuki S., Romero E. et al. 2019. Mechanical
processes in soils. Geotechnical Testing Journal, 38(1), 61-71. characteristics and water retention properties of unsaturated
Villar M.V. & Lloret A. 2004. Influence of temperature on the hydro- 'bentonite-based engineered barrier materials' based on controlled-
mechanical behaviour of a compacted bentonite. Applied Clay suction tests. Journal of Japan Society of Civil Engineers 75(3), 257-
Science 26(1-4), 337-350. 272.
Villar M.V. 1999. Investigation of the behaviour of the bentonite by Yang C., Cui Y.J., Pereira J.M. & Huang M.S. 2008. A constitutive model
means of suction-controlled oedometer tests. Engineering Geology for unsaturated cemented soils under cyclic loading. Computers and
54, 67-73. Geotechnics 35(6), 853-859.
Villar M.V. 2005. MX-80 Bentonite. Thermal-Hydro-Mechanical Yang F., Hingerl F.F., Xiao X., Liu Y. et al. 2015 Extraction of pore-
Characterisation Performed at CIEMAT in the Context of the morphology and capillary pressure curves of porous media from
Prototype Project. Informes Técnicos CIEMAT 1053 synchrotron based tomography data. Sci. Rep., 5, 10635.
Villar M.V., Martín P.L., Bárcena I., García-Siñeriz J.L. et al. 2012. Long- Yang H., Rahardjo H., Leong E.C. & Fredlund D.G. 2004. A study of
term experimental evidences of saturation of compacted bentonite infiltration on three sand capillary barriers. Can. Geotech. J. 41(4),
under repository conditions. Engineering Geology 149, 57-69. 629-643.
Vitali D., Leung A.K., Feng S., Knappett J.A. & Ma L. 2021. Ce Yang Z., Cai H., Shao W., Huang D., Uchimura T., Lei X., .Tian H. &
ntrifuge modelling of the use of discretely spaced energy pile Qiao J. 2019. Clarifying the hydrological mechanisms and
row to reinforce unsaturated silt. Géotechnique. thresholds for rainfall-induced landslide: in situ monitoring of big
Vo T., Taiebat H. & Russell A.R. 2016. Interaction of a rotating rigid data to unsaturated slope stability analysis. Bulletin of Engineering
retaining wall with an unsaturated soil in experiments. Géotechnique Geology and the Environment 78(4), 2139-2150.
66(5), 366-377. Yang, Z.X., Li, X.S., & Yang, J. 2008. Quantifying and modelling fabric
Wang M., Li X., Liu Z., Liu J. & Chang D. 2020. Application of PIV anisotropy of granular soils. Géotechnique 58(4), 237-248.
Technique in Model Test of Frost Heave of Unsaturated Soil. Journal Ye W.M., Borrell N.C., Zhu J.Y., Chen B. & Chen Y.G. 2014. Advances
of Cold Regions Engineering 34(3), 04020014. on the investigation of the hydraulic behavior of compacted GMZ
Wang Q., Cui Y.J., Tang A.M., Barnichon J.D., Saba S. & Ye W.M. 2013. bentonite. Engineering Geology 169, 41-49.
Hydraulic conductivity and microstructure changes of compacted Ye W.M., Wan M., Chen B., Chen Y.G. et al. 2013. Temperature effects
bentonite/sand mixture during hydration. Engineering Geology 164, on the swelling pressure and saturated hydraulic conductivity of the
67-76. compacted GMZ01 bentonite. Environmental Earth Sciences 68(1),
Wang Q., Wang S., Su W., Pan D., Zhang Z. & Ye W.M. 2021. 281-288.
Interpretation of grouting characteristics in unsaturated sand from Ye, W.M., Lai, X.L., Wang, Q., Chen, Y.G., Chen, B., & Cui, Y. J. 2014.
the perspective of water–air interface. Acta Geotechnica 1-12. An experimental investigation on the secondary compression of
Wang, S., Yang, P., & Yang, Z. J. 2018. Characterization of freeze–thaw unsaturated GMZ01 bentonite. Applied Clay Science 97, 104-109.
effects within clay by 3D X-ray Computed Tomography. Cold Yesiller N., Miller C.J., Inci G. & Yaldo K. 2000. Desiccation and
Regions Science and Technology, 148, 13-21. cracking behavior of three compacted landfill liner soils.
Webb P.A. & Orr C. 1977. Analytical methods in fine particle technology. Engineering Geology 57(1-2), 105-121.
Micromeritics Instrument Corp, Norcross. Yildiz A., Graf F., Rickli C. & Springman S.M. 2018. Determination of
Wheeler S.J. & Sivakumar V. 1995. An elasto-plastic critical state the shearing behaviour of root-permeated soils with a large-scale
framework for unsaturated soil. Géotechnique 45(1), 35-53. direct shear apparatus. Catena 166, 98-113.
Wheeler S.J. 1996. Inclusion of specific water volume within an elasto- Yin C., Jiang L., Sun K., Sun W. & Liang B. 2022. Influence of degree
plastic model for unsaturated soil. Canadian Geotechnical Journal of saturation on electrokinetic remediation of unsaturated soil.
33(1), 42-57. Korean J. Chem. Eng.
Wheeler S.J., Gallipoli D. & Karstunen M. 2002. Comments on use of Yuan S., Liu X & Buzzi O. 2021. A microstructural perspective on soil
the Barcelona Basic Model for unsaturated soils. International collapse. Géotechnique 71(2), 132–140.
journal for numerical and analytical methods in Geomechanics Yuan S., Liu X., Romero E., Delage P. & Buzzi O. 2020. Discussion on
26(15), 1561-1571. the separation of macropores and micropores in a compacted
Wheeler S.J., Sharma R.S. and Buisson M.S.R. 2003. Coupling of expansive clay. Géotechnique Letters 10(3), 454-460.
hydraulic hysteresis and stress–strain behaviour in unsaturated soils. Yuan S., Liu X. & Buzzi O. 2017. Calibration of a coupled model to
Géotechnique 53(1), 41-54. predict the magnitude of suction generated by osmotic technique
Wickland B.E., Wilson G.W., Fredlund D.G. & Wijewickreme D. 2006. with PES membranes and temperature effect. Geotechnical Testing
Unsaturated Properties of Mixtures of Waste Rock and Tailings. Proc. Journal 40 (1), 144–149.
4th Int. Conf. Unsaturated Soils, Carefree, Arizona, US, 883-893. Zeng H., Tang C.-S., Zhu C., Cheng Q., Lin Z.-z. & Shi B. 2022.
Wijaya M. & Leong E.C. 2016. Performance of high-capacity Investigating soil desiccation cracking using an infrared thermal
tensiometer in constant water content oedometer test. Int. J. Geo- imaging technique. Water Resources Research 58, e2021WR030916.
Engineering 7(1), 13. Zerhouni M.I. 1991. Application des réseaux de Petri continus à l’analyse
Wildenschild, D., Hopmans, J. W., Rivers, M. L., & Kent, A. J. R., 2005. dynamique des systèmes de production- These do Doctorat do l’lNP
Quantitative analysis of flow processes in a sand using synchrotron- Grenoble, France
based X-ray microtomography. Vadose Zone Journal, 4(1), 112-126. Zhan T.L., Li H., Jia G.W., Chen Y.M. & Fredlund D.G. 2014. Physical
Williams J. & Shaykewich C.F. 1969. An evaluation of polyethylene and numerical study of lateral diversion by three-layer inclined
glycol (P.E.G.) 6000 and P.E.G. 20,000 in the osmotic control of soil capillary barrier covers under humid climatic conditions. Can.
water potential. Canadian Journal of Soil Science 102(6), 394-398. Geotech. J. 51(12), 1438-1448.
Wong C., Pedrotti M., El Mountassir G. & Lunn R.J. 2018. A study on Zhang D., Wang J., Chen C. & Wang S. 2020. The compression and
the mechanical interaction between soil and colloidal silica gel for collapse behaviour of intact loess in suction-monitored triaxial
ground improvement. Engineering Geology 243, 84-100. apparatus. Acta Geotechnica, 15(2), 529-548.Zhang C.B., Chen
Wong H., Morvan M. & Branque D. 2010. A 13-parameter model for L.H., Liu Y.P., Ji X.D. & Liu X.P. 2010. Triaxial compression test of
347
soil root composites to evaluate influence of roots on soil shear and Geotechnical Engineering 5(2), 162-167.
strength. Ecol Eng. 36(1), 19–26. Zur, B. 1966. Osmotic control of the matric soil-water potential. Soil
Zhang F. & Ikariya T. 2011. A new model for unsaturated soil using Science 102(6): 394–398.
skeleton stress and degree of saturation as state variables. Soils and
Foundations 51(1), 67-81.
Zhang F., Ye B., Noda T., Nakano M., & Nakai K. 2007. Explanation of
cyclic mobility of soils: approach by stress-induced anisotropy. Soils
and Foundations 47(4), 635–648.
constitutive model for partially saturated soils. International journal
for numerical and analytical methods in geomechanics 32(14), 1715-
1735.
Zhang X., Li L., Chen G. & Lytton R. 2015. A photogrammetry-based
method to measure total and local volume changes of unsaturated
soils during triaxial testing. Acta Geotechnica 10(1), 55–82.
Zhang Y., Zhou A., Nazem M. & Carter J. 2019. Finite element
implementation of a fully coupled hydro-mechanical model and
unsaturated soil analysis under hydraulic and mechanical loads.
Computers and Geotechnics 110, 222-241.
Zhang, H.W., Qin, J.M., Sanavia, L., & Schrefler, B.A. 2007. Some
theoretical aspects of strain localization analysis of multiphase
porous media with regularized constitutive models. Mechanics of
Advanced Materials and Structures, 14(2), 107-130.
Zhang, H.W., Sanavia, L., & Schrefler, B.A. 1999. An internal length
scale in dynamic strain localization of multiphase porous media.
Journal on Experiments, Modelling and Computation
Zhao C.F., Salami Y., Hicher P.Y. & Yin Z.Y. 2019. Multiscale modeling
of unsaturated granular materials based on thermodynamic
principles. Continuum Mechanics and Thermodynamics 31(1), 341-
359.
Zhao C.G., Liu Y., & Gao F.P. 2010. Work and energy equations and the
principle of generalised effective stress for unsaturated soils. Int. J.
Numer. Anal. Methods Geomech. 34, 920–936.
Zhou A. & Sheng D. 2015. An advanced hydro-mechanical constitutive
model for unsaturated soils with different initial densities.
Computers and Geotechnics 63, 46-66.
Zhou A., Wu S., Li J. & Sheng D. 2018. Including degree of capillary
saturation into constitutive modelling of unsaturated soils.
Computers and Geotechnics 95, 82-98.
Zhou A.N., Sheng D., Sloan S.W. & Gens A. 2012a. Interpretation of
unsaturated soil behaviour in the stress-saturation space, I: volume
change and water retention behaviour. Comput. Geotech. 43, 178-
187.
Zhou A.N., Sheng D., Sloan S.W. & Gens A. 2012b. Interpretation of
unsaturated soil behaviour in the stress–saturation space: II:
constitutive relationships and validations. Computers and
Geotechnics 43, 111-123.
Zhou A., Yang F. Wen-Chieh C. & Junran Z. 2019. A fractal model to
interpret porosity-dependent hydraulic properties for unsaturated
soils.” Advances in Civil Engineering 2019: 3965803.
Zhou C. & Ng C.W.W. 2014). A new and simple stress-dependent water
retention model for unsaturated soil. Computers and Geotechnics
62: 216–222.
Zhou C., Ng C.W.W. & Chen R. 2015. A bounding surface plasticity
model for unsaturated soil at small strains. International Journal for
Numerical and Analytical Methods in Geomechanics 39(11), 1141-
1164.
Zhou Z., Leung A.K., Karimzadeh A.A., Lau C.H. & Li K.W. 2021.
Infiltration through an Artificially Hydrophobized Silica Sand
Barrier. Journal of Geotechnical and Geoenvironmental Engineering
147(6), 06021006.
Zhou, Z., Leung. A. K., Zhu, W. J. & Li Y. 2021. Hydromechanical
behaviour of unsaturated artificially-hydrophobized sand:
compression, shearing and dilatancy. Engineering Geology, 291,
106223.
Zielinski M., Sanchez M., Romero E. & Atique A. 2014. Precise
observation of soil surface curling. Geoderma 226–227, 85-93.
Zienkiewicz, O.C. & Cormeau, I.C. 1974. Visco-plasticity - plasticity and
creep in elastic solids - a unified numerical solution approach.
International Journal for Numerical Methods in Engineering 8, 821-
845.
Zornberg, J. G., LaFountain, L., & Caldwell, J. A. 2003. Analysis and
design of evapotranspirative cover for hazardous waste landfill. J. of
Geotechnical and Geoenvironmental Engineering, 129(6): 427-438.
Zou L., Wang S. & Lai X. 2013. Creep model for unsaturated soils in
sliding zone of Qianjiangping landslide. Journal of Rock Mechanics
348
State of the Art 6: Earthquake Geotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Nonlinear dynamic analyses involving soil liquefaction or cyclic softening
Analyses dynamiques non linéaires impliquant la liquéfaction du sol ou l’adoucissement cyclique
Department of Civil and Environmental Engineering, University of California at Davis, USA, [email protected]
ABSTRACT: The state of the art for nonlinear dynamic analyses (NDAs) of geotechnical structures and soil-structure systems
affected by soils susceptible to liquefaction or cyclic softening during earthquake loading is reviewed. The state of the art for NDAs
has continued improving through extensive efforts across each of its components. Discussion topics include the components of
seismic performance evaluations, numerical methods, constitutive models for sands and clays (developments, calibration, validation,
and challenges), NDA procedures (components, validation, geologic variability, ground motions, and challenges), and future needs
for advancement.
RÉSUMÉ : L’état de l’art pour les analyses dynamiques non linéaires (NDA) des structures géotechniques et des systèmes de structure
du sol affectés par les sols sensibles à la liquéfaction ou à l’adoucissement cyclique pendant la charge sismique est examiné. L’état de
l’art pour les NDA a continué de s’améliorer grâce à des efforts considérables dans chacune de ses composantes. Les sujets de discussion
comprennent les composantes des évaluations de la performance sismique, les méthodes numériques, les modèles constitutifs pour les
sables et les argiles (développements, étalonnage, validation et défis), les procédures NDA (composants, validation, variabilité
géologique, mouvements du sol et défis) et les besoins futurs en matière d’avancement.
KEYWORDS: numerical, constitutive, earthquake, liquefaction, validation.
1 INTRODUCTION NDA challenges. Lastly, future directions and needs in
advancing the state of the art are discussed.
Nonlinear dynamic analyses (NDAs) have continued to advance
in capabilities and use for evaluating the seismic performance of
geotechnical structures and soil-structure systems affected by 2 EVALUATING SEISMIC PERFORMANCE
potentially significant ground deformations, such as may be
expected when there is the potential for liquefaction of Evaluating seismic performance for a geotechnical structure or
cohesionless (sand-like) soils or cyclic softening of cohesive soil-structure system, deterministically or probabilistically,
(clay-like) soils. NDAs generally provide an improved basis for involves synthesizing information from the site characterization
estimating dynamic response and ground deformation over (e.g., stratigraphy, material properties), seismic hazard
simpler analysis methods that do not account for dynamics or are characterization, past performance observations (e.g., during
limited to idealized geometries or conditions. Applications span construction or past earthquakes), dynamic response analyses,
a broad range of geotechnical structures and soil structure damage assessment (e.g., cracking, need for repairs or
systems, which may be collectively referred to as geotechnical replacement), and consequence assessment (e.g., operations
infrastructure for brevity. Examples include foundation systems, downtime, life loss, economic loss, environmental impact). The
transportation networks, ports, locks, underground structures, uncertainties associated with each of these tasks propagate
levees, embankment dams, tailings impoundments, slopes, through the evaluation process and manifest as bias or dispersion
offshore structures, lifelines, and ground improvements. NDAs in the results. The evaluation process is often iterative, with
are particularly valuable for performance-based evaluations. individual tasks refined whenever the potential impact on final
The quality of results from an NDA study depends on several decisions, given the performance objectives, warrants the
technical and nontechnical factors, including the numerical additional engineering time and cost. These components and
methods, constitutive models, modeling procedures, site steps in the seismic evaluation of a geotechnical structure or soil-
characterization, input ground motions, analyst expertise, and structure system are illustrated in Figure 1.
review processes. The state of the art for NDAs therefore reflects Analysis models for computing seismic demands on
the state of the art in each of its components and the way they are geotechnical structures and soil-structure systems affected by
integrated together. potentially significant ground deformations may range in
This paper examines the state of the art for NDAs, with an complexity from equivalent static procedures to NDAs with
emphasis on aspects important for geotechnical infrastructure various degrees of sophistication. Equivalent static procedures
affected by soil liquefaction or cyclic softening leading to for estimating ground deformations and associated kinematic
potentially significant ground deformations. The components of loads on embedded structures have significant limitations
seismic performance evaluations and the role of NDAs are regarding the mechanisms they can model and the level of insight
discussed first, followed by a brief review of available numerical into system responses they can provide, as well as a potential for
methods. Recent constitutive model developments for sands and greater bias or greater dispersion in predictions of performance.
clays are covered in greater detail, followed by discussion of their NDAs with simpler constitutive models offer some improvement
calibration and validation, including (i) general considerations, over equivalent static procedures but also can have significant
(ii) soil characterization versus model calibration tasks, (iii) key limitations and greater potential for bias. These simpler
properties, and (iv) examples of constitutive modeling engineering models can, in some cases, still be sufficient for
challenges. Recent developments related to NDAs are reviewed reaching sound decisions if their limitations are recognized and
and discussed, including (i) components of NDA procedures, (ii) reasonably accounted for through appropriate conservatism.
validation efforts using case histories and physical models, (iii) However, NDAs with more realistic constitutive models are
geologic variability, (iv) ground motions, and (v) examples of
351
Loading
(static, hydrologic,
seismic)
Analysis model Damage model Loss model Decisions
(simple to complex)
Site characterization Propagation of uncertainty?
(geologic model, site Past performance Value of extra effort?
investigation, properties) observations
Figure 1. Components and steps in evaluating seismic performance for geotechnical infrastructure
(after Boulanger & Ziotopoulou 2018).
generally preferable for significant projects and generally require is reasonable for geotechnical earthquake engineering problems
only modest additional effort for experienced analysts. involving soil liquefaction or cyclic softening. The u-p
The planning and execution of an NDA study requires a clear formulation is common, as illustrated by its use in FLAC and
understanding of the components of a seismic performance PLAXIS and availability in OpenSees (Arduino 2021). Effective
evaluation (Figure 1). The analysis team needs to understand the stress based analyses also require numerical methods formulated
basis for, and uncertainties in, the analysis inputs (e.g., to avoid or minimize adverse artifacts (e.g., mesh locking, hour-
stratigraphy, properties, loading), be familiar with how the glassing) and provide efficient and accurate solutions for
results will be used in procedures for estimating damages and strongly nonlinear constitutive equations, pore fluid coupling,
losses, and be clear on the performance criteria being used to and large deformations (e.g., McGann et al. 2012, 2015).
inform decision making. Estimating damages and losses may Numerical platforms that get widely used tend to have some
require evaluating secondary failure modes that are not directly common attributes, including (i) readily accessible as either an
simulated in the NDAs. For example, the slumping of a levee or open source or commercial product, (ii) appropriately complete
embankment dam may be accompanied by significant soil constitutive models for sands and clays along with adequate
longitudinal or transverse cracking, with the potential for structural modeling components, (iii) automated routines for
transverse cracking accentuated by significant variations in the tasks common to geotechnical applications, like specifying a
foundation stratigraphy and associated slumping displacements compliant base, imposing free-field lateral boundaries, and
along the embankment alignment. Seepage through any initializing stresses, (iv) robust numerical implementation, (v)
transverse cracks that extend below the adjacent water levels adequate documentation, manuals, and version controls, (vi)
(either at the time of the earthquake or later if the water levels graphical user interface for efficient model development,
rise before the cracks are repaired) may lead to progressive analysis execution, interpretation, quality control, and
erosion and eventual failure, depending on the nature of the soils, presentation of results, and (vii) established user community and
any internal zoning, severity of cracking, and other factors. record of prior validation studies. A relatively small number of
NDAs cannot directly simulate this type of secondary failure numerical platforms have become dominant for NDAs involving
mode, but an appropriate parametric study and informed liquefaction or cyclic softening as evident in the reviews
interpretation of the analysis results can provide valuable insights presented later or in various communities of practice (e.g.,
for evaluating such failure modes (USSD 2022). Accordingly, Deierlein & Zsarnoczay 2021, USSD 2022). These platforms
parametric studies are essential for identifying the sensitivity of include FLAC (Itasca 2019), FLIP (FLIP Consortium 2022),
the system response to uncertainties in the analysis inputs and for LIQCA (LIQCARI 2021), OpenSees (McKenna et al. 2010), and
presenting results in a format that informs the subsequent steps PLAXIS (Brinkgreve et al. 2017). Today's dominant platforms
of damage and loss assessment. The development of an do not include several of the more proprietary research codes
appropriate NDA model and an effective and adaptive program prominent in earlier NDA developments (e.g., see Finn 2000).
of parametric analyses similarly benefits from the early The dominant NDA platforms have shown continuous
establishment of the performance objectives. improvement over time, but these improvements require
sustained efforts and considerable resources. For this reason, the
3 NUMERICAL METHODS AND PLATFORMS breadth of advanced numerical methods that continue to be
developed for efficiently and robustly addressing complex
Numerical platforms based on the finite element method (FEM) phenomena (e.g., localizations, cracking, large strains, large
and finite difference method (FDM) are both used for effective deformations, multi-physics coupled processes) are not
stress-based, nonlinear dynamic analyses of geotechnical and necessarily readily available in the same numerical platforms that
soil-structure systems. Effective stress based analyses must have the other components essential for performing NDAs of
account for the coupled response of the soil skeleton and any pore geotechnical infrastructure. Over time, the progressive merging
fluid. Zienkiewicz & Shiomi (1984) described formulations for of the most sophisticated numerical methods and soil modeling
coupled responses in saturated soil, building on work by Biot capabilities within common NDA platforms offers the potential
(1941, 1960). The most general formulation is the u-p-U to better address several NDA limitations discussed later.
formulation, which discretizes the soil skeleton displacement, Structural modeling capabilities can be a key consideration
pore fluid pressure, and pore fluid displacement. The u-U in selecting the NDA platform(s) to be used in evaluating seismic
formulation is an irreducible solution derived for compressible performance of some soil-structure systems. For example, an
fluids. The u-p formulation is a solution derived by assuming the advantage of the OpenSees platform is its advanced structure and
fluid acceleration relative to the soil skeleton is negligible. interface modeling capabilities (e.g., fiber-section beam-column
Zienkiewicz et al. (1980) used an example problem to provide elements) which enable NDAs of soil-structure systems to be
guidance for evaluating the limits of validity for this latter performed with a balanced level of sophistication for the soil,
assumption based on the soil permeability and loading frequency. structure, and interface components (e.g., Shin et al. 2007,
Their solutions can be used to confirm that the u-p formulation
352
Khosravifar et al. 2014, Qiu et al. 2020). NDAs for soil-structure
systems involving significant ground deformations can be
approached several ways. Three-dimensional analyses provide
direct coupling of the soil and structural responses (e.g., Elgamal
et al. 2008; Figure 2), with interface elements used to account for
slip or gap formation. Two-dimensional analyses with tributary
out-of-plane dimensions can also provide direct coupling of the
soil and structural responses, although this approach requires
connecting the structural elements to the soil mesh through
interface elements that provide force transfer while
accommodating local relative movements. Analyses using a
large out-of-plane dimension for a 2D or 1D soil mesh can
essentially uncouple the soil response from the structural
response (due to the large difference in mass), in which case the
interface elements connecting structural elements to the soil
mesh must simulate total relative movements (i.e., the interface
forces produce negligible deformation within the soil mesh). For
example, interface elements along a pile for the latter case
correspond to conventional p-y, t-z, and q-z springs (e.g.,
Boulanger et al. 1999), whereas these interface elements need to
be stiffer for the former case because some of the relative
movement is accommodated by deformation within the soil mesh Figure 3. CEL simulations of retrogressive sliding and runout in sensitive
(e.g., Armstrong et al. 2013). Numerical platforms with simpler clay during earthquake loading (Islam et al. 2018).
structure and interface modeling capabilities are still commonly
used to evaluate demands imposed on soil-structure systems, runout processes for evaluating the potential consequences of
although this often requires performing more sophisticated, slope failures.
separate analyses of the structure itself to examine the impacts of Discrete element modeling (DEM) of particulate systems is
those demands. Accordingly, developing more balanced soil and providing unique and exciting capabilities for exploring
structural modeling capabilities in common numerical platforms fundamental soil behaviors. El Shamy et al. (2010) used PFC3D
would offer significant advantages for evaluating seismic (Itasca) to perform DEM simulations of liquefaction-induced
performance of certain types of geotechnical infrastructure.
lateral spreading in a 1g shaking table test; the results showed the
soil would densify near the base, as expected, but would also
loosen near the surface as a consequence of the pore pressure
diffusion patterns. Kuhn et al. (2014) used an open-source code
to perform DEM simulations of monotonic and cyclic loading
tests and develop scalar measures for predicting severity of
different loading sequences (Figure 4). Yang & Wu (2016) used
DEM simulations to demonstrate that a unique critical state line
is obtained irrespective of shearing mode and initial fabric
anisotropy. Kawamoto et al. (2016) outlined a level set DEM for
simulating systems of particles with arbitrary shapes, used it to
simulate triaxial loading of virtual specimens generated from
Figure 2. OpenSees finite element model of Humboldt Bay Bridge for X-Ray Computed Tomography (XRCT) images of triaxial
simulating nonlinear dynamic soil-structure interaction with liquefaction specimens, and showed reasonable agreement between
in foundation soils (Elgamal et al. 2008; courtesy Ahmed Elgamal). computed and experimental stress-strain responses. Wang et al.
Analyses for extremely large deformations like in (2019) used 2D DEM simulations to describe fabric anisotropy
penetration, landslide, or runout problems require more recently and evolution of strains under principal stress rotation. Yang et
developed numerical methods. Soga et al. (2016) reviewed al. (2021a) used 3D DEM simulations to study the microstructure
alternative approaches and applications for modeling landslides, and evolution of different measures of fabric before and after
including mesh-based methods [FEM, FDM, arbitrary liquefaction triggering in unidirectional cyclic shearing. Yang et
Lagrangian Eulerian (ALE), and coupled Eulerian–Lagrangian al. (2021b) extended that study to multidirectional cyclic
(CEL)] and mesh-free methods [material point method (MPM), shearing. The results of DEM studies like these few examples are
smooth particle hydrodynamics (SPH), particle finite element providing invaluable insights on fundamental behaviors and
method (PFEM), FEM with Lagrangian integration points, and informing ongoing improvements in continuum constitutive
element-free Galerkin (EFG)]. Soga et al. (2016) subsequently models.
used MPM with two approaches for soil-pore fluid coupling to
simulate complex and progressive failure mechanisms in a model
test and a field case study. Arduino (2018) reviewed the 4 CONSTITUTIVE MODELS
modeling of soils after liquefaction triggering, including the use
An overview of recent developments for constitutive modeling
of MPM, SPH and DEM for simulating large strain and large
of sands and clays in NDAs is provided in this section. The
deformation problems. Islam et al. (2018) used ABAQUS to
development of a constitutive model for use in NDAs
perform total stress Eulerian analyses of earthquake-induced
encompasses constitutive theory, numerical implementation, and
progressive failure mechanisms in sensitive clay slopes
verification, each of which is a complex subject of its own. This
(Figure 3). Cuomo et al. (2019) used MPM to simulate
section focuses on reviewing the different constitutive models
liquefaction and retrogressive slope failures with the hypoplastic
that have become available in recent years, while referring
constitutive model by Wolffersdorff (1996). Advances in these
readers to the original sources for details on the theory,
numerical methods are providing greatly improved capabilities
for many challenging problems, including the evaluation of slope
353
This model was modified and implemented as a dynamic link
library for FLAC by Boulanger & Ziotopoulou (2013, 2017)
and Ziotopoulou & Boulanger (2013, 2016). It was
subsequently implemented in OpenSees by Chen & Arduino
(2021) and PLAXIS by Vilhar et al. (2018).
Iai et al. (2011) expanded on the strain space, multiple
mechanism (cocktail glass) model of Iai et al. (1992) by
developing a theoretical link between micromechanical and
macroscopic strain energies and using the results to establish
the dilatancy formulation. This model is implemented in FLIP.
Beaty & Byrne (2011) revised the UBCSAND model and
implemented it as version 904aR for use with FLAC. Later,
Galavi et al. (2013) implemented a related, but different,
version called UBC3D-PLM for use in PLAXIS.
Li & Dafalias (2012) developed a bounding surface plasticity
Figure 4. 3D DEM modeling of cyclic undrained simple shear loading model based on a proposed anisotropic critical state theory that
for assembly of bumpy compound clusters (from Kuhn et al. 2014). incorporates an anisotropic fabric variable in defining critical
state.
implementation, and verification work. Calibration and Liu & Muraleetharan (2012) coupled a soil-water
validation of these models are discussed in Section 5. characteristic curve with an elastoplastic constitutive model to
enable modeling the cyclic loading response of partially
4.1 Sands saturated soils.
Constitutive models that provide reasonably realistic Kan et al. (2014) modified the mapping rule for the UNSW
approximations of sand responses important to NDAs have bounding surface plasticity model and obtained improved
simulations for a range of cyclic loading conditions.
increased in number and become more widely accessible,
Taborda et al. (2014) started with the elastoplastic NTUA-
although simpler constitutive models are still common in some
practices. For example, there are a number of uncoupled cycle- Sand model by Papadimitriou & Bouckavalas (2002) and
demonstrated how changes to the expressions for the critical
counting models (e.g., Dawson et al. 2001, Beaty & Byrne 2008)
that focus on getting the first-order effects of liquefaction state line, hardening modulus, and yield surface improved
triggering on strength and stiffness. These simpler models cannot certain simulation capabilities.
simulate cyclic mobility or other complex responses and can Wang et al. (2014) developed a plasticity model with
produce unrealistic responses in certain situations; e.g., cases reversible and irreversible dilatancy to better simulate cyclic
where the transmission of shear stresses during cyclic mobility mobility responses. This model was implemented as
CycLiqCPSP in OpenSees.
after triggering of liquefaction is important. These simpler
Jefferies et al. (2015) modified Norsand to account for
constitutive models may still be useful in certain applications if,
principal stress rotation and provide some ability to
as noted previously, their limitations are recognized and
approximate cyclic mobility responses.
reasonably accounted for through appropriate conservatism. The
Dafalias & Taiebat (2016) presented a zero elastic range sand
wider adoption of more realistic constitutive models, with their
increased potential for generalized application, will benefit from plasticity model within the SANISAND family of models
(e.g., Taiebat & Dafalias 2008) and proposed a general
the further development of productivity tools that reduce the
overshooting correction scheme for bounding surface
effort involved in their calibration.
plasticity models.
Several new constitutive models and improved versions of
Tasiopoulou & Gerolymos (2016a) presented a critical state
existing constitutive models for sand have been produced over
compatible model (Ta-Ger sand) based on a new theoretical
the past decade. The following list of twenty eight studies, while
not exhaustive, provides a sense of these recent efforts. Some framework combining perfect elastoplasticity with a Bouc-
Wen motivated mapping rule. Tasiopoulou & Gerolymos
older constitutive models that have been directly modified in
(2016b) reformulated the model to account for intrinsic and
more recent studies or continue to be widely used are included.
stress-induced anisotropy.
Elgamal et al. (2002, 2003) and Yang et al. (2003) presented a
Dashti et al. (2017) presented a multi-laminate model with 17
pressure-dependent, multiple yield surface plasticity model
predefined planes for simulating fabric anisotropy under
that emphasized simulating post-triggering cyclic mobility
monotonic and cyclic loading.
responses and the associated accumulation of shear strains.
Nie et al. (2017) presented a constitutive model based on the
The model was implemented as PDMY02 in OpenSees.
Bouc-Wen model and endochronic theory to simulate cyclic
Dafalias & Manzari (2004), in extending their prior work,
presented a critical state, bounding surface plasticity model triaxial tests under isotropic and anisotropic consolidation
conditions.
with a fabric-dilatancy related tensor quantity to account for
Bin et al. (2018) developed a constitutive model that
the effect of fabric change during plastic dilation on the
transitions from solid (described by an elastoplastic relation)
contractive response upon reversal of loading direction. This
to fluid (described by Bingham relation) for unified modeling
work formed the basis for the SANISAND class of models
of soil behavior before and after flow liquefaction develops.
(Taiebat & Dafalias 2008).
Zhang et al. (2007) extended the elastoplastic model of Asaoka Khosravifar et al. (2018) produced the PDMY03 model by
modifying its predecessor in OpenSees to better approximate
et al. (2002) to better approximate stress-induced anisotropy
the variation of cyclic resistance ratio (CRR) with number of
and changes in overconsolidation. This model was later
uniform loading cycles, overburden stress ( 'vc), and initial
implemented in DBLEAVES, used in studies by Bao et al.
static shear stress.
(2012) and Kheradi et al. (2018), and further generalized in
Petalas et al. (2018, 2020) extended the SANISAND models
later work by Ye et al. (2012).
by incorporating evolving fabric anisotropy and the ensuing
Boulanger (2010) presented the critical state, bounding surface
effect of noncoaxiality on the dilatancy and plastic modulus,
plasticity, PM4Sand model that modified and extended the
and named the resulting models SANISAND-F and
Dafalias & Manzari (2004) model to improve modeling of
responses important for earthquake engineering applications. SANISAND-FN.
354
Figure 5. Simulations of the cyclic triaxial stress-strain and stress path responses for Ottawa F-65
sand at a relative density of about 72% using nine different constitutive models (modified from
Manzari et al. 2020).
Wang et al. (2019) developed a thixotropic fluid model for soil Pan et al. (2022) extended the Yang et al. (2018) model, which
liquefaction. follows the anisotropic critical state theory by Li & Dafalias
Liu et al. (2019, 2020) added a memory surface to the critical (2012), with fabric-dependent dilatancy and plastic hardening
state, bounding surface SANISAND framework to improve and an accumulated loading index in the plastic modulus
simulations of cyclic loading under large numbers of smaller relationship.
stress cycles, including the transition from cyclic ratcheting to Closely related advances have been also made in constitutive
shakedown for drained loading conditions. models for cemented sands under dynamic loading.
Barrero et al. (2020) extended the SANISAND models with an Andrianopoulous et al. (2016) used the established NTUA-
internal strain liquefaction factor that evolves at low mean Sand model to simulate laboratory and centrifuge model tests
effective stresses and controls plastic stiffness in the semi- of colloidal silica treated sand by re-calibrating the model
fluidized state. This model was named SANISAND-Sf. parameters or decreasing the pore fluid modulus.
Sturt et al. (2021) implemented a model in LS-DYNA based Rahimi et al. (2016) extended a critical state bounding surface
on the SANISAND framework and indicated that further plasticity model with cohesion and fabric anisotropy for
developments were planned to address identified limitations in simulating monotonic and cyclic loading tests of weakly
the constitutive responses. cemented sand.
Fuentes et al. (2020) developed an ISA-Hypoplasticity model Zhang et al. (2021) extended the Dafalias & Manzari (2004)
by extending the hypoplastic model of Wolffersdorff (1996) model to cemented sands by incorporating a bonding strength
with intergranular strain anisotropy (ISA) to account for cyclic variable and demonstrated its capabilities by simulating
mobility effects. monotonic and cyclic tests performed on cemented sand.
Cheng & Detournay (2021) developed the critical state, Several of the recently developed models produce reasonably
bounding surface plasticity P2P model, building off the realistic stress-strain and stress-path responses for conventional
Dafalias & Manzari (2004) model, and implemented it for use undrained cyclic loading conditions, while others produce less
with FLAC3D. than satisfactory responses. For example, the responses of
Yang et al. (2022) developed SANISAND-MSf, which several constitutive models were evaluated against laboratory
incorporates both the memory surface and semi-fluidized state element tests as part of the LEAP-2017 project (Manzari et al.
concepts into the Dafalias & Manzari (2004) model, 2020), with the results for a cyclic triaxial test in Figure 5 being
significantly improving the modeling of liquefaction representative. The first five models compared in Figure 5b-f
resistance and cyclic mobility responses. (Cocktail glass model, PM4Sand, SANISAND, PDMY02,
Elbadawy et al. (2022) modified the flow rule to better control CycLiqCPSP) produced reasonably realistic cyclic mobility
the contraction rate in the PDMY02 model in OpenSees and responses despite not replicating the observed differences in
subsequently obtained improved simulations for a series of strain accumulation in extension versus compression for this
centrifuge model tests for slightly sloping ground. triaxial test; the differences in computed versus measured strain
accumulation responses may partly reflect that this is just one of
355
Figure 6. Stress-strain and stress path responses for Cloverdale clay in undrained cyclic triaxial
loading (left plots) with SANICLAY-B simulations (right plots) (Seidalinov & Taiebat 2014).
several tests used in calibrating the models and that some models Seidalinov & Taiebat (2014) introduced a bounding surface
were formulated with a focus on plane-strain applications (hence formulation and damage mechanism to the simple anisotropic
their more symmetric strain accumulation). The Dafalias & SANICLAY model of Dafalias et al. (2006) and incorporated
Manzari model accumulated strains strongly biased to extension the destructuration mechanism of Taiebat et al. (2010). The
for this triaxial simulation (Figure 5g), and is known to lock-up resulting SANICLAY-B model was used for simulating
in repeating stress-strain loops for symmetric simple shear undrained cyclic loading responses of clay.
loading (Ziotopoulou et al. 2014). The Hypoplastic model Kimoto et al. (2014) developed a cyclic elasto-viscoplastic
(Figure 5h) and ISA-Hypoplasticity model (Figure 5i) did not model for clay with kinematic hardening and structural
reach zero or near zero effective stress (ru 100%), did not degradation that simulates cyclic softening and strain
produce the associated transient large reductions in shear softening responses.
stiffness observed with the other models, and accumulated strains Boulanger & Ziotopoulou (2018a, 2019) developed the critical
strongly biased to extension. The Bowl model (Figure 5j) state, bounding surface plasticity PM4Silt model for clays and
accumulated strains at excessive rates after the first triggering of plastic silts, building on the framework of the PM4Sand
ru 100%. Satisfactory modeling of cyclic mobility behavior in models, and implemented the model for use with FLAC. It was
conventional undrained cyclic tests is a necessary early step in subsequently implemented in OpenSees (Chen & Arduino
validating a model, after which reasonably successful models 2021) and PLAXIS (Bentley (2021).
must still be evaluated against a broader range of loading Qiu & Elgamal (2020) extended an existing multi-yield
conditions as discussed in the following section. surface plasticity model with a strain softening logic for
representing post-peak strain softening in clays, with the
4.2 Clays model implemented in OpenSees.
Stoecklin et al. (2020) presented a multi-surface model with
Constitutive models that provide reasonably realistic kinematic hardening for combined static and dynamic loading
approximations of clay responses important to NDAs have of clays, with the model implemented in ABAQUS.
similarly continued to improve and become more widely In general, the newer models produce reasonably realistic stress-
available, although simpler constitutive models also continue to strain and stress-path responses for undrained cyclic loading
be common in practice. For example, total stress models that conditions, as illustrated in Figures 6 and 7 for two of these recent
include nonlinear hysteretic responses prior to yielding at the constitutive models. The simulation of the undrained cyclic
specified shear strength are common (e.g., UBCHYST by loading response of Cloverdale clay using SANICLAY-B shown
Naesgaard et al. 2015; hysteretic damping models in FLAC). in Figure 6 (Seidalinov & Taiebat 2014) illustrates that model's
These simpler total stress models have inherent limitations, ability to produce reasonable hysteresis loops, generate excess
including the inability to simulate pore pressure generation, pore pressure, and rapidly accumulate shear strains as the clay
cyclic softening, or post-peak strain softening. These simpler progressively softened/degraded. The simulation of the
constitutive models may still be useful in certain applications, as undrained cyclic loading response of Fraser River silt using
discussed for simpler sand models, provided their limitations are PM4Silt shown in Figure 7 (Boulanger & Wijewickreme 2019)
recognized and reasonably accounted for. Wider adoption of illustrates that model's ability to also produce reasonable
more realistic constitutive models for clays, like those for sands, hysteresis loops, generate excess pore pressure, and
will benefit from productivity tools that reduce the effort progressively accumulate shear strains, with the overall response
involved in their calibration. similar to the cyclic mobility responses observed with dense-of-
New constitutive models and improved versions of existing critical sands. More thorough comparisons of these types of
constitutive models for clay are fewer than for sand. The constitutive models also require examining a wide range of
following five studies, while not exhaustive, are illustrative of different loading conditions, as discussed in the next section on
recent developments. calibration and validation.
356
Figure 7. Stress-strain and stress path responses for Fraser River silt in undrained cyclic simple shear
loading (upper plots) with PM4Silt simulations (lower plots) (Boulanger & Wijewickreme 2019).
5 CALIBRATION AND VALIDATION OF CONSTITUTIVE something that should be checked. This approach is tractable for
MODELS simulations of laboratory and physical model tests using
reconstituted soils under well-controlled conditions, but is
5.1 General considerations generally not tractable for applications to field conditions. In
contrast, a generalized approach to calibration may involve (1)
The goal in selecting, calibrating, or validating a constitutive synthesizing information from laboratory tests, in-situ tests,
model is to be able to simulate, as accurately as possible, the correlations between in-situ test measurements and soil
stress-strain responses of primary interest for a specific properties, and typical responses and empirical trends from
application using parameters that can be determined with experimental literature, (2) using the synthesized information to
reasonable effort. Meeting this goal requires establishing the set target responses and properties, and (3) calibrating the model
parameter and loading space over which the model has been to produce reasonable approximations of the target responses
validated, including identifying limitations and biases so the across all the loading conditions considered important for a
model is used appropriately. Consequently, the utility of a specific geotechnical system/structure. This approach is broadly
constitutive model increases with its ability to work across the tractable and offers significant advantages for NDAs of
broad range of conditions encountered in various geotechnical geotechnical systems. This approach benefits from clearly
infrastructure systems. separating the soil characterization and model calibration tasks,
The breadth of a constitutive model's validation space as discussed below.
generally improves with the soundness of its theoretical
framework, recognizing that some theoretical aspects are still 5.2 Separating soil characterization and model calibration
open to debate and all theoretical frameworks have limitations at
some level of detail. For example, the classical critical state Separating the soil characterization and model calibration steps
framework is an excellent framework for understanding and has the advantage of clearly separating the problems that arise
representing a broad range of soil behaviors, but the assumption due to data inconsistencies from the problems that arise in model
of a unique critical state line has its own limitations. Li & calibration (Figure 8). The soil characterization step should
Dafalias (2012) and Dafalias & Taiebat (2014) added fabric produce estimates of key properties using the best means
dependence to the critical state condition in their constitutive available with appropriate allowance for the limitations in each
models for sand and clay, respectively. Furthermore, the critical data source and should explicitly resolve any inconsistences
state line may evolve (shift) during loading that involves across data sources. The model calibration step can then focus on
extremely large strains or stresses (e.g., cone penetration or pile determining the input parameters that make the constitutive
installation) that can cause progressive particle damage (e.g., model approximate the target soil properties/responses as closely
crushing, splitting) with an associated evolution of the grain size as possible. The model calibration will therefore depend on the
distribution or progressive de-structuring of clay macro/micro data sources that were prioritized and should not be expected to
fabric combined with preferential orientation of platy particles resolve the data inconsistencies previously identified. The
(e.g., within shear zones and slickensides in clay). Despite these advantages of this separation of tasks are discussed below with
and other possible complications, critical state frameworks in consideration of the common types of available data and data
continuum constitutive models are extremely valuable for inconsistencies for sands and clays.
producing more realistic responses across a range of loading
conditions.
The approaches used to select, calibrate, and validate
constitutive models can differ significantly across the spectrum
of research and application practices because of differences in
available data and pragmatic tradeoffs. A traditional mechanics
of materials approach to calibration may emphasize laboratory
testing to determine fundamental soil parameters and stress-
strain responses, followed by calibration of the model via direct
simulation of the laboratory element tests. This approach
implicitly assumes the calibrated model will produce reasonable
stress-strain responses for the broader set of loading conditions Figure 8. Separating the soil characterization and constitutive model
produced within a geotechnical system/structure, which is calibration tests.
357
A key consideration for liquefiable soils is that engineering generation, strain accumulation, and strain-softening, but they
properties are predominantly estimated using in-situ test data often require a more qualitative comparison of these features
(e.g., SPT, CPT, Vs) and property correlations, rather than from across sets of tests, with recognition of how the target
laboratory tests given the strongly detrimental effects of sample strengths were selected from the broader site characterization
disturbance caused by commonly available sampling methods dataset.
(e.g., Yoshimi et al. 1994). Laboratory testing of field or The separation of soil characterization and model calibration
reconstituted samples may still be used with certain soils (e.g., steps can be important for efficiently identifying challenges to be
silty or clayey sands) and in certain situations (e.g., new overcome and decisions to be made. Conflating the two steps can
construction) where the effects of sample disturbance may be make it more complicated to resolve challenges when there are
manageable given the loading paths to be imposed. significant inconsistencies in a soil characterization, as is
Consider the task of selecting target properties for an alluvial common for challenging soil types where the in-situ test
sand stratum that has been characterized primarily using in-situ correlations are not well established and the effects of sampling
tests. The cyclic resistance ratio (CRR) may be estimated using disturbance are less well understood (e.g., silty and clayey sands,
SPT, CPT, Vs, or other in-situ testing data, combined with more calcareous sands, coal combustion residuals, pumice).
than one liquefaction triggering correlation for each type of data.
The result will be a range of CRR estimates, with the range 5.3 Key properties for calibration
sometimes great enough to be considered inconsistent. For
example, significant differences can develop from systematic Constitutive models for soils that are susceptible to liquefaction
errors/biases in the in-situ test measurements, from differences or cyclic softening are more versatile if they can approximate the
in how users select representative values from spatially variable stress-strain properties and responses that tend to be most
in-situ test measurement data, or from systematic biases between important to the performance of geotechnical structures. For
correlations for the different in-situ tests. The user must many systems, the key properties fall into four categories.
recognize the limitations in each estimate of CRR, identify and Stiffness and damping: Dynamic stiffness and damping are
resolve inconsistences to the extent possible, and make decisions important for dynamic response and thus directly affect the cyclic
on the range of values to be considered in the numerical analyses. loads/demands imposed on a system. These properties are
The same discussion applies to estimating values or relationships typically described by combining values for Gmax (e.g., from Vs
measurements) with relationships for the secant shear moduli
for maximum (small-strain) shear modulus (Gmax); secant shear
ratio (G/Gmax) and equivalent damping ratio versus cyclic strain
modulus ratio (G/Gmax) and equivalent damping ratio versus
amplitude.
cyclic strain amplitude; CRR versus number of uniform loading Cyclic strength: The CRR for liquefiable soils and cyclic
cycles; cyclic mobility characteristics (rate of strain strength ratio ( cyc/su) for clays are defined by failure criteria
accumulation); and residual shear strength. After these (e.g., strain or ru based) which generally correspond to the onset
characterization steps are completed, the user can calibrate the of more rapid strain accumulation. Cyclic strengths depend on
constitutive model to best approximate the target set of the failure criterion, number and amplitude of undrained loading
properties. Thus, the calibration process will depend on what data cycles, overburden stress, initial static shear stress, stress history,
and correlations the user prioritizes and will not resolve data or and other factors.
correlation inconsistencies. Cyclic mobility: Cyclic mobility refers to how rapidly strains
Another example is selecting cyclic strengths for a accumulate during cyclic loading after liquefaction/cyclic
sedimentary stratum of silty clay that has been characterized by softening has occurred (e.g., shear strains exceeding 3%). This
a combination of laboratory testing of field samples and in-situ aspect of behavior is often important for estimating deformation
tests. The undrained shear strength (su) may be estimated from in systems that remain stable after earthquake loading.
CPT soundings, field vane shear tests, empirical correlations Post-earthquake shear resistance: Estimating the post-
(given an estimated stress history), and laboratory tests on field earthquake shear resistance or residual shear strength (Sr) of a
samples (e.g., simple shear or triaxial). Cyclic strengths may be liquefied sand-like soil is complicated by the effects of sample
estimated from the su values (using empirical correlations) or disturbance for laboratory tests, limitations in experimental
cyclic laboratory tests. The user must recognize systematic equipment, potential void redistribution in the field (a system
differences between strengths obtained using different test types response problem), and insufficient number and quality of
(e.g., simple shear versus triaxial), limitations of each set of documented case histories. Estimating the post-earthquake shear
strength estimates, resolve inconsistences to the extent possible, resistance of a sensitive clay or plastic silt is complicated by
and select the values to be used in analyses. For example, a estimating the rate of post-peak strain-softening, the potential for
common discrepancy is that laboratory test measurements shear localization, the thickness of shear localization intervals,
produce monotonic stress-strain curves that are far softer than and the strains within shear localizations.
implied by the combination of in-situ Vs measurements and Other stress-strain responses, such as post-liquefaction
empirical G/Gmax relationships (e.g., Boulanger & Wijewickreme reconsolidation strains, shearing under partially drained loading
2019). In addition, strength values from standard laboratory tests conditions, or volumetric strains under drained cyclic loading can
be equally important for some geotechnical infrastructure.
at common reference strain rates may be adjusted (increased) for
Permeabilities, and particularly permeability contrasts (i.e.,
strain rates effects in earthquake loading (e.g., 10-30% increase
between strata/units), can also have a strong effect on the extent
depending on soil characteristics) and, if the dynamic analyses
and consequences of liquefaction triggering as discussed later.
will not include 3D loading, adjusted (decreased) for multi- However, a constitutive model's general utility is almost
directional loading effects (e.g., 5-10% decrease depending on certainly limited by strong biases or limitations in how it
soil characteristics). After these characterization steps are approximates the key properties listed above.
completed, the user can calibrate the constitutive model to best Evaluation of any constitutive model calibration should
approximate a target set of properties/responses. Consequently, include suites of single element simulations covering the range
the calibration process is seldom well served by attempting to of loading paths important to system level responses. These
match the calibration to any single laboratory test result because suites of simulations, given any baseline model calibration, can
no single test is likely to accurately represent the target be compared to the trends predicted by empirical correlations to
properties. Calibration comparisons to laboratory test results are identify areas of significant differences or inconsistencies. A first
still essential, particularly for features of pore pressure level evaluation is whether there are strong differences between
358
simulated and expected stress-strain behaviors that could lead to
strongly biased NDA results, such that the model might be
considered invalidated for such loading conditions. A second
level of evaluation is quantifying and documenting lesser biases
for different types of loading conditions. The documentation of
representative stress-strain behaviors or biases relative to
expected behaviors or correlations across a broad range of
loading conditions is valuable for guiding users in selecting and
calibrating constitutive models for specific applications and
guiding developers in improving constitutive models.
5.4 Examples of constitutive modeling challenges
5.4.1 Stress overshooting
Duque et al. (2021) illustrated the artifact of stress overshooting
responses for four different constitutive models in simulations of
monotonic and cyclic undrained triaxial tests on Toyoura sand
shown in Figures 9a and 9b from Ishihara (1996). Simulations
with SANISAND-MSf produced modest stress overshooting
during cyclic loading (dashed lines in Figures 9c and 9d),
whereas the same model combined with the overshooting
correction of Dafalias & Taiebat (2016) eliminated the stress
overshooting (Figures 9e and 9f). Simulations with the hypo-
plasticity (HP) model of Wolffersdorff (1996) with the inter-
granular strain (IS) extension by Niemunis & Herle (1997)
produced excessive overshooting (HP+IS in Figures 9g and 9h),
whereas the model with the inter-granular strain anisotropy (ISA)
extension by Fuentes et al. (2020) produced modest
undershooting (HP+ISA in Figures 9i and 9j). The results for HP-
IS illustrate why users must not assume that constitutive models
will perform reasonably for loading paths that are more complex
than commonly used for calibration, and should instead include
more complex paths in evaluating model responses, as well as
diligently screen NDA results for any signs of unusual responses.
The results for the HP-IS model are serious enough to invalidate
its general use in NDA applications, and thus its documentation
is both necessary and appropriate. Results for the other models
shown in Figure 9 demonstrate progressive improvements,
illustrating the value of continually examining a model's
performance under more complex loading paths.
5.4.2 Cyclic strength versus number of cycles
The relationship between undrained cyclic strength and number
of uniform loading cycles (N) can often be approximated with a
power law, with the negative slope described by an exponent b.
For clean sands, the exponent b tends to increase with increasing
relative density (DR), with typical values of 0.08-0.15 for loose
sands, 0.15-0.3 for medium sands, and 0.3-0.5 for dense or very
dense sands (Boulanger & Idriss 2015). For clays and plastic
silts, the exponent b is typically 0.1-0.14 for a range of fines
contents and plasticity indices. For many constitutive models, the
slope of this relationship may be largely controlled by the
functional forms of the constitutive equations (plastic modulus, Figure 9. Evaluation of stress overshooting comparing (a,b) monotonic
dilatancy, etc.) such that only minor adjustments may be possible and cyclic triaxial tests on Toyoura sand from Ishihara (1996) with
by calibration. simulations using (c,d) SANISAND MSf, (e,f) SANISAND-MSf with
For sands, simulations of cyclic strength curves have shown overshooting correction by Dafalias & Taiebat (2016), (g,h) HP+IS, and
(i,j) HP+ISA (rearranged from Duque et al. 2021).
mixed degrees of agreement with experimental data. The
Dafalias & Manzari (2004) model tends to produce CRR versus
model (Elgamal et al. 2003) tends to produce slightly steep cyclic
N curves that are considerably steeper (b = 0.51 to 0.99) than the
strength curves (b values on the order of 0.5 to 0.7 depending on
range observed experimentally, as illustrated by the simulations
DR), whereas the modified PDMY03 model (Khosravifar et al.
from Duque et al. (2021) shown in Figure 10a. These steep CRR
2018) produces reasonable slopes with b values of 0.33-0.34 for
versus N curves for the Dafalias & Manzari (2004) model are
the simulations shown in Figure 10c.
consistent with the b values of 0.55 to 0.61 obtained with an
Reasonable slopes for cyclic strength curves are an essential
implementation of this model by Ziotopoulou & Boulanger
aspect of constitutive modeling for NDAs which need to address
(2012). Steep CRR versus N curves consistent with b values of
a range of seismic hazards with short to long-duration loadings.
0.8 to 1.1 were also obtained by Karamitros (2010) using the
Furthermore, reasonable slopes are essential for approximating
related NTUA-Sand model. In contrast, the related SANISAND-
pore pressure and strain accumulation during the irregular
MSf and PM4Sand models produce reasonable slopes with
loading produced by an earthquake. In this regard, the equivalent
b values of 0.21 to 0.29 and 0.25 to 0.30 for the simulations
number of uniform loading cycles that might be used to represent
shown in Figures 10b and 10d, respectively. The PDMY02
359
0.5 0.5
Dafalias & Manzari (2004) model: SANISAND MSf model:
Triaxial, p' o = 100 kPa, SA = 2.5% Triaxial, p' o = 100 kPa, SA = 2.5%
0.4 (Duque et al. 2021) 0.4 (Duque et al. 2021)
0.3 0.3
DR = 81%
DR = 81%
0.2 63% 0.2 63%
26% b = 0.29
26%
b = 0.99 b = 0.22
0.1 0.1
b = 0.73
b = 0.51 b = 0.21
(a) (b)
0 0
1 10 100 1 10 100
Number of uniform cycles Number of uniform cycles
0.5 0.5
PDMY03 model: PM4Sand model (ver 3.1):
(N1)60 = 25
DSS, 'vo = 100 kPa, SA = 3% (N1)60 = 25 DSS, 'vo = 100 kPa, SA = 2.5%
0.4 (Khosravifar et al. 2018) 0.4 (Boulanger & Ziotopoulou 2017)
15
0.3 0.3
b = 0.30
b = 0.34 15
0.2 5 0.2
b = 0.34 5 b = 0.29
0.1 0.1
b = 0.33 b = 0.25
(c) (d)
0 0
1 10 100 1 10 100
Number of uniform cycles Number of uniform cycles
Figure 10. Cyclic strength versus number of uniform loading cycles for different relative densities: (a)
DM04 model, (b) SANISAND-MSf model, (c) PDMY03 model, and PM4Sand model.
an irregular CSR time series is fundamentally related to the slope shear stress reversals (i.e., producing a temporary isotropic state
of the cyclic strength curve for a specific sand. If a constitutive of stress). Cyclic strength for a sand that is significantly dense of
model has an overly steep cyclic strength curve and is calibrated critical state increases with increasing because the sand is more
to a cyclic strength based on approximately 15 uniform loading resistant to strain accumulation, and the cyclic strength for these
cycles (generally associated with an M=7.5 earthquake and a loading conditions is commonly defined using a strain-based
common reference point in SPT/CPT based triggering failure criterion (e.g., 3% peak shear strain is used for Figure 11).
correlations), then the model will significantly underestimate the For sands, simulating the effects of on cyclic strength has
soil's resistance to the long-duration loading associated with proven challenging for many models and been the incentive for
subduction zone events. Recalibration of a constitutive model for model improvements in certain cases (e.g., Ziotopoulou et al.
different numbers of equivalent uniform cycles is undesirable 2014, 2016). Simulations using UBCSAND 904a produced K
and thus an overly steep cyclic strength curve can be a barrier to responses (Figure 11c) that are inconsistent with empirical trends
a constitutive model's adoption in practice. (Figures 11a and 11b), which motivated the development of
For clays, simulations of cyclic strength versus number of UBCSAND 904aR with its more reasonable responses (Figure
loading cycles are less variable, which is consistent with cyclic 11d). Simulations using PM4Sand version 2 produced relatively
degradation effects being less significant for clays than for sands flat K responses (Figure 11e), which informed development of
(i.e., b values are smaller for clays). Simulations by Kimoto et al. PM4Sand version 3 with its improved responses (Figure 11f).
(2014) using their model for clays produced cyclic strength Simulations using PDMY02 produced relatively poor K
curves consistent with b values of 0.13-0.17. Simulations with responses (Figure 11g) which informed the modifications in
PM4Silt by Boulanger & Ziotopoulou (2018a) produced b values PDMY03 leading to its improved responses (Figure 11h). The
of about 0.15. These slopes are reasonably consistent with poor responses obtained with the earlier versions of these three
experimental data for clays and plastic silts and are a significant constitutive models were immediately evident upon their
improvement over simpler total stress models that can produce comparisons to empirical relationships, did not require soil-
essentially flat relationships (i.e., b = 0.0). specific laboratory testing to identify, and were not stress-strain
behaviors that could be fixed by calibration (i.e., the limitations
5.4.3 Initial static shear stress were inherent to the model formulations). These three examples
For sands, the effect of an initial (sustained) static shear stress illustrate how comparing simulations to empirical relationships
ratio (e.g., = hv/ 'vc for simple shear conditions) on cyclic can help identify modeling limitations and promote model
strength is commonly represented as a K relationship, where K improvements.
= CRR /CRR =0. Compilations of cyclic test data (Figures 11a For normally consolidated or lightly overconsolidated clays,
and 11b) show that the effects of depend on the state of the the presence of an initial static shear stress will generally increase
sand, as represented by DR for 'vc < 300 kPa in Figure 11a and su (a result of anisotropic consolidation) and decrease the cyclic
by relative state parameter index ( R) in Figure 11b. The presence strength ratio. Simulations with PM4Silt (Boulanger &
of reduces cyclic strengths for sands that are loose of critical Ziotopoulou 2018a) and the multi-surface, kinematic hardening
state or only slightly dense of critical state, whereas it increases model of Stoecklin et al. (2020) produced responses that are
cyclic strengths for sands that are significantly dense of critical generally consistent with these empirical trends. Other
state. Cyclic loading with a static shear bias causes progressive constitutive models exhibit responses that suggest they will
accumulation of shear strains in the direction of the static bias, reasonably approximate these trends, although this needs to be
even though a maximum ru of 100% may not develop without systematically explored and documented.
360
K
K
Figure 12. Post-triggering strain accumulation under small levels of
cyclic loading in undrained direct simple shear loading with a small static
shear bias: (a) UBCSAND 904aR, and (b) PM4Sand version 3
(simulations from Boulanger & Ziotopoulou 2018).
K
K
UBCSAND at low levels of cyclic loading, which helped explain
why the use of these two models in a system level simulation had
produced dramatically different deformations during a long-
duration subduction zone earthquake motion. The problematic
responses obtained with UBCSAND manifested most strongly
for conditions of small initial static shear stress ratios at higher
overburden stresses, and thus had not been observed in other
studies involving steeper slopes, smaller overburden stresses, or
shorter duration loadings. Tsuboi et al. (2015) evaluated a
different elastoplastic constitutive model and showed that it also
overestimated deformations in liquefiable soils during low-level
K
K
long-duration cyclic loading when calibrated by conventional
procedures. Laboratory test data for irregular, long-duration
loading conditions are relatively limited, but the existing data are
sufficient for providing some bounds on reasonable behaviors
(e.g., Tsuboi et al. 2015, Ziotopoulou & Boulanger 2016).
Additional laboratory tests approximating a wider range of
Figure 11. Effect of static shear stress ratio on cyclic resistance ratio of irregular, long-duration loading conditions are needed for
sand to 3% peak shear strain in undrained cyclic direct simple shear evaluating constitutive models.
loading: (a, b) empirical relationships, (c, d) simulations using
UBCSAND 904a and 904aR, respectively, (e, f) simulations using 5.4.5 Partially drained loading
PM4Sand versions 2 and 3, respectively, and (g, h) simulations using Field loading involves partially drained conditions as excess pore
PDYM02 and PDYM03, respectively. pressures diffuse, whether during or after earthquake shaking.
Potential consequences include local densification (net outflow)
5.4.4 Long-duration irregular loading and local loosening (net inflow) in deposits with lower-
The accumulation of shear strain during long-duration permeability interlayers. This process is illustrated schematically
earthquake shaking, including the low-level coda motions after for an infinite slope in Figure 13, wherein upward seepage driven
the strongest shaking, can vary significantly across constitutive by excess pore pressure gradients causes the sand just below the
models for sands. Boulanger & Ziotopoulou (2018b) compared low permeability lens to loosen sufficiently to render the slope
the responses of UBCSAND 904aR and PM4Sand (version 3) unstable. The response of soils to partially drained conditions
under stepped cyclic loading with a small initial to illustrate with imposed volumetric strain paths (net outflow/inflow) during
this issue. The simple shear simulations shown in Figure 12 are monotonic or cyclic loading have been studied experimentally by
for the two constitutive models calibrated to the same target Boulanger & Truman (1996), Vaid & Eliadorani (1998, 2000),
properties (e.g., Gmax, CRR15cycles, and DR). A uniform cyclic Sento et al. (2004), and Yoshimine et al. (2006). Kamai &
stress ratio (CSR) of 0.12 was imposed until a peak shear strain Boulanger (2012) used PM4Sand with FLAC to simulate the
of 3% was triggered, after which 5 uniform cycles at a reduced effects of partial drainage on monotonic, cyclic, and post-cyclic
CSR = 0.015 were applied. The low-level loading generated an responses of liquefied sand. The PM4Sand simulations of the
additional 16% shear strain (3.2% per cycle) with UBCSAND volumetric strain-controlled, constant shear stress experiments
versus an additional 2.5% shear strain (0.5% per cycle) with by Sento et al. (2004) shown in Figure 14 illustrate how the net
PM4Sand. These and other simulations identified a limitation in inflow of water to specimens of different initial DR were
361
plasticity-based constitutive models to predict reconsolidation
volumetric strains from a condition of ru=100% that are only a
fraction of one percent (Ziotopoulou & Boulanger 2013b, Howell
et al. 2014), whereas experimental data show values ranging
from one to four percent depending on DR. Boulanger &
Ziotopoulou (2017) included in PM4Sand an option for
improved simulation of reconsolidation strains during the post-
strong shaking portion of a simulation. This option involves
reducing the elastic moduli to increase reconsolidation strains,
compensating for the sedimentation strains that are not directly
modeled. Basu et al. (2022) analyzed three centrifuge tests and
two shaking table tests using PM4Sand with its post-shaking
option and concluded that reconsolidation settlements were
under-predicted by an average of about 15% (predicted values
were 60-125% of measured values) and that a function increasing
the vertical permeability with increasing ru improved simulation
of dissipation rates. Several other studies have identified
Figure 13. Schematic stress paths of elements near the top (point A) and limitations in simulating reconsolidation settlements, which
bottom (point B) of a liquefiable layer overlain and underlain by low suggests there is a need to systematically evaluate constitutive
permeability soil, in which diffusion of earthquake-induced excess pore models for their ability/inability to simulate reconsolidation
pressures causes sufficient loosening at point A to trigger slope instability behavior as observed in laboratory tests.
(after Kulasingam et al. 2004).
5.4.7 Other constitutive challenges
An overall challenge for constitutive models is broadening the
parameter and loading spaces over which models perform
reasonably well, while clearly defining those spaces where they
perform less satisfactorily or their performance is unknown.
Validation spaces for some constitutive models may be
broadened (or better defined) through a combination of more
detailed evaluations and possible model modifications. Such
efforts can anticipate that constitutive models are likely to have
significant challenges approximating stress-strain responses for
various loading conditions that differ significantly from those
emphasized during model development. Examples include (1)
irregular cyclic undrained loading with static bias, (2) large
ranges of overburden stress where, for example, it is desirable for
sand models to produce CRR values consistent with empirical K
relationships, (3) rotational loading paths such as those produced
by waves, (4) multiple loading events with intermittent
consolidation, (5) volumetric strains during drained cyclic
loading from earthquakes, and (6) volumetric strains during
Figure 14. Results of volumetric strain-controlled, constant shear stress drained or partially drained loading with large numbers of
simulations using PM4Sand (solid lines) compared to experimental data smaller amplitude cyclic loads, such as produced by wind, wave,
(dashed lines) by Sento et al. (2004) (Kamai & Boulanger 2012) or traffic loading. Constitutive models also have limitations and
uncertainties (due to lack of data) in modeling soils that differ
significantly from the clean sands and sedimentary/reconstituted
accompanied by large increases in void ratio and shear strain.
clays that have been studied most extensively, including (1)
Lade & Liggio (2014) also performed experiments and
intermediate soils like silty and clayey sands, (2) well-graded
simulations (using a single hardening constitutive model) for
mixtures like silty gravels with sand, (3) crushable soils like
conditions with imposed volume strains during monotonic
calcareous sands, coal combustion residuals, and pumices, (4)
shearing.
lightly cemented soils, and (5) aged soils. Some of the above
Partially drained loading paths can be a good test of
challenges and limitations may be overcome by modifications to
constitutive model performance and warrant broader inclusion in
existing constitutive models, whereas as others may require new
validation studies for constitutive models. Modeling partially
approaches and frameworks to overcome.
drained conditions in NDAs is another area in need of study, as
discussed in the next section, because it has major implications
for how the residual shear strength of a liquefied soil can be 6 NDA PROCEDURES
affected by system-level diffusion processes (i.e., it is not a
unique function of the pre-earthquake void ratio). 6.1 Components of an NDA procedure
5.4.6 Reconsolidation strains An NDA procedure comprises several components that
Volumetric strains that develop during post-liquefaction collectively affect the quality and repeatability of analysis results
reconsolidation of sand are difficult to model using stress ratio- (Figure 15). These components include the constitutive models
based constitutive models and the conventional separation of and numerical methods, which are the hardwired tools (with
strains into elastic and plastic components. The problem is that a required verification) that are used to perform the analyses. They
large portion of post-liquefaction reconsolidation strains are due also include the protocols used for calibrating the constitutive
to sedimentation effects (i.e., volume reductions while the models and numerically modeling the system (e.g., choice of
effective stresses remain close to zero) which are not easily boundary conditions, discretization, initialization of seepage and
modeled using most sand constitutive models as calibrated for stress conditions, specification of stress-dependent soil
earthquake loading. For example, it is common for many properties, method of input motion application, numerical
362
measurements using numerical and experimental response
surfaces (Figure 16), as initially explored in Kutter et al. (2020a)
for the international Liquefaction Experiment and Analysis
Projects (LEAP) efforts and further developed by Carey & Kutter
(2022). Comparisons of numerical and experimental response
surfaces, as illustrated by the displacement-DR-PGAeff surfaces
in Figure 16 for the uniform sand slope used in the LEAP studies,
provide a more thorough evaluation of an NDA procedure's
ability to predict responses and their sensitivity over a broader
range of conditions.
Validation studies for NDAs have been the subject of
considerable efforts this past decade, with efforts ranging from
project-specific studies to coordinated group efforts. Additional
coordinated efforts like those for LEAP but expanded to include
Figure 15. Validation applies to a reproducible set of NDA modeling
components, including the constitutive models, numerical methods,
more complex, holistic geotechnical infrastructure are needed for
calibration protocols, numerical protocols, and the user and reviewers. further advances (NASEM 2016, Bray et al. 2017). The
following synopses of more than fifty select studies, while not
exhaustive, provide a sense of efforts this past decade.
damping, step sizes, tolerances), which are the ways that the tools Validation efforts focused on one-dimensional (1D) site
are used. Lastly, they include the expertise of the user (analyst) response have been performed using NDA software developed
and reviewers, which dictates the artistry of the work, the solely for 1D loading as well using NDA software capable of
thoroughness of the quality control, and the judgment exercised two-dimensional (2D) or three-dimensional (3D) analyses.
in interpreting the results. Shahir et al. (2012) used the Dafalias & Manzari (2004) model
Validation of an NDA procedure is only meaningful if others to perform 1D FEM simulations, with two relationships for
can apply the procedure in ways consistent with those used in the describing permeability variations during the process of
validation studies. The reproducibility of NDA procedures liquefaction, of the centrifuge model tests of level uniform
depends on documentation of the calibration and numerical liquefiable soil performed by Gonzalez et al. (2002).
modeling protocols, which is sometimes referred to as an NDA Kramer et al. (2015) evaluated four nonlinear effective-stress
modeling procedures document. Similarly, a thorough review based site response analysis methods (PM4Sand with FLAC;
process depends on transparent and thorough documentation of cocktail glass model with FLIP; PDMY02 with OpenSees; and
the analysis results. Boulanger & Beaty (2016) addressed this 1D PSNL) against the Wildlife array case history, concluding
latter need by presenting a checklist of questions (discussed later that all four produced consistent responses up to the point of
in Section 7) that a reviewer of an NDA study should consider liquefaction triggering but that the responses after triggering
asking as part of any internal or external review process. were highly variable.
The inputs to an NDA (Figure 1), which include the site Montgomery et al. (2017) used PM4Sand with FLAC to
characterization (e.g., stratigraphy, properties) and ground perform 2D site response analyses of the Wildlife array to
motions, can strongly affect the results obtained in applications illustrate how spatial variability in the liquefiable strata may
or validation studies. Validation studies using centrifuge model have affected site response, liquefaction triggering, and
tests have the advantage of largely removing uncertainties diffusion of excess pore pressures from looser to denser zones.
associated with these analysis inputs, whereas validation studies Pender et al. (2016) used the PDMY models by Elgamal et al.
using case histories need to consider these uncertainties in (2003, 2006) in Cyclic1D to evaluate the role of pore pressure
interpreting the results. Recent validation efforts are discussed in dissipation in gravelly soils for explaining certain field
the following section. observations.
Ramirez et al. (2018) evaluated the ability of three numerical
6.2 Validation efforts: case histories and physical models modeling teams using PDYM02 in OpenSees and
SANISAND in both OpenSees and FLAC for simulating 1D
Validation of an NDA procedure necessarily relies on a site response, with validation against a series of centrifuge
combination of case histories and physical models. Field models of layered liquefiable profiles.
observations are vital, but insufficient on their own for Zhang & Muraleetharan (2018) used the coupled elastoplastic
supporting ongoing advances because (i) design-level events are CM4USS model with pore water and pore air pressure as nodal
rare and unpredictable, (ii) geotechnical infrastructure systems variables in 1D finite element simulations of the centrifuge
are incredibly diverse, (iii) ground motions have complex model tests of unsaturated sand (38%, 66%, and 68% degrees
characteristics, and (iv) maintaining field instrumentation for of saturation) performed by Mirshekari & Ghayoomi (2017).
decades is resource intensive. Component-level tests are vital, Cubrinovski et al. (2018) used the elastoplastic, state-concept
but also insufficient on their own because they do not capture the based, stress-density model (Cubrinovski et al. 1998) in 1D
coupling of different physical processes. Physical models at 1g finite element simulations of the response of representative
can be valuable, but generally are too small to realistically model profiles in the Canterbury Earthquake Sequence to illustrate
most geotechnical infrastructure. Accordingly, hypergravity key mechanisms of system response.
scaled modeling using geotechnical centrifuges are invaluable Reyes et al. (2019) used the Dafalias & Manzari (2004) model
for advancing fundamental understanding and validating NDA in FLAC3D to explore bidirectional shaking effects after
methods for geotechnical infrastructure. validation against four centrifuge model tests with
Current NDA procedures for geotechnical infrastructure bidirectional base excitation performed by Su & Li (2008).
affected by soils susceptible to liquefaction or cyclic softening Tsaparli et al. (2020) used a bounding surface plasticity model
have limited ability to represent generalized conditions and thus with the Imperial College finite element program (ICFEP) to
cannot be reliably used without validation for geotechnical perform 1D site response analyses of a strong motion station
infrastructure having similar soil, structure, and earthquake recording from the 2011 Christchurch Earthquake.
loading conditions. These validation studies benefit strongly Hutabarat & Bray (2021) using PM4Sand in FLAC to perform
from use of multiple datasets covering different geometries, site response analyses of two liquefaction case history sites
properties, and loadings. For datasets from centrifuge model test from Christchurch.
series, simulation results can be compared to the physical
363
Figure 16. Numerical response surface from NDAs using PDMY02 in OpenSees (upper left) and
experimental response surface from LEAP centrifuge model tests (upper right) for the submerged uniform
sand slope used in the LEAP studies (bottom) (surfaces courtesy T. J. Carey & B. L. Kutter; centrifuge
model sections from Kutter et al. 2019).
Chen & Arduino (2021) used PM4Sand in OpenSees to attributable to variability in both the centrifuge experiments (e.g.,
simulate the site responses and liquefaction observed at the initial DR, uniformity) and NDA procedures, including
Wildlife and Port Island arrays. differences in the selected soil properties, constitutive models,
Several other studies have used constitutive models that do not calibration protocols, user mistakes/misunderstandings, and
include the ability to simulate cyclic mobility and thus have other factors. Lessons and findings from these validation projects
focused on evaluating responses up to triggering only (e.g., are discussed in the above referenced compilations.
Chiaradonna et al. 2020, Olson et al. 2020); note that site For LEAP 2017, Kutter et al. (2020a) organized eight
response models that lack the dilatancy components necessary to simulation teams in a sensitivity study comparing numerical and
simulate cyclic mobility cannot produce realistic ground motions experimental response surfaces for the mid-slope surface
after triggering. displacement (see Figure 16 for model configuration). Five
The LEAP-2015, 2017, and 2019 projects engaged a large sensitivity cases (NS-1 to NS-5) were evaluated with the initial
number of researchers from academia, industry, and agencies DR and base motion PGAeff values listed in Table 1. The expected
across several countries in an effort to improve repeatability of experimental displacement for each case was estimated using
centrifuge model tests and validate soil liquefaction analysis two experimental response surfaces fit to the body of
methods. The contributions from these collaborative efforts are experimental results; the two surfaces gave similar values and
compiled in special issues (Manzari et al. 2018, Zeghal et al. thus the average was taken as the expected value. The mid-slope
2021) and proceedings (Kutter et al. 2020b). The centrifuge displacements from the NDAs ( NDA) and experimental response
model testing focused on a submerged, gentle slope of uniform
surface ( Experiment) are listed in Table 1. Differences between
sand in a rigid container (Figure 16). NDAs were performed by
experimental and NDA values (i.e., residuals) were evaluated in
several teams using constitutive models and numerical platforms
natural logarithm space so that equal weighting is given to results
that included the CycLiq model in OpenSees, Dafalias &
across the full range of observed displacements. The mean
Manzari model in OpenSees, PDMY model in OpenSees,
differences and root mean square error (RMSE) for the results by
PM4Sand model in OpenSees and FLAC, UBCSAND model in
each simulation team are listed in the table. The mean difference
FLAC, SANISAND model in FLAC3D, cocktail glass model in
is a measure of bias, whereas the RMSE reflects total error (bias
FLIP Rose and FLIP TULIP, Bowl model in HiPER, hypoplastic
and dispersion combined). The first three simulations teams
model in PLAXIS, and ISA-hypoplasticity model in ABAQUS.
listed in the table used PM4Sand with either FLAC or OpenSees.
Stress-strain and stress-path responses from calibrations for
The first team listed (Montgomery & Ziotopoulou 2019)
several of these constitutive models, as shown previously in
produced the smallest bias (mean difference closest to zero) and
Figure 5, illustrated limitations for some of the models. Zeghal
second smallest RMSE of the eight simulation teams. Their mean
et al. (2018) compared stress-strain curves and stress-paths
difference of -0.23 indicates a bias toward overestimating slope
obtained by inverse analyses of several LEAP centrifuge model
displacements by 26% (i.e., e0.23 = 1.26). The next two teams
tests to those obtained by numerical simulation models, with
listed (Chen et al. 2019, Tsiaousi et al. 2019) produced
results also illustrating that some aspects of behavior were
significantly greater differences, including stronger biases
consistent and others were conflicting. The NDA results by
toward overestimating displacements (i.e., factors of e0.45 = 1.57
different teams sometimes showed large differences in key
and e1.22 = 3.39), despite having used the same constitutive
response measures (e.g., slope displacement), which were partly
364
Table 1. Mid-slope displacements from NDA simulations and experimental response surface for five sensitivity study cases from the LEAP 2017
program (after Kutter et al. 2020a)
Numerical sensitivity study cases
NS-1 NS-2 NS-3 NS-4 NS-5
Base motion PGAeff (g) 0.146 0.146 0.146 0.243 0.107
Sand DR (%) 65 50 75 65 65
Surface displacement at mid-slope (m): Mean ln Experiment
RMS ln Experiment
Experimental and NDA results NDA NDA
Experimental response surface
0.049 0.127 0.024 0.300 0.010 -- --
(Kutter et al. 2020)
PM4Sand in FLAC
0.042 0.610 0.010 0.540 0.010 -0.23 0.84
(Montgomery & Ziotopoulou 2019)
PM4Sand in OpenSees
0.033 0.889 0.013 1.004 0.011 -0.45 1.07
(Chen et al. 2019)
PM4Sand in FLAC
0.070 4.800 0.010 2.890 0.020 -1.22 1.98
(Tsiaousi et al. 2019)
PDMY02 in OpenSees
0.110 0.230 0.070 0.240 0.030 -0.68 0.83
(Qui & Elgamal 2019)
DM04 in OpenSees
0.231 0.494 0.135 0.374 0.021 -1.13 1.26
(Chen et al. 2019)
Cocktail glass in FLIP Rose
0.160 0.160 0.161 0.199 0.140 -1.12 1.56
(Ozutsumi 2019)
SANISAND in FLAC-3D
0.310 0.969 0.087 0.453 0.036 -1.38 1.49
(Yang et al. 2019)
UBCSAND in FLAC
0.490 6.360 0.020 1.390 0.020 -1.66 2.17
(Tsiaousi et al. 2019)
model. The difference in results by these three teams illustrates Oka et al. (2012) used LIQCA2D11 with an elastoplastic
how validation of an NDA procedure pertains to more than just model for sands (Oka et al. 1999) and an elasto-viscoplastic
the tools (Figure 15). The fourth team listed (Qui & Elgamal model for clays (Kimoto et al 2014) to simulate damage
2019) used PDMY02 in OpenSees and had the third smallest bias patterns for river embankments from the 2011 Tohoku
and smallest RSME. The remaining four teams produced earthquake.
relatively large negative mean differences (biased to Chakrabortty & Popescu (2012) used DYNAFLOW to
overestimation) and large RMSEs. The sixth team listed analyze centrifuge models of liquefiable sand deposits of
(Ozutsumi 2019) predicted displacements of 0.140-0.199 m uniform and heterogeneous density.
across all five events, which did not track the experimental trend Chaloulos et al. (2013) used NTUA-Sand in FLAC3D to
of displacements varying from 0.01-0.300 m depending on DR simulate the centrifuge model tests of piles within laterally
and PGAeff. Comparing numerical and experimental responses spreading ground performed by Gonzalez et al. (2009).
across a range of conditions, such as illustrated by the study Armstrong et al. (2013) used UBCSAND and Armstrong &
summarized in Table 1 (after Kutter et al. 2020a) or the response Boulanger (2015) used PM4Sand, both with FLAC, to
simulate centrifuge model tests of pile groups in bridge
surfaces shown in Figure 16 (Kutter & Carey 2022), provides an
approach embankments over liquefiable soils.
improved approach for systematically evaluating NDA
Galavi et al. (2013) used UBC3D-PLM in PLAXIS 2D to
procedures, as it recognizes that good agreement for any one
model a quay wall response to the 1995 Kobe earthquake.
experiment does not necessarily ensure good agreement for more Chian et al. (2014) used the Wang et al. (1990) model in FLAC
generalized conditions. to simulate centrifuge models of liquefaction-induced flotation
Other validation efforts for 2D and 3D NDAs have been of underground structures.
performed for a wide range of geotechnical infrastructure, Howell et al. (2014) used PDMY02 in OpenSees to simulate
loading conditions, and numerical simulation methods. centrifuge tests of laterally spreading slopes with and without
Shahir & Pak (2010) used the Dafalias & Manzari model in prefabricated vertical drains. Simulations were performed
OpenSees to perform 3D analyses of the centrifuge model tests with a 2D model of the full centrifuge model geometry and
of shallow foundations over liquefiable soil profiles performed both 2D and 3D unit cell models of a single drain.
by Hausler (2002). Ferdosi et al. (2015) used UBCSAND 904aR in FLAC to
Cubrinovski (2011) used the state index-based, stress-density simulate a 1g shake table test of a level profile of saturated
(S-D) model (Cubrinovski & Ishihara 1998) to perform 2D tailings.
simulations of the recorded responses of levees at Nakashimo Marasini & Okamura (2015) used LIQCA-2D to simulate
during the 2003 Miyagi-Hokubu earthquake (Takahashi & dynamic responses of centrifuge models de-saturated by air
Sugita 2009). injection; good results were obtained using an appropriately
Cubrinovski (2011) compared results of 3D simulations using reduced bulk modulus for the pore fluid.
the S-D model in DIANA-J (results in Cubrinovski et al. 2008) Bhatnagar et al. (2016) used UBC3D-PLM in PLAXIS 3D to
and the elastoplastic model by Oka et al. (1999) in LIQCA simulate the centrifuge model tests of an embankment over
(results in Uzuoka et al. 2008) for shaking table tests of pile liquefiable sand performed by Adalier et al. (1998).
foundations behind a sheet pile wall embedded in liquefiable Wang et al. (2017) used the PDMY02 and PIMY soil models
sand. Both NDA procedures produced good agreement with and the pile-interface springs (Brandenberg et al. 2013) in
the pile foundation responses, but underestimated the sheet OpenSees to connect a soil column with a structural model to
pile wall's large displacements because geometric nonlinearity simulate the centrifuge model tests of pile foundations in
was not considered in the analyses. laterally spreading ground by Singh et al. (2000) and
Ming et al. (2011) used the critical state, bounding surface Brandenberg et al. (2003). These analyses largely replicated
model by Li (2002) in SUMDES2D to simulate the response those of Chang et al. (2013) but with an alternative approach
of the Upper San Fernando Dam.
365
for simulating the shear localization that developed between curating and archiving of data from laboratory tests, physical
the overlying clay crust and underlying liquefied sand. models, and case histories greatly benefits the community by
Karimi & Dashti (2016) used the PDMY02 model in expanding the breadth of datasets openly available for validating
OpenSees to perform 3D simulations of centrifuge model tests NDA procedures; enabling comparisons of simulations by
of shallow foundations over liquefiable deposits. Bullock et al. multiple teams or multiple NDA procedures for common
(2019) used subsequent parametric analyses with the validated datasets; increasing the impact of the experimental work, thereby
model to derive functional forms, after which they used a case providing increased justification for its cost and effort; and
history database to develop a regression model for predicting promoting collaborations.
shallow foundation settlements. Validation studies for 2D or 3D NDAs of systems affected
Vytiniotis & Whittle (2017) used the Dafalias & Manzari by soils susceptible to cyclic softening have been less numerous,
constitutive model in OpenSees to perform 2D simulations, with the following three examples utilizing case history data.
with line elements to represent prefabricated vertical drains, of Boulanger (2019) used PM4Silt to represent moderately
the centrifuge model test by Kamai et al. (2007) that comprised compacted clayey sands and clayey gravels in FLAC
treated and non-treated gentle slopes with a clay crust over simulations of the response of Austrian Dam in the 1989 Loma
liquefiable sand. Prieta earthquake.
He et al. (2017) used PDMY02 models in OpenSees to Abdellaziz et al. (2020) used total stress models with and
perform 3D simulations of a large shaking table test of piles in without strain-softening capabilities in FLAC to simulate the
laterally spreading ground. Saint-Adelphe landslide in sensitive plastic clay after the 1988
Li et al. (2018) used the PMY02 and Dafalias & Manzari Saguenay earthquake.
models in OpenSees for 2D simulations of centrifuge tests of Kiernan & Montgomery (2020) used PM4Silt in FLAC to
level and gentle sloping liquefiable sites treated with granular simulate the 4th Avenue landslide in Bootlegger Cove Clay
columns; for the treated side of the centrifuge models the during the 1964 Great Alaska Earthquake.
lateral displacements were overestimated using the Dafalias & Future validation efforts for systems affected by cyclic softening
Manzari model and underestimated using the PDMY02 model. are likely to make greater use of physical model tests, as has been
Su et al. (2019) used PDMY02 in OpenSees to simulate a the case for liquefiable soils. Examples of datasets available at
shake table test of a pile supported quay wall subject to lateral DesignSafe for reuse in validation studies include centrifuge
spreading ground. experiments of single piles and pile groups in soft clay profiles
Tasiopoulou et al. (2019) used the Ta-Ger sand model, with and without local soil-cement improvements (e.g., Taghavi
implemented for use with FLAC and FLAC3D, to perform 2D & Muraleetharan 2012) and centrifuge tests involving three
and 3D simulations of centrifuge model tests of shallow different structures in soil profiles with soft, low-plasticity fine-
foundations over liquefiable soils (tests by Dashti et al. 2010). grained soil layers (Buenker et al. 2021).
Tasiopoulou et al. (2020) proposed a procedure for estimating
post-triggering cyclic strain accumulation rates and used the 6.3 Spatial variability
procedure with PM4Sand in FLAC to simulate centrifuge
models tests of a submerged subway tube with liquefiable The results of an NDA study for a field problem involving soils
backfill (performed by Chuo et al. 2011) and a sheet-pile quay susceptible to liquefaction or cyclic softening, whether for
wall with liquefiable backfills (performed by Sato et al. 2001). application or validation, can depend strongly on site
Demir & Ozener (2019, 2020) used UBC3D-PLM in PLAXIS characterization details, including how the spatial variability is
and PDMY02 in OpenSees, respectively, to perform 2D and either directly modeled or indirectly accounted for. Geostatistical
3D analyses of centrifuge model tests (performed by approaches for modeling spatial variability in soil types or
Rayamajhi et al. 2015) of liquefiable soil treated with soil- properties are well developed for a range of geotechnical
cement columns. applications (Baecher & Christian 2003, Fenton & Griffiths
Esfeh & Kaynia (2020) used SANISAND in FLAC3D to 2008, Joint TC205/TC304 Working Group 2017). Recent
simulate centrifuge model tests of wind turbine monopole and examples of NDA studies examining spatial variability effects
caisson foundations in liquefiable soils. for systems affected by liquefaction or cyclic softening are
Kassas et al. (2020) used PM4Sand in FLAC to simulate the discussed in this section.
centrifuge model tests by Adamidis & Madabhushi (2018) of The majority of NDA studies use uniform analysis models
strip footings over liquefiable sands. They obtained good wherein individual soil units are assigned representative
agreement with pore pressures and settlements but some properties. A number of NDA studies using both uniform and
aspects of the dissipation phenomena were not accurately stochastic (random field) liquefiable layers have examined how
reproduced. representative properties should be selected. Popescu et al.
Luque & Bray (2020) used PM4Sand in FLAC to simulate the (1997, 2005) performed 2D and 3D NDAs of level soil profiles
response of two buildings affected by liquefaction in the with stochastic and uniform soil properties and concluded that
Canterbury earthquake sequence; they showed that the details the pore pressure generation and triggering of liquefaction in a
of the stratigraphy affected the directions of the liquefaction- portion of a stochastic profile was best approximated using
induced building tilts. uniform models assigned the 20th percentile property values.
Zalachoris et al. (2021) used the PM4Sand, UBCSAND, and Chakrabortty & Popescu (2012) performed NDAs of centrifuge
URS/Roth models in FLAC to simulate performance of quay tests which demonstrated that heterogeneous deposits generate
walls in the 2014 Cephalonia earthquakes. excess pore pressures more rapidly than uniform deposits.
Vargas et al. (2022) presented a validation exercise focused on Dakoulas et al. (2018) performed NDAs of quay walls with
quantifying median responses and associated variability. The spatially variable foundation and backfill properties and
validation exercise was performed using the strain space, concluded spatial variability had a 11-16% effect on wall
cocktail glass model (Iai et al. 1992, 2011) in FLIP ROSE for displacements, whereas the selection of time histories had a
simulating lateral spreading in clean sands based on a diverse 50% effect on displacements. Montgomery & Boulanger (2017)
set of laboratory element and LEAP centrifuge model tests. performed 2D NDAs of infinite slopes and showed that the
Many of these validation studies have re-used experimental data representative percentile (Prep) property for the liquefiable layer
that was archived for public access (e.g., DesignSafe in the US, to estimate the median value of lateral spreading displacement
www.designsafe-ci.org) or shared by experimentalists at the from stochastic models generally ranged from the 30th to 70th
same institution or across institutions (e.g., LEAP). Open-access percentile. The smaller Prep values corresponded to thicker crust
366
Figure 17. NDA results for a baseline stochastic realization at the end of shaking: (a) numerical model;
(b) ru within the saturated portion of Unit C; (c) shear strains within the saturated portion of Unit C; (d)
horizontal displacements; and (e) end-of-shaking smoothed horizontal strains ( xx) along the surface from
baseline case. Results from all additional realizations showing extensional and compressional patterns
are included in (e). (Pretell et al. 2020).
layers, thicker liquefiable layers, greater slope angles, and the 45-50% for the liquefiable alluvial layer provided reasonable
lower range of the imposed shaking intensities; these conditions estimates of the median crest settlement and median
appeared to enable shear deformations to more easily develop embankment slope displacements, (2) using Prep 30% provided
through interconnected networks of looser lenses. Boulanger & reasonably conservative estimates of these displacements for
Montgomery (2016) performed 2D NDAs of a 45 m high most cases (e.g., displacements exceeded in less than 16% of the
embankment on a 3 m or 12 m thick alluvial layer, and showed cases), and (3) the level of conservatism obtained using Prep
that the Prep properties for the alluvium to estimate the median 30% may be less for cases involving localized deformations such
values of dam crest settlement or embankment shell deformation as at embankment toes for cases with x/B > 0.4. In all cases,
from stochastic models generally also ranged from the 30th to 70th uniform analysis models cannot reproduce the complexity or
percentile. The smaller Prep values corresponded to the thicker variability in local and global deformation mechanisms and this
sand layer and the lower range of the imposed shaking limitation must be recognized when uniform models are used to
intensities; these analyses used a horizontal scale of fluctuation assess the consequences of liquefaction-induced deformations.
( x) of 20 m, which is a small fraction of the base width of the NDA studies of field cases with interbedded deposits of
embankment. Paull et al. (2020, 2021) performed 2D NDAs of distinctly different soil types have been examined using
5, 10, 25, and 45 m tall embankments on liquefiable foundation transition probability methods. Boulanger et al. (2019) and
layers with unconditional and conditional random fields. They Pretell et al. (2020) performed 2D NDAs of Çark Canal in the
observed that embankment deformation patterns for stochastic 1999 Kocaeli earthquake and Balboa Boulevard in the 1994
models became increasingly variable between realizations as x Northridge earthquake (Figure 17), respectively, using a 2D
became larger than about a third of the embankment's base width conditional transition probability method for the interbedded
(B) because deformations became increasingly dependent on the sand and clay strata (i.e., a two category model) to better estimate
location and connectivity of the looser zones in the foundation. seismic deformations experienced at each site. These two studies
Deformation variability was relatively small for cases with x demonstrated that realistic modeling of the relative proportions
less than about a seventh of B because there was more averaging (i.e., sills) of the liquefiable and nonliquefiable soil types was
of shear resistances across looser and denser zones. Paull et al. necessary for obtaining agreement with seismic lateral spreading
(2020) concluded that, for the conditions examined, (1) using Prep deformations observed in the field. Bassal et al. (2021)
367
Figure 18. Realization of Stratum B at a lateral spreading site at Wufeng during the 1999 Chi-Chi earthquake
using 3D transition probability geostatistics for three soil categories (sand, silty/clayey sand, clay) (courtesy
P. Bassal; related details in Bassal et al. 2021).
Figure 19. Shear strain contours from FLAC simulations of Perris Dam over alluvium represented
by 2D slices from 3D Markov-based conditional transition probability realizations: (a) a baseline
realization, (b) a realization from a case with double the mean lengths, and (c) a deterministic
interpretation (Paull et al. 2022).
performed 2D NDAs of a lateral spreading site in Wufeng during Development of input time series for NDA applications is a
the 1999 Chi-Chi earthquake using cross-sections from 3D complicated subject beyond the scope of this paper, but a few
realizations developed using 3D conditional transition brief remarks are in order. Input time series for the horizontal and
probability methods with three soil categories (sand, silty/clayey vertical components need to be developed on the basis of their
sand, clay) in the key strata (Figure 18). Paull et al. (2021) consistency with the primary earthquake scenarios contributing
performed 2D NDAs of Perris Dam using cross-sections from 3D to the seismic hazard and key intensity measures (e.g., response
realizations developed using 3D conditional transition spectra, significant duration, Arias intensity, cumulative absolute
probability methods with three soil categories (liquefiable sand, velocity). The use of conditional mean spectra, hybrid
nonliquefiable sand, clay) in each of two alluvial strata. The conditional mean spectra, or scenario earthquake spectra depend
shear strain contours for Perris Dam shown in Figure 19 illustrate on the geotechnical system, seismic setting, and study objectives.
how deformations coalesced along the weaker interlayers in the The target response spectra for the input motions should be
spatially variable model (Figure 19a), how deformation patterns developed for site conditions (i.e., Vs values) consistent with the
were affected by changes in the mean lengths for the interlayers properties of the base layer in the NDA model, with the base
(Figure 19b), and how uniform analysis models did not
layer located at a significant impedance contrast if possible. Input
reproduce this complexity (Figure 19c).
motions can be applied to a compliant base or they can be applied
Geostatistical methods for representing stratigraphic
to a fixed base with deconvolution. For estimating mean
uncertainty can have significant limitations in approximating
certain depositional environments (e.g., nonstationary properties, deformation responses using NDA models, it is common to
variably inclined bedding, deposits with multiple modes of require three or more spectrally matched motions or seven or
sediment transport and deposition) or may be challenging to more motions scaled to the frequency of interest for each
apply with insufficient exploration data, indicating the need for earthquake scenario. A larger number of motions is required for
further development of these methods for geotechnical estimating the variability in deformations. A recent review of
applications. challenges in, and recommendations for, developing ground
motions for NDAs of geotechnical infrastructure is provided by
6.4 Ground motion time series Thavaraj et al. (2018).
Ground motions are a major source of uncertainty for estimating
deformations using NDAs, as noted by many studies.
368
6.5 Examples of NDA challenges as well as by spatial and temporal variations in soil properties
that are difficult to characterize (e.g., seasonal variations in
6.5.1 Post-liquefaction settlements partial saturation above a water table).
Estimating post-liquefaction settlement is complicated by
constitutive model limitations in simulating reconsolidation 6.5.3 Void redistribution
strains and the inability to simulate loss of subsurface soil as Physical model tests and numerical simulations have shown that
surface ejecta. Constitutive models generally are poor at pore pressure diffusion in liquefiable sands with lower-
estimating reconsolidation strains because continuum models are permeability interlayers can lead to void redistribution-induced
not well suited for representing sedimentation strains, as strength loss and localized deformations. The mechanism of void
discussed in the earlier section on constitutive modeling redistribution was previously illustrated in Figure 13, along with
challenges. While Basu et al. (2022) concluded, as previously discussion of the constitutive modeling challenges in simulating
noted, that using PM4Sand with its post-shaking option for the partially drained loading paths that can develop within a field
reconsolidation modeling was able to predict post-liquefaction profile. The mechanism of void redistribution is illustrated
settlements from three centrifuge tests and two shaking table tests further by the centrifuge model and numerical simulation results
with an average under-prediction of only about 15% (predicted shown in Figure 20, wherein the upward seepage was trapped
values were 60-125% of measured values), the portion of surface beneath an overlying clay crust layer (Figure 20, top), leading to
settlement that could be attributed to surface ejecta (i.e., sand local slip of the clay crust over the underlying liquefied sand
boils) in those tests was relatively small compared to what has (photographs in Figure 20, middle). The numerical simulations
been observed in some case studies, such as in Christchurch using PM4Sand with FLAC predicted loosening of the sand in
during the 2010-2011 Canterbury Earthquake Sequence or the row of elements immediately below the clay crust and the
Urayasu during the 2011 Tohoku earthquake. NDA studies of associated concentration of shear deformations near that
building settlements associated with liquefaction have identified interface (Figure 20, bottom), but did not accurately reproduce
how settlements are affected by various structural, soil, and the distribution or magnitude of interface deformations observed
ground motion characteristics, after which adjustments must be in the centrifuge model. These and other NDAs have
made for the effects of sedimentation strains and soil ejecta (e.g., demonstrated that deformations or instability caused by pore
Macedo & Bray 2018, Bullock et al. 2019). Calibration of NDA pressure diffusion can occur after strong shaking has ended,
findings against case history databases offers one approach for which may explain some past observations of delayed failures.
indirectly accounting for soil ejecta effects (Bullock et al. 2019), Examples of nonlinear simulations of void redistribution leading
although it does not explicitly identify situations where soil to large deformations or localizations include work by Yoshida
ejecta could be a major factor. Hutabarat & Bray (2021) & Finn (2000), Yang & Elgamal (2002), Naesgaard and Byrne
developed an Ejecta Potential Index (EPI) for relating the results
of 1D effective stress site response analyses to the potential for
surface ejecta to develop. Bassal & Boulanger (2021) performed
2D NDAs of Palinurus Road in Christchurch and concluded that
the diffusion of excess pore pressures around a laterally
discontinuous stratum of finer-grained soil explained why sand
boils were concentrated on one side of the site and not observed
on the other side. These and other studies have provided valuable
insights and approaches for estimating post-liquefaction
settlements, while also demonstrating that direct modeling of
sedimentation strains and soil ejecta effects remains beyond
current capabilities.
6.5.2 Localizations
Ground deformations are often accompanied by various
localizations, including those visible from the ground surface
(e.g., cracks, grabens, ejecta pathways) or internal to the
deforming mass (e.g., shear bands). Borja (2012) performed
simulations with biased/unbiased meshes to illustrate abilities
and challenges in predicting localization details from laboratory
test specimens characterized by computed tomography and
digital image processing. Borja further discussed different modes
of localized deformation (continuum strain localization versus
separation mode) and the use of enhanced strain and extended
finite element methods for accommodating a discontinuity
passing through the interior of an element. Islam et al. (2018)
evaluated mesh effects in their coupled Eulerian-Lagrangian
finite element modeling with a total stress, strain-softening
constitutive model of earthquake-induced retrogressive
landslides and runouts in sensitive clay (Figure 3); they showed
that the regularization tactic of scaling the post-peak strain to
mobilize a given shear resistance as inversely proportional to the
thickness of the finite element (analogous to setting a fixed Figure 20. Centrifuge model of a layered soil profile treated with
localization length scale) significantly reduced, but did not prefabricated drains on one side (top), photographs of the localized slip
eliminate, mesh effects. Localizations can also develop due to that occurred between the clay crust and liquefied sand layers due to void
diffusion of earthquake-induced excess pore pressures, which is redistribution processes (middle), and contours of post-shaking relative
discussed separately below. Accurately simulating localizations density from simulations using PM4Sand with FLAC that predicted
in NDAs of geotechnical infrastructure can be limited by the localized loosening and deformation in the top row of sand elements on
the untreated side of the centrifuge model (bottom) (Kamai & Boulanger
computational software (e.g., the combination of numerical
2013, Boulanger et al. 2014).
methods and constitutive models available in a given program),
369
(2007), Seid-Karbasi & Byrne (2007), Kamai & Boulanger
(2013), and Boulanger et al. (2014). Simulations of void
redistribution are complicated by limitations in the constitutive
modeling of reconsolidation strains; limitations in the numerical
modeling of localizations, cracking of crust layers, water film
formation, venting, permeability changes, and large
strains/deformations; and limitations in the accuracy with which
subsurface conditions can be sufficiently characterized for
simulating these mechanisms.
6.5.4 Soil-cement ground improvements
Ground improvement techniques often produce systems with
complex geometries, materials, and loading conditions, for
which the local details of their earthquake loading responses can
be challenging to simulate. One example is the use of soil-cement
grids, walls, or lattices to protect structures or limit ground
deformations for a variety of infrastructure systems (Boulanger
& Shao 2021). Challenges in design can include the effects of
progressive damage to the soil-cement elements under strong
loading and the tradeoffs between fully 3D simulations versus
2D simulations that use composite properties for the treatment
zones.
Challenges with soil-cement wall improvements and 2D
simulations are illustrated by the centrifuge model test and
simulation results in Figure 21. The centrifuge model comprised
an embankment underlain by liquefiable sand, with a row of soil-
cement walls near the embankment toe (Figure 21, top). The soil-
cement walls experienced extensive and irregular cracking with
offsets during the strongest shaking event, as seen in the
photographs of the walls during model dissection/excavation
(Figure 21, middle). 2D NDAs were performed with composite
properties for the soil-cement treatment zone, following common
design practices that did not include strain softening (Boulanger
et al. 2018) and did include strain softening (Boulanger & Shao
2021). Simulations by either approach were in reasonable Figure 21. Centrifuge model of an embankment over liquefiable soil
agreement with the recorded dynamic responses (accelerations, treated with soil-cement walls (top), photographs of the damaged soil-
excess pore pressures, embankment displacements) and correctly cement walls during post-testing model excavation (middle), and
contours of shear strain from 2D simulations with composite strain-
predicted that the soil-cement walls would shear through during softening properties for the soil-cement wall treated zone (bottom)
the strongest shaking event (Figure 21, bottom). The results of (Boulanger et al. 2018; Boulanger & Shao 2021).
these comparisons provided support for the use of the 2D NDA
procedures in practice, but also illustrated a number of local and Uzuoka et al. 2008 for 3D models; Zeghal et al. 2018 for 2D
mechanisms not captured by the simulations. Deformation of the models; Kramer et al. 2015, Ramirez et al. 2018, and Chen &
treatment zone included a component of the enclosed liquefied Arduino 2021 for 1D models). If results of NDAs using two or
sand displacing relative to the soil-cement walls (i.e., extruding more constitutive models or NDA procedures are reasonably
between the walls), with the extrusion effect being greatest consistent, the outcome is usually increased confidence in the
toward the toe region. Cracking and offsets varied irregularly reasonableness of the estimated performance. If NDA results are
along the length of the walls and between adjacent walls. Offsets dramatically different, the outcome is usually reduced
along these undulating crack surfaces were likely accompanied confidence until the root cause for the difference has been
by local fluctuations in normal stress and hence pore pressure. identified. The latter situation provides opportunities to identify
Excess pore pressures in the soil-cement cracks during shearing specific limitations in the constitutive models or the ways they
would be expected to be negative, such that excess pore pressures are used (e.g., Boulanger et al. 2015), which in turn provides
in the enclosed liquefied sand would diffuse into the cracks and important feedback to the model developers and users.
progressively reduce the shear resistance mobilized in the walls. For example, Montgomery & Abbaszadeh (2017) used
In addition, the shear resistance provided by the walls must have PM4Sand (version 3) and UBCSAND (version 904aR) with
degraded progressively during shaking as the cracks and offsets FLAC to simulate a 45 m tall embankment dam underlain by a
grew. The ability to simulate these local details of damage and liquefiable alluvium foundation layer. The NDAs were repeated
deformation will require more advanced 3D NDA models, which for cases with the shells assigned SPT (N1)60cs values of 25 or 35,
can be similarly evaluated against the datasets archived from this the alluvium assigned (N1)60cs of 10, 15, or 20 and with 3 different
centrifuge model test. input motions scaled to peak outcrop accelerations of 0.5 or
0.75g. Crest and shell displacements were reasonably similar for
6.5.5 Evaluating model uncertainty the two constitutive models (e.g., within 30% as illustrated in
Comparing NDA results from use of different NDA procedures Figure 22), while noting that the PM4Sand analyses tended to
can promote advancements and provide an indication of give larger crest and shell displacements when the alluvium had
modeling uncertainty for specific structures. Many studies have the higher blow counts and the UBCAND analyses tended to give
examined the effects of different constitutive models within the larger displacements when the alluvium had the lower blow
same numerical platform (e.g., Armstrong et al. 2013, Armstrong counts. Other comparative studies using the above constitutive
& Boulanger 2015, Li et al. 2018, and Zalachoris et al. 2021), models have produced different trends (e.g., Armstrong et al.
while studies using entirely different numerical platforms for 2D 2013, Armstrong and Boulanger 2015, Hadidi et al. 2017,
or 3D problems are less common (e.g., Cubrinovski et al. 2008 Martinelli & Elkadi 2018), illustrating that results are conditional
370
constitutive models and protocols for numerically representing
systems and their loading, which are the ways that the tools are
used, and (3) expertise of the user and quality of review
processes, which dictate the artistry of the work (Figure 15).
Some needs for further advancing the state of art for NDA studies
are discussed below.
Constitutive models and numerical methods: Advancing the
state of the art will require ongoing improvements in constitutive
models and numerical methods to address various limitations, as
discussed previously, but would also strongly benefit from
having these capabilities integrated in commonly available
software platforms. Adoption of new NDA tools depends on their
verification and validation, which is an effort more efficiently
performed if the broader community can contribute. Thus,
integrating promising new models and numerical methods in
open-source or commercial software platforms will continue to
be important for efficiently advancing the states of art and
practice.
Calibration protocols and numerical modeling protocols:
Validation of NDA procedures depends on repeatable processes
for calibrating constitutive models and performing the numerical
simulations. Advances in NDA procedures, for both the state of
art and practice, would benefit from greater sharing of lessons
Figure 22. Comparison of NDA results for an embankment on alluvium learned and the establishment of best practices documents with
using UBCSAND and PM4Sand (results from Montgomery & detailed descriptions of recommended procedures and protocols.
Abbaszadeh 2017) User expertise and review processes: The quality of NDA
results depends on the expertise of the user (analyst) and the
quality of the review processes. The NDA results need to be
on the calibration protocols and cannot be generalized to other documented in sufficient detail that internal and external
project configurations, loading conditions, or user practices. reviewers can independently evaluate all aspects of the analyses
As a counter example, recent NDA studies using these same and evaluate potential errors. Boulanger & Beaty (2016)
two constitutive models found that using UBCSAND produced discussed the need for improved documentation practices for
ground displacements that were several times those obtained NDA studies and provided a list of questions that a reviewer may
using PM4Sand. Common features of the systems being reasonably ask as part of an internal or external review process.
analyzed were larger-than-typical depths of liquefiable soils, An abbreviated version of the key aspects and issues they listed
gently sloping ground, and long-duration subduction zone for review and documentation is provided in Table 2.
motions. Detailed examination of response time series showed Visualization tools: Visualization tools and graphical user
that significant amounts of pore pressure generation and ground interfaces are essential for efficient interpretation and quality
displacement with the UBCSAND model occurred during the control of NDA results. Visualization tools can facilitate
long-duration, low-level shaking that followed the period of understanding of complex interactions that are otherwise
strongest shaking for these subduction zone motions. The root difficult to infer, and enable systematic examination of the
cause of the difference in NDA results was clarified by single- analysis results at various stages, so the user/reviewer can
element simulations with various stepped cyclic loading confirm whether all aspects of the analyses are functioning
sequences, of which an example was presented previously in appropriately. For example, Chen et al. (2011) used frequency
Figure 12 (Boulanger & Ziotopoulou 2018b). These large decomposition and resampling with direct-volume rendering
differences, compared to the more consistent results shown in and/or deformable implicit surfaces to view 3D NDA results for
Figure 22, further illustrate that comparisons of NDA results the HBMC bridge previously shown in Figures 2. Their
using different constitutive models are conditional on the project visualization results, illustrated by the examples in Figure 23,
configurations and loading conditions, and cannot be broadly allowed viewing of motions in different frequency bands (e.g.,
generalized. low-frequency for lateral spreading, high-frequency for
Experiences like these have demonstrated the importance and structural vibration), enabling a better comprehension of the
value of repeating analyses using two or more NDA procedures, complex interactions between the soil masses, bridge
whether for research or application. Significant differences foundations, and bridge superstructure. Future developments in
between NDA results can be indicative of more serious visualization techniques can be expected to help advance the
limitations in the constitutive models, calibration procedures, or state of art and facilitate improved communication across users,
numerical modeling procedures. Detailed comparisons of system reviewers, and consumers of NDA studies.
and single-element simulations can be used to isolate and identify Validation protocols: Evaluation and validation of an NDA
the potential limitations and thus promote improvements. modeling procedure are processes that improve with increasing
number and variety of physical models and cases histories
analyzed. Multiple evaluation studies are necessary because the
7 ADVANCING THE ART accuracy of an NDA procedure is both system- and loading-
dependent. Good agreement between simulated and measured
The state of art for NDA studies reflects the state of the art in its
responses for one case history or physical model test does not
components and the way they are integrated. An NDA study
ensure good agreement for any other system or for the same
integrates information from site characterization and ground
system under a different loading. The accumulation of
motion studies and produces results that inform the estimation of
experience from multiple evaluation studies provides a basis for
damages and consequences in support of decision making
quantifying potential biases and dispersion levels in NDA
(Figure 1). The NDA procedure comprises (1) constitutive
predictions, while also helping to identify areas where the NDA
models and numerical methods, which are the tools used to
modeling procedures can be improved. Advancing the state of art
perform the analyses, (2) protocols for calibrating the
371
calls for systematic validation efforts, like the international
collaborative LEAP projects, but with extension to more holistic
systems and dynamic loadings. These types of systematic
validation efforts would benefit from expanded reuse of archived
case history and physical model test data, increased archiving of
NDA models, increased automation of validation processes, and
improved data management systems.
Table 2: Key aspects and issues for review and documentation of NDA
studies (abbreviated from Boulanger and Beaty 2016)
Aspect 1: Seismic failure or damage modes
Performance objectives
Potential failure/damage modes
Stress-strain behaviors important to failure/damage modes
Limitations in ability to simulate key behaviors
Aspect 2: Validation for NDA procedure
NDA procedure's record of validation for similar systems
Constitutive model's stress-strain responses compared to
empirical trends/relationships for broad range of conditions
Aspect 3: Site characterization
Basis for all material parameters
Uncertainty in stratigraphy and key material properties
Aspect 4: Calibration of the constitutive model
Calibration protocol and summary of input parameters
Demonstration of key stress-strain responses under
representative initial stress and loading conditions
Aspect 5: Numerical modeling methods and protocols Figure 23. Volume visualization for HBMC bridge (Chen et al. 2011). (a)
Numerical methods Displacement low pass (y)—the vertical drifting (permanent) motion. (b)
Analysis steps for static, dynamic, and post-shaking phases Displacement high pass (x)—the horizontal shaking (cyclic) movement.
Boundary conditions, discretization, zoning Yellow-brown isosurfaces near the approach ramps show downward
Aspect 6: Input ground motion time series settlement, and reddish surfaces in the middle of the bridge show upward
Significant characteristics of the input motions motions. (c) Both low and high-pass parts are composited together. (d)
Consistency with seismic hazard and the numerical model Deformable implicit surfaces along the vertical direction with the high-
Method for applying motion frequency portion superimposed with the low-pass volume. (e) Motion-
Aspect 7: Initial static stress conditions blur effect is applied to convey the high-pass movements.
Distributions for effective stress, Ko, and
Pore pressures consistent with piezometer data
Aspect 8: Dynamic response 8 CONCLUDING REMARKS
Acceleration and displacement time series and response spectra
The state of the art for nonlinear dynamic analyses (NDAs) of
for key points; stress paths and stress-strain responses for
geotechnical infrastructure affected by soils susceptible to
representative elements; and deformed shapes with contours of
liquefaction or cyclic softening has improved significantly
strain, pore pressure ratio, and displacement
through a broad range of efforts over the past two decades.
Consistency with reasonable expectations, any recordings from
Recent developments and needs related to the components of
past earthquakes, and responses of similar systems
NDA procedures were discussed, including: (1) constitutive
Aspect 9: Post-shaking deformations models and numerical methods, which are the tools used to
Approach for representing strength losses not included in perform the analyses, (2) protocols for calibrating constitutive
dynamic response phase (e.g., residual strength, void models and numerically modeling systems, which are the ways
redistribution effects, cyclic degradation, cracking, remolding) that the tools are used, and (3) expertise of the user and the
Aspect 10: Parametric analyses quality of the review processes, which dictate the artistry of the
Effects of key soil properties, system characteristics, procedural work.
issues, or other parameters on computed responses Further advancing the state of the art for NDA studies would
Aspect 11: Uncertainties and limitations benefit from approaches that facilitate broad community
Uncertainty in predicted performance given modeling engagement and participation, including: (1) integrating ongoing
limitations, model uncertainty, validation record, and other improvements in constitutive models and numerical methods in
considerations commonly available open-source or commercial software
Overall tendency for optimistic or pessimistic selection of platforms, (2) greater sharing of lessons learned and the
inputs at each step of the process establishment of best practices documents with detailed
Aspect 12: Reasonableness of conclusions descriptions of recommended calibration protocols and
Conclusions and recommendations drawn from the NDAs, with numerical modeling procedures, (3) documentation practices that
due consideration for the geology, material characteristics, are sufficiently detailed to support critical reviews, and (4)
system configuration, seismic hazard, estimated response, systematic validation efforts spanning a broader range of case
potential consequences, and key sources of uncertainty history and physical model datasets. System responses that
Potential for additional information (e.g., explorations or depend on highly nonlinear, multi-physics, coupled processes are
analyses) to change the conclusions or recommendations inherently difficult to predict accurately, such that quantification
of potential bias and uncertainty from broader validation studies
are important goals. Repeating analyses using two or more NDA
procedures, whether for research or application, can promote
372
improvements in the state of the art and practice. Challenges with centrifuge and shake table tests. Soil Dynamics and Earthquake
NDAs in application are essentially boundless given the diversity Engineering, 153: 107089, 10.1016/j.soildyn.2021.107089.
and complexity of geotechnical infrastructure, geologic Beaty, M.H. and Byrne, P.M. 2008. Liquefaction and deformation
materials, and earthquake loading, which warrants continued analyses using a total stress approach. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 134(8): 1059-1072,
efforts toward fundamentally sound procedures that can be more 10.1061/(ASCE)1090-0241(2008)134:8(1059).
broadly generalized. Beaty, M. H., and Byrne, P. M. 2011. UBCSAND constitutive model:
version 904aR. Documentation report: UBCSAND constitutive
model on Itasca UDM Web Site, February.
9 ACKNOWLEDGEMENTS Bentley. 2021. User defined soil models: PM4Silt – A silt plasticity
model for earthquake engineering. PLAXIS, Connect Edition
Valuable suggestions and comments regarding different parts of V21.00, January 18.
the manuscript were provided by Professors Pedro Arduino, Bhatnagar, S., Kumari, S., and Sawant, V.A. 2016. Numerical analysis
Misko Cubrinovski, Ahmed Elgamal, I. M. Idriss, Mahdi of earth embankment resting on liquefiable soil and remedial
Taiebat, and Katerina Ziotopoulou. Professors Trevor Carey and measures. International Journal of Geomechanics, 16(1): 04015029,
Bruce Kutter provided the figure comparing numerical and 10.1061/(ASCE)GM.1943-5622.0000501.
experimental response surfaces. The author greatly appreciates Bin, Y., Xueqian, N., Yu, H., and Feng, Z. 2018. Unified modeling of
the above assistance and suggestions. soil behaviors before/after flow liquefaction. Computers and
Geotechnics 102, 125–135, 10.1016/j.compgeo.2018.06.011.
Biot, M.A. 1941. General theory of three dimensional consolidation.
Journal of Applied Physics, 12: 155-164.
10 REFERENCES Biot, M.A. 1960. Mechanics of deformation and acoustic propagation in
Abdellaziz, M., Karray, M., Hussien, M. N., Delisle, M.-C., Ledoux, C., porous media. Journal of Applied Physics, 33: 1483-1498.
Locat, P., Mompin, R., and Checkired, M. 2020. Experimental and Borja, R.I. 2012. Computational plasticity for geotechnical applications.
numerical investigation of the Saint-Adelphe landslide after the 1988 Geotechnical Engineering State of the Art and Practice.
Saguenay earthquake. Canadian Geotechnical Journal, 57: 1936– Geocongress, ASCE, 34-53.
1952, 10.1139/cgj-2018-0629. Boulanger, R. W. 2003. Relating K to relative state parameter index.
Adalier, K., Elgamal, A.W., and Martin, G.R. 1998. Foundation Journal of Geotechnical and Geoenvironmental Engineering, ASCE,
liquefaction countermeasures for earth embankments.” Journal of 129(8), 770-773.
Geotechnical and Geoenvironmental Engineering, Boulanger, R.W. 2010. Sand plasticity model for nonlinear seismic
10.1061/(ASCE)1090-0241(1998)124:6(500), 500–517. deformation analyses. Fifth International Conference on Recent
Adamidis, O. and Madabhushi, S. P. G. 2018. Deformation mechanisms Advances in Geotechnical Earthquake Engineering and Soil
under shallow foundations on liquefiable layers of varying thickness. Dynamics, San Diego, CA, paper IMI-6.
Géotechnique 68, No. 7, 1–13, 10.1680/jgeot.17.P.067. Boulanger, R.W. 2019. Nonlinear dynamic analyses of Austrian Dam in
the 1989 Loma Prieta Earthquake. Journal of Geotechnical and
Andrianopoulos, K.I., Agapoulaki, G.I., and Papadimitriou, A.G. 2016.
Simulation of seismic response of passively stabilised sand. Geoenvironmental Engineering, ASCE, 145(11): 05019011,
Geotechnical Research, 2016, 3(2), 40–53, 10.1680/jgere.16.00003. 10.1061/(ASCE)GT.1943-5606.0002156.
Arduino, P. 2018. On the modeling of soils after the onset of liquefaction. Boulanger, R.W., and Beaty, M.H. 2016. Seismic deformation analyses
Proc., Geotechnical Earthquake Engineering and Soil Dynamics V, of embankment dams: A reviewer's checklist. Proceedings,
Geotechnical Special Publication 290, S. J. Brandenberg and M. T. Celebrating the Value of Dams and Levees – Yesterday, Today and
Manzari, eds., ASCE, 1-16. Tomorrow, 36th USSD Annual Meeting and Conference, United
Arduino, P. 2021. Section 2.2: Geotechnical Systems. State of the Art in States Society on Dams, Denver, CO, 535-546
Boulanger, R.W., and Idriss, I.M. 2015. Magnitude scaling factors in
Computational Simulation for Natural Hazards Engineering (Second
Edition). Deierlein and Zsarnóczay (eds), NHERI SimCenter, Report liquefaction triggering procedures. Soil Dynamics and Earthquake
No. 2021-01, 10.5281/zenodo.4558106. Engineering, 79, 296-303, 10.1016/j.soildyn.2015.01.004.
Armstrong, R.J., and Boulanger, R.W. 2015. Numerical simulations of Boulanger, R.W., and Montgomery, J. 2016. Nonlinear deformation
liquefaction effects on piled bridge abutments. 6th International analyses of an embankment dam on a spatially variable liquefiable
Conference on Earthquake Geotechnical Engineering, November 1 deposit. Soil Dynamics and Earthquake Engineering, 91, 222-233,
4, Christchurch, New Zealand. 10.1016/j.soildyn.2016.07.027.
Armstrong, R.J., Boulanger, R.W., and Beaty, M.H. 2013. Liquefaction Boulanger, R.W., and Shao, L. 2021. Liquefaction mitigation with deep
mixing. Proceedings, Deep Mixing 2021, Deep Foundations
effects on piled bridge abutments: Centrifuge tests and numerical
analyses. Journal of Geotechnical and Geoenvironmental Institute, 1146-1202.
Engineering, ASCE, 139(3), 433-443. Boulanger, R.W., and Truman, S.P. 1996. Void redistribution in sand
Asaoka, A., Noda, T., Yamada, E., Kaneda, K. and Nakano, M. 2002. An under post-earthquake loading. Canadian Geotechnical Journal,
elasto-plastic description of two distinct volume change mechanisms 33:829–34.
of soils, Soils and Foundations, 42 (5), 47–57. Boulanger, R.W., and Wijewickreme, D. 2019. Calibration of a
Baecher, G.B., and Christian, J.T. 2003. Reliability and statistics in constitutive model for the cyclic loading response of Fraser River
geotechnical engineering. Wiley, Chichester, U.K., 619. Delta Silt. Proc. 7th International Conference on Earthquake
Geotechnical Engineering, Silvestri and Moraci (eds), Associazione
Bao, Y., Ye, G., Ye, B., and Zhang, F. 2012. Seismic evaluation of soil-
foundation-superstructure system considering geometry and Geotecnica Italiana, Italy, ISBN 978-0-367-14328-2, 121-137.
material nonlinearities of both soils and structures. Soils and Boulanger, R.W., and Ziotopoulou, K. 2013. Formulation of a sand
Foundations, JGS, 52:257-278, 10.106/j.sandf.2012.02.005. plasticity plane-strain model for earthquake engineering
Barrero, A.R., Taiebat, M., and Dafalias, Y.F. 2020. Modeling cyclic applications. Journal of Soil Dynamics and Earthquake Engineering,
shearing of sands in the semifluidized state. International Journal for Elsevier, 53, 254-267, 10.1016/j.soildyn.2013.07.006.
Numerical and Analytical Methods in Geomechanics, 44(3), 371- Boulanger, R. W., and Ziotopoulou, K. 2017. PM4Sand (Version 3.1): A
388. sand plasticity model for earthquake engineering applications.
Report No. UCD/CGM-17/01, Center for Geotechnical Modeling,
Bassal, P.C., and Boulanger, R.W. 2021. System response of an
interlayered deposit with spatially preferential liquefaction Department of Civil and Environmental Engineering, University of
manifestations. Journal of Geotechnical and Geoenvironmental California, Davis, CA, 113 pp.
Engineering, ASCE, 147(12): 05021013, 10.1061/(ASCE)GT.1943- Boulanger, R. W., and Ziotopoulou, K. 2018a. PM4Silt (Version 1): A
5606.0002684. silt plasticity model for earthquake engineering applications. Report
Bassal, P.C., Boulanger, R.W., DeJong, J.T., and Ziotopoulou, K. 2021. No. UCD/CGM-18/01, Center for Geotechnical Modeling,
Calibration of post-liquefaction shear deformation for a fluvial Department of Civil and Environmental Engineering, University of
deposit in the Chi-Chi earthquake. 17th World Conference on California, Davis, CA, 108 pp.
Boulanger, R.W., and Ziotopoulou, K. 2018b. On NDA practices for
Earthquake Engineering (17WCEE), Sendai, Japan.
Basu, D., Montgomery, J., and Stuedlin, A.W. 2022. Observations and evaluating liquefaction effects. Proc., Geotechnical Earthquake
challenges in simulating post-liquefaction settlements from
373
Engineering and Soil Dynamics V, Geotechnical Special Publication 2021/02, Pacific Earthquake Engineering Research Center,
290, S. J. Brandenberg and M. T. Manzari, eds., ASCE, 1-20. University of California, Berkeley.
Boulanger, R.W., and Ziotopoulou, K. 2019. A constitutive model for Chen, L., Ghofrani, A., and Arduino, P. 2019. Prediction of LEAP-UCD-
clays and plastic silts in plane-strain earthquake engineering 2017 centrifuge test results using two advanced plasticity sand
applications. Soil Dynamics and Earthquake Engineering, 127: models. In B. Kutter et al. (Eds.), Model tests and numerical
105832, 10.1016/j.soildyn.2019.105832. simulations of liquefaction and lateral spreading: LEAP-UCD-2017.
Boulanger, R.W., Curras, C.J., Kutter, B.L., Wilson, D.W., and Abghari, New York: Springer.
A. 1999. Seismic soil-pile-structure interaction experiments and Cheng, Z., and Detournay, C. 2021. Formulation, validation and
analyses. Journal of Geotechnical and Geoenvironmental application of a practice-oriented two-surface plasticity sand model.
Engineering, ASCE, 125(9), 750-759. Computer and Geotechnics, 132:103984.
Boulanger, R.W., Kamai, R., and Ziotopoulou, K. 2014. Liquefaction Chian, S.C., Tokimatsu, K., Madabhushi, S.P.G. 2014. Soil liquefaction–
induced strength loss and deformation: Simulation and design. induced uplift of underground structures: Physical and numerical
Bulletin of Earthquake Engineering, Springer, 12: 1107-1128, modeling. Journal of Geotechnical and Geoenvironmental
10.1007/s10518-013-9549-x. Engineering, 140(10): 04014057, 10.1061/(ASCE)GT.1943-
Boulanger, R.W., Montgomery, J., and Ziotopoulou, K. 2015. Nonlinear 5606.0001159.
deformation analyses of liquefaction effects on embankment dams. Chiaradonna, A., Flora, A., d'Onofrio, A., and Bilotta, E. 2020. A pore
Perspectives on Earthquake Geotechnical Engineering, A. Ansal and water pressure model calibration based on in-situ test results. Soils
M. Sakr, eds., Geotechnical, Geological and Earthquake Engineering and Foundations, 60(2): 327-341, 10.1016/j.sandf.2019.12.010.
37, 247-283, Springer, 10.1007/978-3-319-10786-8_10. Chou, J.C., Kutter, B.L., Travasarou, T., and Chacko, J.M. 2011.
Boulanger, R.W., Khosravi, M., Khosravi, A., and Wilson, D.W. 2018. Centrifuge modeling of seismically induced uplift for the BART
Remediation of liquefaction effects for an embankment using soil- Transbay tube. J, Geotechnical and Geoenvironmental Engineering,
cement walls: Centrifuge and numerical modeling. Soil Dynamics 137 (8):754–765, 10.1061/(ASCE)GT.1943-5606.0000489.
and Earthquake Engineering, 114, 38-50, Cubrinovski, M. 2011. Seismic effective stress analysis: Modeling and
10.1016/j.soildyn.2018.07.001. application. 5th International Conference on Earthquake
Boulanger, R.W., Munter, S.K., Krage, C.P., and DeJong, J.T. 2019. Geotechnical Engineering (5ICEGE), Santiago, Chile.
Liquefaction evaluation of interbedded soil deposit: Çark Canal in Cubrinovski, M., Ishihara, K., 1998. State concept and modified
1999 M7.5 Kocaeli Earthquake. Journal of Geotechnical and elastoplasticity for sand modelling. Soils Found. 38, 213–225.
Geoenvironmental Engineering, ASCE, 145(9): 05019007, https://doi.org/10.3208/sandf.38.4_213.
10.1061/(ASCE)GT.1943-5606.0002089. Cubrinovski, M., Uzuoka, R., Sugita, H., Tokimatsu, K., Sato, M.,
Brandenberg, S.J., Chang, D., Boulanger, R.W., and Kutter, B.L. 2003. Ishihara, K., Tsukamoto, Y., and Kamata, T. 2008. Prediction of pile
Behavior of piles in laterally spreading ground during earthquakes: response to lateral spreading by 3-D soil–water coupled dynamic
Centrifuge data report for SJB03. Rep. No. UCD/CGMDR-03/03, analysis: Shaking in the direction of ground flow. Soil Dynamics and
Univ. of California, Davis, CA. Earthquake Engrg. 28:421–435, 10.1016/j.soildyn.2007.10.015.
Brandenberg, S.J., Zhao, M., Boulanger, R.W., and Wilson, D.W. 2013. Cubrinovski, M., Rhodes, A., Ntritsos, N., and van Ballegooy, S. 2018.
p-y plasticity model for nonlinear dynamic analysis of piles in System response of liquefiable deposits. Soil Dynamics and
liquefiable soil. Journal of Geotechnical and Geoenvironmental Earthquake Engrg., 124:212–229, 10.1016/j.soildyn.2018.05.013.
Engineering, ASCE, 10.1061/(ASCE)GT.1943-5606.0000847, Cuomo, S., Ghasemi, P., Martinelli, M., and Calvello, M. 2019.
1262–1274. Simulation of liquefaction and retrogressive slope failure in loose
Bray, J.D., Boulanger, R.W., Cubrinovski, M., Tokimatsu, K., Kramer, coarse-grained material. International Journal of Geomechanics,
S.L., O’Rourke, T., Rathje, E., Green, R.A., Robertson, P.K., and 19(10): 04019116, 10.1061/(ASCE)GM.1943-5622.0001500.
Beyzaei, C.Z. 2017. U.S.-New Zealand-Japan International Dafalias, Y.F., and Manzari, M.T. 2004. Simple plasticity sand model
Workshop, Liquefaction-induced ground movement effects. PEER accounting for fabric change effects. Journal of Engineering
Report 2017/02, Pacific Earthquake Engineering Research Center, Mechanics, ASCE, 130(6), 622-634.
University of California, Berkeley, CA. Dafalias, Y.F., and Taiebat, M. 2014. Rotational hardening with and
Brinkgreve, R.B.J., Engin, E. and Swolfs, W.M. 2017. PLAXIS 2D without anisotropic fabric at critical state. Géotechnique, 64(6): 507-
manual. Rotterdam, Netherlands, Balkema. 511, 10.1680/geot.13.T.035.
Buenker, J.M., Brandenberg, S.J., and Stewart, J.P. 2021. Centrifuge Dafalias, Y.F., and Taiebat, M. 2016. Sanisand-Z: Zero elastic range sand
testing of soil–structure interaction effects on cyclic failure potential plasticity model. Géotechnique 66, 999–1013,
of fine-grained soil. Earthquake Spectra, 37(2):1177-1198, 10.1680/jgeot.15.P.271.
10.1177/8755293020981978. Dafalias, Y.F, Manzari, M.T., and Papadimitriou, A.G. 2006.
Bullock, Z., Karimi, Z., Dashti, S., Porter, K., Liel, A.B., and Franke, SANICLAY: simple anisotropic clay plasticity model. International
K.W. 2019. A physics-informed semi-empirical probabilistic model Journal for Numerical and Analytical Methods in Geomechanics
for the settlement of shallow-founded structures on liquefiable 2006; 30(12):1231–1257.
ground. Géotechnique 69(5), 406–419, 10.1680/jgeot.17.P.174. Dakoulas, P., Vazouras, P., Kallioglou, P., and Gazetas, G. 2018.
Carey, T.J., and Kutter, B.L. 2022. Validation of models used to predict Effective-stress seismic analysis of a gravity multi-block quay wall.
the effects of lateral spreading by comparison of experimental and Soil Dynamics and Earthquake Engineering 115: 378–393,
numerical response surfaces. Soil Dynamics and Earthquake 10.1016/j.soildyn.2018.08.032.
Engineering, in review. Dashti, S., Bray, J.D., Pestana, J.M., Riemer, M.R. & Wilson, D. 2010.
Chakrabortty, P. and Popescu, R. 2012. Numerical simulation of Centrifuge testing to evaluate and mitigate liquefaction induced
centrifuge tests on homogeneous and heterogeneous soil models. building settlement mechanisms. Journal of Geotechnical and
Computers and Geotechnics, 41, 95–105, Geoenvironmental Engineering, ASCE, 136(7):918–929.
10.1016/j.compgeo.2011.11.008. Dashti, H., Sadrnejad, S.A., and Ganjian, N. 2017. Multi-directional
Chaloulos, Y.K., Bouckovalas, G.D., and Karamitros, D.K. 2013. modeling for prediction of fabric anisotropy in sand liquefaction.
Analysis of Liquefaction Effects on Ultimate Pile Reaction to Lateral Computers and Geotechnics 92: 156–168,
Spreading. Journal of Geotechnical and Geoenvironmental 10.1016/j.compgeo.2017.08.002.
Engineering, 140(3): 04013035, 10.1061/(ASCE)GT.1943- Dawson, E.M., Roth, W.H., Nesarajah, S., Bureau, G., and Davis, C.A.
5606.0001047. 2001. A practice oriented pore pressure generation model.
Chang, D., Boulanger, R.W., Brandenberg, S., and Kutter, B.L. 2013. Proceedings, 2nd FLAC Symposium on Numerical Modeling in
FEM analysis of dynamic soil-pile-structure interaction in liquefied Geomechanics. Oct. 29-31, Lyon, France.
and laterally spreading ground. Earthquake Spectra, EERI, 29(3), Deierlein, G.G., and Zsarnóczay, A. (eds). 2021. State of the Art in
733-755. Computational Simulation for Natural Hazards Engineering (Second
Chen, C.-K., Ho, C., Correa, C., Ma, K.-L., and Elgamal, A. 2011. Edition). NHERI SimCenter, Report No. 2021-01,
Visualizing 3D earthquake simulation data. Computing in Science 10.5281/zenodo.4558106.
and Engineering, 13(6): 52-63, 10.1109/MCSE.2010.120. Demir, S., and Ozener, P. 2019. Numerical investigation of seismic
Chen, L., and Arduino, P. 2021. Implementation, verification, and performance of high modulus columns under earthquake loading.
validation of the PM4Sand model in OpenSees. PEER Report No. Earthquake Engineering and Engineering Vibration, 18(4): 811–822,
10.1007/s11803-019-0537-2.
374
Demir, S., and Ozener, P. 2020. Parametric investigation of effectiveness Howell, R., Rathje, E.M., and Boulanger, R.W. 2014. Evaluation of
of high modulus columns in liquefaction mitigation. Soil Dynamics simulation models of lateral spread sites treated with prefabricated
and Earthquake Engineering, 139: 106337, vertical drains. Journal of Geotechnical and Geoenvironmental
10.1016/j.soildyn.2020.106337. Engineering, ASCE, 141(1), 04014076, 10.1061/(ASCE)GT.1943-
Duque, J., Yang, M., Fuentes, W., Masin, D., and Taiebat, M. 2021. 5606.0001185.
Characteristic limitations of advanced plasticity and hypoplasticity Hutabarat, D., and Bray, J.D. 2021. Effective stress analysis of
models for cyclic loading of sands. Acta Geotechnica, liquefiable sites to estimate the severity of sediment ejecta. Journal
10.1007/s11440-021-01418-z. of Geotechnical and Geoenvironmental Engineering, 147(5):
Elbadawy, M. A., Zhou, Y-G., and Liu, K. 2022. A modified pressure 0402102, 10.1061/(ASCE)GT.1943-5606.0002503.
dependent multi-yield surface model for simulation of LEAP-Asia- Iai, S., Matsunaga, Y., Kameoka, T. 1992. Strain space plasticity model
2019 centrifuge experiments. Soil Dynamics and Earthquake for cyclic mobility. Soils and Foundations, 32:1–15,
Engineering, 154: 107135, 10.1016/j.soildyn.2021.107135. 10.3208/sandf1972.32.2_1.
Elgamal, A., Yang, Z., and Parra, E. 2002. Computational modeling of Iai, S., Tobita, T., Ozutsumi, O., Ueda, K. 2011. Dilatancy of granular
cyclic mobility and post-liquefaction site response. Soil Dynamics materials in a strain space multiple mechanism model. Int. Journal
and Earthquake Engineering, 22: 259-271. Numerical Analytical Methods in Geomechanics, 35(3): 360-392.
Elgamal, A., Yang, Z., Parra, E. and Ragheb, A. 2003. Modelling of Ishihara, K. 1996. Soil behaviour in earthquake geotechnics. Oxford
cyclic mobility in saturated cohesionless soils. International Journal University Press, New York.
of Plasticity 19(6): 883–905. Islam, N., Hawlader, B., Wang, C., and Soga, K. 2018. Large-
Elgamal, A., Yang, Z. and Lu, J. 2006. Cyclic1D: a computer program deformation finite-element modelling of earthquake-induced
for seismic ground response, Report No. SSRP-06/05. San Diego, La landslides considering strain-softening behaviour of sensitive clay.
Jolla, CA, USA: Department of Structural Engineering, University Canadian Geotechnical Journal, 56:1003–1018, 10.1139/cgj-2018-
of California. See http://soilquake.net/cyclic1d/index.html. 0250
Elgamal, A., Yan, L., Yang, Z., and Conte, J.P. 2008. Three-dimensional Itasca. 2019. Fast Lagrangian Analysis of Continua (FLAC), release 8.1.”
seismic response of Humboldt Bay bridge-foundation-ground Itasca Consulting Group, Inc., Minneapolis, MN.
system. Journal of Structural Engineering, 134(7): 1165-1176, Jefferies, M., Shuttle, D., and Been, K. 2015. Principal stress rotation as
10.1061/(ASCE)0733-9445, 134:7(1165). cause of cyclic mobility. Geotechnical Research, 2(2), 66–96,
El Shamy, U., Zeghal, M., Dobry, R., Thevanayagam, S., Elgamal, A., 10.1680/gr.15.00002.
Abdoun, T., Medina, C., Bethapudi, R., and Bennett, V. 2010. Joint TC205/TC304 Working Group. 2017. Discussion of
Micromechanical aspects of liquefaction-induced lateral spreading. statistical/reliability methods for Eurocodes – Final Report. 5th
International Journal of Geomechanics, 10(5):190-201, International Symposium on Geotechnical Safety and Risk,
10.1061/(ASCE)GM.1943-5622.0000056. International Society for Soil Mechanics and Geotechnical
Esfeh, P.K., and Kaynia, A.M. 2020. Earthquake response of monopiles Engineering, Rotterdam, Netherlands, 2015.
and caissons for Offshore Wind Turbines founded in liquefiable soil. Kamai, R. and Boulanger, R.W. 2012. Single-element simulations of
Soil Dynamics and Earthquake Engineering 136:106213, partial-drainage effects under monotonic and cyclic loading. Journal
10.1016/j.soildyn.2020.106213. of Soil Dynamics and Earthquake Engineering, 35, 29-40.
Fenton, G.A., and Griffiths, D.V. 2008. Risk Assessment in Geotechnical Kamai, R., and Boulanger, R.W. 2013. Simulations of a centrifuge test
Engineering. John Wiley & Sons, Inc., New York. with lateral spreading and void redistribution effects. Journal of
Ferdosi, B., James, M., and Aubertin, M. 2015. Numerical simulations of Geotechnical and Geoenvironmental Engineering, ASCE, 139(8),
seismic and post-seismic behavior of Tailings. Canadian 1250-1261.
Geotechnical Journal, 53:85–92, 10.1139/cgj-2014-0345. Kamai, R., Kano, S., Conlee, C. Marinucci, A., Rathje, E., Boulanger, R.
Finn, W.D.L. 2000. State-of-the-art of geotechnical earthquake and Rix, G. 2007. Evaluation of the effectiveness of prefabricated
engineering practice. Soil Dynamics and Earthquake Engineering, vertical drains for liquefaction remediation: Centrifuge data report
20: 1-15. for SSK01. Report No. UCD/CGMDR. Center for Geotechnical
FLIP Consortium. 2022. Finite element analysis program of liquefaction Modeling, University of California, Davis. Curated at NEES,
process. Kyoto, Japan, https://www.flip.or.jp/en/flip_english.html. Dataset, dx.doi.org/10.4231/D32J6839J.
Fuentes, W., Wichtmann, T., Gil, M., and Lascarro, C. 2020. ISA- Kan, M.E., Taiebat, H.A., and Khalili, N. 2014. Simplified mapping rule
Hypoplasticity accounting for cyclic mobility effects for liquefaction for bounding surface simulation of complex loading paths in
analysis. Acta Geotechnica, 15:1513–1531, 10.1007/s11440-019- granular materials. International Journal of Geomechanics, 14(2):
00846-2. 239-253, 10.1061/(ASCE)GM.1943-5622.0000307.
Galavi, V., Petalas, A., and Brinkgreve, R.B.J. 2013. Finite element Karimi, Z. and Dashti, S. 2016. Seismic performance of shallow founded
modelling of seismic liquefaction in soils. Geotechnical Engineering structures on liquefiable ground: validation of numerical simulations
Journal, SEAGS AGSSEA, 44(3), 55–64. using centrifuge experiments. Journal of Geotechnical and
Gonzalez, L., Abdoun, T., and Sharp, M.K. 2002. Modelling of Geoenvironmental Engineering, 142(6), 04016011.
seismically induced liquefaction under high confining stress. Karamitros, D. 2010. Development of a numerical algorithm for the
International J. of Physical Modelling in Geotechnics, 2(3), 01-15. dynamic elastoplastic analysis of geotechnical structures in two (2)
González, L., Abdoun, T., and Dobry, R. 2009. Effect of soil and three (3) dimensions. PhD Dissertation, National Technical
permeability on centrifuge modeling of pile response to lateral University of Athens, Greece.
spreading. Journal of Geotechnical and Geoenvironmental Kassas, K., Adamidis, O., Gerolymos, N., and Anastasopoulos, I. 2020.
Engineering, 10.1061/(ASCE)1090-0241, 135: 1(62), 62–73. Numerical modelling of a structure with shallow strip foundation
Hadidi, R., Dinsick, A., Wahl, D., Hu, J., and Perez, A. 2017. during earthquake-induced liquefaction. Géotechnique,
Characterization and seismic performance of Bouquet Canyon Dam 10.1680/jgeot.19.P.277.
No. 1. Proc., Performance-based Design in Earthquake Geotechnical Kawamoto, R., Ando, E., Viggiani, G., and Andrade, J.E. 2016. Level set
Engineering, PBD-III Vancouver, M. Taiebat et al., eds., ISSMGE discrete element method for three-dimensional computations with
Technical Committee TC203, paper 110. triaxial case study. Journal of the Mechanics and Physics of Solids,
Harder, L.F., and Boulanger, R.W. 1997. Application of K and K 91: 1-13, 10.1016/j.jmps.2016.02.021.
correction factors. Proceedings of the NCEER Workshop on Kheradi, H., Nagano, K., Nishi, H., and Zhang, F. 2018. 1-g shaking table
Evaluation of Liquefaction Resistance of Soils, T. L. Youd and I. M. tests on seismic enhancement of existing box culvert with atrial
Idriss, Eds., Technical Report NCEER-97-0022, National Center for ground-improvement method and its 2D dynamic simulation. Soils
Earthquake Engineering Research, SUNY, Buffalo, 167-190. and Foundations, 58: 563-581.
Hausler, E.A. 2002. Influence of ground improvement on settlement and Khosravifar, A., Boulanger, R.W., and Kunnath, S.K. 2014. Effects of
liquefaction: a study based on field case history evidence and liquefaction on inelastic demands on extended pile shafts.
dynamic geotechnical centrifuge tests. PhD dissertation, University Earthquake Spectra, EERI, 30(4), 1749-1773,
of California, Berkeley. 10.1193/032412EQS105M.
He, L., Ramirez, J., Lu, J., Tang, L., Elgamal, A., and Tokimatsu, K. Khosravifar, A., Elgamal, A., Lu, J., and Li, J. 2018. A 3D model for
2017. Lateral spreading near deep foundations and influence of soil earthquake-induced liquefaction triggering and post-liquefaction
Permeability. Canadian Geotechnical Journal, 54:846–861, response. Soil Dynamics and Earthquake Engineering, 110: 43-52.
10.1139/cgj-2016-0162.
375
Kiernan, M., and Montgomery, J. 2020. Numerical simulations of Fourth injection on liquefiable foundation soil of light structures. Soils and
Avenue Landslide considering cyclic softening. Journal of Foundations, 55(6): 1388-1399.
Geotechnical and Geoenvironmental Engineering, 146(10): Martinelli, M., and Elkadi, A. 2018. Numerical modeling of liquefaction
04020099, 10.1061/(ASCE)GT.1943-5606.0002349. under sloped ground condition. 16th European Conference on
Kimoto, S., Khan, B.S., Mirjalili, M., and Oka, F. 2014. Cyclic Earthquake Engineering, Thessaloniki, June.
elastoviscoplastic constitutive model for clay considering nonlinear McGann, C.R., Arduino, P., and Mackenzie-Helnwein, P. 2012.
kinematic hardening rules and structural degradation. International Stabilized single-point 4-node quadrilateral element for dynamic
Journal of Geomechanics, 15(5): A4014005, analysis of fluid saturated porous media, Acta Geotechnica: 7(4):
10.1061/(ASCE)GM.1943-5622.0000327. 297-311.
Kramer, S.L., Asl, B.A., Ozener, P. and Sideras, S. 2015. Effects of McGann, C.R., Arduino, P., and Mackenzie-Helnwein, P. 2015.
liquefaction on ground surface motions. Perspectives on Earthquake Stabilized single-point 8-node hexahedral element for dynamic
Geotechnical Engineering, A. Ansal, M. Sakr (eds.), Geotechnical, analysis of fluid saturated porous media, Computers and
Geological and Earthquake Engineering 37, DOI 10.1007/978-3- Geotechnics, 66: 126-141.
319-10786-8_11. McKenna, F., Scott, M.H., and Fenves, G.L. 2010. Nonlinear finite-
Kuhn, M.R., Renken, H.E., Mixsell, A.D., and Kramer, S.L. 2014. element analysis software architecture using object composition.
Investigation of cyclic liquefaction with discrete element Journal of Computing in Civil Engineering,
simulations. Journal of Geotechnical and Geoenvironmental 10.1061/(ASCE)CP.1943-5487.0000002, 95–107.
Engineering, 140(12): 04014075, 10.1061/(ASCE)GT.1943- Ming, H.Y., Li, X.S., and Zheng, T.Y. 2011. Effects of soil density and
5606.0001181. earthquake intensity on flow deformation of the upper San Fernando
Kulasingam, R., Malvick, E.J., Boulanger, R.W., Kutter, B.L. 2004. dam. Géotechnique 61, No. 12, 1019–1034,
Strength loss and localization at silt interlayers in slopes of liquefied 10.1680/geot.9.P.004.3682.
sand. Journal of Geotechnical and Geoenvironmental Engineering, Mirshekari, M., and M. Ghayoomi. 2017. Centrifuge tests to assess
ASCE, 130(11), 1192-1202. seismic site response of partially saturated sand layers.” Soil
Kutter B.L. et al. 2020a. Numerical sensitivity study compared to trend Dynamics and Earthquake Engineering, 94: 254–265,
of experiments for LEAP-UCD-2017. In: Kutter B., Manzari M., 10.1016/j.soildyn.2017.01.024.
Zeghal M. (eds) Model Tests and Numerical Simulations of Montgomery, J., and Abbaszadeh, S. 2017. Comparison of two
Liquefaction and Lateral Spreading. Springer, 219-236, constitutive models for simulating the effects of liquefaction on
10.1007/978-3-030-22818-7_11. embankment dams. Proceedings, 37th Annual USSD Conference,
Kutter, B.L., Manzari, M., and Zeghal, M. (eds) 2020b. Model tests and USSD, Denver, CO.
numerical simulations of liquefaction and lateral spreading. Montgomery, J., and Boulanger, R.W. 2016. Effects of spatial variability
Springer, 10.1007/978-3-030-22818-7_11. on liquefaction-induced settlement and lateral spreading. Journal of
Lade, P.V. and Liggio, C.D.2014. Stability and instability of granular Geotechnical and Geoenvironmental Engineering, ASCE, 2017,
materials under imposed volume changes: Experiments and 143(1), 04016086, 10.1061/(ASCE)GT.1943-5606.0001584.
predictions. International Journal of Geomechanics, 14(5): Montgomery, J., and Ziotopoulou, K. 2019. Numerical simulations of
04014020, 10.1061/(ASCE)GM.1943-5622.0000352. selected LEAP centrifuge experiments with PM4Sand in FLAC. In
Li, P., Dashti, S., Badanagki, M., and Kirkwood, P. 2018. Evaluating 2D B. Kutter et al. (Eds.), Model tests and numerical simulations of
numerical simulations of granular columns in level and gently liquefaction and lateral spreading: LEAP-UCD-2017. New York:
sloping liquefiable sites using centrifuge experiments. Soil Springer.
Dynamics and Earthquake Engineering, 110: 232–243, Montgomery, J., Boulanger, R.W., and Ziotopoulou, K. 2017. Effects of
10.1016/j.soildyn.2018.03.023. spatial variability on the seismic response of the Wildlife
Li, X.S. 2002. A sand model with state-dependent dilatancy. Liquefaction Array. Proc., Performance-based Design in Earthquake
Géotechnique 52(3): 173–186, 10.1680/geot.2002.52.3.173. Geotechnical Engineering, PBD-III Vancouver, M. Taiebat et al.,
Li, X.S. and Dafalias, Y.F. 2012. Anisotropic critical state theory: role of eds., ISSMGE Technical Committee TC203, paper 533.
fabric. Journal of Engineering Mechanics, 138(3): 263–275. Naesgaard, E., and Byrne, P.M. 2007. Flow liquefaction simulation using
LIQCARI. 2021. LIQA Liquefaction Geo-Research Institute, Kyoto, a combined effective stress: total stress model. In: 60th Canadian
Japan, https://liqca.org/. geotechnical conference. Canadian Geotechnical Society, Ottawa.
Liu, C., and Muraleetharan, K.K. 2012. Coupled hydro-mechanical Naesgaard, E., Byrne, P.M. and Amini, A. 2015. Hysteretic model for
elastoplastic constitutive model for unsaturated sands and silts. I: non-liquefiable soils (UBCHYST5d). UDM Version 5d, on Itasca
Formulation. International Journal of Geomechanics, 12(3): 239- UDM website.
247, 10.1061/(ASCE)GM.1943-5622.0000146. NASEM. 2021. State of the art and practice in the assessment of
Liu, H.Y., Abell, J.A., Diambra, A., and Pisano, F. 2019. Modelling the earthquake-induced soil liquefaction and its consequences. National
cyclic ratcheting of sands through memory-enhanced bounding Academies of Sciences, Engineering, and Medicine. Washington,
surface plasticity. Géotechnique 69 (9): 783–800, DC: The National Academies Press. 10.17226/23474.
10.1680/jgeot.17.P.307. Nie, Z., Chi, S., and Gong, S. 2017. Numerical Modeling of Cyclic
Liu, H.Y., Abell, J.A., Diambra, A., and Pisano, F. 2020. Memory- Triaxial Experiments for Granular Soil. International Journal of
enhanced plasticity modeling of sand behavior under undrained Geomechanics, 17(6): 04016, 10.1061/(ASCE)GM.1943-
cyclic loading. Journal of Geotechnical and Geoenvironmental 5622.0000832.
Engineering, 146(11): 04020122, 10.1061/(ASCE)GT.1943- Niemunis, A., Herle, I. 1997. Hypoplastic model for cohesionless soils
5606.0002362. with elastic strain range. Mechanics of Cohesive-Frictional
Luque, R., and Bray, J.D. 2020. Dynamic soil-structure interaction Materials, 2(4):279–299.
analyses of two important structures affected by liquefaction during Oka, F., Yashima, A., Tateishi, A., Taguchi, Y., Yamashita, S. 1999. A
the Canterbury earthquake sequence. Soil Dynamics and Earthquake cyclic elasto-plastic constitutive model for sand considering a
Engineering 133: 106026, 10.1016/j.soildyn.2019.106026. plastic-strain dependence of the shear modulus. Géotechnique 49(5):
Macedo, J., and Bray, J.D. 2018. Key trends in liquefaction-induced 661-680.
building settlement. Journal of Geotechnical and Geoenvironmental Oka, F., Tsai, P., Kimoto, S., and Kato, R. 2012. Damage patterns of river
Engineering, 144(11): 04018076, 10.1061/(ASCE)GT.1943- embankments due to the 2011 off the Pacific Coast of Tohoku
5606.0001951. Earthquake and a numerical modeling of the deformation of river
Manzari, M., Kutter, B.L., and Zeghal, M., 2018. LEAP 2015. Special embankments with a clayey subsoil layer. Soils and Foundations,
Issue, Soil Dynamics and Earthquake Engineering. 52(5): 890-909.
Manzari, M.T., et al. 2020. LEAP-2017 Simulation exercise: Calibration Olson, S.M., Mei, X., and Hashash, Y.M.A. 2020. Nonlinear site
of constitutive models and simulation of the element tests. In: Kutter response analysis with pore-water pressure generation for
B., Manzari M., Zeghal M. (eds) Model Tests and Numerical liquefaction triggering evaluation. Journal of Geotechnical and
Simulations of Liquefaction and Lateral Spreading. Springer, 165- Geoenvironmental Engineering, 146(2): 04019128.
185, https://doi.org/10.1007/978-3-030-22818-7_11. Ozutsumi, O. 2019. LEAP-UCD-2017 numerical simulation at Meisosha
Marasini, N.P., and Okamura, M. 2015. Numerical simulation of Corp. In B. Kutter et al. (Eds.), Model tests and numerical
centrifuge tests to evaluate the performance of desaturation by air simulations of liquefaction and lateral spreading: LEAPUCD-2017.
New York: Springer.
376
Pan, K., Xu, T.T., Liao, D., and Yang, Z.X. 2022. Failure mechanisms of Dynamics and Earthquake Engineering, 125: 105665,
sand under asymmetrical cyclic loading conditions: experimental 10.1016/j.soildyn.2019.05.004.
observation and constitutive modelling Géotechnique 72(2): 162– Sato, M., Tazoh, T., and Ogasawara, M. 2001. Reproduction of lateral
175, 10.1680/jgeot.20.P.004. ground displacement and lateral flow—Earth pressure acting on pile
Papadimitriou, A.G., and Bouckovalas, G.D. 2002. Plasticity model for foundations using centrifuge modeling.” In Proc., 4th Int. Conf. on
sand under small and large cyclic strains: a multiaxial formulation. Recent Advances in Geotechnical Earthquake Engineering and Soil
Soil Dynamics and Earthquake Engineering, 22(3):191–204. Dynamics Symp., edited by S. Prakash. Rolla, MO: Univ. of
Paull, N.A., Boulanger, R.W., and DeJong, J.T. 2020. Accounting for Missouri–Rolla.
spatial variability in nonlinear dynamic analyses of embankment Seidalinov, G., and Taiebat, M. 2014. Bounding surface SANICLAY
dams on liquefiable deposits. Journal of Geotechnical and plasticity model for cyclic clay behavior. Int. Journal for Numerical
Geoenvironmental Engineering, ASCE, 146(11): 04020124, and Analytical Methods in Geomechanics. 38:702–724,
10.1061/(ASCE)GT.1943-5606.0002372. 10.1002/nag.2229.
Paull, N.A., Boulanger, R.W., DeJong, J.T., and Friesen, S.J. 2021. Using Seid-Karbasi, M., and Byrne, P.M. 2007. Seismic liquefaction, lateral
conditional random fields for a spatially variable liquefiable spreading, and flow slides: a numerical investigation into void
foundation layer in nonlinear dynamic analyses of embankments. redistribution. Canadian Geotechnical Journal, 44(7):873–89.
Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Sento, N., Kazama, M., Uzuoka, R., Ohmura, H., Ishimaru, M. 2004.
147(11): 04021134, 10.1061/(ASCE)GT.1943-5606.0002610. Possibility of postliquefaction flow failure due to seepage. Journal
Paull, N.A., Boulanger, R.W., DeJong, J.T., and Friesen, S.J. 2022. of Geotechnical and Geoenvironmental Engineering, 130:707–16.
Nonlinear dynamic analyses of Perris Dam using transition Shahir, H., and Pak, A. 2010. Estimating liquefaction-induced settlement
probability to model interbedded alluvial strata. Journal of of shallow foundations by numerical approach, Computers and
Geotechnical and Geoenvironmental Engineering, ASCE, 148(1): Geotechnics, 37: 267–279
05021015, 10.1061/(ASCE)GT.1943-5606.0002663. Shahir, H., Pak, A., Taiebat, M., and Jeremic , B. 2012. Evaluation of
Pender, M.J., Orense, R.P., Wotherspoon, L.M., and Storie, L.B. 2016. variation of permeability in liquefiable soil under earthquake
Effect of permeability on the cyclic generation and dissipation of loading. Computers and Geotechnics, 40, 74–88.
pore pressures in saturated gravel layers. Géotechnique 66(4): 313– Shin, H., Arduino, P., and Kramer, S.L. 2007. Performance-based
322, 10.1680/jgeot.SIP.15.P.024. evaluation of bridges on liquefiable soils. Research Frontiers at
Petalas, A.L., Dafalias, Y.F., and Papadimitriou, A.G. 2018. Structures Congress, ASCE, 10.1061/40944(249)41.
SANISAND-FN: An evolving fabric-based sand model accounting Singh, P., Subramanian, P.K., Boulanger, R.W., and Kutter, B.L. 2000.
for stress principal axes rotation. Interrnational Journal of Numerical Behavior of piles in laterally spreading ground during earthquakes.
and Analytical Method in Geomechanics, 43, 97-123. Rep. No. UCD/CGMDR-00/05, Univ. of California, Davis, CA.
Petalas, A.L., Dafalias, Y.F., and Papadimitriou, A.G. 2020. Soga, K., Alonso, E., Yerro, A., Kumar, K., and Bandara, S. 2016. Trends
SANISAND-F: Sand constitutive model with evolving fabric in large-deformation analysis of landslide mass movements with
anisotropy. International Journal of Solids and Structures, 188-189, particular emphasis on the material point method. Géotechnique
12-31. 66(3): 248–273, 10.1680/jgeot.15.LM.005.
Popescu, R., Prevost, J.H., and Deodatis, G. 1997. Effects of spatial Stoecklin, A., Friedli, B., and Puzrin, A.M. 2020. A multisurface
variability on soil liquefaction: some design recommendations. kinematic hardening model for the behavior of clays under combined
Géotechnique, 47(5), 1019–1036. static and undrained cyclic loading. Int J Numer Anal Methods
Popescu, R., Prevost, J.H., and Deodatis, G. 2005. 3D effects in seismic Geomech. 2020;1–30.
liquefaction of stochastically variable soil deposits. Géotechnique, Sturt, R., Cengiz, C., Huang, Y., Go, J., Bandara, S., and Pillai, A. 2021.
55(1), 21–31. Modelling liquefaction of soils wi ht LS-DYNA usiong a
Pretell, R., Ziotopoulou, K, and Davis, C.A. 2020. Liquefaction and SANISAND-based material model. 13th European LS-DYNA
cyclic softening at Balboa Boulevard during the 1994 Northridge Conference, Ulm, Germany.
Earthquake. Journal of Geotechnical and Geoenvironmental Su, D., and Li, X.S. 2008. Impact of multidirectional shaking on
Engineering, ASCE, 147(2): 05020014, 10.1061/(ASCE)GT.1943- liquefaction potential of level sand deposits. Géotechnique,
5606.0002417. 58(4):259–67.
Qiu, Z., and Elgamal, A. W. 2019. Numerical simulations of LEAP Su, L., Wan, H.-P., Abtahi, S., Li, Y., and Ling, X.-Z. 2019. Dynamic
dynamic centrifuge model tests for response of liquefiable sloping response of soil–pile–structure system subjected to lateral spreading:
ground. In B. Kutter et al. (Eds.), Model tests and numerical shaking table test and parallel finite element simulation. Canadian
simulations of liquefaction and lateral spreading: LEAP-UCD-2017. Geotechnical Journal, 57: 497–517, 10.1139/cgj-2018-048.
New York: Springer. Taborda, D.M.G., Zdravkovic, L., Kontoe, S., and Potts, D.M. 2014.
Qiu, Z., and Elgamal, A. 2020. Three-dimensional modeling of strain- Computational study on the modification of a bounding surface
softening soil response for seismic-loading applications. Journal of plasticity model for sands. Computers and Geotechnics, 59:145–160.
Geotechnical and Geoenvironmental Engineering, 146(7): Taghavi, A., and Muraleetharan, M. 2012. Seismic and Cyclic Lateral
04020053, 10.1061/(ASCE)GT.1943-5606.0002282. Load Tests of Pile Groups and Single Piles - Centrifuge Model No.
Qiu, Z., Ebeido, A., Almutairi, A., Lu, J., Elgamal, A., Shing, P. B., and 4. DesignSafe-CI [publisher], doi: 10.4231/D39P2W61H.
Martin, G. 2020. Aspects of bridge ground seismic response and Taiebat, M., Dafalias, Y.F. 2008. SANISAND: Simple anisotropic sand
liquefaction induced deformations. Earthquake Engineering and plasticity model. Int. J. Numer. Anal. Meth. Geomech. 32, 915–948.
Structural Dynamics, 49(4), 375-393. https://doi.org/10.1002/nag.651.
Rahimi, M., Chan, D., Nouri, A., and Rasouli, R. 2016. Effects of Taiebat, M., Dafalias, Y.F., and Peek, R. 2010. A destructuration theory
inherent fabric anisotropy and intermediate principal stress on and its application to SANICLAY model. International Journal for
constitutive behavior of uncemented and cemented sands. Numerical and Analytical Methods in Geomechanics, 34(10), 1009–
Computers and Geotechnics, 80: 237–247. 1040.
Ramirez, J., Barrero, A.R., Chen, L., Dashti, S., Ghofrani, A., Taiebat, Takahashi, A. and Sugita, H. 2009. Behaviour of SCP-improved levee
M., and Arduino, P. 2018. Site response in a layered liquefiable during 2003 Miyagiken-Hokubu Earthquake. Proc. IS-Tokyo2009,
deposit: Evaluation of different numerical tools and methodologies Earthquake Geotechnical Case Histories for Performance-Based
with centrifuge experimental results. Journal of Geotechnical and Design, Kokusho, (ed), Taylor & Francis, pp. 177-184.
Geoenvironmental Engineering, 144(10): 04018073, Tasiopoulou, P., and Gerolymos, N. 2016a. Constitutive modeling of
10.1061/(ASCE)GT.1943-5606.0001947. sand: Formulation of a new plasticity approach. Soil Dynamics and
Rayamajhi, D., Tamura, S., Khosravi, M., Boulanger, R.W., Wilson, Earthquake Engineering, 82:205-221.
D.W., Ashford, S.A., and Olgun, C.G., 2015. Dynamic centrifuge Tasiopoulou, P., and Gerolymos, N. 2016b. Constitutive modelling of
tests to evaluate reinforcing mechanisms of soil-cement columns in sand: a progressive calibration procedure accounting for intrinsic
liquefiable sand. Journal of Geotechnical and Geoenvironmental and stress-induced anisotropy. Géotechnique, 6(9):754-770.
Engineering, ASCE, 141(6), 04015015, 10.1061/(ASCE)GT.1943- Tasiopoulou, P., Chaloulos, Y., Gerolymos, N., Giannakou, A., and
5606.0001298. Chacko, J. 2019. 3D and 2D simulations of liquefaction-induced
Reyes, A., Adinata, J., and Taiebat, M. 2019. Impact of bidirectional settlements of shallow foundations using Ta-Ger model. Earthquake
seismic shearing on the volumetric response of sand deposits. Soil Geotechnical Engineering for Protection and Development of
Environment and Constructions, Silvestri and Moraci (Eds),
377
Associazione Geotecnica Italiana, Rome, Italy, ISBN 978-0-367- Yang, M., Taiebat, M., Mutabaruka, P., Radjaï, F., 2021a. Evolution of
14328-2, 5241-5248. granular materials under isochoric cyclic simple shearing. Physical
Tasiopoulou, P., Ziotopoulou, K., Humire, F., Giannakou, A., Chacko, Review E 103, 032904, 10.1103/PhysRevE.103.032904.
J., and Travasarou, T. 2020. Development and implementation of Yang, M., Taiebat, M., Mutabaruka, P., Radjaï, F., 2021b. Evolution of
semiempirical framework for modeling postliquefaction shear granular media under constant-volume multidirectional cyclic
deformation accumulation in sands. Journal of Geotechnical and shearing. Acta Geotechnica, 1–24, 10.1007/s11440-021-01239-0.
Geoenvironmental Engineering, 146(1): 04019120, Yang, M., Taiebat, M., Dafalias, Y. 2022. SANISAND-MSf: a memory
10.1061/(ASCE)GT.1943-5606.0002179. surface and semifluidized state enhanced sand plasticity model for
Thavaraj, T., Sy, A., Williams, J., and Liu, R. 2018. Challenges in the undrained cyclic shearing. Géotechnique, 10.1680/jgeot.19.P.363
selection and scaling of records for seismic analyses: A Geotechnical Yang, Z., and Elgamal, A.W. 2002. Influence of permeability on
Engineer’s perspective. Geohazards 7, Canmore, Canadian liquefaction-induced shear deformation. Journal of Engineering
Geotechnical Society. Mechanics, 128(7):720–729.
Tsaparli, V., Kontoe, S., Taborda, D.M.G. and Potts, D.M. 2020. A case Yang, Z., Elgamal, A., and Parra, E. 2003. Computational model for
study of liquefaction: demonstrating the application of an advanced cyclic mobility and associated shear deformation. Journal of
model and understanding the pitfalls of the simplified procedure. Geotechnical and Geoenvironmental Engineering, 129(12), 1119-
Géotechnique 70(6): 538–558, 10.1680/jgeot.18.P.263. 1127.
Tsiaousi, D., Ugalde, J., and Travasarou, T. 2019. LEAP-UCD-2017 Yang, Z.X. and Wu, Y. 2016. Critical state for anisotropic granular
simulation team Fugro. In B. Kutter et al. (Eds.), Model tests and materials: A discrete element perspective. International Journal of
numerical simulations of liquefaction and lateral spreading: LEAP- Geomechanics, 17(2): 04016054, 10.1061/(ASCE)GM.1943-
UCD-2017. New York: Springer. 5622.0000720.
Tsuboi, S., Ohsumi, T., Uzuoka, R., and Sento, N. 2015. Applicability of Yang, Z.X., Xu, T.T., and Chen, Y.N. 2018. Unified modeling of the
effective stress analysis for prediction of deformation during strong influence of consolidation conditions on monotonic soil response
motion with long duration. Computer Methods and Recent Advances considering fabric evolution. Journal Engineering Mechanics,
in Geomechanics, Oka, Murakami, Uzuoka, and Kimoto (eds), 144(8): 04018073.
Taylor and Francis Group, 683-686, ISBN 978-1-138-00148-0. Ye, B., Ye, G., and Zhang, F. 2012. Numerical modeling of changes in
USSD. 2022. Analysis of seismic deformations of embankment dams. anisotropy during liquefaction using a generalized constitutive
Prepared by Committee on Earthquakes, United States Society on model. Computers and Geotechnics, 42: 62–72.
Dams, Denver, CO (in press). Yoshida, N., and Finn, W.D.L. 2000. Simulation of liquefaction beneath
Uzuoka, R., Cubrinovski, M., Sugita, H., Sato, M., Tokimatsu, K., Sento, an impermeable surface layer. Soil Dynamics and Earthquake
N., Kazama, M., Zhang, F., Yashima, A., and Oka, F. 2008. Engineering, 19:333–338.
Prediction of pile response to lateral spreading by 3-D soil–water Yoshimi, Y., Tokimatsu, K., and Ohara, J. 1994. In situ liquefaction
coupled dynamic analysis: Shaking in the direction perpendicular to resistance of clean sands over a wide density range, Géotechnique,
ground flow. Soil Dynamics and Earthquake Engineering 28: 436– 44(3), 479–94.
452, 10.1016/j.soildyn.2007.08.007. Yoshimine, M., Nishizaki, H., Amano, K., Hosono, Y. 2006. Flow
Vaid, Y.P., Eliadorani, A.A. 1998. Instability and liquefaction of granular deformation of liquefied sand under constant shear load and its
soils under undrained and partially drained states. Canadian application to analysis of flow slide of infinite slope. Soil Dynamics
Geotechnical Journal, 35:1053–62. and Earthquake Engineering, 26:253–64.
Vaid, Y.P., Eliadorani, A.A. 2000. Undrained and drained (?) stress– Zalachoris, G., Zekkos, D., Athanasopoulos-Zekkos, A., and Gerolymos,
strain response. Canadian Geotechnical Journal, 37:1126–30. N. 2021. The role of liquefaction on the seismic response of quay
Vargas, R.R., Tang, Z., Ueda, K., and Uzuoka, R. 2022. Validation of walls during the 2014 Cephalonia, Greece, Earthquakes. Journal of
numerical predictions for liquefaction phenomenon – Lateral Geotechnical and Geoenvironmental Engineering, 147(12):
spreading in clean sands. Soils and Foundations 62: 101101, 04021137, 10.1061/(ASCE)GT.1943-5606.0002662.
10.1016/j.sandf.2021.101101. Zeghal, M. Goswami, N., Kutter, B.L., Manzari, M.T., Abdoun, T., et al.
Vilhar, G., Laera, A., Foria, F., Gupta, A., and Brinkgreve, R.B.J. 2018. 2018. Stress-strain response of the LEAP-2015 centrifuge tests and
Implementation, validation, and application of PM4Sand model in numerical predictions. Soil Dynamics and Earthquake Engineering,
PLAXIS. Proc., Geotechnical Earthquake Engineering and Soil 113: 804–818.
Dynamics V, Geotechnical Special Publication 292, 200-211. Zeghal, M., Kutter, B.L., Manzari, M., Tobita, T., and Ueda, K. 2021.
Vytiniotis, A., and A. J. Whittle. 2017. Analysis of PV drains for The Liquefaction Experiments and Analysis Projects: LEAP 2017 &
mitigation of seismically induced ground deformations in sand 2019. Special Issue, Soil Dynamics and Earthquake Engineering.
slopes. Journal of Geotechnical and Geoenvironmental Engineering, Zhang, A., Dafalias, Y.F., and Jiang, M. 2021. A bounding surface
143(9): 04017049, 10.1061/(ASCE)GT.1943-5606.0001722. plasticity model for cemented sand under monotonic and cyclic
Wang, R., Zhang, J.M., Wang, G. 2014. A unified plasticity model for loading. Géotechnique, 10.1680/jgeot.20.P.275.
large post-liquefaction shear deformation of sand. Computers and Zhang, B. and Muraleetharan, K.K. 2018. Liquefaction of level ground
Geotechnics, 59:54–66. unsaturated sand deposits using a validated fully coupled analysis
Wang, R., Fu, P., Zhang, J.-M., and Dafalias, Y.F. 2019. Deformation of procedure. International Journal of Geomechanics, 04018118,
granular material under continuous rotation of stress principal axes. 10.1061/(ASCE)GM.1943-5622.0001230.
International Journal of Geomechanics, 19(4): 04019017, Zhang, F., Ye, B., Noda, T., Nakano, M., and Nakai, K. 2007.
10.1061/(ASCE)GM.1943-5622.0001383. Explanation of cyclic mobility of soils: approach by stress-induced
Wang, X., Luo, F., Su, Z., and Ye. A. 2017. Efficient finite-element anisotropy. Soils and Foundations, 47(4):635–48.
model for seismic response estimation of piles and soils in liquefied Zienkiewicz, O.C., and Shiomi, T. 1984. Dynamic behavior of saturated
and laterally spreading ground considering shear localization. porous media; The generalized Biot formulation and its numerical
International Journal of Geomechanics, 17(6): 06016039. solution. International Journal for Numerical and Analytical
Wang, Z., Ma, J., Gao, H., Stuedlein, A.W., He, J., and Wang, B. 2019. Methods in Geomechanics, 8: 71-96.
Unified thixotropic fluid model for soil liquefaction. Géotechnique, Zienkiewicz, O.C., Chang, C.T., and Bettess, P. 1980. Drained,
10.1680/jgeot.17.P.300. undrained, consolidating and dynamic behavior assumptions in soils.
Wang, Z.-L., Dafalias, Y.F., and Shen, C.-K. 1990. Bounding surface Géotechnique, 30(4): 385-395.
hypoplasticity model for sand. Journal Engineering Mechanics, Ziotopoulou, K., and Boulanger, R. W. 2012. Constitutive modeling of
10.1061/(ASCE)0733-9399(1990)116:5(983), 983–1001. duration and overburden effects in liquefaction evaluations. 2nd
Wolffersdorff, A. 1996. A hypoplastic relation for granular materials International Conference on Performance-Based Design in
with a predefined limit state surface. Mech Cohes Frict Mater Earthquake Geotechnical Engineering, ISSMGE, Taormina, Italy,
1(3):251–271. May 28-30, paper no. 03.10, 467-482.
Yang, M., Barrero, A. R., and Taiebat, M. 2019. Application of a Ziotopoulou, K., and Boulanger, R.W. 2013a. Calibration and
SANISAND model for numerical simulations of the LEAP 2017 implementation of a sand plasticity plane-strain model for
experiments. In B. Kutter et al. (Eds.), Model tests and numerical earthquake engineering applications. Soil Dynamics and Earthquake
simulations of liquefaction and lateral spreading: LEAP-UCD-2017. Engineering, 53, 268-280, 10.1016/j.soildyn.2013.07.009.
New York: Springer. Ziotopoulou, K., and Boulanger, R.W. 2013b. Numerical modeling
issues in predicting post-liquefaction reconsolidation strains and
378
settlements. Proc., 10th Int. Conf. on Urban Earthquake Engineering, Ziotopoulou, K., Maharjan, M., Boulanger, R.W., Beaty, M.H.,
Tokyo Institute of Technology, Tokyo, 469–475. Armstrong, R.J., and Takahashi, A. 2014. Constitutive modeling of
Ziotopoulou, K., and Boulanger, R.W. 2016. Plasticity modeling of liquefaction effects in sloping ground. Tenth U.S. National
liquefaction effects under sloping ground and irregular cyclic Conference on Earthquake Engineering, Frontiers of Earthquake
loading conditions. Soil Dynamics and Earthquake Engineering, 84, Engineering, July 21-25, Anchorage Alaska.
269-283, 10.1016/j.soildyn.2016.02.013.
379
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
381
p q
p q
p p q q
p
K
p
K
lK nK
p p
K rK K K K
p p
rK lK
K
p
nK
p
p
p p
p p
p p
lK
382
q
lK n
Sc S p lK
lK K
n
Sc S p lK
K
Sc p p p q p rK K
K n
q q
p p l q q
lK
lK
p p
p rK K
lK
lK p
lK p
lK nK p
383
p
q p
384
p e e
p e
p
e p p
F
e F F
p
q
q p l
lK lK
q q p
q q q
q
p q l
r lK lK
q
p
p
h
r
h
r
r
r
q
r
q
r
h
r r
h
385
p
p
k k
k
k k
386
387
p
388
Site characterization
r
h h
q q
389
p
p
p
k k
kS
p
q
rK
lK
k
p
K rK lK G h
r r q q S c
390
Time histories of delayed flow failure
391
392
393
Delayed flow failure associated with void redistribution
394
r
h h
r
q q
tress-strain path associated with delayed flow failure
395
Deformation mode of surface crust induced by delayed
flow failure
396
397
398
Discussion of back-calculated residual strength from case
histories
q q
399
400
Developments in Earthquake Geotechnics
Science Journal of
Geotechnical Engineering
Canadian Geotechnical Journal
Canadian Geotechnical Journal
Geotechnical
Earthquake Engineering and Soil Dynamics V
Developments in Earthquake Geotechnics
Bulletin of the Seismological
International Journal for Numerical and Analytical Methods Society of America
in Geomechanics Proc 11th International
Conference on Soil Mechanics and Foundation Engineering
Soils and
Foundations
International Journal for Numerical and Analytical
Methods in Geomechanics
Soils and Foundations
Soils and
Foundations
Journal of
Geotechnical and Geoenvironmental Engineering, ASCE
Soils and Foundations
Soils and Foundations
Journal of Geotechnical and
Geoenvironmental Engineering, ASCE
Soils and Foundations
Canadian Geotechnical Journal
Soils and Foundations
VII
International Conference on Earthquake Geotechnical
Engineering, Rome
401
State of the Art 7: Biogeotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
State of the Art: MICP soil improvement and its application to liquefaction hazard
mitigation
Etat de l’art : Amélioration de sols par biocalcification (MICP) et applications pour la prévention
des risques de liquéfaction
Jason T. DeJong
Civil and Environmental Engineering, University of California, Davis, CA, USA, [email protected]
Michael G. Gomez
Civil and Environmental Engineering, University of Washington, Seattle, WA, USA, [email protected]
Alexandra C.M. San Pablo
Civil and Environmental Engineering, University of California, Davis, CA, USA
Charles M.R. Graddy & Douglas C. Nelson
Microbiology and Molecular Genetics, University of California, Davis, CA, USA
Minyong Lee
Civil and Environmental Engineering, University of Washington, Seattle, WA, USA
Katerina Ziotopoulou & Maya El Kortbawi
Civil and Environmental Engineering, University of California, Davis, CA, USA
Brina Montoya
Civil, Construction, and Environmental Engineering, North Carolina State University, Raleigh, NC, USA
Tae-Hyuk Kwon
Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea
ABSTRACT: The field of biogeotechnics has emerged from the realization that processes intrinsic to natural systems can provide
new approaches and inspiration through which the efficiency, sustainability, and functionality of geotechnical systems can be
improved. Of these processes, microbially induced calcite precipitation (MICP) has advanced the most rapidly with the use of
ureolytic microbial activity providing an opportunity to control the precipitation of calcium carbonate minerals throughout a soil
matrix, thereby significantly improving soil engineering behaviors. The process affords increases in soil stiffness, strength, and
dilatancy, with utility across a breadth of geotechnical and geoenvironmental applications, including mitigation of earthquake-
induced soil liquefaction. This state of the art paper first covers: (1) enabling scientific processes, (2) treatment methods, and (3)
monitoring techniques, which are broadly useful for different engineering applications. The second part focuses on how MICP can:
(1) improve engineering behaviors at the element scale, (2) be modeled at the particle- and continuum-scales, (3) be applied at the
field-scale, and (4) improve the resistance to liquefaction triggering and reduce the consequences when it does occur.
RÉSUMÉ : La biogéotechnique a émergé de la prise de conscience que des processus intrinsèques du vivant peuvent apporter de
nouvelles approches et inspirations grâce auxquelles l'efficacité, la durabilité et la fonctionnalité de procédés géotechniques peuvent être
améliorées. Parmi eux, la précipitation de calcite induite par voie biologique (MICP) basée sur la voie uréolytique a progressé très vite
pour contrôler la précipitation de carbonate de calcium dans une matrice de sol et améliorer ainsi ses propriétés. Le procédé permet
d'augmenter la rigidité, la résistance et la dilatance du sol, pour une grande diversité d'applications géotechniques et
géoenvironnementales, dont la réduction des risques de liquéfaction de sol sous séisme. Cet état de l’art couvre dans un premier temps
(1) les principes fondateurs scientifiques, (2) les méthodes de traitement et (3) les techniques de contrôle, utilisés largement pour
différentes applications de traitement de sol. La seconde partie se concentre sur la façon dont la biocalcification peut (1) améliorer les
comportements du sol à l'échelle élémentaire, (2) être modélisée à l'échelle des particules et du continuum, (3) être appliquée sur site, et
(4) améliorer le seuil de déclenchement de la liquéfaction et en réduire les conséquences le cas échéant.
KEYWORDS: biogeotechnics, microbially induced calcite precipitation, ground improvement, biocementation
405
1 INTRODUCTION improvements are primarily realized through the binding of
particle contacts, densification of soil matrixes due to addition of
Biogeotechnics, specifically biomediated and bioinspired precipitated solid masses, and increase in particle surface
geotechnical engineering, has matured rapidly over the past two roughness and angularity. A range of different technologies can
decades, becoming one of the fastest growing subdisciplines achieve this end state condition, of which variations in the
within geotechnical engineering. Academics and practitioners process may include different biogeochemical pathways (i.e.,
exploring this field are motivated in part by the notion that urea hydrolysis, denitrification, sulfate reduction, iron reduction,
natural processes have been optimized to some degree to address carbonic anhydrase), varying precipitated minerals (calcium
biological constraints including limited material and energy carbonate, dolomite, ankerite and other amorphous and mixed
resources and can thereby yield new insights regarding how mineral phases), ranges in treatment times, and use of augmented
geotechnical processes may be accomplished in a more efficient or stimulated whole cells versus free enzymes (Ciurli et al.
and sustainable manner. The natural world provides endless 1991, Ferris et al. 1997, Benini et al. 1999, Warthmann et al.
opportunities to identify alternative resources and inspire 2000, Bachmeier et al. 2002, Roden et al. 2002, Karatas et al.
innovation in the geotechnical engineering community. This is a 2008, DeJong et al. 2010, van Paassen et al. 2010b, Chu et al.
contrast to current practices for which geotechnologies have been 2011, Hamdan et al. 2011, Gat et al. 2011, Mortensen et al. 2011,
almost exclusively inspired by human ingenuity with practical Weaver et al. 2011, Burbank et al. 2011, 2012, Gomez et al.
constructability constraints, capital cost considerations, and 2014, 2016, Dhami et al. 2017, Graddy et al. 2018, Kavazanjian
public safety being of highest priority. As new technologies and Hamdan 2015, Khodadi Tirkolaei et al. 2020, Yu et al. 2021).
developed, energy and material resource limitations as well as Of these various approaches, microbially induced calcite
environmental quality were largely not considered, and broader precipitation (MICP) via microbial ureolysis is the most
issues related to climate change and sustainability remained technologically mature biomediated soil improvement process
outside the profession. Consequently, our profession seeks new and will be the focus of this paper.
opportunities to realize a more sustainable, less impactful, Bio-gas generation processes yield gas bubbles in soil pore
geotechnical engineering practice in the future with spaces, typically N2 or CO2 gases, which do not alter particle
biogeotechnical solutions providing one path forward to realize structures, but rather increase pore fluid compressibility. The
this vision. resulting effect is best realized on loose of critical specimens
with increases in monotonic strengths and cyclic resistances to
1.1 Biomediated & bioinspired geotechnics liquefaction under undrained conditions and decreases in soil
The envisioned opportunities for biogeotechnics are broad, as permeabilities. Perhaps the most prominent of such methods is
discussed by several researchers (DeJong et al. 2013, Pacheco- the generation of nitrogen gases and soil desaturation via
Torgal & Labrincha 2013, Mujah et al. 2016, Baveye et al. 2018, denitrification activity commonly referred to as microbial
El Mountassir et al. 2018, Terzis & Laloui 2019, Almajed et al. desaturation and precipitation or MIDP (van Paassen et al. 2010a,
2021, Lai et al. 2021, Martinez et al. 2021, Assadi-Langroudi et He & Chu 2014, O’Donnell et al. 2017b, van Paassen et al. 2018,
al. 2022), with engineering applications spanning across the Pham et al. 2018, Hall et al. 2018). While bio-gas is primarily
geotechnical, environmental, water, transportation, and energy generated, alkalinity generated from this process can also enable
sequestration sectors. Geotechnical specific applications carbonate precipitation; however, mineral production rates are
identified include, but are not limited to, liquefaction mitigation, orders of magnitude slower than ureolysis.
building settlement reduction, dam and levee safety, tunneling, Bio-film production involves the formation of an
foundations and anchorage systems, bluff and slope stabilization, agglomeration of microbial cells, extracellular polymeric
impermeable and reactive barriers, drilling technologies, and site substances (EPS), and organic and inorganic materials on soil
characterization and monitoring tools (DeJong et al. 2013, particle surfaces in aqueous environments (Cunningham et al.
DeJong 2020, Martinez et al. 2021). 1991, Proto et al. 2016). The accumulation of biofilms can be
accomplished through the introduction of growth factors which
1.1.1 Biomediated geotechnologies permit the formation of biofilms in the presence of augmented
A generalized workflow for biomediated geotechnologies is and/or indigenous microbial communities. Hydraulic
shown in Figure 1. In these technologies, geochemical reactions conductivity reductions exceeding 100-fold can be achieved in
in the soil are most frequently controlled by biological processes saturated sands due to the coating and bridging of adjacent
which determine the timing, rate, and distribution of some particles, which can be maintained through continued nutrient
realized change oftentimes consisting of modification of soil injections. The permanence of biofilm structures is dependent on
pore fluids, generation of inorganic minerals, or production of both continued nutrient availability and hydraulic conditions,
organic polymers. Such changes can be tracked using various however, and affords the unique advantage of reversibility when
process monitoring tools including compression and shear wave reductions are no longer needed (Phradichith et al. 2020, Zamani
propagation, electrical conductivity, and measurements of et al. 2022).
solution species, with changes to mechanical properties
oftentimes varying several orders of magnitude. Such processes 1.1.2 Bioinspired geotechnologies
can be further refined to achieve specific performance objectives The workflow process for bioinspired geotechnologies can be
(i.e., predictable increase in improvement per treatment) through similarly described. In the bioinspired approach, inspiration from
modification of implementation details. Finally, upscaling of the various biological systems can be leveraged in conjunction with
treatment methodology to field-scale must address treatment current understandings of soil behavior, soil-structure
uniformity, material permanence, and byproduct management, interaction, and system design to yield new technologies
with identification of the means by which treatment success and (Martinez et al. 2021). The development of bioinspired
quality can be effectively verified. engineered solutions requires appropriately defining and
Three examples of biomediated technologies on which contextualizing both the engineering problem to be solved and
substantial research has been performed include bio- abstracting the forms, behaviors, and principles of the biological
cementation, bio-gas generation, and bio-film formation in sands analog(s) used for inspiration (DeJong et al. 2017). Hypothesis
(DeJong 2020). Bio-cementation coats soil particles and testing and refinement is performed iteratively throughout this
produces stiff, brittle bonds at particle-particle contacts, which
increases initial shear stiffness, peak shear strength, and
decreases hydraulic conductivity and compressibility. Such
406
Biological Process Upscale
Mediation Monitoring to Field
Chemical Mechanical Field
Reaction Properties Applications
f
Soil
Figure 1. Generalized workflow for biomediated soil improvement geotechnologies.
process with upscaling challenges including physical constraints Nation’s 17 sustainability goals (United Nations 2019) indicates
encountered at both stress and length scales. Although not fully that at least 10 of these goals can be directly impacted by
refined, some bioinspired solutions have already been found to geotechnical practices, with the balance oftentimes indirectly
afford performances that are orders of magnitude better than affected (DeJong 2020). In recognition of these broader potential
existing technologies (e.g., Burrall et al. 2020). impacts, it is increasingly critical to realize the benefits that a
Three examples of bioinspired technologies for which given geotechnical technology may afford with respect to
substantial research has been performed include snake-skin sustainability and the role that continued development may have
inspired anisotropic surfaces, self-burrowing probes, and root- towards furthering such benefits. Biogeotechnical technologies
inspired anchorage systems (DeJong 2020, Martinez et al. 2021). are particularly well suited for such developments with their
Snake-skin inspired anisotropic surfaces, inspired by the ventral potential to improve the sustainability of geotechnical practice.
scales of snakes, can mobilize directionally-dependent interface To this end, life cycle sustainability assessments (LCSA) can
friction (Martinez et al. 2019). Such surfaces have potential be used to quantitatively evaluate the impacts of both existing
applications for piles, soil nails, geomembranes, and other and future technologies. LCSA integrates the three-pillars of
interfaces (O’Hara & Martinez 2020, Gray & Lissmann 1950, sustainably, namely, environmental, economic, and social
Marvi et al. 2013, Baum et al. 2014) and have shown promise for impacts, through environmental life cycle assessment (LCA), life
upscaling in recent centrifuge testing (Martinez & O’Hara 2021). cycle costing assessment (LCCA), and social life cycle
Self-penetrating probes, inspired by the burrowing and assessment (S-LCA) (Guinée 2016). LCAs provide quantitative,
movement strategies of marine bivalves (i.e., clams), earth benchmarked evaluations of technologies and their impact on the
worms, and other organisms, may hold significant promise environment over an entire life cycle. Raymond et al. (2021)
towards improving the energy efficiency and exploration recently developed further criteria specifically for geotechnical
capabilities of current subsurface techniques. Such probes have evaluations. The use of LCA in the geotechnical profession has
been investigated through cavity expansion analysis, DEM been limited (Kendall et al. 2018), however, efforts are growing
simulations, and laboratory experiments (Cortes & John 2018, and now include evaluations of existing technologies against
Huang & Tao 2018, 2020, Khosravi et al. 2018, Chen et al. 2020, which emerging biogeotechnical technologies can be compared,
Martinez et al. 2020, Borela et al. 2021, Vego et al. 2021). Robot including deep foundations (Misra 2010, Giri & Reddy 2014, Lee
prototypes have also been used to study such mechanisms and & Basu 2018), various soil improvement methods (Spaulding et
have been inspired by razor clam locomotion (Tao et al. 2019, al. 2008, Pinske 2011, Shillaber et al. 2016, Raymond et al.
Huang et al. 2020) and marine peristaltic body motion (Ortiz et 2021), and site characterization methods (Purdy et al. 2021).
al. 2019), among other processes. More recently, LCA evaluations have been performed for
Novel foundation and anchorage systems, inspired by the developing biogeotechnical technologies, including MICP
architecture of tree root systems, have the potential to provide (Raymond et al. 2022) and EICP (Martin et al. 2020, Raymond
significant increases in stiffness and capacity compared to et al. 2020b). Such individual evaluations can be effective in
current foundation systems when normalizing achieved identifying the primary components contributing to sustainability
performance by material requirements. Mechanical performance impacts and therefore prioritizing changes that can be most
of root and root-inspired systems has been investigated through effective in improving their sustainability. Further, comparative
field pull-out studies of trees (Burrall et al. 2020), laboratory LCSAs can be used to provide early and updated assessments of
testing of root analogs (Dyson and Rognon 2014, Mickovski et how beneficial a new technology may be relative to existing
al. 2007, Schwarz et al. 2011), centrifuge testing of root systems practice. Raymond et al. (2020b) outlined an approach by which
and analogs (Liang et al. 2017, Zhang et al. 2022b), numerical a technology can be re-evaluated and design considerations re-
modeling (Dupuy et al. 2007, Vego et al. 2021, Yang et al. 2018), prioritized during this process in order to guide future field
and analytical methods (Blackwell et al. 1990, Cohen et al. 2011, implementation.
Mallett 2019). The bioinspired process provides new The benefits and challenges of performing individual and
opportunities to realize sustainability and performance objectives comparative LCSA evaluations frequently throughout the
in a manner that is very different than biomediated technologies development of a technology can be captured by considering the
but equally important. “Collingridge dilemma” and the “Gartner hype cycle”
paradigms. The Collingridge dilemma (Collingridge 1980),
1.2 Sustainability through innovation shown in Figure 2a, contrasts the ability for reducing
As biogeotechnical technologies mature there is an opportunity environmental impacts during a technologies development with
the increasing cost of realizing those changes as the technology
to maximize sustainability gains relative to current practices.
matures. Early in development, when a technology is poorly
This opportunity is abundantly clear for biomediated processes,
constrained, there are significant opportunities for change that
which aim to improve upon existing soil improvement
technologies, which are generally highly material and energy
intensive (Jefferis 2008, Raymond 2020). A review of the United
407
a) b)
Figure 2. Conceptual schematics illustrating the (a) Collingridge dilemma and (b) Gartner hype cycle paradigms.
inflict minimal additional costs, but for which the current technology is occurring as intended and that the final improved
knowledge base may be insufficient to rationally assess the condition can meet project requirements.
consequences of such decisions. As a technology matures The second part, building on the previously established base
towards implementation, however, available processes and knowledge, advances specifically into how the engineering
knowledge increase, but at the expense of decreased flexibility behaviors of sandy soils can be improved by MICP in the context
with the design becoming increasingly less nimble. Thus, of behaviors most relevant to liquefaction hazard mitigation. This
without continued evaluation of a technology during its portion begins at the element scale, addressing how MICP
development, maximum sustainability benefits will almost treatment affects low-strain dynamic properties including
certainly remain unrealized. The potential benefits of frequent stiffness degradation and damping, monotonic behaviors, and
and transparent evaluation can be further considered within the undrained cyclic responses. Leveraging these experiment
Gartner hype cycle (Gartner 2018), shown in Figure 2b. As understandings, advances in the ability to model such behaviors
shown, oscillations in perceptions commonly observed during at the discrete and continuum scales are further addressed.
technology development, including inflated expectations and Finally, upscaling of the MICP treatment technology itself for
subsequent disillusionment, can be tempered through initial and field scale implementation is addressed through large upscaling
ongoing quantitative evaluations of a technology. experiments, centrifuge modeling, and a synthesis of experience
In this context, it is important to understand and examine how and guidance from planned and executed field implementations.
biogeotechnical technologies and all related design decisions Conclusions and priorities for continued maturation of
including treatment solution compositions, realized and desired biogeotechnologies, and MICP in particular, are then discussed.
improvement levels, and implementation details may require
continuous evaluation, refinement, and optimization in order to 2 MICP TREATMENT PROCESS
more fully realize the greatest sustainability gains while still
meeting financial constraints and holding societal safety 2.1 Chemical reaction network for ureolytic MICP
paramount.
Biocementation is enabled by the biological production of
1.3 Objectives and scope carbonate species in the presence of dissolved calcium salts.
There are a number of pathways through which carbonate species
The focus of this paper is microbially induced calcite can be biologically generated (Section 2.2) however the
precipitation (MICP) via microbial ureolysis, the biomediated fundamental microbial processes capable of enabling MICP must
process that has arguably achieved the greatest academic and yield carbonate species, hydroxide ions capable of elevating
industrial interest, research investment, and realized progress to surrounding solution pH, or both. Urea hydrolysis (ureolysis),
date. While this paper is specifically focused on the applicability the most well-studied pathway through which biocementation
of this process towards liquefaction mitigation, the paper is has been achieved, involves the hydrolysis of urea, a reaction
organized in two parts recognizing the broader interest in this catalyzed by the enzyme urease, which yields carbonic acid and
particular technology and the applicability of more fundamental ammonia (Equation 1). Ammonia, under pH conditions that are
insights to other applications. representative of most soil systems (pH < 9.5), will have some
The first part of this paper is designed to be relevant and fraction protonated to form ammonium, leading to a reduction in
applicable to all with interest in the MICP process, regardless of protons and thus a net increase in pH (Equation 2). The
the specific engineering application considered. This portion distribution of carbonate species is pH dependent and governed
begins by describing the more fundamental microbiological and by carbon dioxide equilibria (Equations 3 & 4); therefore, pH
biogeochemical processes by which MICP can be achieved, the increases resulting from the formation of ammonium enables
structure of resulting biocemented materials, and other more generated carbonic acid to shift increasingly towards bicarbonate
fundamental considerations including material permanence. This and carbonate (Equation 4). Due to the relatively low solubility
is followed by detailed coverage of the entire treatment process, product (Ksp) of calcite and other calcium carbonate (CaCO3)
from establishment of biological activity through either polymorphs, even when small amounts of aqueous carbonate are
stimulation of indigenous bacteria or augmentation of non-native
formed in the presence of sufficient calcium, calcium carbonate
strains, to cementation treatments, through byproduct
mineral precipitation can be favored (Equation 5). The
management strategies, and finally consideration of first-order
environmental variables. Finally, details and recommendations mineralogy and morphology of the generated mineral product
regarding methods and techniques that can enable process can be variable and has been shown to depend on many factors
monitoring and verification including those relevant to
microbiological aspects, geochemical changes, and mechanical
enhancements are further provided to ensure that the treatment
408
(Section 2.5). Although generated minerals may include calcite,
aragonite, and vaterite, among others, the authors refer to the
precipitate resulting from MICP throughout this paper as either
calcite or calcium carbonate, and do so interchangeably. This
subtlety may be not strictly correct, but is supported by
experimentation demonstrating that the predominant mineral in
almost all MICP experiments is calcite (Burdalski 2020). The net
reaction for ureolytic MICP presented in Equation 6 is simplified
to neglect pH-dependent ammonium speciation. During CaCO3
precipitation, free protons are generated as carbonate is
consumed, making precipitation an acidogenic process. Thus,
without a continuing source of alkalinity, the precipitation would
acidify and self-limit, reaching equilibrium without fully
utilizing the provided calcium. In the existing literature, it is
frequently discussed that MICP requires substantially alkaline
conditions to be successful. However, the low solubility of
CaCO3 minerals permits substantial pH latitude in order to
maintain super-saturation (Section 2.5). For example, only 3.3
pM CO32- is needed to supersaturate solutions with respect to
calcite when 1 mM Ca2+ is present. Indeed, the steady state pH
when both urea hydrolysis and carbonate precipitation are
occurring in MICP experiments is generally between 7.5 and 7.8
(Gomez et al. 2018b) with only the final pH altered by supplied Figure 3. Solution pH in time during a representative cementation
injection using solutions with 250 mM calcium show the ability to
urea (Figure 3). Thus, while the precipitation can and does occur
control final pH by varying urea concentrations (adapted from San Pablo
at more neutral pH, it does require constant hydroxide and et al. 2022a).
carbonate species generation to maintain super-saturation, both
of which are provided by ureolysis. rates for recently isolated ureolytic strains range from
approximately 80% down to 1% of the corresponding rate for the
(1) S. pasteurii type. However, even the lowest measured rates for
ureolytic isolates from biostimulated MICP are near the highest
known rates for the other metabolic processes. Ureolytic
(2) representatives from other genera such as Bacillus,
Lysinibacillus, and Alkalibacterium (Cheng & Cord-Ruwisch
2012, Kang et al. 2016, Cheng et al. 2017b, Gowthaman et al.
(3) 2019, Dikshit et al. 2020, Ezzat & Ewida 2021, Mukherjee et al.
2021) have also been used to stabilize soils via MICP.
Of the most promising alternate methods, microbially induced
desaturation and precipitation (MIDP) uses denitrification to
(4) generate both alkalinity for precipitation as well as dinitrogen gas
bubbles, which desaturate soils (DeJong and Kavazanjian 2019).
(5) Data from Figure 4 clearly suggest a much slower rate of MICP
with denitrification, which was demonstrated by soil column
experiments (van Paassen et al. 2010a) wherein achieving a
relatively low calcite content (~2% by mass) required 100 days.
(6) In comparison, this same amount of biocementation was
achieved after 4 daily cementation treatments by both
2.2 Alternate MICP pathways bioaugmented or biostimulated ureolytic MICP (Graddy et al.
2021), and alternate higher concentration treatment formulas
Urea hydrolysis by Sporosarcina pasteurii (S. pasteurii) has could be realized within 48 hours. It should be noted that field
been the most frequently employed biological process and implementations of MICP are likely to impose practical time
organism combination to drive CaCO3 biocementation. constraints on the treatment frequency and overall duration of
Nonetheless, other biological processes, namely sulfate- projects. To this end, much attention has been given to achieving
reduction, iron-reduction, and denitrification have also been the highest production rates of carbonate and alkalinity; however,
proposed (Castanier et al. 1999, DeJong et al. 2010). In Graddy it is unknown what effect, if any, a lower rate of calcite
et al. (2018), kinetic parameters for whole cell urease from 6 precipitation may have on bio-cemented soil behavior. At a
MICP isolated Sporosarcina strains and the S. pasteurii type minimum, lower production rates can present an opportunity to
strain were reported. Furthermore, alkalinity and carbonate improve the spatial uniformity and extent of treatment as
generation rates (in CO2 equivalents) were compared between discussed later in Section 3.2.3.3 & 7.2.
these ureolytic strains involved in both biostimulated and The use of alternative alkalinity generating processes,
bioaugmented MICP (Sections 2.3 & 2.4) to rates in the available however, may present further drawbacks than simply slower
literature for denitrifying, iron-reducing, and sulfate-reducing precipitation rates. For example, iron-reduction requires the use
bacteria in pure culture. Figure 4 presents the ranges of of insoluble iron (III) compounds and sulfate-reduction can
production rates for Sporosarcina strains compared to produce toxic hydrogen sulfide gases, both of which would likely
heterotrophic bacteria respiring anaerobically with alternative present field implementation challenges. A possible advantage
electron acceptors. As shown, the S. pasteurii type strain for denitrification-driven MICP is that it may not leave
produces alkalinity and carbonate species at rates that are roughly remediation-requiring waste products in the soil; however,
100-fold greater than even the most active of the other proposed comprehensive nitrogen balance studies have not yet been
biological processes, as studied in pure cultures. Furthermore, completed for applications of MIDP in natural soils. Thus, the
409
possibility exists that a portion of added nitrate may be instead toxifying its environment to achieve a competitive advantage.
respired to ammonia by native bacteria, remain unreacted, or be Additionally, it has been suggested that urea hydrolysis provides
reduced to intermediate phases other than dinitrogen gas (e.g., cellular energy to S. pasteurii adequate to fuel anoxic growth
NO2-, N2O; Rütting et al. 2011). (Cheng et al. 2017a). Because it is a hydrolysis and not an
A number of non-ureolytic bacterial strains or enrichments oxidation/reduction reaction or a fermentation, there is no
have been successfully employed in bioaugmented or literature precedent for such a reaction playing a sustained role
biostimulated MICP studies, wherein the explanations for the in energy generation by a free-living bacterium. The literature
mechanisms by which alkalinity and carbonate are generated are does, however, provide two short-term, biophysical studies that
sometimes unclear. The common feature of most of the systems demonstrated urea-driven production of ATP and a proton
is the alkalinization deriving from organic acid uptake. That is, motive force in S. pasteurii and Ureaplasma urealyticum (Smith
any time a metabolizable weak organic acid is added as a salt, et al. 1993, Jahns 1996).
e.g., sodium formate (pKa = 3.74) or sodium lactate (pKa = 3.88), During the last two decades, literature involving
bacterial utilization of that substrate as an uncharged molecule, bioaugmented MICP has grown rapidly and pertained to a wide
for subsequent respiration or incorporation, will include uptake variety of applications. The discussion of the more fundamental
of protons, thereby rendering the medium more alkaline. This factors involving augmentation, however, will be confined to
mechanism is explicitly recognized as the cause of MICP driven conditions applicable to liquefaction mitigation applications in
by aerobic formate oxidation (Ganendra et al. 2014). saturated soils at nominal depths up to 30 meters. Foremost, the
Acetoclastic methanogenesis, which is based on acetate preparation of a bulk quantity of cells by growth under well-
consumption, almost certainly resulted in the alkalinity increase aerated conditions must occur prior to augmentation. The ATCC
that drove carbonate precipitation in a consortium feed sodium recommended medium (2% yeast extract, 75 mM (NH4)2SO4,
acetate (Su & Yang 2020). Likewise, mechanisms of non- 130 mM Tris, pH 9) or a variety of other formulations found in
ureolytic MICP observed for aerobic growth of Bacillus the literature can be used for cultivation, but minimum
alkalinitrilicus and Lysinibacillus sp. strain YS11 (Nielsen et al. requirements include the presence of at least 15 mM NH4+, some
2020, Lee et al. 2017) strongly implicate growth at the expense complex source of amino acids, nucleotides, and other trace
of, respectively, lactate or acetate anions. Yet another mechanism nutrients, and an initial pH in the range of 8.0 to 9.0. Urea is not
for alkalinity generation driving MICP based on deamination of required for bulk cultivation, but if it is present, controlling the
amino acids has also been proposed (Lee et al. 2017). pH of the surrounding medium once growth has commenced is
largely unnecessary as intrinsic buffering will occur from urease-
2.3 Ureolytic augmentation generated ammonia/ammonium and carbonate species.
Bioaugmentation is the introduction of non-native Following preparation of cells in some medium, the most basic
microorganisms to soils to initiate some process of interest. augmentation process involves rinsing and resuspending cells in
Augmentation has been both the first and most commonly some medium, injecting cells into soils, initiation of strategies to
employed method for ureolytic MICP. Although other strains can distribute and retain cells, and proceeding with the subsequent
be used, S. pasteurii (strain 11859; American Type Culture biocementation process. In most studies the cell rinsing and
Collection, ATCC) is the most commonly augmented resuspension process has largely been omitted, with grown
augmented bacteria introduced into soils in a solution containing
bacterium. Despite some suggestions in the MICP literature to
urea, ammonium salts, and organic constituents (unconsumed
the contrary, bacterial ureases are typically cytoplasmic, with any
organics and waste products present in spent growth medium and
apparent extracellular activity resulting from cell lysis (Mobley
possibly defined additions of yeast extract, nutrient broth, etc.).
& Hausinger 1989). The canonical role long assigned to urease, Variations on this treatment scheme may also include, as
which is active in many types of bacteria, is to alleviate nitrogen examples, providing a fixation solution after augmenting soils,
starvation by providing assimilable ammonia. However, given removing generated carbonate species after augmentation using
the high ammonium environment that emerges during ureolytic a flush treatment, repeated augmentations during the cementation
MICP treatments, the constitutive and persistent urease process, and alternating directions of pumping (van Paassen
expression suggests it may also play an ecological role in 2009, Al Qabany & Soga 2013, Gomez et al. 2016, Cheng et al.
2017b, El Mountassir et al. 2018, Zamani et al. 2019b).
Perhaps most fundamental to the augmentation process are
the mechanisms controlling successful distribution of injected
cells into initially untreated soils. Bacterial cells introduced into
porous media as a flowing suspension may adsorb to individual
soil particles and/or be subject to colloidal filtration, dependent
on soil particle size, the result typically being higher augmented
bacterial cell densities adjacent to injection sources, which result
in higher reactivity and ultimately greater localization of
precipitation at these locations provided that the flow conditions
are not exceedingly fast. Whether this phenomenon can be solely
attributed to non-uniform augmentation or abiotic precipitation
events (resulting from the mixing of high calcium and high
carbonate solutions near the injection source) is unclear.
Uniformity of cell densities, however, is one critical factor
influencing whether or not more uniform cementation can be
achieved. In some of the earliest MICP studies, augmented cells
supplied to columns filled with sand (Kantzas et al. 1992,
Figure 4. Rates of whole-cell production of CO2 and alkalinity (expressed Mahawish et al. 2019b) produced highly localized porosity
as OH-), i.e. the drivers of carbonate production essential for MICP for reductions, indicative of non-uniform augmented cell densities.
denitrification, sulfate reduction, iron reduction and urea hydrolysis. This non-uniformity was attributed to issues with cell
Data and equations used to calculate production rates for CO2 and OH- distribution as governed by colloidal filtration and cell retention
include the contribution of acetate uptake on alkalinity production as governed by electrostatic repulsion.
(adapted from Graddy et al. (2021).
410
The impact of colloidal filtration is well exemplified by three such cells may be attached to soil particle surfaces and or
studies wherein bacteria were augmented in different sands. In embedded in generated CaCO3 minerals. Recent experiments by
San Pablo et al. (2020), a poorly graded, alluvial sand (D10 = 0.23 the authors involving 1.7 m diameter tanks included one
mm, D60 = 1.5 mm, 1% fines) contained in a 3.7 m long column specimen that was initially bioaugmented and one that was
was augmented over a 3 hour period with 1.5 pore volumes (PV) biostimulated; however, daily cementation treatments included
of a dilute S. pasteurii culture (density of 9.4 x 107 cells/mL). organic nutrients (Gomez et al. 2016). In this study, the added S.
After four daily 1.5 PV injections of a non-nutritive cementation pasteurii type strain was shown to be outcompeted by enriched
solution, substantial cementation ( Vs ~800 m/s) was achieved native ureolytic strains, as evaluated by cultured strains, high
at a distance of 0.3 m with no improvement detected at more throughput sequencing, and sequencing of full length 16S clones
distal locations. A second augmentation injection consisting of 9 (Graddy et al. 2018, 2021). Thus, in the presence of nutrients, the
PV was applied followed by 5 additional cementation injections final composition of each initially augmented bacterial
with, yet again, more minimal improvement observed at further community converged largely to that of the corresponding
distances. In contrast, a similar stimulated 3.7 m column,
stimulated one. Furthermore, the augmented type strain, which
achieved significant improvement at distances exceeding 3.4 m.
was dominant on day 5, following inoculation on day 4 as judged
This study comes closest to highlighting the primary challenges
by isolates (Figure 5a), was barely detected by day 8 by DNA
associated with interactions between augmented bacteria and
surrounding soils, namely those related to electrostatic sequencing (Figure 5c), and not detected by day 12 at
interactions and colloidal filtration, as the potential for experiment’s end in either method. These results suggest that, at
unintended biostimulation (discussed later in Section 2.3.2) was a minimum, augmented microbial populations may be transient,
kept to a minimum by two procedural choices. First, organic although it is less clear if similar trends may be observed for
constituents that were residual in the bulk culture were diluted augmented soils wherein nutrients are not supplied.
approximately 20-fold before augmentation and injected cells
were resuspended and injected in a 150 mM NaCl solution. 2.3.2 Unintended biostimulation
Second, the organics were eluted within 24 hours by the first As previously discussed, a clear distinction should be made
cementation treatment. In contrast, two other studies (Whiffin et between augmented experiments wherein nutrients are not
al. 2007, van Paassen et al. 2010b) which showed significant supplied with augmented bacteria (deemed “pure” or “non-
improvements at more than 10 times this distance (4.5 and 7.5 m, nutritive” augmentation) and those experiments involving
respectively) also used a finer poorly-graded sand (D10 = 0.110 introduction of cells with sufficient nutrients wherein the
mm, D90 = 0.275 mm, no fines) and augmentation. However, in potential for stimulation of indigenous bacteria should be
their treatments, unamended stationary phase cultures were acknowledged (deemed “nutritive” augmentation). The use of
injected and CaCl2 fixation solutions were applied 18 to 24 hours nutritive augmentation dominates the literature regarding studies
later. While this offers a possible explanation for the observed of bioaugmented MICP and may result from the often-cited
differences, another important difference in the protocols of these assumption that bacteria grown to a high density for use in MICP
two studies from that of San Pablo et al. (2020) is that enter a stationary phase because all readily available nutrients
augmentation was performed using dense undiluted cultures of have been consumed from the medium (Whiffin et al. 2007,
S. pasteurii with correspondingly higher concentrations of
residual nutrients. The residual nutrients in combination with
slower pumping rates and longer residence times may have
provided between 5 and 15 days for unintended biostimulation to
have occurred (Section 2.3.3).
The initial retention of bacteria following augmentation is
thought to be substantially controlled by electrostatic forces (El
Mountassir et al. 2018). The enhancement of bacterial cell
attachment to soil particles requires a reduction of electrostatic
repulsion and can be accomplished by providing a solution of
sufficient ionic strength, such as synthetic (artificial)
groundwater, or by providing urea in augmentation solutions in
which augmented cells can generate their own increases in ionic
strength. Harkes et al. (2010) provides an early example
demonstrating the utility of a fixation solution. In this study, a
dense cell culture was applied to a saturated fine sand with higher
ionic strength fixation fluids provided to retain cells. Their
findings suggested that a fixation solution consisting of 50 mM
CaCl2 achieved greater fixation than 150 mM NaCl; however, no
cell retention was reported when fixed with deionized water.
Variations on similar strategies have been employed in many
other successful bioaugmented studies. Furthermore, other more
elaborate schemes have been used including applying alternating
< 1 PV injections of bacterial suspensions and fixation solutions
(Cheng & Cord-Ruwisch 2012, 2014).
2.3.1 Realized bacterial populations over time
Studies examining changes in soil bacterial populations in
natural soils over time following non-nutritive bioaugmented
MICP are currently non-existent. If and when such studies are Figure 5. Microbial community estimates based on the (a,b) relative
completed, representative bacterial sampling will remain abundance of isolates and (c, d)16S rRNA fragments in (a,c) augmented
challenging as recent observations suggest that ~90% of the and (b, d) stimulated soils. Dotted line indicates the augmentation
active bacteria during MICP are not readily sampled as treatment for (a) and (c), and the shaded regions in (b) and (d) represent
suspended cells (Graddy et al. 2021). This is presumably because stimulation treatments. (adapted from Graddy et al. 2018, 2021).
411
Harkes et al. 2010, Mahawish et al. 2019a). Based on this of ~108 per gram of soil. Whether or not stimulation occurred in
assumption, augmentation cultures are commonly introduced in past augmented experiments may not be fully known;
what is termed “spent” growth medium. If not fully consumed, nonetheless, it is clear that there are important nuances between
however, dilute residual organics could stimulate indigenous soil employed augmentation procedures meriting further
bacteria. In this spent growth medium it is possible that the consideration and investigation with the boundary between
growth of bacteria in cultures to be augmented may be limited stimulation and augmentation being less clear than perhaps
not by nutrient depletion, but rather by the accumulation of toxic originally thought.
metabolites (i.e., waste products) and oxygen transfer
limitations. This nutrient exhaustion assumption can be 2.4 Ureolytic stimulation
examined by considering the experimental details of Whiffin et Biostimulation is a second approach by which ureolytic cells can
al. (2007). In this study, a medium containing 20 g/L of yeast be generated to complete the MICP process. Rather than growth
extract, yielded an OD600 of 1.6, which translates into about 1 x and injection of non-native bacteria, stimulation instead involves
109 cells/mL based on the authors’ best estimates (Figure 32). To the injection of a growth medium that imposes selective
estimate the fractional use of yeast extract in producing bacterial environmental conditions intended to enrich indigenous ureolytic
cells, the following well accepted assumptions can be made, bacteria in natural soils. Successful stimulation of native
which are essential and accurate to approximately 10%: (1) the ureolytic bacteria requires careful consideration of the needs of
elemental composition of yeast extract and bacterial dry matter the target microorganism and how the growth of competing
are approximately 50% carbon by mass, (2) a single S. pasteurii microorganisms can be minimized. Stimulation approaches have
cell contains approximately 3 x 10-13 g of organic carbon, and (3) largely involved the application of solutions containing organic
when growing aerobically, a bacterium utilizes organic carbon carbon, trace nutrients, and urea. Progressive modifications to
roughly equally between catabolism (i.e., energy generation) and these formulations have been made to increasingly select for
anabolism (i.e., generation of biomass). From these assumptions, alkali-tolerant and ammonium-tolerant soil microorganisms of
20 g/L yeast extract would be expected to yield S. pasteurii at which a larger fraction are thought to be ureolytic. A multitude
about 1.7 x 1010 cells/mL, 17 times the observed final cell density of treatment solution factors may impact enriched microbial
in Whiffin et al. (2007). This suggests that ~95% of the growth communities with our current understanding of individual factors
potential provided by yeast extract might still present in “spent” described later in Section 3.2.
medium; however, to the authors’ knowledge this has not yet Initial attempts at ureolytic stimulation largely involved
been tested. To put such concentrations into the context of what attempting to stimulate ureolytic activity simply by adding
is needed for stimulation, it should be noted that a very low yeast substantial urea to a general growth medium. By supplementing
extract concentration (0.04 g/L) enabled successful stimulation solutions with urea, such efforts presumably hypothesized that
over meter-scale distances in San Pablo et al. (2020). Thus, even ureolytic microorganisms may gain some benefit from urea
if “spent” medium and bacteria were diluted 20-fold prior to degradation such as the alleviation of nitrogen limitations.
introduction into sand columns, there might still exist ~25 times Recognizing that stimulation media typically include
more yeast extract than was needed for enrichment in these nutritionally complex ingredients like nutrient broth or yeast
stimulated experiments. extract that are rich in amino nitrogen and are supplemented with
In light of these observations, the authors suggest that earlier ammonium (both of which satisfy the nitrogen requirements of
augmented studies could be divided into three groups, dependent growing bacteria), it is unlikely that such a benefit may be related
on the particular augmentation protocols employed. Although to increased nitrogen availability. More recent experiments,
exceedingly rare, the first category deemed “pure” or “non- discussed below, suggest that the bacteria selected for are closely
nutritive” augmentation, involves the use of rinsed cell pellets of related to S. pasteurii, and may share its capacity for urea
ureolytic bacteria that are introduced into a soil in strictly hydrolysis linked ATP synthesis (Jahns 1996). Although
inorganic medium with later cementation solutions lacking any stimulated MICP was originally investigated by Ferris et al.
added nutrients (e.g., Gollapudi et al. 1995, Ferris et al. 1997, (1997), a multitude of other studies have contributed toward the
Stocks-Fischer et al. 1999, Peng and Liu 2019, Nafisi et al. development of stimulation techniques and current
2020b, Ezzat and Ewida 2021, Burdalski 2020, Burdalski & understandings of the biostimulation process (e.g., Fujita et al.
Gomez 2020). In a second group, augmented bacteria are added 2008, Tobler et al. 2011, Burbank et al. 2013, Gomez et al. 2014,
to soils with nutrients provided at some point, either by Gat et al. 2016, Gomez et al. 2016, Dhami et al. 2017). The
suspending cells in fresh or “spent” growth media or providing current understanding of ureolytic enrichment involves the
cementation solutions including organics, thus presenting a targeted selection for microorganisms in soils with unregulated
possibility for unintended stimulation (e.g., Whiffin et al. 2007, ureolytic activity and elevated ammonia and pH tolerance. This
Harkes et al. 2010, Jiang et al. 2017, Jiang & Soga 2017, 2019, is accomplished by supplying stimulation solutions with
Nafisi et al. 2019, Song & Liu 2020, San Pablo et al. 2020, Lee concentrations of ammonia and hydroxide that are tolerated by
et al. 2021). Importantly, these studies provided relatively short the targeted microorganisms, but suppress the growth of non-
residence times (< 36 hours), however, the time courses observed ureolytic bacteria that would otherwise compete for the limiting
for urea degradation experiments with sequential enrichments media components. Stimulation by this approach has always
(Figure 20 & 21), suggest that stimulated ureolytic bacteria could been successful in application to 14 different sands or sites,
have still made some contribution to MICP in these scenarios. A including repeated trials with a single sand tested 7 times over an
third group allow for organic constituents to be present for 8-year span and one site tested at three different depths ranging
several days during the cementation phase and therefore offer a from 2 to 12 m (Gomez et al. 2014, Gomez et al. 2016, Gomez
clear opportunity for stimulated MICP to occur (e.g., van Paassen & DeJong 2017, Gomez et al. 2018a, b, Lee et al. 2019a, b, San
et al. 2010b, Al Qabany et al. 2012, Gomez et al. 2016, Cheshomi Pablo et al. 2020, Graddy et al. 2021, San Pablo et al. 2022a,
& Mansouri 2019, Jiang et al. 2019). One clear example of this Yanez et al. 2022). Although it may be possible that certain soils
potential is highlighted in Mukherjee at al. (2021) where an have an uncooperative natural population of bacteria that stymies
organic resuspension solution with no added bacteria was mixed efforts to stimulate, no such circumstances have been
into a silty soil column experiment. Their results suggested that encountered thus far.
this single organic amendment triggered stimulated MICP that From a practical perspective, stimulation offers the ability to
achieved nearly 25% of the strength improvement observed in avoid previously discussed challenges related to colloidal
other columns with added organics and bacteria at a cell density filtration of injected cells by enriching microorganisms that are
412
already present. This differentiation can yield important benefits 2.4.1 Stimulated microbial communities
with respect to improving cementation spatial uniformity and The stimulation treatment method (Section 3.2) is strongly
control through the elimination of issues related to the selective for a narrow group of closely related microorganisms
localization of cells near injection locations and poor distribution (Figure 5b, d). Nevertheless, within this clade, there is also a
of ureolytic activity and reactants over large treatment distances. surprising amount of diversity given the repeated application of
For field scale applications, such advantages may be a decisive stimulation solutions, which impose stringent selection. Recent
factor allowing for the MICP process to become practically findings suggest that enrichment techniques used in the authors’
superior to other technologies, allowing for treatment of soils studies yield at least 96% Firmicutes at the end of stimulation
over large injection distances, including those potentially needed when solutions are applied with artificial ground water (AGW)
when addressing existing infrastructure, as well as permitting or deionized water as the solution base (Figure 6a-c), 93% when
associated reductions in treatment implementation costs. using stimulation solutions with natural seawater (NSW) as a
Additional benefits of stimulation include the elimination of the base (Figure 6d, e), or 90% using a combination of 40% NSW
need for sterile cultivation and transportation of dense cell and 60% AGW (Figure 6f), of which 85%, 92%, and 89%,
suspensions to project sites and the reduction in environmental
respectively, were of the genus Sporosarcina. Similar selection
and soil ecological impacts that follow. From a treatment
appears in both freshwater alluvial sands (Concrete Sand) and
efficacy standpoint, stimulation also yields a persistent ureolytic
marine sands (Beach Sand) when applying solutions with either
microbial community that requires minimal additional inputs to
maintain ureolytic capacity. In effect, this precludes activity artificial ground water or natural sea water, indicating that this
losses associated with cell death and entombment, thereby selection is robust and the requisite bacteria are ubiquitously
avoiding the periodic need for reintroduction of augmented distributed. While enrichment for Firmicutes is not surprising,
ureolytic cells. The stimulation process does have some different given that they include a number of halo- and alkali-tolerant
requirements, however. In particular, biostimulation requires bacteria, the consistent and stringent selection for a single genus
is remarkable.
Figure 6. Microbial community composition by 16S rRNA analysis of (a, c, d) alluvial Concrete Sand or (b, e, f) Beach Sand using treatment solutions
based in (a,b) artificial ground water (AGW; Ferris et al. 2004), (d,e) natural seawater (NSW), (c) deionized water (from Graddy et al. 2021), or (f) a
40% mixture of NSW with 60% AGW. Biostimulation treatments were 350 mM urea, 100 mM ammonium chloride, 42.5 mM sodium acetate, 0.2 g/L
yeast extract, pH 9 and the cementation treatments were 350 mM urea, 250 mM calcium chloride, 12.5 mM ammonium chloride, 42.5 mM sodium
acetate, 0.2 g/L yeast extract, at an incidental pH of around 8.3 in all but (c), where the treatments were the same save 12.5 mM ammonium, 0.1 g/L
yeast extract with no pH adjustment throughout. Shaded regions indicate stimulation treatments. Data are missing for the initial microbial communities
in (a) and (d) (adapted from Yanez et al. 2022).
repeated applications of a selective growth medium to soils and Sporosarcina are an example of such haloalkalitolerant
incubation thereafter in order to achieve sufficient ureolytic bacteria, but are also known for their preference for urea-
activity. This might involve the addition of several days to a nitrogen, constitutive ureolytic activity, and ammonia tolerance
given treatment program before cementation and improvement (Mörsdorf & Kaltwasser 1989, Leejeerajumneanet al. 2000,
of soils can commence. Given the significant practical Trujillo et al. 2015). These stimulated species are therefore
performance benefits that stimulation affords including its high presumably strongly selected for on the basis of ionic strength,
rate of success and ease of application, however, the authors pH, and urea-hydrolysis end products. It remains unknown if
almost exclusively employ stimulated MICP with specific they are the best suited for these conditions or simply the first to
techniques described later in Section 3. dominate these enrichments and, if given enough time, more
efficient but more slowly growing bacteria may displace them
from this niche. If their success and competitive edge is the result
of more rapid growth, increasing treatment application frequency
413
may maintain selection for fast growing organisms through the microorganisms to overcome abiotic reaction activation
repeated elution of resident bacteria. Given these repeated energies. For a given solution condition, the thermodynamic
growth and elution cycles (analogous to a traditional serial favorability for a given mineral phase to either precipitate or
enrichment), it is rather surprising that 7 species of Sporosarcina dissolve can be expressed using saturation state
appear to be present at significant abundance ( > 10%) in these Equation 7, wherein Ksp is the solubility product of a given
environments (Graddy et al. 2021). This outcome further mineral polymorph, (Ca2+) is the activity of calcium in solution,
suggests that a number of Sporosarcina species are frequently and (CO32-) is the activity of carbonate in solution. The product
present in natural soils and capable of contributing to the function of (Ca2+) and (CO32-) is commonly referred to as the ion activity
of stimulated MICP, rendering stimulation success less product (IAP) and represents the availability of these species in
susceptible to the presence or absence of a particular species in solution to participate in reactions that either form or dissolve
any given soil. This diversity also suggests that there may be minerals.
niches present for each of these species, perhaps at different
points during a treatment cycle. (7)
The relative abundance of Sporosarcina has been found to
range from remaining steady to decreasing substantially over the
course of repeated stimulated MICP treatments. In one instance,
Sporosarcina was found to be partially displaced by mine
Lysinibacillus (Figure 5d “Other Firmicutes”, Graddy et al. are undersaturated with respect to a given mineral and dissolution
2021). In this case, frequent pore fluid sampling over the course
of the experiment likely desaturated portions of the column and neither dissolution nor precipitation would be expected. For the
increased oxygen availability to the microbial community. The same geochemical conditions (i.e., aqueous Ca2+ and CO32-
appearance of Lysinibacillus was not seen in other analogous activities) solutions may be either undersaturated or
sub-culturing (Graddy et al. 2018) or comparable AGW based supersaturated, depending on the solubility of the CaCO3
enrichments independent of the parent material; better control of polymorph. Activity ratio diagrams, such as that shown in Figure
desaturation maintained strict selection for Sporosarcina using 7 can be used to compare existing solution conditions to the
both freshwater alluvial sand and marine sand (Figure 6a, b). mineral solubilities of different CaCO3 polymorphs expected
Therefore, increased oxygen availability was thought to have during MICP. In this diagram, solution Ca2+ and CO32- activities
enabled the shift in microbial community composition, though 3
no apparent decrease in ureolytic activity was observed, despite polymorphs to determine conditions under which mineral
Lysinibacillus isolates not manifesting clear ureolytic activity precipitation or dissolution would be expected. Solution
(Graddy et al. 2021). When treated with natural seawater,
activities plotting above respective lines indicate supersaturation
substantial diversity was observed following the transition from
and a tendency for the solution to precipitate a given CaCO3
stimulation to cementation treatments. This was partly attributed
polymorph with values plotting below lines indicative of
to the use of natural seawater that provided a consistent source
of exogenous bacteria with each solution injection as well as the undersaturation and a tendency for dissolution. As shown for the
reduction in selective conditions employed in cementation
treatments (reduced initial pH and reduced added ammonium).
This may also be partially due to the fact that marine bacteria are
naturally more salt and alkalitolerant than freshwater soil
bacteria, making the primary mechanism for negative selection
the presence of ammonia. The selectivity of total ammonium
additions in turn is largely attenuated by the decreased pH
experienced during cementation relative to stimulation.
Regardless of these distinct community differences, there was no
correlation between this increase in diversity and a decrease in
ureolytic rate or achieved engineering performance, as discussed
later in Sections 3.2.3.4 and 3.7.3 (Figures 23 & 30).
2.5 CaCO3 mineralogy & morphology
Microbially-induced calcite precipitation is commonly referred
to as generating calcite minerals, however, the generated calcium
carbonate (CaCO3) may exist as a variety of different mineral
polymorphs, each with identical elemental compositions but
Figure 7. Activity ratio diagram illustrating differences in equilibrium
important differences in crystalline structures and physical
lines ( = 1) between CaCO3 polymorphs (at 20°C and 1 atm) resulting
properties (Bladh et al. 2001, APHA 2005). CaCO3 polymorphs from differences in mineral solubility products (Ksp values) (adapted
generated during ureolytic biocementation include vaterite (Ksp from Burdalski 2020).
= 10-7.91), aragonite (Ksp = 10-8.34), and calcite (Ksp = 10-8.48),
presented here in order of decreasing solubility in water at 20º C three CaCO3 polymorphs considered, solution conditions can
(Plummer & Busenberg 1982, Gal et al. 1996, APHA 2005). exist wherein values can remain supersaturated with respect to
CaCO3 can also exist in an amorphous phase with no definitive lower solubility polymorphs (e.g., calcite) yet simultaneously
structure and as a result with varying solubilities and undersaturated with respect to higher solubility phases (e.g.,
morphologies. The solubilities of these three primary vaterite). Thus, the tendency for various CaCO3 polymorphs to
polymorphs vary significantly and govern the CaCO3 minerals form during MICP is governed by aqueous geochemical
that can be formed under different aqueous conditions. It is conditions present during and after precipitation.
helpful to consider expected differences in mineral formation Differences between CaCO3 polymorphs extend beyond their
from the perspective of abiotic mineral stability although chemical solubilities and include important morphological
biogenic mineral formation oftentimes cannot be fully captured differences. For example, calcite crystals most frequently exhibit
using thermodynamics alone due to the ability of a trigonal (rhombohedral) lattice structure, aragonite exhibits an
414
orthorhombic crystal structure, and vaterite has a hexagonal the ripening of spherical vaterite-like crystals to rhombohedral
crystal structure and often appears in a spherical form (Bladh et calcite-like crystals was observed during stop-flow treatment
al. 2001). Although perhaps less relevant to MICP applications, residence periods as reagent concentrations were consumed
at temperatures near freezing, CaCO3 can also exist as ikaite, a (Figure 8). Similarly, they also observed the appearance of
polymorph with a monoclinic crystal structure and a spike-like smaller and presumably more soluble crystals early during
morphology. Amorphous calcium carbonate has no definitive injections with later dissolution and recrystallization of these
crystalline structure and thus can exhibit even larger differences minerals to form larger more thermodynamically stable crystals.
in morphology. Although many researchers have characterized Burdalski (2020) performed small volume plate experiments that
the mineralogy of ureolytic biocementation generated under were destructively sampled at various points during the
specific experimental conditions, a clear understanding of how biocementation process to examine changes in mineralogy with
differences in achieved minerals may translate to changes in soil reaction extent using semi-quantitative x-ray diffraction analyses
mechanical behaviors has remained limited. Nonetheless, these (XRD). In experiments involving relatively high reactant
morphological differences and their spatial distribution have
concentrations (1250 mM urea and Ca2+), the formation of
important implications with respect to the engineering behavior
significant quantities of vaterite were observed early during
of biocemented soils with anticipated effects with respect to
reactions with progressive reductions in vaterite and associated
interparticle friction, volumetric responses, and initial shear
stiffnesses, among other critical behaviors. increases in calcite fractions as the treatment process proceeded
(Figure 9). Zambare et al. (2020) performed MICP droplet
2.5.1 Reaction progression experiments and noted similar transitions in mineralogy and
The realized biocementation mineralogy will be influenced by morphology in time during reactions, with the appearance of
the extent of reactions that occur during treatment injections as poorly-structured amorphous calcium carbonate as a precursor to
well as by the treatment solutions and subsurface environmental vaterite and some limited amounts of calcite. Collectively, these
factors. In particular, recent studies have shown that while higher studies suggest that the mineralogy of biocementation is
solubility CaCO3 polymorphs (e.g., aragonite and vaterite) may significantly influenced by both chemical conditions during and
precipitate initially during biocementation in the presence of after precipitation with the potential for ripening of more soluble
higher IAP values, such minerals may subsequently dissolve and mineral phases towards more stable CaCO3 polymorphs in time
reprecipitate later in the treatment process as activities of when reactions are allowed to proceed to completion and water
involved species (Ca2+, CO32-) decrease below that needed to remains present.
maintain supersaturation for these higher solubility polymorphs.
This process, more formally known as Ostwald ripening, is not 2.5.2 Treatment solution composition
unique to MICP and suggests that the mineralogy and structure Although researchers have considered a wide variety of reactant
of MICP may be transient during the treatment process, concentrations, chemical additives, and bacterial densities in
influenced by not only reaction geochemical conditions but also biocementation treatments, the relationship between these
reaction progression. Wang et al. (2019) captured this formulations and resulting mineralogy has not been
a) b) c)
T = 0 hr. T = 1 hr. 10 min T = 1 hr. 20 min
d) e) f)
T = 1 hr. 30 min T = 4 hr. T = 12 hr.
Figure 8. Microscope images (a) 0 hr., (b) 1 hr. 10 min, (c) 1 hr. 20 min, (d) 1 hr. 30 min, (e) 4 hr., and (f) 12 hr. after injections show the initial
appearance and dissolution of spherical vaterite-like CaCO3 crystals with the appearance of rhombohedral calcite-like CaCO3 crystals later in time
(adapted from Wang et al. 2019).
phenomenon visually using microfluidic experiments wherein systematically examined. This has primarily resulted from
415
differences in treatment formulations, soil materials, and 2.5.3 Environmental conditions
treatment application methods between researchers which have Treatment solutions can impose specific geochemical conditions
prevented attributing specific mineralogical differences to during MICP, however, remaining uncertainties include
isolated variables. Depending on the particular study, unavoidable environmental conditions including soil types, trace
cementation solution urea and calcium concentrations can vary ions, pH, and other factors, which may alter geochemical
by orders of magnitude with specific additives such as sodium conditions present during biocementation. While differences in
bicarbonate, ammonium chloride, and various growth factors environmental conditions present nearly endless possibilities
frequently included or excluded. “Legacy” treatment with respect to the geochemical conditions, experiments by
formulations have persisted likely out of concern for unintended Burdalski et al. (2022a) have examined a limited range of
consequences if formulations are modified. Nonetheless, it is environmental factors expected to be of importance for practical
clear that the decision to modify applied reactant concentrations field applications. In these experiments, notable differences in
and include or omit specific additives should be carefully biocementation mineralogy and morphology were observed in
considered within the context of the minerals that are being the presence of seawater ions with magnesium concentrations
generated. Burdalski (2020) aimed to further improve found to significantly inhibit ureolytic activity. When variations
understanding of these relationships and examined the impact of in parent soils were considered, including different natural
isolated factors including applied urea and Ca2+ concentrations, poorly-graded sands and clay minerals, large differences in
augmented cell densities, and common treatment solution reaction kinetics were observed. This was likely due to
additives (e.g., NH4Cl, NaHCO3) on biocementation mineralogy differences in exchangeable soil ions, however, large
and morphology in high-throughput plate experiments. These mineralogical differences were not observed, with calcite
investigations yielded important insights regarding how remaining the predominant mineral polymorph (Figure 11).
treatment formulation changes can have significant While questions remain regarding a multitude of factors, similar
consequences on the produced mineral phases. For example, in high-throughput experiments may allow for the impact of site-
experiments wherein augmented Sporosarcina pasteurii cell specific environmental conditions to be more readily examined,
densities were varied by over one order of magnitude, large quantified, and understood. Finally, although the impact of
differences in urea hydrolysis reaction rates were observed, but mineralogical and morphological differences on soil mechanical
only small differences in mineralogy and morphology resulted responses and permanence have yet to be fully understood, it is
(Figure 10). In contrast, when growth factors (i.e., yeast extract, clear that such differences will likely have important practical
nutrient broth) were supplied to treatment solutions with a implications and thus require more extensive characterization
Sporosarcina pasteurii cell density of ~7.5 x 107 cells/mL moving forward.
dramatic differences in mineralogy and morphology were
observed with increasing quantities of more soluble vaterite 2.6 Microstructure & microscale properties
minerals detected with increases in growth factor concentrations
While the microstructure of MICP is inherently linked to the
(Burdalski et al. 2022c). Interestingly, both increases in
treatment processes used to generate such cementation, the
augmented cell densities and growth factor concentrations have
effects of CaCO3 precipitation can be considered to more broadly
been used in past studies to achieve faster reaction rates, and
include the following primary microstructural changes: (1) the
therefore these results suggest that such changes may have had
bonding of adjacent particles through cohesive cementation
significant consequences for the produced minerals.
bridges that generate tensile capacities and cohesive forces, (2)
the coating of particle surfaces with CaCO3 crystals that alter
a) b) c)
5% 20% 100%
Figure 9. SEM images of biocemented specimens from 1250 mM urea and Ca2+ plate experiments after (a) 5, (b) 20, and (c) 100% reaction completion
show the appearance of significant vaterite early on during reactions with progressive dissolution of vaterite and reprecipitation as calcite in time as
conditions approached equilibrium (adapted from Burdalski et al. 2022b).
a) b) c)
1.7 x 10 7 cells/mL 7.4 x 10 7 cells/mL 3.4 x 10 8 cells/mL
Figure 10. SEM images of biocemented specimens from 250 mM urea and Ca2+ plate experiments with augmented Sporosarcina pasteurii at cell densities
of (a) 1.7 x 107, (b) 7.4 x 107, and (c) 3.4 x 108 cells/mL (adapted from Burdalski et al. 2022c).
416
surface roughness, angularity, and interparticle friction, and (3) imaging, clearly show the coating of soil particle surfaces with
the densification of soil particle matrices through the filling of generated CaCO3 and the resulting reductions in void ratios
voids with solid mineral phases. As a direct outcome of more (DeJong et al. 2006). DeJong et al. (2010) conceptualized the
microbiological and geochemical fundamental processes cementation distribution and expected failure mechanisms,
discussed earlier in Section 2, Figure 12 presents examples of considering the proportional distribution of cementation at
SEM images of biocemented sand particles for which noticeable particle-particle contacts and on open surfaces and whether
differences in cementation morphologies and distributions can be breakage would occur within the cementation precipitate or at the
observed. Even without knowledge of their mechanical particle-precipitate interface. Subsequent studies have refined
behaviors, such differences would be expected to manifest as these hypotheses and have provided quantitative evaluations. For
differences in engineering responses dependent on not only the example, further study is still necessary to confirm that
magnitude of cementation, but also its distribution at the particle- preferential cementation occurs at contact locations; while it is
scale, crystal morphology, and association with particle surfaces clear that cementation preferentially forms near bacterial cells
(wherein carbonate species are generated) and that bacterial cells
and contacts.
are oftentimes attached to soil particle surfaces, the extent to
Further characterizations including those presented in Figure
which flow stagnation zones and colloidal filtration may
13, which involved the impregnation of cemented sand samples
with epoxy, cross-sectioning, and subsequent polishing before
a) b) c)
Montmorillonite Fraser River Sand Ottawa F-65 Sand
Figure 11. SEM images of biocemented specimens from 250 mM urea and Ca2+ plate experiments involving (a) montmorillonite, (b) Fraser River Sand,
and (c) Ottawa F-65 sand materials (adapted from Burdalski et al. 2022a).
Figure 12. SEM images from the authors’ experiments involving both stimulation and augmentation as well as varying treatment techniques highlight
the breadth of possible microstructures that can result from the MICP process.
EDS Elemental Mapping
a) SEM Image b) Carbon c) Silicon d) Calcium
Figure 13. SEM images and EDS elemental mapping of a cross-section of biocemented sand with elemental concentrations of carbon silicon, and calcium
shown (adapted from DeJong et al. 2006). It is noted that the carbon map reflects some influence of the carbon mounting tape.
417
concentrate cells near particle-particle contacts (El Mountassir et 2.6.3 Characterization of cemented soil matrices
al. 2018) is still debated. X-ray computed tomography (X-ray CT) has been used to non-
destructively visualize spatial distributions of generated
2.6.1 Mechanisms of improvement cementation (DeJong et al. 2008, Terzis & Laloui 2018, Baek et
How the primary mechanisms of improvement, (1) cohesive al. 2022) and quantify the microstructure of cemented soil
bonds, (2) particle angularity and surface roughness changes, and matrices including bond frequencies, orientations, and how such
(3) bulk densification of soils through the filling of void volumes, characterizations may inform predictions of mechanical and
alter the engineering behaviors of soils, depends on the applied hydraulic behaviors (Roy 2021), These CT images are similar to
loading. The presence of interparticle bonds in biocemented soils that shown in Figure 16, which presents cross-sections through
has been established through a variety of means including the poorly graded soils with D50 values of 0.18, 0.51, 1.39, and 2.74
direct measurement of particle contact tensile capacities to the mm which have been treated to approximately 20% calcite by
qualitative imaging of particle bonds. These particle bonds mass.
provide cohesive forces between particles, however, brittleness
results in improvements observed at smaller shear strains (< 2.6.4 Distribution of contact-level cementation and effects on
~0.5%) being diminished at higher strains due to bond fracturing. macroscale responses
This is most easily observed in specimens treated to low levels Improved understandings regarding cementation distributions
of cementation wherein initial stiffnesses and peak strengths are and contact mechanics have been investigated by comparing
improved, with more minimal improvement post-peak (Section anticipated behaviors from simplified models to experimental
5). data sets. For example, Lin et al. (2020) examined how
The coating of soil particle surfaces inevitably results in soil differences in shear and compression wave velocities of
particles with apparent increases in shape and surface roughness. biocemented sands from the literature compared to expectations
Recent imaging of Ottawa F-65 sands at varying cementation from models which assumed either (1) localization of
levels (Figure 14a) clearly shows the progression of these cementation on contacts, (2) uniform coating of grains with
changes. Image analysis and quantitative evaluation of roundness cementation, or (3) void infilling by cementation with no particle
(Wadell 1932, 1933, 1935) and sphericity (S = length/width) associations. Figure 17 presents an example of the model and
following Zheng and Hryciw (2015) show that roundness can
capture such changes, particularly when comparing untreated
particles to particles cemented to 5% calcite by mass, while
sphericity shows no strong correlation (Figure 14b). These shape
and surface roughness increases would be expected to influence
behaviors after cemented bonds have fractured and the frictional
resistance and volumetric behaviors of the soil are mobilized
(shear strains > ~0.5%). At larger strains, it is expected that
sufficient cementation may also alter critical state behaviors via
increases in constant volume friction angles.
The addition of precipitate solids further densifies soils,
which can also alter volumetric behaviors. The addition of
mineral masses would be expected to alter shearing behaviors
primary at strains corresponding to peak strength and beyond
where the volumetric tendencies of a soil are mobilized (shear
strains > ~0.5%). While the density of calcite mineral is similar
to that of quartz (GS ~ 2.71), the volume occupied by cementation
may be larger due to the presence of voids and imperfections.
Thus, progressive crushing of cementation during mechanical
loadings, may further alter the volume occupied by CaCO3
during shearing, thereby resulting in cemented soils becoming
less dilatant at higher shear strains. Aside from mechanical
behaviors, the addition of mineral solids may also alter important
hydraulic characteristics as discussed later in Section 4.
2.6.2 Mechanical testing of single particle contacts
In an effort to better understand the mechanisms of enhancement
at the particle level, researchers have examined the behavior of
single particle contacts to characterize the strength and stiffness
of such bonds, their modes of failure, and inform input
parameters for the modeling of cemented materials using discrete
elements. Ham et al. (2022) examined the response of cemented
glass beads in tension and shear (Figure 15). Results showed the
failure of cemented bonds to depend on cementation level, with
the internal fracturing of bonds observed at high cementation,
and the debonding of bonds and particles as well as combined
failures exhibited at lower cementation levels. The tensile
capacity of such bonds appeared to exceed the shear capacity of
similar bonds by a factor of two. Interestingly, debonding failure
mechanisms also appear to achieve the largest tensile and shear Figure 14. (a) Particle images of Ottawa F-65 sand comparing an
strength when compared to internal and combined failure untreated soil grain and a specimen treated with 10% calcite content and
mechanisms. (b) roundness and sphericity for particles at varying cementation levels
calculated using parameters from Zheng and Hryciw (2015).
418
experimental data comparisons for shear wave velocity 2013, Liu et al. 2019, Chen & Achal 2020), freeze-thaw cycles
measurements. Their results suggested that experimental trends (Cheng et al. 2013, Liu et al. 2019), wet/dry cycles (Liu et al.
were best approximated by considering a combination of particle 2019), and unloading/reloading events (Porter et al. 2018) have
coating and void infilling mechanisms. yielded practical insights regarding specific scenarios, but
broader understandings regarding the long-term permanence of
2.7 Permanence & reversibility biocementation under conditions more representative of
liquefaction mitigation applications (i.e., saturated natural sand
The resilience of biocemented soils to extreme mechanical and
deposits) has remained limited. As biocementation advances
chemical stressors, including acid rain exposure (Cheng et al.
towards field-scale deployment, estimating the permanence of
biocemented soils will be critical to identify favorable field
applications, quantify life-cycle environmental impacts, and
predict long-term in situ mechanical performances. Although
few studies have examined the dissolution of CaCO3-based
biocementation specifically (Ribeiro & Gomez 2022a, Ribeiro &
Gomez 2022b), other studies considering the dissolution
behavior of naturally-occurring and abiotically-generated CaCO3
minerals (Sjöberg 1978, Morse and Arvidson 2002, Cubillas et
al. 2005, Gehlen et al. 2005, Colombani 2016, and others) have
yielded important insights regarding the factors that control
CaCO3 mineral stability, including their dissolution rates,
magnitudes, and how such phenomena can be captured using
geochemical kinetic frameworks. Although the chemical
stability of biocementation is still being investigated, recent
insights and preliminary results from ongoing studies are
presented here.
2.7.1 Tendency for dissolution
The tendency for minerals to dissolve over time is controlled by
the chemical composition of the pore fluid. More specifically, the
stability of existing CaCO3 biocementation is primarily
controlled by the activity product (IAP) of carbonate and calcium
ions in solution, and its relation to the solubility product (Ksp) of
the particular CaCO3 mineral(s) present. The state of this system
Section 2.5. When solutions contain low concentrations of
aqueous carbonate and calcium, the saturation ratio can be less
than 1, suggesting that dissolution is favored. The greater the
undersaturation, the closer the saturation ratio will be to zero, and
the more CaCO3 mass that must be dissolved to achieve
equilibrium. Since the solubility of different CaCO3 minerals can
vary by orders of magnitude, both the biocementation
mineralogy and pore fluid chemistry directly determine which
Figure 15. (a) Conceptual schematics and images of debonding, mixed, minerals dissolve most readily and the magnitudes of dissolution
and internal bond failure mechanisms, (b) tensile force responses for needed to achieve equilibrium. Although biocementation
example cemented particle tests involving the three different failure
mechanisms (adapted from Ham et al. 2022).
dissolution can be intuitively assessed under some conditions,
such as in the presence of aggressive groundwater or acid mine
Figure 16. Example of X-ray CT images of a cemented soil volume wherein various volumes have been removed with distinctions between sand particles
(red), pore fluids (blue), cementing CaCO3 minerals (green) (adapted from Baek et al. 2022).
419
occur with rates dependent on many variables including the
physical properties of the minerals and groundwater. In this
context, biocementation dissolution is a reactive transport
problem, governed by the transport of groundwater through
biocemented soils and its interactions with mineral surfaces, the
mineralogy of biocementation and its specific surface area
among other characteristics, the rate at which ions can be
exchanged between mineral surfaces and undersaturated
solutions, and other aspects related to the groundwater chemistry.
While the dissolution behavior of biocementation has only been
recently investigated (Ribeiro & Gomez 2022a), past studies
regarding CaCO3 minerals have suggested that rates may be
strongly correlated to surrounding solution chemistry and in
particular the solution pH and temperature (Sjöberg & Rickard
1984). Sjöberg & Rickard (1984) showed that when the pH of
surrounding solutions exceeded ~5.5, CaCO3 dissolution rates
were saturation-controlled with dissolution rates being
proportional to the solution’s saturation index (Equation 8). In
contrast, under more acidic conditions where pH values were
near and below 4, CaCO3 dissolution rates were increasingly
diffusion-controlled due to limitations in how fast ions can
exchange at mineral surfaces with dissolution rates primarily
dependent on the pH of solutions (Equation 9). Equation 8
presents the saturation-controlled dissolution kinetic equation,
wherein rc is the dissolution rate, kc is a dissolution rate constant,
Ac is the mineral specific surface area, n is the reaction order, Ksp
is the solubility product of the mineral, and (Ca2+) and (CO32-)
are the activities of calcium and carbonate ions, respectively
(Morse & Berner 1972). Equation 9 presents the diffusion-
controlled kinetic equation, wherein rc, kc, n, and Ac are similar
to Equation 8 and (H+) is the activity of hydrogen ions (Sjöberg
& Rickard 1984). Although understanding of biocementation
Figure 17. Comparison between an extensive set of shear wave velocity dissolution remains somewhat limited, these studies have
measurements from biocemented sands in the literature and model
expectations for simplified cementation distributions shows best
provided key insights and suggest that the surrounding solution
agreement between the experimental data and matrix-supporting and chemistry is an important factor altering dissolution rates and our
grain-coating models (adapted from Lin et al. 2020). ability to model and predict such processes.
drainage, in many cases where the groundwater may be slightly
undersaturated and/or near equilibrium, accurate characterization (8)
of solution chemistry may be critical towards assessing stability
and predicting longevity over years or decades. While accurate (9)
prediction of dissolution may be of high priority for some
projects and sites, it should be recognized that dissolution may 2.7.3 Current understanding of dissolution
also not occur at some sites. For example, typical surface Recent studies have provided new insights regarding
seawater includes ~10 mM calcium and ~2 mM bicarbonate with biocementation dissolution rates, the spatial and temporal
a pH near 8.1, resulting in saturation states ranging between ~2 progression of biocementation degradation when exposed to
and 7 (Dickson and Goyet 1994). Under these conditions undersaturated solutions, and the ability of existing kinetic
representative of coastal locations, supersaturation of solutions frameworks to capture such processes. Ribeiro and Gomez
with respect to CaCO3 may exist, suggesting the favorability for (2022a) completed a series of batch experiments where
subsequent precipitation events after MICP treatments. biocemented poorly-graded sands were subjected to different
Therefore, dissolution may not be of primary concern if aqueous acidic solutions to investigate the magnitude and rate of CaCO3
conditions are initially characterized and deemed favorable. dissolution as a function of acid type, concentration, initial pH,
Even when conditions are undersaturated, if the groundwater is and other factors. The results suggested that although acid type
not readily exchanged, the low solubility of CaCO3 minerals may and concentration can alter dissolution patterns, such trends can
require only minimal mass losses to achieve equilibrium and be captured using existing saturation-controlled kinetic models,
maintain further stability. Although questions regarding especially when surrounding solutions are strongly buffered.
biocementation dissolution and long-term stability will always Figure 18 presents an example of dissolution batch experiment
remain site-specific, the frameworks for assessing mineral results where biocemented samples were subjected to 50 mM
stability are available today and can be used, requiring only acetic acid with and without 100 mM acetate, with both solution
collection and characterization of site groundwater samples and calcium and pH changes monitored in time to assess dissolution
evaluation of generated CaCO3 mineralogy. progression. A kinetic model was fit to the achieved data
following Equation 8 by varying unknown parameters kcAc and
2.7.2 Rate of dissolution n. As CaCO3 dissolves, calcium ions are released into solution,
Predicting how biocementation will dissolve over time is resulting in increasing calcium concentrations and more alkaline
complicated by the temporal progression of dissolution. That is, solution pH due to the release of carbonate ions and their
even when solutions are undersaturated, dissolution will not speciation as other carbonate species (i.e., bicarbonate) in
occur instantaneously, but instead a slow kinetic process will accordance with solution pH. Following calibration, dissolution
420
Figure 19. Progression of biocementation dissolution during experiments
wherein repeated dissolution injections were applied over time to
biocemented soil columns and specimens were analyzed after 0, 20, and
50 dissolution injections. Trends were further captured using the model
presented in Equation 2.8 following calibration of kcAc and n (adapted
following Ribeiro & Gomez 2022b).
3 MICP TREATMENT METHOD
3.1 Development of augmentation approaches
Bioaugmentation, the injection of specialized non-native
microbial strains to complete some process of engineering
Figure 18. Progression of (a) solution pH and (b) calcium concentration interest, has been leveraged in the literature for a variety of
changes in time during biocementation dissolution batch experiments different environmental applications, of which perhaps
involving biocemented sand samples subjected to 50 mM acetic acid bioremediation of contaminated soils and pore fluids has been
solutions with 100 mM acetate (50A_100Act) and without acetate
most notable (Ellis et al. 2000,
additions (50A). Trends were further captured using the model presented
in Equation 2.8 following calibration of kcAc and n (adapted following 2017). With respect to biocementation, augmentation has been
Ribeiro & Gomez 2022a). the method of choice for almost all prior literature, with
Sporosarcina pasteurii ATCC strain 11859, a constitutively
kinetic frameworks can be implemented in reactive transport ureolytic bacterium with high specific activity, being the most
models to capture spatial and temporal changes in soil commonly augmented microorganism. In this process, ureolytic
biocementation when subjected to undersaturated solutions. microorganisms are prepared using sterile culturing techniques,
Preliminary results from Ribeiro and Gomez (2022b) quantified, and injected into soils to enable urea hydrolysis and
demonstrate the potential of such approaches to capture the biocementation process. This process in the context of MICP
dissolution progression at the element scale. Figure 19 presents has been extensively investigated in numerous past studies,
measured soil CaCO3 contents during progressive dissolution which have been comprehensively reviewed in prior works (El
injections applied to soil column experiments along with Mountassir et al. 2018, Rahman et al. 2020, Bhutange & Latkar
modeled distributions from a calibrated reactive transport model. 2020, Tang et al. 2020); therefore, this section provides a brief
As evident, such calibrated models can reasonably predict synopsis of the augmentation process and primary considerations
dissolution induced changes and could be used to forward predict when employing the technique.
dissolution for field-scale applications for more complicated
flow regimes and biocementation distributions. While still being 3.1.1 History of augmentation approaches
refined, it is likely that either site specific column or batch testing Although microorganisms have been used to reduce borehole
will be necessary to calibrate dissolution frameworks to enable hydraulic conductivity since the early 1980’s (Updegraff et al.
the use of more advanced reactive transport programs (e.g., 1983, Finnerty & Singer 1983, Jenneman et al. 1984), specific
PHT3D, COMSOL) for forward prediction of biocementation use of S. pasteurii (formerly Bacillus pasteurii) to catalyze what
longevity and spatial and temporal changes following initial is now called augmented MICP appeared nearly one decade later
implementation. As the technology achieves more widespread (Kantzas et al. 1992). Shortly thereafter, S. pasteurii was further
application, project sites may also provide further opportunities used to drive CaCO3 precipitation to enable hydraulic
to calibrate models to field data when permanence is monitored conductivity reductions for sandy soils (Gollapudi et al. 1995,
over time. Ferris et al. 1997, Stocks-Fischer et al. 1999). These studies
represent an early stage in the use of what is now broadly
recognized as augmented MICP to alter soil engineering
properties. They are also among the very few examples of “pure”
421
augmented MICP, as they involved no carryover of residual 3.1.2.2 Non-nutritive augmentation
organic compounds from cultured growth media, which could As discussed earlier, the potential for unintended enrichment of
have unintentionally enriched other native ureolytic strains native ureolytic microorganisms and/or contaminants remains
(Section 2.3.2). In its most pure form, the augmentation process when augmented microorganisms are supplied with added
involves growth of S. pasteurii in some growth medium, nutrients. These considerations and their implications for current
commonly ATCC 1376 medium [yeast extract 20g/L, (NH4)2SO4 and past studies were further discussed in Sections 2.1.1 and
10g/L, 0.13 M Tris buffer, pH= 9.0], harvesting cells by 2.3.2. It should be mentioned that in future applications, care
centrifugation, washing cells to remove most medium should be taken to minimize carryover nutrients and establish
constituents, resuspending in some alternative medium, injecting sterile controls if observed phenomena are intended to be
cell-dense suspensions, sometimes also introducing other high attributable to specific augmented strains. For some applications,
ionic strength constituents into soils to facilitate improved such efforts may not be critical nor desired, and thus the potential
attachment and retention, and subsequently providing sequential for enrichment of alterative communities should be explicitly
injections containing urea and calcium to permit CaCO3 acknowledged. While other approaches can be used, most pure
biocementation. In the literature, such processes oftentimes augmented experiments have involved culturing of S. pasteurii
include supplying cells in a resuspending medium that contains in a sterile ATCC medium, rinsing of cells through multiple
significant nutrients, whether intentionally added or as carryover cycles with sterile saline (9 g/L NaCl), and final resuspension in
from growth medium, that could enable enrichment of similar sterile saline solutions, which can then be injected into
indigenous bacteria. Indeed, this consideration, along with heat (or otherwise) sterilized soils. Specific procedures may be
others, may hinder our ability to understand the true distribution critical to this effort and are further outlined later in Section
of augmented bacteria from the majority of past augmented 3.6.1.
MICP experiments (Section 2.3.1). That being said, it is clear that
augmented microorganisms can be used as an approach to 3.1.2.3 Colloidal filtration
achieve ureolytic activity particularly over smaller length scales When introduced into soil systems, the distribution of augmented
wherein other physical limitations can be largely overcome. bacterial cells can be significantly altered by interactions with
While augmentation is not the authors’ suggested approach for surrounding soil particles including electrostatic
field applications requiring the improvement of large soil repulsion/attraction and colloidal filtration. The net effect is that
volumes, when relying on augmented microbes for either field or the successful distribution of bacterial cells in porous media
lab applications, the following specific considerations should be depends on the properties of the provided medium, the
made. surrounding soil surfaces, and the filtration attributes of the
porous media (e.g., pore size and pore distributions). When
3.1.2 Specific considerations for augmentation developing augmentation protocols, such interactions should be
The application of non-native cell suspensions to soils during considered and may require more formal investigation of the
augmented MICP presents some unique physical challenges effects of augmentation media on bacterial cell and soil mineral
including ensuring that cells (1) remain the primary catalyst of surface charges through more involved measurements (i.e., zeta
MICP, (2) can be effectively transported through porous media, potential) or other means (Tominaga et al. 2009). Similarly,
which can act to filter out and concentrate cells spatially, and (3) differences in injection volume, flow, and cell density should be
can successfully attach to soil particles following initial considered, which can promote improved cell uniformity and
introduction. Although not exhaustive, such considerations are delivery over larger injection distances even when filtration
mentioned briefly below with reference to studies by a multitude challenges are present. Although the authors refrain from
of authors who have examined such mechanisms recommending specific augmentation protocols here, such
comprehensively. details are more comprehensively described elsewhere (El
Mountassir et al. 2018). Interactions should be carefully
3.1.2.1 Augmented culture validation considered if pure augmentation protocols are used for field
Efforts should be made to ensure that the bacterium to be applications. For laboratory applications involving centimeter-
augmented remains uncontaminated and true to its nomenclature scale specimens, however, more uniform distributions and
and genetic profile. Cell lines that are broadly available ought to retention of cells can be accomplished through brute-force
be used where possible to facilitate reproducibility. For example, strategies wherein large, augmented volumes are injected in
the most frequently used S. pasteurii strain is available through multiple directions and cycled to promote uniformity. The
the American Type Culture Collection as strain 11859, but the authors’ own protocols (e.g., Darby et al. 2019, Lee et al. 2021,
same strain is available in other culture collections and databases San Pablo et al. 2022b) have used similar injection techniques
under other designations including DSM 33, NCTC 4822, CCM and may be admittedly suboptimal in comparison to others
2056, NCIB 8841, etc. New cell lines require initial (Section 2.3) who have been more dedicated in their
characterization to generate a familiarization with their investigation of such processes. Furthermore, practical soil
appearance and behavior such that deviations, which represent treatability concerns applicable to augmented cells are discussed
possible contamination or mutation, can be detected. Thus, its later in Section 3.7, with specific examples illustrating the
recommended to carefully note colony morphology, liquid considerable challenges presented by colloidal filtration
culture description, growth rate, and specific activity of newly processes discussed earlier in Section 2.3.
acquired cell lines to serve as a reference. Frozen glycerol stocks
should be prepared from new cell lines and returned to at the start 3.1.2.4 Cell fixation
of a new round of experiments or when any changed behavior is Following the introduction of augmented cells into porous
observed in a culture to guard against genetic drift or media, fixation solutions have been used to promote attachment
contamination. Details of specific methods, while beyond the and retention of supplied cells (Harkes et al. 2010, Cheng &
scope of this paper, may be readily found in a volume on Cord-Ruwisch 2012, 2014). The primary mechanism of such
microbiological methods (e.g., Gerhardt et al. 1994). fixation solutions is thought to provide sufficient ionic strength
that net attractive interactions between largely negatively-
charged soil minerals and bacterial cells can be achieved. While
an understanding of these resultant forces requires more involved
422
analysis of surface ion interactions in these systems (e.g., DLVO experimentation demonstrated similar enrichments at meter-
theory; Hermansson 1999), numerous researchers have scale (Gomez et al. 2016, Graddy et al 2018) and the repeatability
demonstrated the merit of such process (Section 2.3). While these of such processes in a variety of materials (Gomez & DeJong
other researchers favor use of CaCl2, in the authors’ current 2017). Recognizing that enrichment selectivity could be
approach to augmentation, similar increases in ionic strength improved, significant stimulation solution enhancements were
have occurred via introduction of saline, urea, NH4Cl, or other demonstrated in Gomez et al. (2018b), which included
salts with augmented cells; however, such additions have been adjustment of initial solution pH to more alkaline values and
less than intentional from a fixation standpoint, but rather have increases in solution initial ammonium concentrations, the
instead attempted to mitigate cell osmotic stress. Although the fundamental mechanisms for which were discussed in Section
importance of this effect is hard to separate in such 2.4. Current formulations have retained similar selective
experimentation, at a minimum, the authors agree that prolonged pressures, but have included additional modifications for control
retention periods for augmentation solutions can improve cell of bulk ureolytic rates and reductions in supplied reactant
retention during subsequent cementation injections, an outcome concentrations during stimulation injections, primarily to
which may result from improved attachment via electrostatic improve spatial uniformity and material efficiency (San Pablo et
effects or other means. al. 2020).
3.2 Development of stimulation approaches 3.2.2 Recommended stimulation solution composition
While other stimulation solutions can likely be successful, the
Similar to bioaugmentation, biostimulation, the enrichment of authors’ current approach involves the application of repeated
indigenous microorganisms with particular enzymatic or stimulation solution treatments containing urea, along with
metabolic capabilities to complete a process of interest, was ammonium chloride, sodium acetate, and yeast extract
originally investigated for bioremediation applications. With concentrations between ~0.04 and 0.2 g/L (San Pablo et al. 2020,
respect to biocementation applications, the process has been Lee et al. 2022). Such injections have proceeded in past
intentionally investigated by researchers (Fujita et al. 2008, experiments on a daily or alternating-day basis (depending on the
Tobler et al. 2011, Burbank et al. 2013, Gomez et al. 2014, Gat targeted activity and the potential of the specific soil material)
et al. 2016, Gomez et al. 2016, Dhami et al. 2017) and perhaps until a desired ureolytic rate was achieved, as assessed by various
unintentionally involved in a considerable fraction of monitoring methods discussed in Section 4. Key to selecting a
experiments that are mistakenly deemed augmented. The particular stimulation protocol is an understanding of the reactive
fundamental science behind the stimulation process as it pertains transport problem of interest. This includes identifying the
to MICP is covered in detail in Section 2; however, the distance over which successful treatment delivery is desired, the
development of the recommended approach and considerations hydraulic conductivity and flow regime expected in the treated
for stimulation treatments are summarized here. zone, including the flow rates that can be reasonably achieved in
that material, and the impact of supplied nutrients on bulk
3.2.1 History of stimulation approaches
ureolytic activity for a specific soil and groundwater
Ferris et al. (1997) provided the first attempt of ureolytic
composition. The recommended stimulation solutions attempt to
stimulation wherein an enrichment medium [333 mM urea, 187
enforce highly selective conditions for ureolysis (high
mM ammonium chloride, 25 mM sodium bicarbonate, and 3 g/L
ammonium, high pH) that are maintained during the entire
nutrient broth, at a pH of 6] was supplied to a sandy soil and
stimulation period. During these treatments, supplied complex
incubated anaerobically for several weeks. Such efforts achieved
organics (e.g., yeast extract) can be used to control bulk ureolytic
minimal ureolytic activity and precipitation as monitored
rates. When higher ureolysis rates are desired, supplying yeast
indirectly by soil hydraulic conductivity changes. While
extract concentrations near 0.2 g/L may be appropriate.
hydraulic conductivity is an insensitive measure of ureolytic
Conversely, when slower rates of ureolysis are desired,
enrichment, the success of this approach likely suffered from
supplying lower yeast extract concentrations near 0.02 g/L may
protracted anoxic incubation, limited selection due to the use of
be needed. Such concentrations are based upon experimental
a single enrichment treatment, and off-target enrichment
observations; however, it remains likely that lower and/or higher
resulting from initially low pH conditions. Later, Fujita et al.
yeast extract concentrations can be used and further
(2008) achieved successful ureolytic stimulation during a field
modifications in rates achieved accordingly. Other stimulation
trial involving treatment of an aquifer sand layer for the purpose
solution components can be further altered to change the
of strontium sequestration. In their study, a 50-fold increase over
selectivity of a given stimulation solution and/or modify
baseline ureolytic activity was observed in the characterized
provided carbon and electron acceptors, and current
community, which, while notable, was approximately six orders
understandings of these factors as they pertain to stimulation
of magnitude less than what can be currently achieved with more
success are further described below. However, once successful
modern biostimulation approaches. An opportunity to apply the
enrichment and the establishment of sufficient ureolytic activity
technique for fracture sealing and geotechnical applications was
for MICP have occurred, soils can be transitioned to the
later recognized by Tobler et al. (2011) and Burbank et al.
cementation phase. In the authors’ current approach, supplied
(2011), respectively. The former demonstrated enrichment
cementation solutions include reduced ammonium
success using groundwater collected from a sandstone aquifer
concentrations, increased urea concentrations, and the addition
and the latter demonstrated detectable CaCO3 and increases in
of calcium when compared to earlier stimulation treatments. This
the genetic potential of urease by 2.5 orders of magnitude (as
transition is particularly critical and thus specific considerations
measured by ureC copy number) in experiments involving
are also provided in Section 3.2.3.5.
natural alluvial sands. While similar studies followed (Gat et al.
2016, Dhami et al. 2017), the authors’ intentional stimulated 3.2.3 Specific considerations for stimulation
experimentation started with Gomez et al. (2014) wherein ten A number of factors exist that govern the success and behavior
successive stimulation treatments were applied to soil column of stimulation approaches. While the success of an enrichment
experiments containing eight different poorly-graded natural
could be judged by the ability to achieve high bulk ureolytic
sands. The use of successive treatments differed from past work activity, under some conditions, rapid urea degradation may pose
and ensured more vigorous enrichment for ureolysis by additional challenges with regard to spatial control and treatment
compounding the selection within each treatment. Further
423
extent. It is recommended, instead, that stimulation approaches impact of increasing initial pH and added yeast extract, both of
not be generalized, but rather contemplated as part of some which increased bulk ureolytic rates by increasing selectivity for
particular treatment design, with the requirement that some ureolysis and increasing total cell densities, respectively. Figure
minimal ureolytic activity be achieved. Important factors that 20c shows a comparison of urea degradation in similar soil
may affect the behavior of a given enrichment as they are columns receiving stimulation solutions containing varying
currently understood are presented here. ammonium (12.5 mM to 200 mM) that were adjusted to an initial
pH of 9 as well as the response of a single column receiving a
3.2.3.1 Selectivity of medium previous stimulation solution formulation that was not pH
Generally, stimulation solutions have been designed to maximize adjusted and contained 12.5 mM ammonium following solutions
the selective pressure for ureolysis, thereby allowing for soils similar to that used in Gomez et al. (2016). As shown, urea
resistant to ureolytic enrichment to have the best opportunity to degradation is observed more rapidly when ammonium
stimulate. Initial solution pH and ammonium concentrations can concentrations exceed 12.5 mM and increase to 100 mM;
effectively serve as the primary mechanisms that enforce this however, with no further improvement observed at the higher
selection process. These factors are thought to stress or kill concentration (200 mM). Following these results, current
bacteria intolerant of urea hydrolysis end-products before they stimulation solutions have both included an initial pH of 9.0, an
utilize provided nutrients for further growth. Added stimulation ammonium concentration of 100 mM, and have proven
solution ammonium concentrations between 12.5 mM and 200 successful for a variety of natural soils. For more recalcitrant
mM as well as initial pH increases from 7.0 to 9.0 have been soils, however, increasing the availability of membrane-
examined and shown to impact selectivity as measured by permeable ammonia concentrations through an increase in pH (to
observed urea hydrolysis magnitudes (Gomez et al. 2018b). > 9) or the addition of substantially higher ammonium
Figure 20. Select data from stimulated soil columns receiving stimulation solutions with (a,b) varying yeast extract and initial pH conditions and (c)
varying ammonium concentrations with alkaline pH adjustment highlight the benefits afforded by higher yeast extract additions, alkaline pH adjustment,
and higher initial ammonium concentrations. Plots (a) and (c) present residual urea remaining after treatment residence periods versus daily treatment
injection number and plot (b) presents numbers of treatments required to achieve 250 mM urea degradation versus yeast extract concentration from the
earlier data set present in (a) (adapted from Gomez et al. 2018b).
Figure 21. Select data from stimulated soil columns receiving stimulation solutions with varying urea during stimulation and cementation highlighting
the impact of added urea concentrations during stimulation on the efficacy of degradation during cementation. Data presents urea concentrations
measured at the start and end of each treatment retention period versus treatment number during both stimulation and cementation. Data is from columns
receiving (a) 0 mM, (b) 50 mM, and (c) 350 mM urea during stimulation and 350 mM during cementation whereas (d) received 350 mM urea during
stimulation but only 250 mM urea during cementation. Injected urea concentrations are indicated by dashed lines and stimulation treatments are indicated
by shading. (adapted from San Pablo et al. 2022a).
Figure 20a and 20b present a subset of these data highlighting the concentrations could apply greater selective pressure and
424
increase the probability of stimulation success. Such cost, this excess is unlikely to be detrimental and may or may not
modifications in employed solutions present unavoidable merit further optimization.
tradeoffs with respect to material use and byproduct production In addition to acetate, other complex sources of organic
and therefore should be carefully contemplated. carbon are provided in stimulation solutions, of which the
preferred form is presently yeast extract. Complex sources of
3.2.3.2 Supplied urea concentrations organic carbon are preparations unable to be chemically defined
Although somewhat related to the selectivity of the medium, that are used to provide a broad array of organic molecules for
variations in supplied urea concentrations during stimulation uptake and utilization. Classically, these complex carbon
may be a critical factor controlling stimulation success. Figure additions can satisfy the growth requirements of all but the most
21 presents results from a series of experiments by San Pablo et fastidious organisms ensuring that the medium satisfies
al. (2022a) where four soil columns received either 0 mM, 50 fundamental macro and micronutrient requirements. If media
mM, or 350 mM urea during stimulation with 350 mM urea were to be provided with defined organic carbon alone (e.g.,
during cementation or 350 mM urea during stimulation but only acetate), enriched microorganisms in natural soils would be
250 mM urea during cementation. All stimulation solutions subject to rapid limitations by nitrogen, sulfur, phosphorous, or
contained 100 mM ammonium chloride, 42.5 mM sodium other essential nutrients. The addition of yeast extract in
acetate, 0.2 g/L yeast extract, and the pHinitial was 9. All solutions alleviates those limitations and allows for increased
cementation solutions contained 250 mM calcium chloride, 12.5 rates and magnitudes of growth, which assist in the generation of
mM ammonium chloride, 42.5 mM sodium acetate, and 0.2 g/L bulk microbial activity. By the same token, reducing the complex
yeast extract with a non-adjusted pHinitial of 8.3. The column carbon in stimulation solutions can impose limits to microbial
receiving no urea during stimulation (Figure 21a) exhibited growth and, by extension, can regulate achieved bulk ureolytic
minimal ureolytic activity upon the initial transition to activity. Recent experiments have shown that increases in
cementation. In contrast, the columns that received 50 mM supplied yeast extract can increase the speed at which ureolytic
(Figure 21b) and 350 mM urea (Figure 21c, d) during stimulation activity is achieved in biostimulated soils, decreasing the time
both achieved sufficient ureolytic activity to hydrolyze all and number of treatments required to achieve sufficient
provided urea in the ~23.5 hour residence of the first and enrichment (Figure 20a, b). Conversely, decreases in supplied
subsequent cementation treatments. A reduction in urea during yeast extract have been shown to reduce bulk ureolytic activity
cementation also appeared to have minimal impacts on ureolytic (Figure 22) and may help reduce non-uniformity that emerges
activity during cementation (Figure 21d). Collectively, these from reactions during transport (San Pablo et al. 2020, Lee et al.
results suggest that even with highly selective stimulation 2022). The spectrum of yeast extract concentrations ranging from
solutions, the addition of some amount of urea during stimulation a lower bound that may be soil- and condition-specific to ~0.2
may be critical to enrichment success. Notably, the response of g/L provides the greatest latitude with respect to ureolytic rate
the column receiving 50 mM during stimulation was similar to control. As previously mentioned, and more fully discussed in
the column receiving 350 mM urea during stimulation, Section 7.2, the impact of ureolytic rate on biocementation
suggesting that even low concentrations near 50 mM may be distribution is substantial so the types and amounts of organics
sufficient. Such results have in part provided the basis for used may be of primary importance.
reductions in urea during simulation, as demonstrated by San
Pablo et al. 2020 and the authors’ more recent studies. 3.2.3.4 Supplied electron acceptors
As described earlier, the proposed stimulation protocols have
3.2.3.3 Supplied organic carbon attempted to minimize the use of fermentable carbon sources to
Organic carbon serves two essential functions in biological avoid the potential for acid formation. The direct outcome being
systems, providing both an energy source and a source of carbon that the stimulation protocols are strongly dependent upon the
needed for biosynthesis. It is suggested that non-fermentable energy generated from respiration. The authors’ current
sources of organic carbon be provided in stimulation solutions to stimulation solutions also do not explicitly include alternate
minimize the potential for acidogenesis and the reversal of electron acceptors, and therefore it is presumed that they rely
generated biocementation. While glucose and other sugars are heavily upon equilibrium oxygen dissolved concentrations to
prototypical sources of organic carbon in microbial media, under enable energy generation. It should be acknowledged, however,
anoxic conditions such additions can be readily fermented by a that small quantities of such acceptors can be found in the
host of soil microbes to furnish organic acids, which decrease complex organic source provided, although they are likely at
solution pH. Many other carbohydrates can likewise be negligible concentrations. It may then be essential to ensure
fermented to produce organic acids. Molasses, as an example, is atmospheric equilibration of stimulation solutions prior to
a relatively inexpensive carbon source that has seen frequent injection to maximize the oxygen available for cellular growth
application for stimulated MICP. Molasses is composed of and utilization of the supplied growth medium. This is especially
approximately 96% sugar by effective dry mass (Clarke 2003, important for saturated soils where oxygen diffusion is limited
USDA ARS 2021), giving it substantial fermentative potential if and dissolved oxygen will rapidly become limiting. When and if
not used sparingly. Acetate is often provided as a non- dissolved oxygen requirements prove limiting, alternate electron
fermentable carbon source, though lactate is an alternative with acceptors may need to be considered to maintain respiration and
comparable geomicrobiological significance. When supplied in energy generation after oxygen has been depleted. It remains
stimulation solutions, only 125 µM acetate is needed to consume unclear, however, if ureolytic bacteria that are enriched using our
the 250 µM dissolved oxygen concentration measured in currently recommended solutions, which do not contain
atmosphere-equilibrated media, rendering them anoxic. additional electron acceptors, have the respiratory machinery to
However, given that a substantial portion of acetate is likely to utilize other electron acceptors in order to extend their growth in
be used for biosynthesis rather than energy, suppling millimolar the absence of oxygen. A recent study employing varying
quantities are likely more than sufficient for efficient aerobic percentages of natural seawater as the treatment solvent (Yanez
growth. The proposed stimulation solutions provide acetate at a et al. 2022) suggests that ureolytic enrichment was not
concentration of 42.5 mM, which may be in 1.5-2 orders of substantially affected by the presence of up to ~28 mM sulfate in
magnitude in excess of what is strictly required for aerobic the applied media (Figure 23), suggesting that the utility of other
growth. While perhaps in excess, aside from the added material alternative electron acceptors may be less than clear.
425
Furthermore, the small but consistent decrease in ureolytic maintain a steady-state biomass rather than fuel substantial
activity observed with increasing seawater concentrations may growth. Additionally, since ammonium is readily generated upon
be indicative of other inhibitory species present in seawater (e.g., introduction of cementation treatments due to hydrolysis of the
Mg2+) which have been noted in purely augmented MICP supplied urea, cementation solutions similarly include reduced
experiments (Burdalski 2020). initial concentrations of ammonium and reduced initial pH with
the assumption that ureolytic activity during the treatment
3.2.3.5 Cementation processes following stimulation residence period will increase selectivity in time. Cementation
Upon the transition from stimulation to cementation, special solutions, then, apply reduced selective pressure by lessening the
considerations must be made to enable the continued selection ammonium additions and foregoing pH adjustment and
and growth of a select subset of soil bacteria. Aside from furthermore achieve more minimal pH rise during the treatment
allowing for generation of CaCO3, cementation solutions applied period due to simultaneous carbonate consumption during
following stimulation must provide conditions that allow for CaCO3 precipitation. This can be controlled in part by varying
microbial growth to replace cells lost during precipitation and the urea concentration supplied during cementation. As shown in
washed out during subsequent injections, while also ensuring Figure 24, when urea concentrations were varied between 250
that the achieved microbial community maintains sufficient mM and 350 mM and exceeded the concentration of supplied
ureolytic activity. Cementation solutions, then, attempt to calcium (which was constant at 250 mM), increases in final pH
maintain similar conditions experienced during stimulation values could be achieved during the cementation process. While
treatments, but include enough calcium and urea to achieve final pH values during cementation are clearly reduced from
efficient CaCO3 precipitation. That said, some additional values observed during stimulation, which approach 9.5, excess
urea concentration variations provided the opportunity to vary
final pH values between 8 and 9 and accordingly apply more
selective pressure for ureolysis. Although not encountered in the
authors’ experiments (e.g., Figure 21d), if insufficient selective
pressure were to be provided during cementation, it is possible
that a microbial community could shift and undesirable
reductions in ureolytic activity and incomplete precipitation
could occur, thus urea concentrations may provide a means by
which this outcome could be mitigated. Additionally, as
discussed in Section 3.3.2.5, the transition from biostimulation to
cementation typically includes a flush treatment intended to
mitigate abiotic calcium carbonate precipitation during the
introduction of high calcium. This flush almost certainly reduces
Figure 22. Initial bulk rates determined from columns receiving Figure 23. Initial bulk urea hydrolysis rates obtained from experiments
stimulation treatments (S) containing 50 mM urea and varying yeast performed on Beach Sand (BS) and an alluvial Concrete Sand (CS)
extract concentrations and cementation treatments (C) containing similar receiving stimulation and augmentation treatments of varying seawater
yeast extract concentrations with either (a) 350 mM or (b) 250 mM urea. compositions. Initial rates were determined from first-order fitting of
Provided letters and numbers refer to treatment type and injection urea degradation curves. As shown, minimal decreases in ureolysis rates
numbers, respectively. Initial rates were determined from first-order were observed with increasing seawater concentrations (adapted from
fitting of urea degradation curves (data from Lee et al. 2022). Yanez et al. 2022).
changes are merited to meet the different requirements of the the established microbial densities by eluting some portion of the
cementation process. It is thought that less selective pressure is resident bacterial population. While this may render the system
required for maintenance of established stimulated ureolytic more vulnerable to secondary, competitive enrichment, wherein
populations than that required to initially convert and develop the there has been detectable ureolytic activity established following
ureolytic microbial community. Thus, the proposed cementation stimulation, issues regarding the transition from stimulation to
solutions, as discussed in Section 3.3, typically use less organics cementation have not been encountered. Instead, while a
in comparison to stimulation solutions, as they are intended to reduction in ureolytic activity has been observed immediately
426
3.3.1 History of cementation approaches
Although the first cementation formulations were proposed by
Ferris et al. (1997), applied solutions have been continuously
modified by researchers who have changed urea to calcium
concentrations, added and removed treatment additives, and
shifted the types of microorganisms used and requiring
sustainment during the process. Perspective on cementation
solution formulation evolution is evident in Table 1, which
includes the original MICP cementation formulation developed
by Ferris et al. (1997) and subsequently used by Stocks-Fischer
et al. (1999) and others, three other formulations showing how
such treatments solutions have been further modified, and two
more recent formulations currently used by the authors for most
of their research (Gomez et al. 2016, San Pablo et al. 2020). As
shown, early formulations (Stocks-Fischer et al. 1999, Bang et
al. 2001, DeJong et al. 2006) remained relatively unchanged
from Ferris et al. (1997) with a transition to cementation
solutions consisting of purely urea and calcium salts at generally
much higher concentrations by Whiffin et al. (2007), van Paassen
et al. (2010b), Chu et al. (2012), and others. As time progressed,
cementation formulations appear to generally fall into two
camps: (1) modification of the Ferris et al. (1997) formulation
and (2) modification of pure urea and calcium formulations
Figure 24. Select data from stimulated soil columns receiving stimulation similar to that by Whiffin et al. (2007). Formulas by Martinez et
solutions with 50 mM urea and cementation solutions with urea varying al. (2013) and Montoya and DeJong (2015) represent minor
between 250 mM and 350 mM and calcium at 250 mM (constant), departures from Ferris et al. (1997) while those by other
highlight the impact of urea concentrations during cementation on final researchers including Tobler et al. (2018) remain similar to the
pH conditions (adapted from San Pablo et al. 2022a).
Whiffin et al. (2007) formulation. Perhaps the most significant
after the flush treatment, this activity generally rebounds during departure from both formulations has occurred as the result of the
cementation to exceed what was observed during the final transition to stimulation. Although stimulation of MICP was first
stimulation treatments. Nevertheless, the transition from intentionally investigated by Ferris et al. (1997) and later
stimulation to cementation treatments has the potential to be successfully demonstrated by Fujita et al. (2008) and others for
perilous and it is recommended to maintain similar selective strontium immobilization, similar strategies for geotechnical
pressures during cementation as were present during stimulation, applications resulted in treatment formulations that started
with modifications made only once the stability of a given deviating most significantly following Burbank et al. (2011),
ureolytic population enriched in some soil material is better who proposed a cementation formulation similar to Ferris et al.
understood. (1997) but included additional substrates likely intended to
maintain enriched ureolytic cell populations throughout the
3.3 Development of cementation approaches cementation process. Building from these formulations,
modifications by the authors presented in Gomez et al. (2014,
Through over two decades of development MICP treatment
2017, 2018b) and others have led to further optimization of these
solutions have been progressively modified to enhance
cementation solutions for both enhanced material-efficiency,
cementation mineral products, improve control of the
retention of ureolytic activity during the cementation period, and
geochemical process, and minimize energy, materials, and
improved control of precipitation location. Eventually, this has
environmental impacts among other objectives. While one would
led to the authors’ most recent formulations presented in San
be hard pressed to suggest that “optimal” cementation solutions
Pablo et al. (2020), which will be discussed in greater detail in
and techniques exist, continuous advances with respect to the
Section 7.2.
collective understanding of the MICP process have manifested in As cementation solution formulations have been modified, so
a variety of modifications to the way researchers perform too has an understanding of the role of specific additives. For
biocementation, including those pertaining to differences in example, while early formulations include the addition of sodium
treatment injection sequences, cementation solution bicarbonate with the intention for such additions to serve as a pH
formulations, and even the types of microorganisms used to buffer, the addition of bicarbonate also increases the potential for
hydrolyze urea and catalyze the biocementation process. In some abiotic precipitation by eliminating the ability of urea hydrolysis
cases, these modifications have been driven by improved to solely control carbonate generation. Similarly, ammonium
understandings of the fundamental science behind the MICP chloride has been included at high concentrations, as it was
process including those pertaining to precipitation saturation perceived to slow urea hydrolysis reaction rates through product
state. In other cases, researchers may not have been able to inhibition. While such concentrations were shown to inhibit free
clearly state the mechanisms responsible for specific differences, enzymes dramatically (Fidaleo & Lavecchia 2003), in other
but have provided empirical evidence supporting hypotheses and
their preferred treatment techniques. Rather than provide a fully
inclusive summary of all past treatment solutions, in this section,
the authors instead provide a brief history of cementation
solution formulations, followed by a discussion of the
modifications made to their treatment processes including the
rationale for these changes and recommendations for
cementation treatment solution formulations. Lastly, some more
poorly understood factors will be discussed along with recent
results that highlight the unexpected impact of solution changes.
427
Table 1. Progression of cementation solution development.
Augmentation or
Year Reference Cementation Solution Formulation Commentary
Stimulation
333 mM of Urea; 25.2 mM Calcium
Augmented with
1997 Ferris et al. Chloride; 187 mM NH4 Cl; 25 mM Sodium -
Bicarbonate; 3 g/L Nutrient Broth
333 mM of Urea; 25.2 mM Calcium
Augmented with
1999 Stocks-Fischer et al. Chloride; 187 mM NH4 Cl; 25 mM Sodium Identical to Ferris et al.
Bicarbonate; 3 g/L Nutrient Broth
Augmented with Removal of Additives; Use of
2007 Whiffin et al. 1100 mM Urea; 1100 mM Calcium Chloride
Higher Concentrations
Augmented with Removal of Additives; Use of
2010b van Paassen et al. 1000 mM Urea; 1000 mM Calcium Chloride
Higher Concentrations
333 mM Urea; 100 mM Calcium Chloride; Similar to Ferris et al.; Modification
Augmented with
2013 Martinez et al. 374 mM NH4 Cl; 25.2 mM Sodium of Urea-to-Calcium Ratios &
Bicarbonate Additive Concentrations
350 mM Urea; 250 mM Calcium Chloride; Transition to Stimulation; Addition
2017 Gomez et al. Stimulation 12.5 mM NH4 Cl; 42.5 mM Sodium Acetate; of Growth Factors and Modification
0.1 g/L Yeast Extract of Urea-to-Calcium Ratios
250 mM Urea; 250 mM Calcium Chloride;
Further Optimization of Growth
2020 San Pablo et al. Stimulation 12.5 mM NH4 Cl; 42.5 mM Sodium Acetate;
Factors and Urea-to-Calcium Ratios
0.2 g/L Yeast Extract
studies involving whole cells, such effects appear to be much recent years. These choices have been made for several main
more minimal (Lauchnor et al. 2015). Thus, from a cementation reasons: (1) the transition from the use of pure augmentation to
perspective, increasing ammonium chloride concentrations may almost solely stimulated microbial communities to complete
only increase generated byproducts, albeit with some potentially precipitation reactions, (2) the desire to remove superfluous
significant impacts on the activity of augmented and stimulated additives and reagents to minimize environmental impacts,
ureolytic bacteria (Section 2). Similar to cementation solutions financial costs, and byproduct concentrations, and (3) the desire
proposed by Whiffin et al. (2007), the only essential components to enhance spatial uniformity and control of the process. The
of a cementation solution are sufficient urea and calcium. Of most significant differences can be summarized as follows: (1)
course, urea is necessary to enable the generation of carbonate reductions in calcium and urea concentrations, (2) reductions in
species and calcium is an essential component of calcium urea-to-calcium ratios, (3) addition of nutrients and trace growth
carbonate precipitation. While the other additions early on were factors in cementation treatments and the elimination of
primarily intended to control reaction speed and pH, it is the augmentation, (4) elimination of additives including sodium
authors’ opinion that for purely augmented experiments, many of bicarbonate and reductions in ammonium chloride, and finally
these additives could have been removed with little consequence. (5) the addition of “flush” injections to avoid spontaneous abiotic
This is clear in the works of Whiffin et al. (2007), van Paassen et precipitation events. Each of is described briefly below, along
al. (2010b), Chu et al. (2012), and many others who with the justification for why such decisions were made.
demonstrated successful cementation using simple solutions
following augmentation with ureolytic bacteria. When 3.3.2 Specific considerations for cementation
considering the formation of cementation, the addition of As presented above, specific factors were considered in the
calcium carbonate forming reagents is key, but so too is the development of the current cementation solutions used. While
viability of the ureolytic cells completing the process. This was cementation treatment formulations are presented largely in the
recognized by Ferris et al. (1997) and others who included context of stimulated MICP, specific considerations for and
various components such as nutrient broth and yeast extract modifications to cementation solutions have occurred over time
likely intended to maintain or increase densities of augmented and the authors’ understanding of these factors are described
cells. Interestingly, such additives may have also unintentionally briefly below.
led to stimulation of indigenous bacteria (Section 2.3.2). As the
potential of enriched native ureolytic bacteria to complete this 3.3.2.1 Urea and calcium concentrations
process has been increasingly recognized, these additives have The selection of urea and calcium concentrations, although rarely
started to play a more critical role during the cementation explained, is an aspect of the cementation solution meriting
process, to both provide nutrients needed to maintain these careful consideration. Concentrations of urea approximately 333
populations but also poise the system such that enriched bacteria mM have long persisted in cementation solutions most likely out
can remain dominant in otherwise mixed bacterial cultures. It is of deference to those first used by Ferris et al. (1997). At the
the authors’ opinion that this aspect is where most of the novelty
higher end of the spectrum, concentrations at and above 1 M urea
regarding cementation solutions has occurred and has led to the
and calcium similar to those used by Whiffin et al. (2007) are
primary differences proposed in the solution recommendations
also commonly found in the literature. Beyond preference,
discussed later.
While the understanding of the cementation process is researchers using higher concentrations have likely done so to
continuously evolving, the authors have made several significant achieve higher magnitudes of cementation per injection, thus
departures from even their own past treatments techniques in potentially minimizing implementation costs and reducing
428
treatment time. Alternatively, higher concentrations can also include such factors in cementation solutions. While the specific
result in more unpredictable urea hydrolysis behavior due to at organic additives used have changed over time (e.g., yeast extract
least two competing effects. Ureolytic rates increase with versus nutrient broth, the addition of sodium acetate), the more
increases in urea concentration based on the KM values of S. fundamental objective has remained the same – to allow for
pasteurii and commonly enriched species (Graddy et al. 2018, sustained microbial activity throughout the cementation process.
Lauchnor et al. 2015). But at very high concentrations, the It seems intuitive that in the absence of organic substrate renewal,
resultant ammonia may be inhibitory or cytotoxic and result in ureolytic activity will decrease over time, due to an inability of
reduced ureolytic activity. For example, only S. pasteurii and one cells to meet their metabolic needs and the precipitation of
other strain among the 26 Gram positive bacterial species tested CaCO3, which can encapsulate and stress cells over progressive
grew at 1.4 M total ammonium (Leejeerajumnean et al. 2000), injections. Thus, treatment additives such as yeast extract and
though the upper limit for maintaining ureolytic activity for S. sodium acetate have been included during the cementation
pasteurii or other stimulated strains has not yet been determined. process to maintain ureolytic activities generated in the
Aside from impacts on ureolysis, higher calcium concentrations enrichment period. When modified during cementation
can also increase precipitation rates, which are oftentimes treatments involving enriched microorganisms, however, the
assumed to be proportional to saturation states. Although the literature has only begun to understand the utility of such
authors admit the lower concentrations used in their formulations concentrations to control ureolytic cell densities and alter
(~250 mM urea and calcium) have been selected primarily out of reaction rates. From this context, the concentrations of organic
experience, higher concentrations present the possibility of carbon, energy sources, and growth factors should be carefully
inhibiting enriched microbial growth, altering urea hydrolysis selected to target some ureolytic activity that will be reached
and calcite precipitation kinetics thus impacting spatial control, once the system approaches a steady condition. Additionally,
generating effluent with elevated by-product concentrations, and when such nutrients are provided in gross excess, the potential
even influence of the mineralogy and morphology of achieved for fermentation of supplied carbohydrates may become a
precipitates with more soluble phases including amorphous significant possibility, thereby potentially reversing generated
CaCO3 and vaterite becoming increasingly possible at higher precipitates through acid-driven dissolution. While there is much
reagent concentrations. Although the impacts of altering such to understand about the potential of such substrates in the
concentrations may be specific to the formulations employed, the stimulation process, it is clear that this treatment decision may
use of higher concentrations (> 500 mM urea and calcium) may have important implications with respect to both maintaining
present more risks than benefits. ureolytic activity process and altering the spatial control of the
process (Section 3.2.3.3).
3.3.2.2 Reductions in urea to calcium ratios
The notion is pervasive throughout MICP literature that pH rise 3.3.2.4 Removal of sodium bicarbonate and high ammonium
is needed in order for CaCO3 precipitation to occur. Following chloride
the discussion in Section 2, however, it should be clear that this Although many different treatment additives exist, two of the
is not needed, but rather the product of the activities of calcium most consistent additions to cementation solutions are sodium
and carbonate simply must exceed the solubility product of some bicarbonate and ammonium chloride. Sodium bicarbonate,
CaCO3 mineral phase. Through extensive modeling of the MICP although included by some researchers to act as a pH buffer
reaction network and physical experimentation, it can be (Ferris et al. 1997, Stocks-Fischer et al. 1999, Al Qabany & Soga
concluded that providing urea at concentrations exceeding that 2013), does have the potential to undermine biological mediation
of calcium provides almost no measurable benefit with respect to of the MICP process due its abiotic reactivity with added calcium
precipitate formation provided that the supplied urea is fully resulting in abiotic precipitation. This could limit distribution of
hydrolyzed. The authors have tended to use solutions with reactants during injections among other potential issues such as
concentrations of urea exceeding calcium particularly when injection source clogging. Secondarily, the addition of sodium
performing stimulated experiments in order to maintain higher bicarbonate likely provides very little additional buffering as
pH and ammonium concentrations during cementation and such solutions will undoubtedly become dominated by much
therefore maintain selective pressure favoring ureolytic higher ammonium and carbonate concentrations produced during
microorganisms. Recent experiments suggest, however, that urea hydrolysis, although it is acknowledged that such additions
excess urea may not be needed and communities can seemingly could play some microbiological role. However, given these
remain stable over 10 daily cementation treatments even when additional limitations and added material costs, bicarbonate
large pH rises do not occur (Figure 21d, 24). It is the authors’ additions may be largely superfluous. In addition to sodium
opinion that equimolar concentrations of urea and calcium be bicarbonate, ammonium chloride is also commonly provided in
used to minimize byproduct production and urea consumption, cementation solutions. While the authors believe that
and that the lack of excess urea has minimal impacts on bulk concentrations exceeding 150 mM such as those provided by
ureolytic activity and precipitate magnitudes. There are special Ferris et al. (1997) and others may be excessive, such
conditions, however, where excess urea could be advantageous concentrations likely do play a critical role in applying selective
including: (1) acidic soils requiring greater alkalinity generation pressure for ureolysis and may be critical toward maintaining
to make sufficient concentrations of carbonate available, (2) ureolytic activity. Given that ammonium is generated as a by-
conditions where pH spikes indicative of Ca2+ consumption may product of urea hydrolysis, the authors have shown that supplied
be favorable for process monitoring (Section 4), and (3) when ammonium chloride concentrations may need to be higher
additional selective pressure is needed to ensure that stimulated initially to maintain greater selection for ureolysis, but may be
ureolytic microorganisms remain dominant. reduced to values of 12.5 mM or below once sufficient ureolytic
activity is observed and the enrichment becomes more stable.
3.3.2.3 Addition of nutrients & elimination of augmentation Since ammonium byproducts are undesirable from an
While early formulations by Ferris et al. (1997) and others environmental perspective, any reductions in ammonium
recognized the need to include growth factors to maintain chloride concentrations in treatment solution may be beneficial
augmented cell populations throughout the cementation process and aid in the post-treatment rinsing and waste-management
if re-augmentation is to be avoided, the recent use of indigenous processes.
ureolytic microorganisms has only further solidified the need to
429
3.3.2.5 Addition of flush treatments formulations that have been identified following the
The application of urea-rich solutions during stimulation results advancements described earlier. It should also be mentioned that
in the enrichment of native ureolytic bacteria with the production certainly other treatment solution formulations can and have
of alkaline ammonium- and carbonate- rich effluent. This been used with success and such recommended formulations are
presents problems with the application of the first cementation specific to stimulated treatment techniques. Table 2 presents the
treatment wherein calcium is first introduced into an otherwise composition of the authors’ baseline cementation solution
calcium-limited system. If not mitigated, cementation solution formulation along with comments regarding the potential
injections will react with carbonate-rich stimulation effluent modifications that researchers may want to consider with respect
upon mixing, leading to substantial abiotic calcium carbonate to this composition. As shown, the main decision points may
precipitation. This precipitation oftentimes occurs in free include: (1) the modification of yeast extract concentrations to
solution thereby blocking soil pore spaces and well filters, among target terminal ureolytic rates optimal for some treatment
other problems, and results in a decrease of soluble calcium that application (i.e. flow rate, treatment residence period, desired
may reduce cementation at further distances. Although not cementation distribution), (2) the modification of urea and
strictly related to the cementation solution itself, transitional calcium concentrations to higher values while still maintaining
treatment solutions, termed “flush” solutions, have been urea to calcium ratios near 1, and (3) potential increases in
employed to preclude these behaviors. First employed in Gomez ammonium chloride concentrations when the potential for
et al. (2016) and later used in Gomez et al. (2018b) and San Pablo enrichment failure is high (e.g., when cell densities of competing
et al. (2020), flush treatment solutions are similar to the microorganisms are high and threaten to undermine the
composition of cementation solutions but lack calcium (and enrichment for ureolysis). Past research suggests that the most
sometimes other amendments). These solutions are applied dynamic changes in these behaviors may be observed when yeast
immediately after stimulation and just before applying extract is varied between ~0.02 and 0.2 g/L and ammonium
cementation solutions to remove high aqueous carbonate species chloride is varied between ~12.5 mM and 100 mM, while further
present following the end of stimulation. Figure 25 illustrates the recommending that care should be taken when using urea and
impact of the flush treatment on solution carbonate as indicated calcium concentrations in excess of 500 mM to prevent potential
in an activity ratio diagram (Gomez et al. 2019) highlighting the inhibition by high concentrations and unintended consequences
reduction in carbonate activities and saturation states achieved with respect to biological function and mineral products (i.e.,
with a flush injection, thereby minimizing the potential for mineralogical and morphological differences).
abiotic precipitation. When augmentation is performed using
urea-containing cell suspensions, similar flush solutions can be 3.4 Ammonium byproduct management and removal
employed prior to cementation to avoid practical issues related
The produced ammonium resulting from microbial urea
to clogging and losses in cementation from spontaneous
hydrolysis may present serious human health and environmental
precipitation of CaCO3 upon the introduction of calcium.
impacts if left untreated in subsurface soils (DeJong et al. 2013,
3.3.2.6 Recommended cementation solution composition Gat et al. 2017). Although few researchers have investigated
While decisions regarding treatment solution formulations ammonium removal following MICP, it has been proposed that
should be made with reference to the specific applications generated ammonium may be removed through a variety of
considered, the authors do have several primary treatment processes including subsequent biological transformations of
Figure 25. Example of solution carbonate and calcium activity changes following stimulation/augmentation, a single flush treatment, and a single
cementation treatment as monitored at various locations in a large meter-scale tank experiments. As shown, at locations receiving concentrated solutions
(Locations A, D) in both the augmented and stimulated tanks, aqueous carbonate activities were reduced substantially after flushing (adapted from Gomez
et al. 2018a).
430
Table 2. Baseline cementation following stimulation (summarized from Recent studies have considered injecting cation-amended
San Pablo et al. 2020). solutions to remove ammonium-rich MICP effluent from
previously biocemented soils via solution advective transport
and cation-exchange with adsorbed ammonium ions. In this
Solution Baseline Cementation process, cations are provided to exchange with ammonium ions
at soil surfaces, thereby reducing adsorbed ammonium
concentrations present following the MICP process and
1
Urea (mM) 250 accelerating aqueous ammonium transport out of treated zones.
Results from Lee et al. (2019a) suggested that interactions
1
Calcium Chloride (mM) 250 between positively charged ammonium ions and negatively
charged, poorly-graded sand surfaces may result in significant
2
Ammonium Chloride (mM) 12.5 retardation of ammonium transport when compared to passive
ions (Figure 26). In addition, a significant fraction of injected
Sodium Acetate (mM) 42.5 3
ammonium was retained within columns suggesting that such
surfaces may retain sorbed ammonium. It was expected that such
Yeast Extract (g/L) 0.2 4 phenomena would likely be influenced by both surrounding pore
fluid conditions as well as the cation exchange capacities of
1
treated soils. Lee et al. (2019a) further demonstrated that the
Urea and calcium magnitudes may be increased to values near
500 mM, however, urea-to-calcium ratios near 1.0 remain
2
Increases in ammonium chloride concentrations may be
considered when there is a potential for instability of the
enrichment (i.e., conditions where other microorganisms may
outcompete ureolytic bacteria for supplied nutrients).
3
Recent results suggest that the addition of sodium acetate may not
be critical.
4
Yeast extract concentrations can be modified to control ureolytic
rates (higher concentration = higher cell densities = higher ureolytic
rates).
generated nitrogen species in situ and removal of ammonium-
rich solutions using post-treatment injections with later
remediation of extracted ammonium occurring ex situ.
Numerous past studies have shown that generated ammonium
concentrations following MICP are approximately twice that of
the applied urea in accordance with reaction stoichiometry. Such
concentrations may surpass values above 500 mM, which far
exceed expected groundwater concentrations and some
regulatory limits. Although total ammonium limits have not been
established for drinking water in the U.S., the U.S.
Environmental Protection Agency recommends maximum
ammonium concentrations for aquatic life of ~1 mM and ~0.1
mM at a pH of 7.0 for acute (1 hour) and chronic (30-day)
exposure, respectively (EPA 2013). Furthermore, the European
Union recommends maximum ammonium concentrations for
drinking water of ~0.03 mM (European Union 2020). These
regulations are both location and application specific and reflect
the potential impacts of ammonium concentrations on human
health and environmental quality. For example, in some locations
maximum concentrations are pH-dependent due to differences in
the inherent toxicity of ammonium when speciated as charge-
neutral ammonia. For example, the B.C. Contaminated Sites
Registry (British Columbia Contaminated Sites Regulation Figure 26. Breakthrough curves during ammonium bromide injections in
2020) suggests that acceptable total ammonium levels may be soil column experiments suggest that the transport of ammonium is
higher under more acidic conditions (pH < 8.5) due to the inhibited relative to passive ions such as bromide due to interactions with
increased speciation of total ammonium as ammonium and lower soil mineral surfaces. Limited appearance of ammonium at column
under more basic conditions (pH > 8.5) due to the increased outlets, suggested that appreciable sorption of ammonium occurred in the
speciation of total ammonium as ammonia. While such considered clean sand material (adapted from Lee et al. 2019a).
regulations limit concentrations of aqueous ammonium, it should addition of calcium ions to rinse solutions and adjustment of
also be mentioned that positively charged ammonium ions can rinse solution pH to more alkaline values may result in enhanced
sorb to negatively-charged soil surfaces. Thus, post-treatment ammonium removal likely due to both improved cation exchange
ammonium removal following MICP may require careful and increased speciation of residing ammonium as charge-
monitoring of both aqueous and adsorbed concentrations. neutral ammonia while also limiting the potential for CaCO3
Adsorbed ammonium concentrations may pose long-term dissolution. Although such techniques have been largely
management challenges due to the potential for desorption over developed at the centimeter bench-scale, recent studies have
time when such quantities are not preemptively addressed and demonstrated the effectiveness of post-treatment rinsing
monitored during the remov processes at meter-scale under more practically-relevant
431
treatment conditions. Lee et al. (2019b) characterized suppling cations during rinsing. Although not shown, staged
ammonium removal during post-treatment rinse injections with injections wherein 1 PV was applied for twelve daily injections
a single 12 PV continuous injection in five 3.7-meter-long as opposed to a single 12 PV injection were also shown to
columns containing different poorly-graded sands (San Pablo et improve both aqueous and adsorbed ammonium removal likely
al. 2020). The results suggest appreciable adsorbed ammonium due to increased cation exchange during residence periods.
may persist after rinse injections and may require careful The above advancements have improved understanding of
characterization following field-scale applications. More recent post-treatment ammonium removal. However, further
experiments by Lee & Gomez (2022a) have examined the impact investigation of these techniques remains needed. It is
of changes in supplied cations within rinse solutions as well as recommended that several aspects be monitored related to both
injection sequencing. These experiments suggest that while the progression of ammonium removal and the potential for
deionized water rinse injections can remove similar magnitudes detrimental impacts with respect to post-treatment cementation
of aqueous ammonium as other rinse solutions containing added integrity. For example, ammonium concentrations from both the
cations; adsorbed ammonium removal is significantly reduced aqueous and adsorbed phases should be monitored using both
when cations are not supplied. Potassium (K+), sodium (Na+), fluid and soil samples to assess removal success. Ammonium
and Ca2+ cation additions were explored at varying concentrations in pore fluid samples can be measured using
concentrations within rinse solutions with select results for colorimetric assays similar to that described in Krom (1980).
biocemented columns receiving a single 12 PV continuous rinse Furthermore, soil samples should be preserved at a free-draining
injection with different rinse solutions (200 mM of Na+, K+, Ca2+ water content and soil ammonium extraction should be
added rinse solutions and deionized water) presented in Figure performed following procedures similar to Keeney & Nelson
27. Interestingly, K+ ions appeared to enhance aqueous and (1983) and Lee et al. (2019b) with subsequent aqueous
sorbed ammonium removal when compared to other ions at ammonium measurements completed to account for the presence
similar concentrations. The improved performance of K+ ions of ammonium masses in residual pore fluids. Although pH
relative to other ions suggests it may be favorable for exchange measurements may not provide significant insights, it is also
with adsorbed ammonium. Additionally, deionized water rinses recommended that such measurements be completed to examine
were shown to perform poorly on the basis of reducing sorbed the impact of soil solution buffering during rinsing and other
ammonium, clearly indicating the advantages afforded by unintended reactions resulting from rinse solution cation
additions. Finally, measurements of both soil shear wave
velocities and aqueous calcium concentrations can provide the
opportunity to examine changes in cementation integrity during
rinsing which may reduce shear wave velocity values and release
measurable calcium when such injections are dissolving CaCO3
minerals. While not exhaustive, such measures have been used
in previous studies (Lee et al. 2019b) and may provide the
opportunity to quantify and compare ammonium removal
success, thereby accelerating further developments with respect
to these techniques. Although the effectiveness of cation-
amended rinse solution injections for post-treatment ammonium
removal has been demonstrated, knowledge gaps still exist
regarding a variety of aspects. For example, optimization of rinse
techniques within the context of life-cycle sustainability analyses
(LCSA) is needed to identify the impacts of specific cation
additives and identify the breakpoint at which increasing the
injected volume will increase overall impacts due to more limited
ammonium removal near the end of rinse injections. It is
anticipated that ammonium removal will remain a key aspect of
the MICP treatment process, which may accelerate more
widespread implementation of the technology. From this
perspective, further investigation of approaches which leverage
biological transformations of produced nitrogen may afford new
opportunities to enhance post-treatment remediation using
3.5 Injection scheme
The treatment injection scheme employed in the laboratory or
field is a balance between achieving treatment spatial control and
implementation simplicity, the latter being particularly
challenging and expensive when scaled to field-scale
applications (Section 7). Achieving spatial control of
cementation is primarily dependent on the ureolytic reaction rate
relative to the treatment solution transport rate. Ideally, treatment
solutions would be injected into soils without any reactions
occurring thus allowing for full solution concentrations to be
Figure 27. (a) Effluent aqueous ammonium concentrations in time during delivered into the treatment zone with perfectly spatial
continuous 12 PV injections of alkaline cation-amended solutions in
uniformity achieved by allowing solutions to react in place over
small columns show ammonium removal during rinsing with greater than
97% aqueous ammonium removed after completing injections. (b) some retention period. While this is not practically possible, a
Normalized sorbed ammonium masses along soil column lengths given treatment strategy can be designed to minimize
obtained from soil samples after rinse treatment completion (adapted precipitation during treatment injections thereby achieving
from Lee & Gomez 2022a). improved uniformity (San Pablo et al. 2020).
432
The authors’ preferred injection strategy for treatment 3.6.1 Augmentation treatment approach
uniformity is a stopped flow injection sequence where the The suggested best practices for “pure” bioaugmented MICP as
treatment injection time is a small fraction of the total time differentiated earlier in Section 2.3.2 is described in the below
required for substantial ureolysis of the injected urea steps. It should be mentioned that other approaches have and can
concentration (and associated consumption of calcium), be used with success, however. Thus, these recommendations
followed by a retention time period that meets or exceeds the provide solely the authors’ suggested approach with
required reaction time (Barkouki et al. 2011). In laboratory tests implementation recommendations made as appropriate.
on ~125 cm3 columns treated with 1.25 pore volumes using a Step 1: Grow the desired ureolytic bacteria under aseptic
peristaltic pump, such treatment injections can oftentimes be
conditions in recommended growth medium to a
completed in under 20 minutes. This is followed by a retention
predetermined optical density (OD600) that represents late
period that exceeds the expected time required for cell growth,
exponential phase of growth, as judged by a previously
urea hydrolysis, and/or precipitation dependent on the treatment
determined growth curve. Cell densities corresponding to the
stage with treatment intervals oftentimes being once every 24 or
48 hours, with adaptations possible as reaction activities increase target OD600 should be predetermined as outlined in Section
or decrease. Following stimulation, the authors’ stimulated soil 4.1.1 f
column experiments suggest that consumption of 250 mM urea this step is to obtain a culture that is as dense as possible
and calcium can be completed between ~3 and 48 hours, without substantial death or a reduction in specific ureolytic
dependent on targeted reaction activity. Ideally, injections should activity. Actual achievable cell densities will be limited by
be performed over time periods that exceed no more than 10% of rate of oxygen transfer, build-up of toxic waste products, and
the total time required for full hydrolysis. Aside from more other factors under the selected growth regime.
obvious benefits with respect to uniformity, mineralogical Step 2: Cell densities targeted for the final added cell
transformations over time observed and discussed in Section 2 suspension should be determined based upon project
clearly suggest that ripening of precipitate phases toward more objectives, with special emphasis on colloidal filtration
stable calcium carbonate polymorphs may occur over time with properties of the specific soil and whether or not pure
longer treatment residence periods perhaps affording benefits augmentation is intended. The potential for unintended
with respect to material permanence. While continuous injection stimulation described earlier in Section 2 should be
schemes can also be implemented, their use will almost surely considered and mitigated if specific outcomes are to be
result in large cementation gradients when injection rates are too attributed to augmented cells alone.
slow (and large amounts of reactants are consumed during Step 3: An inert saturation solution should be applied to soil
injections) and poor chemical utilization efficiency (when
specimens (e.g., artificial groundwater) followed by the
injection rates are too fast and reactants are transported out of the
injection of prepared cell suspensions when laboratory tests
treatment zone without being fully consumed). Depending on
are to be performed. Cell suspensions for pure augmentation
particular site conditions field treatments may involve some
combination of both, with stop-flow treatments used to introduce should be prepared with rinsed and pelleted cells and
reagents and existing groundwater gradients moving injected resuspended in some medium of sufficient ionic strength to
solutions over time yielding a predictable flow regime but mitigate cell osmotic stress and potentially enhance cell
perhaps much more complicated reactive transport conditions. cases, this suspension may contain
Beyond these considerations, altering the flow direction added urea to generate ionic strength during the
between subsequent treatment injections has also been explored augmentation process and enable monitoring of urea
to improve treatment uniformity. This was attempted particularly degradation activity. Recirculation of augmented cell
in earlier studies where augmentation was used and the stop-flow suspensions can further promote more uniform distribution
methodology had yet to be fully explored and refined. of cells along the length of a so
Nevertheless, such an approach can be useful when treatment be specific to the advective-dispersive and filtration
distances are relatively large and stop-flow conditions may not properties of the porous media, however, augmented cell
be successful due to other practical limitations (i.e., exceeding breakthrough curves may be used to help aid in this
low soil hydraulic conductivities). Martinez et al. (2013) determination (i.e., monitoring of injected cells at some outlet
demonstrated specifically how alternating injections can improve location to determine extent and densities transported in the
the spatial distribution of augmented cells and its impact on the aqueous phase).
subsequent distribution of precipitated calcite. Additional Step 4: Urease activity should be determined with distance
consideration should be given when attempting to introduce from the cell suspension injection source location when
alternating injections of treatment solutions with excess urea to possible y
calcium (urea to calcium ratios > 1), as the introduction of high of suspended cells, total direct cell counts (Section 4) are
calcium at a previous effluent location upon flow reversal can recommended over OD600 measures, which have been used
lead to significant abiotic precipitation and localized clogging of for this purpose occasionally by others, but can be
injection sources and surrounding soils.
compromised by suspended fines released from the treated
soil. Depending on the time frame needed for bulk ureolytic
3.6 Generalized treatment approaches
activity determinations, aqueous total ammonium appearance
Although the authors have experience with both augmented and and/or urea disappearance can be monitored. Urea
stimulated MICP, considerations of cost, scalability, degradation is the preferred approach due to the volatility of
environmental impacts, and treatment process benefits have led ammonium species at high pH and their soil surface
to primary use of stimulation in the authors’ most recent interactions at low pH.
experimentation. Still, augmented MICP can prove useful for Step 5: Cementation injections proceed once the specimen is
more controlled experimentation wherein more fundamental ready for cementation, as judged by sufficient and uniform
effects are to be elucidated. Two brief examples of treatment urease activity. When urea is included in augmentation
protocols are provided below for either augmentation or solutions, a flush injection of artificial ground water or other
stimulation. solution is advisable to remove much of the carbonate species
and alkalinity generated from urea hydrolysis prior to
cementation. Such methods can help mitigate abiotic CaCO3
433
precipitation upon the introduction of high calcium as growth and maintenance. For a given treatment scheme, both
discussed earlier in Section 3. the flow regime employed (i.e., flow rates) and the targeted
Step 6: Flush treatments, when used, should be followed in reaction rates (i.e., bulk ureolytic activity) should be
rapid succession by the injection of a cementation solution considered together to identify supplied organics. Cell
that is near equimolar in CaCl2 and urea to begin the growth, and by association bulk ureolytic rates, can
significantly alter treatment extent and uniformity. Thus,
applied successively once such reactants are deemed to have specific treatment solutions may vary in composition
been fully utilized during a given residence period. depending on the desired treatment uniformity, treatment
Treatments should proceed until some target treatment level level, and financial and sustainability constraints. Finally, for
is achieved. The authors recommend that improvement be applications wherein evaluating and attributing stimulation
non-destructively monitored during treatments using shear potential to specific soil materials is of highest priority,
wave velocities and other geophysical and geochemical stimulation solutions shall be prepared aseptically. This may
methods as discussed later in Section 4. During this phase, require filter sterilization (i.e., use of 0.2-micron filters) of
modifications in injection schemes may be used to further prepared solutions but could involve other sterilization
control cementation extent and distribution. methods. Autoclaving may not be a suitable method for
Step 7: Soil and pore fluid samples may be destructively sterile solution preparation, however, given the heat labile
obtained for further analysis of achieved microbial nature of urea.
communities following all augmentation and cementation Step 4: Monitoring of ureolytic activity shall be completed
treatments. Alternately, rinse injections should be applied to during stimulation solution treatment residence periods to
remove soluble salts prior to destructive mechanical testing. quantify urea degradation. With knowledge of urea
utilization, appropriate treatment frequencies can be
3.6.2 Stimulation treatment approach
Th have been provided on a daily or alternating-day basis and
should be mentioned that stimulation treatment schemes are have consisted of sufficient volume to fully replace the pore
nuanced and therefore the authors refrain from providing volume of a given sample (~1.5 PV but dependent on soil
universal formulations. Instead, stimulation protocols should be advective-dispersive transport characteristics). Aseptic
designed with consideration of the specific soil materials, site protocols should be maintained throughout the stimulation
environmental factors, and reactive transport conditions present, process, however, such precautions may be most impactful at
among other important factors. These recommendations provide the beginning of the stimulation period during which
an outline for such protocols, with more specific implementation indigenous communities may be most susceptible to
recommendations provided only where generalization may be contamination. At all points during the stimulation and
appropriate. cementation process methods described in Section 4 may be
Step 1: Care should be taken to collect soil samples employed to characterize enriched communities and
aseptically when stimulation is to be evaluated and attributed associated microbiological and geochemical changes in time
to a given soil material ex situ. This can be achieved through during the treatment process.
a variety of different approaches, which include but are not Step 5: Treatments can be transitioned to cementation
limited to collection with sterilized split-spoon samplers, following successful stimulation, as identified by achieving
sonic drilling, and/or other methods. During and following some desired ureolytic capacity (e.g., degradation of enough
soil collection, similar efforts shall be made to minimize urea in 12 hours to efficiently precipitate the calcium salts to
contamination, though such procedures may never fully be employed during cementation). During this transition a
eliminate the potential for contamination. Furthermore, flush treatment should be applied immediately prior to
sterilized glass bead control specimens can be subject to initiating the cementation process. The flush solution may be
similar procedures to evaluate and control for unexpected similar to the composition of the stimulation solution or more
contamination. Gomez et al. (2018b) provides some inert. For example, similar solutions lacking urea and pH
examples of such efforts. adjustment have been used (San Pablo et al. 2020). This flush
Step 2: Minimize the potential for contamination when treatment, however, will elute stimulated cells, thus
preparing soils in particular experiments for a laboratory test monitoring of ureolytic activity following this transitional
program. Efforts should include sterilization of soil period may be important.
containment devices (e.g., columns), tubing, soil placement Step 6: Cementation injections can proceed once flushing has
tools, and all items which may be in contact with soils. been performed, with repeated cementation injections
Sterilization is ideally accomplished using wet or dry heat in continuing until some target treatment level is achieved. The
an autoclave or oven respectively, though chemical agents authors recommendation would be to assess this
(e.g., bleach, ethanol) or a number of other methods may also improvement non-destructively during treatments using
be appropriate. shear wave velocities and other geophysical and geochemical
Step 3: Apply an inert saturation solution (e.g., artificial methods as discussed later in Section 4. During this phase,
groundwater) or stimulation treatment solution following modifications in injection schemes may be used to further
initial soil preparation. For in situ treatments, however, such control cementation extent and distribution. Additionally,
processes may not be necessary. Stimulation solutions should during the cementation process it is recommended that
be designed in consideration of the soil material and solution compositions not differ dramatically from those
surrounding environmental conditions as described earlier in used during stimulation to maintain selective pressure for
Sections 2 and 3 in order to successfully enrich native ureolysis. However, in the auth initial
ureolytic bacteria. This includes consideration of selective ammonium chloride concentrations and pH can be reduced
factors such as provided ammonium concentrations and due to the production of ammonium from urea hydrolysis
initial pH as well as provided nutrients and growth factors. directly and strong buffering of pH by carbonate and
Solutions should at a minimum include some amount of ammonium species during cementation treatments regardless
ammonium, urea, and organic carbon and nutrients for cell of initial conditions.
434
Step 7: Soil and pore fluid samples may be destructively dry state (Sneath 1962). Thus, it remains possible that such
obtained for further analysis of achieved microbial spores may contribute to the success of stimulated MICP in
communities following all augmentation and cementation natural soils, including those which would be expected to have
treatments. Alternatively, rinse injections should be applied limiting particle sizes.
to remove soluble salts prior to destructive mechanical The use of stimulation approaches may extend the limit of
testing. soils amenable to MICP at smaller particle sizes by requiring
only the transport of aqueous chemical concentrations to native
3.7 Soil & environmental considerations ureolytic cells and/or spores already present within soil materials.
By avoiding the physical constraints related to cell transport in
Methods regarding the activation of microbial ureolytic capacity porous media, treatability for stimulation is largely governed by
through applied stimulation treatments, the precipitation of the rate at which a given treatment solution is transported through
CaCO3 through applied cementation treatments, and the removal the soil (i.e., governed by soil hydraulic conductivity) relative to
of ammonium byproducts resulting from MICP through applied the rate at which biologically-mediated chemical reactions can
rinse injections was largely described throughout Section 3 for occur (i.e. governed by microbial populations, activity, and
poorly-graded silica sands under saturated freshwater conditions. reactant availability). As, discussed earlier in Section 3, these
Such methods are also robust and effective across a much considerations have been the primary motivation for developing
broader range of soil and environmental conditions. This section new treatment techniques by which stimulated bulk ureolytic
surveys the broad treatment effectiveness demonstrated to date rates can be controlled.
across variations in soil gradations and mineralogy as well as Mechanical mixing methods have also been used for
environmental conditions. treatment of fine-grained soils for which hydraulic conductivity
limitations prevent advective transport and solution injections.
3.7.1 Soil gradation For example, the treatment of clayey soils has been explored via
MICP treatment effectiveness is primarily governed by the: (1) ex situ mixing of ureolytic bacteria and treatment solutions with
ability of stimulation and augmentation procedures to develop a soils followed by placement and compaction (Li 2015, Cheng &
Shahin 2015, Jiang et al. 2017). Others have attempted mixing
robust microbial community to facilitate CaCO3 precipitation,
solely treatment solutions with soils to stimulate indigenous
(2) ability of treatment methods to deliver stimulation,
bacteria (Islam et al. 2020, Asghari et al. 2021, Mukherjee et al.
cementation, and byproduct constituents through a targeted
2021), however, such results have varied. In particular, the role
treatment zone within a time period compatible with achieved of changes in solution chemistry on clay mineral electrochemical
microbial reaction rates, and (3) magnitude and spatial interactions can remain unclear when not controlled for.
distribution of CaCO3 precipitation generated at particle contacts
and on open particle surfaces. be performed for such studies in order to be able to separate the
Analysis of the spatial compatibility between microbes and effects of cementation from other clay-ion interactions. MICP
soil particle sizes (Mitchell & Santamarina 2005, DeJong et al. treatment of organic soils, such as peat, have also been explored
2010) identified probable limits for in situ augmentation of soils using high water content slurry mixtures and stimulation of
using whole bacterial cells and hypothesized lower bound indigenous bacteria (Phang et al. 2019, Safdar et al. 2021). While
particle size limits for which the effectiveness of such treatments not strictly MICP, Duggan et al. (2017) achieved stabilization of
may be significantly less probable. Rebata-Landa & Santamarina peat via abiotic carbonation wherein peat was oxidized, releasing
(2006) further investigated microbe viability in fine-grained soils CO2 in the presence of calcium ions in the pore fluid, and calcium
as a function of applied overburden stress, and identified carbonate precipitation occurred. Although in this instance peat
conditions for which microbial viability limits likely exist. specimens achieved some level of mechanical improvement, it
The breadth of microbes that can be used to facilitate MICP should be noted that carbonation reduces the pH of surrounding
are now more fully characterized in part due to the transition pore fluids and could lead to degradation of generated CaCO3 if
from augmentation to stimulation. The S. pasteurii strain (ATCC not carefully controlled.
11859) typically used in bioaugmentation is a rod shape with a The upper bound particle size limit for MICP treatments in
coarse-grained soils is unconstrained. However, more practical
Bacterial species identified in stimulated soils (S. pasteurii, S. size limits can be thought of as being reflective of treatment
soli, S. saromensis, S. aquamarina, S. luteola, S. koreensis) are efficiency considerations as opposed to physical flow and
all rod-shaped and have published diameters of 0.5 to transport constraints. While coordination number in a poorly
Tominaga et al. 2009), which are graded soil matrix is independent of particle size, the number of
consistent with direct measurements of MICP isolates of these contacts per unit volume decreases with particle size, resulting in
species. Injection of microbes of this size into soil becomes a each particle-particle contact being subjected to a higher contact
limiting factor due to their ability to move through pore throats stress. Assuming that the proportional distribution of CaCO3
and consequently influences cementation uniformity. Colloidal precipitation between open surfaces and at contacts is
filtration of the microbes during injection typically produces a independent of particle size, for a given CaCO3 content the
log-linear decrease in population with distance under 1D flow effectiveness of cementation toward improving mechanical
conditions (Li et al. 2004) though variations in pumping responses would be expected to reduce with particle size
techniques can produce some improvements (Martinez et al. increases. This is because with larger particles and thus larger
2013). Nonetheless, severe filtration/straining of the bacteria is interparticle forces, the role of interparticle friction is greater and
likely when the D10 of the soil is less than about 10 times the accordingly the impact of added cohesive forces from
microbe length, and effects of filtration may persist when the D10 cementation on soil mechanical responses becomes smaller.
of soil is up to 50 times the microbe length. Hence, filtration is However, there is some indication that the distribution of CaCO3
expected when the D10 of the soil is less than about 50 mm. An precipitation between open particle surfaces and particle contacts
alternative augmentation approach wherein freeze-dried spores does not remain constant as particle sizes increase. In addition,
(typically spherical, diameter ~0.5 to 1.0 mm) are resuspended in in more well-graded natural soils the coefficient of uniformity
some treatment solution and are injected (Dadda et al. 2017, (Cu = D60/D10) tends to increase with average particle size (D50),
Terzis & Laloui 2018, Pungrasmi et al. 2019, and others), may in which case the number of particle contacts per unit volume
be slightly more effective in treating finer-grained soils. would increase relative to more finer and poorly-graded soils,
Additionally, spores of close relatives of Sporosarcina have thus reducing interparticle contact stresses and perhaps making
demonstrated viability in soils stored for hundreds of years in a
435
MICP improvement more effective. Thus, the true impact of a D10 of 0.02 to 5.64 mm. Both poorly- and well-graded soils
particle size differences may be difficult to generalize and instead have been effectively treated, with Cu values ranging from 2.0 to
specific to the treatment techniques employed (i.e., distribution 8.4. It should be noted that treatment effectiveness does not
of cementation) in addition to differences in soil material particle appear to be directly driven by geologic deposition, but rather
sizes and gradations. this is an indirect factor in determining soil gradation. Further
MICP has been shown to effectively improve a broad range extension of treatment success into sand and silt mixtures is clear,
of silica-based granular soils and some silts using both however, success in pure silts is not reported likely due to the low
stimulation and augmentation approaches. Documentation hydraulic conductivity of these soils. The upper bound particle
regarding suitable soils for such treatments is therefore extensive size treatment range, however, is unconstrained and as shown is
in this respect as most publications provide details about tested only reflective of the geotechnical interests of such studies and
soil material gradations. Figure 28 presents a representative can extend even into treatment of isolated rock fractures (Minto
sampling of the range of quartz-based soils that have been et al. 2016).
successfully treated using MICP using direct treatment injections
(exclusive of mixing), with differences in the reported biological 3.7.2 Soil mineralogy
approach (augmentation versus stimulation) and USCS soil Extension of MICP treatments from silica-based soils into other
classification (SP, SM, SW, GP) distinguished. Although the geomaterials with significantly different mineralogical
majority of past MICP studies have reported the use of compositions has been largely successful. The presence of other
augmentation, it remains unknown what fraction of these studies minerals may influence MICP treatment effectiveness if the (1)
may have been inadvertently stimulated due to a carbon source particle surface charge and/or surface roughness differs from
being included in treatment solutions (Section 2.3.2). Thus, such largely quartz-based sands and influences bacteria attachment,
differentiations here are less clear and are only intended to (2) presence of additional dissolved chemical species and surface
highlight the breadth of treatment processes and soil gradations minerals alters precipitation nucleation, growth, morphology,
that can be successfully improved. As shown, the successful and mineralogy, and (3) availability of additional minerals
treatment range has spanned from a D50 of 0.12 to 10.0 mm and catalyzes unanticipated parallel chemical reaction pathways.
Figure 28. Grain size distribution of select MICP research studies illustrates the range of quartz-based soils that have been successfully tested and are
categorized in their respective USCS group symbol.
436
To date natural sands with high concentrations of carbonate al. (2015). Liu (2021) observed that MICP could achieve an
minerals (Xiao et al. 2018, Zhang et al. 2020, Yanez 2022) and appreciable increase in shear wave velocity (from 50 m/s to 325
iron oxides (Mortensen et al. 2011) as well as anthropogenically m/s) indicating the development of a stiffer cemented soil matrix.
produced materials including fly ash (Montoya et al. 2019, Lastly, O’Toole et al. (2022) treated iron and gold mine tailings
Safavizadeh et al. 2019), and iron and gold tailings (Liang et al. and observed undrained shear strength increases up to 4.5 times
2014, Liu 2021, O’Toole et al. 2022) have been effectively following MICP treatments when compared to untreated
treated. The range of gradations of these different soil materials controls.
is provided in Figure 29. MICP treatment of anthropogenically generated coarser soils,
Figure 29. Grain size distribution of select MICP research studies involving different soil mineralogies. Band of primarily quartz-based soils tested with
MICP is also shown for reference (Figure 28).
MICP treatment effectiveness in soils consisting of human- such as waste rock and base aggregate, have also been successful.
induced process waste products, including fly ash, tailings, and Jiang and Soga (2019) investigated the effectiveness of MICP
other waste rocks have shown promise. For example, Montoya et treated gravel-sand mixtures for erosion resistance via one-
al. (2019) demonstrated MICP effectiveness in fly ash by testing dimensional internal erosion tests. This study highlighted the
two different coal ash materials which contained aluminum importance of soil gradation, overburden stress, and MICP
silicate, silicon oxide, aluminum oxide, and iron oxide. treatment method in mitigating internal erosion issues.
Treatment of these fine-grained materials (D50 = 19.9 and 31.6 Furthermore, Mahawish et al. (2018) performed a suite of tests
µm) in modified oedometer tests showed an increased in soil on coarse and fine aggregate soil mixtures in order to investigate
stiffness and resilience of such materials to repeated stress the effect of particle size distribution and its contribution to
loading and unloading cycles as the cementation levels unconfined compressive strength values. The addition of 25%
increased. Safavizadeh et al. (2019) tested the same coal ash fine aggregates into a coarse soil matrix led to a measurable
material as Montoya et al. (2019) in addition to one other coal improvement in strength (i.e., UCS increase of 575 kPa),
ash. This additional coal ash material, however, was not however, higher fines contents above 25% did not provide
effectively cemented with MICP and may have resulted from additional improvement.
inhibitory concentrations of metals and other dissolved species
in these materials.
Researchers have also investigated the effectiveness of MICP
in tailings materials wherein it can be used to accelerate
consolidation (Liang et al. 2015, Liu & Montoya 2020, Liu &
Montoya 2021, Liu 2021) and enable strength improvement for
applications including liquefaction mitigation (Gui et al. 2018,
O’Toole et al. 2022). Results from Liu and Montoya (2021)
showed that MICP can accelerate sedimentation in oil sand
tailings by increasing soil particle sizes with resulting
densification of the material and a reduction in reclamation time.
The same effect was observed in mature fine tailings by Liang et
437
3.7.3 Environmental conditions concentrations could be compensated for through greater cell
MICP treatment effectiveness not only depends on the largely densities in theory, however, the success of this approach is less
physical constraints posed by different soil materials but also certain due to the varying impact of such concentrations on cell
highly depends on surrounding geochemical conditions, the growth which would merit more specific testing.
presence of pore fluids, and other site dependent factors. Of the
infinite possible environmental variations, some variables that
have been explicitly examined to date include changes in water
salinity, surrounding temperatures, specific ion additions, and
partial saturation.
MICP soil improvement feasibility in coastal environments
where either seawater or brackish pore fluids are present has been
shown to be feasible. For example, studies by Mortensen et al.
(2011) and Cheng et al. (2014) examined treatments in saline
water conditions. Recent experiments by the authors’ (Yanez et
al. 2022) considered a broader range of seawater mixtures and
similarly demonstrated treatment effectiveness using
stimulation. Figure 30 presents a subset of these results obtained
in experiments involving applications of MICP treatments
applied in various combinations of natural seawater and fresh
water (100% AGW to 100% NSW) to an alluvial (Concrete
Sand) and marine (Beach Sand) sand. As shown, stimulation was
effective towards achieving sufficient urea hydrolysis activity in
all experiments (Figure 30a) with uniform calcite contents near
5.2% achieved after cementation (Figure 30b) and average
unconfined compressive strengths for all soils near 840 kPa.
The in situ temperature of soils at depths below 2 m in most
environments is between about 8 and 11ºC. However, the vast
majority of MICP research is performed in the laboratory where
the temperature is about 20ºC. Bacterial activity typically
reduces by a factor of 2 to 3 per 10ºC decrease in temperature, as
bacteria processes typically have a Q10 temperature coefficient of
2 to 3 in ideal conditions where growth and other factors are
ignored. De Muynck et al. (2013) performed batch experiments
with S. pasteurii under various temperatures (10, 20, 28, 37°C)
and observed varying rates of urea degradation in the absence of
any added nutrients. Their results showed expected increases in
rates between 10ºC and 20ºC specimens with the 20ºC specimen
achieving roughly twice the ureolysis rate of the 10ºC specimen
for three different pure cultures (i.e., Sporosarcina psychrophila,
Sporosarcina pasteurii, Bacillus sphaericus). Conversely, as
temperatures increased from 28°C to 37°C, minimal effects were
observed. van Paassen (2009) reported an increase in urease
activity near a factor of 3.4 for every 10°C increase. Whiffin
(2004) further observed stable urease activity from 15°C to 25°C
and a linear increase of 0.04 mM urea hydrolyzed/min/OD from
25°C to 60°C. Consequently, the kinetics governing the MICP
reaction network are expected to slow, perhaps significantly,
when comparing responses observed in lab experiments to that Figure 30. (a) urea degradation and (b) calcite contents from stimulated
observed in the field. This decline in reaction rates may be column tests involving two sands and solutions of varying salinity. Full
acceptable and manageable for certain projects, but in other urea degradation is observed after 3 stimulation treatments with sustained
cases, particularly when such effects are not anticipated, may activity after 10 cementation treatments (adapted from Yanez et al.
2022).
necessitate unfavorably long retention and reaction periods
between treatment injections. Compensation for this shift in The vast majority of MICP work to date has been performed
environmental conditions can be addressed in part by using saturated soils, and this remains the focus herein as it
modification of the treatment formulation (Section 3.1). For directly applies to liquefaction mitigation. However, MICP
example, a decrease in single-cell activity due to temperature treatment of free-draining soils susceptible to liquefaction could
reductions encountered in the field could be conceptually be performed if temporary dewatering was implemented.
compensated for by increasing the stimulated microbial cell Conceptually, soil treatment under freely draining conditions
population, resulting in a comparable bulk ureolytic rate to that would be expected to be comparably, or more effective, than the
achievable in the laboratory. Although not discussed here, fully saturated counterpart condition if more specific challenges
changes in ionic strength and specific inhibitory solution species related to treatment uniformity and implementation are not
(such as magnesium ions and other metals) may further alter considered. Free-draining treatments could in theory provide
achieved bulk ureolytic activities. While the authors’ will refrain some benefits as it requires a smaller treatment solution volume
from discussing all of these effects, characterization of soluble (~0.25 to 0.5 of the pore volume) and would allow for
species present in soil material pore fluids and testing of site soils localization of retained solutions at solution menisci existing at
in tandem with other more well characterized soils which can particle-particle contacts where precipitation can be most
serve as a control, may allow for such effects to be identified effective towards improving soil mechanical properties. Cheng
early during the treatment design process. As with temperature, & Cord-Ruwisch (2012) and Cheng et al. (2013) demonstrated
reductions in bulk ureolytic rates resulting from such inhibitory this effect and showed that treatments applied under partially
438
saturated conditions can be more effective than treatments solutions. In contrast, the natural sand is readily stimulated and
applied under saturated conditions when comparing specimens at begins degrading urea in the third cycle of treatment with
similar cementation levels. increasing activity over time.
The efficacy of MICP treatment using stimulation in natural
3.7.4 Reagent purity soils under non-sterile conditions has been sufficiently
While the majority of laboratory experiments have been established in both laboratory and field conditions. Nonetheless,
performed with research-grade chemicals, such sources will in in the laboratory thorough cleaning and sterilization of treatment
many cases be cost prohibitive when applied to field-scale. solutions and testing equipment is still recommended to
Therefore, it will likely be necessary to source chemicals from minimize the likelihood of sequential transfer of microbes and
other commercial sources. The use of such sources will require contamination from one test series to the next. Otherwise, it may
compromise with respect to product purity and particular be incorrectly concluded that a specific soil may support
impurities may be important to recognize and quantify prior to stimulated MICP, when the true source of activity was instead
use. It is the authors’ recommendation that some amount of some contaminant. In the field thorough cleaning of mixing and
testing be performed to evaluate the performance of these pumping equipment is also recommended prior to use,
feedstocks relative to other more well characterized reagent particularly if the equipment was previously used for
grade sources. Specifically, the authors’ suggest that caution be implementation of a different soil improvement method or
taken when selecting urea sources, as some commercial fertilizer environmental remediation technology.
products intentionally contain urease inhibitors to promote the
3.8.2 Stress conditions during MICP
gradual release of ammonium nitrogen for agricultural
The MICP cementation formed is brittle, and susceptible to
applications. When reflecting on past up-scaling experiments,
disturbance from stress and strain perturbations, especially at
the authors’ believe that such chemical grade differences may
cementation levels below about 5% calcite content by mass. This
have played some role in the limited success of a high
susceptibility requires consideration and management of the
concentration MICP field trial plot performed for surface
stabilization as described in Gomez et al. (2013). Therefore, this
aspect among many other considerations should be contemplated
3.8 Additional considerations
Implementation of MICP in the laboratory or field requires
additional process and design considerations, including the
sterility conditions and stress changes during and after treatment.
A summary of these important considerations is provided below,
while acknowledging that other important considerations may
exist for a particular application and/or experimental program.
3.8.1 Sterility during testing
Preparation of soils and MICP treatment solutions in past studies
have ranged from fully sterile laboratory tests designed to
definitively demonstrate the biogeochemical basis of the
technology to field scale implementation where construction- Figure 31. Comparison of ureolytic activity generated by non-sterile
scale equipment is used for preparation and injection of stimulation medium treatments of a sterile soil analog, 0.25-0.42 mm
treatments into natural, non-sterile subsurface conditions. As glass beads, with one composed of a natural alluvial sand. The soil analog
alluded to in Section 2, fully sterile laboratory tests where was oven-sterilized for 72 hours at 180°C prior to application of the same
augmentation has been applied using non-nutritive solutions (and stimulation treatments. Urea concentrations were measured at the start
perhaps even sterilized soils) have clearly defined the science of and end of each treatment. Error bars represent one standard deviation of
triplicate urea quantifications (adapted from Graddy et al. 2021).
MICP. In such tests, augmented S. pasteurii was the primary
microbe performing ureolysis and facilitating precipitation of confining stress applied to specimens during treatment and
calcite, enabling measurements specific to the S. pasteurii strain subsequently during some applied loading. In the laboratory, this
to be made. is relevant when stresses are relaxed (or entirely) released as
However, recognizing that eventual implementation of MICP specimens are transferred from a treatment cell to some other
as a soil improvement technology in the field would inevitably testing apparatus. In the field, this is relevant when considering
be under non-sterile conditions and that stimulation of native the in situ stress during treatment in comparison to the lower or
microbes in the natural environment would be preferrable to higher stresses that may be applied during some project’s service
augmentation, research efforts shifted focus towards verifying life (e.g., due to earthquake or infrastructure loadings).
robustness of the stimulated MICP process under non-sterile Differences in confining stresses between those applied
conditions. Acknowledgement of non-sterile field conditions and during treatments and those applied during subsequent loading
the need for treatment robustness led to the refinements in the treat, and the
stimulation formula that produce temporary environmental oper. Conceptually, calcite precipitation
stressors, which are favorable for ureolytic bacteria to thrive and forms brittle cemented bonds between particles (Section 2.4)
outcompete other bacteria (Section 2). One example of the effect while particle-particle contacts experience contact forces
soil sterility can have on stimulation success is shown in Figure treat, particle size, and mineral-specific
31, which presents urea degradation across multiple treatment compressibility, which can be approximately understood using
cycles applied to a fully sterile glass bead soil and a natural sand Hertzian contact theory. Following cementation, deviations from
material (Concrete Sand) both treated with non-sterile solutions. treat can result in elastic loading or unloading of the contact
As shown, the sterile control does not exhibit any detectable forces, which will produce changes in compressive or tensile
ureolytic activity over seven different treatment cycles, stresses acting at the precipitated contacts. The susceptibility of
indicating that minimal activity resulted from use of non-sterile damage to this loading and unloading is dependent on the amount
439
of precipitation and the magnitude of stress change, however, all 4 PROCESS MONITORING & QA/QC
other factors equal, the susceptibility of a cemented soil to this
disturbance decreases with cementation level increases and as The following section describes a series of methods for
stress changes are minimized. Montoya et al. (2013) documented monitoring of ureolytic MICP. This includes methods for
how cementation integrity (as measured using shear wave determination of microbial activity and abundances,
velocities) can systematically degrade with unloading and geochemical changes, the mineralogy and composition of
reloading cycles for specimens treated to low cementation levels, achieved cementation, and other methods aimed at furthering
but can remain relatively intact when similar stress cycles are understanding of microstructural and mechanical changes
applied to higher cementation levels. This can be explained in afforded by the presence of cementation.
part by the fact that as cementation levels increase, shear
modulus values becomes increasingly stress invariant, exhibiting 4.1 Microbiological process monitoring
behaviors more similar to intact rock. Such behaviors are further
described in Section 5. Several methods have been used for quantifying densities of
treat oper are important to
bacteria used in MICP. Some of these methods come with
recognize as such variations can affect the subsequent significant potential pitfalls, and the most rigorous methods are
engineering behaviors of MICP-treated soils, particularly at low typically not the most readily implemented ones.
cementation levels. This typically manifests as a decrease in
stiffness, peak strength, and changes in dilative tendencies. The 4.1.1 Cell quantification approaches
magnitude of this influence can range from complete degradation Although many methods exist by which total cells can be
to the corresponding uncemented state to more minor effects with quantified, the following sections outline the primary methods
the magnitude being dependent on the cementation level, the for cell enumeration in MICP experiments with a discussion of
specific considerations.
treat - oper), and other unquantifiable handling
disturbances. Per the above discussion, the impact of such
4.1.1.1 Optical density measurements
disturbances will decrease as the cementation level increases due
Quantifying the initial density of S. pasteurii or other bacteria to
to decreased stress dependence. The effect of the stress
be added to soil column or flask experiments is most often
treat - oper), depends on whether or not the
accomplished using a spectrophotometer to measure optical
operational stress exceeds or is less than the treatment stress.
density at some standard wavelength, 600 nm being typical
treat to a lower
(OD600). Bacterial suspensions, which impact optical density in a
oper induces less, sometimes minimal, damage relative to
fundamentally different way than colored solutions, do so by
oper. This can be explained by a smaller
scattering (not absorbing) light. Data collected on six different
elastic deformation being recovered upon unloading and both
randomly selected spectrophotometer models are presented
plastic and elastic deformation occurring when loading to higher
(Figure 32) and show that the same S. pasteurii suspensions
stresses never experienced by the sample. However, given these
scatter light very differently on the various instruments based,
complexities, the effects of stress changes are not necessarily
presumably, on different internal geometries and other factors.
predictable and there are no existing empirical corrections to
As shown, an OD600 of 1.0, near the middle of this range,
account for the change in engineering behaviors induced by
corresponds to a 4-fold range in cell densities of approximately
treat oper.
1.8 to 7.8 x 108 cells/mL. When considering the extremely low
MICP treated specimens prepared in soil columns,
particularly at lower cementation levels, are often even more value of 0.8 x 108 cells/mL, which was derived from the oft-used
susceptible to damage due to the strains imparted on the standard equation of Ramachandran et al. (2001), that equation
specimen during extrusion, movement, handling, and is found to be largely inconsistent with observed trends and the
remounting of specimens between a treatment cell and a testing range of cell densities expands to 10-fold. The use of
device. Again, the magnitude of damage is inconsistent and uncalibrated optical density measurements therefore should be
cannot be reliably corrected for. Thus, treatment of laboratory confined only to assuring that different experiments within a
test specimens following placement in some testing device (e.g., single laboratory are performed at equivalent cell densities. This
triaxial, simple shear) is preferred whenever possible. approach also requires appropriate pre-measurement dilutions to
The consequence of handling-induced damages on the remain within the linear range of the instrument, which can be
subsequently measured engineering behavior is often substantial quite different for various instruments (Figure 32). Only after a
and contributes significantly to exceptionally large variations in specific spectrophotometer has been accurately calibrated, such
measurements reported in the literature. This is most evident in as by measuring total bacterial protein or total direct counts and
reported unconfined compression test strength (UCS) values. For relating these to OD600 values, can the data produced be relied
this specific test, the large variations can be primarily attributed upon to accurately measure and report absolute cell densities.
to the damage induced during extraction of the specimens from Key to reproducing any type of experimental work by another
treatment molds, setup of the specimens in the test device, and researcher is ensuring that such values are transparent and
end cap imperfections, with treatment formulations and injection universally understood. Thus, it is recommended that cell
methods having likely only secondary effects. For this reason, densities be reported in the literature only after accurate
UCS values could be appropriate as a more qualitative metric by conversion from optical density and that raw OD600 values not be
which improvement can be said to have been achieved. However, reported in isolation.
it likely that the actual strength gain achieved is larger (possibly
substantially) than that measured with UCS. Consequently, it is 4.1.1.2 Total bacterial protein (TBP)
recommended that more advanced strength tests be performed Total protein comprises approximately 50% of bacterial dry
using processes where the specimens can be treated at a stress weight (Luria 1960, Ingraham et al. 1983) and is, therefore, an
state treat oper). excellent surrogate for bacterial abundance (Figure 33a). To
convert from protein/mL to cell densities (cells/mL), the
following additional conversions can be employed: (1) the wet
density of whole bacteria is nearly constant at 1.05 g/cm3, and (2)
the ratio of dry weight to wet weight for bacterial cells is
approximately 0.30 (Bakken & Olsen 1983). Thus, one can
440
readily convert total bacterial protein/mL to wet biomass/mL and produces colonies whose density corresponds to that measured
finally to actual cells/mL, if the shape and average dimensions of by TDC or TBP, then such counts can be used as another way to
the bacterium under the employed conditions are known. The calibrate and periodically check an OD600 to total cell density
Coomassie brilliant blue method of Bradford (1976), has long conversions factor, independent of access to an epifluorescence
been employed; representative details for measuring total scope. Unfortunately, for the type strain of S. pasteurii, different
cellular protein (soluble + membrane associated) can be found in types of plate count medium can yield very different recoveries
Nelson & Jannasch (1983). A second widely used dye-binding for the same cell suspension. Especially on ATCC 1376 Medium
method employs bicinchoninic acid (Smith et al. 1985). (Stocks-Fischer et al. 1999), which was used to calibrate an
Commercial kits are available for both and include advice on equation relating OD600 to cell density (Ramachandran et al.
specific protein standards. 2001), recovery was poor and variable (Figure 33a, b) compared
with recovery on BHI (brain heart infusion) agar with urea (75
4.1.1.3 Total direct counts (TDC) mM), pH 8 (Figure 33c). Unknown factors, including how long
Initially devised for enumeration of dilute bacterioplankton in the ATCC 1376 medium was autoclaved, how long plates of
marine environments, the epifluorescence microscopy method of medium were aged prior to serial dilution, and culture condition
Hobbie et al. (1977), which makes individual bacteria visible by appear to contribute to its highly variable recovery. In the
brightly staining their nucleic acid, has been adapted to authors’ experience, BHI gives by far the most reproducible and
accurate results for S. pasteurii as shown by the agreement
between the BHI calibration OD600 with the measured TDC on
two separate cultures (Figure 33a, b). It is therefore exceedingly
important, if reported VPC results are to be an accurate measure
of that cell densities, that they be calibrated ultimately and
regularly against either TDC or TBP.
4.1.2 Ureolytic rate determination approaches
The following sections outline the primary methods by which
ureolytic activity can be quantified in MICP experiments. This
includes methods for determination of ureolytic activity per cell
(termed specific ureolytic activities) as well as the determination
of the collective ureolytic activity of enriched microbial
communities (termed bulk ureolytic rates). Finally, methods by
which these rates can be modelled are further discussed.
Measurements of urea hydrolytic activity are important in the
design, implementation, and analysis of both biostimulated and
bioaugmented ureolytic MICP implementations. These
determinations are applied in vitro to pure culture suspensions of
bacteria, frequently but not exclusively to S. pasteurii, to, for
example, validate and dose the bioaugmentation inoculation or
to evaluate the potential utility of a microbial strain for MICP. In
situ measurements are made on pore fluid samples obtained from
soils during treatments to monitor the progress of urea hydrolysis
in the soil, to determine when enough time has elapsed to apply
a subsequent treatment, as well as to evaluate the efficacy and
uniformity of stimulation or augmentation inoculation.
Figure 32. Spectrophotometric response of a set of washed S. pasteurii Measurements of urea hydrolysis can be made either from the
suspensions on a series of spectrophotometers. Total direct counts (TDC) disappearance of urea directly or through the appearance of
were performed using acridine orange epifluorescence method. Cell reaction products (e.g., total ammonium). While measurements
counts estimated from the measured OD600 using the relationship from of the appearance of end-products is indirect, it is typically a
Ramachandran et al. (2001) are plotted for comparison. Error bars are
more sensitive means of detecting small changes in urea
standard errors of triplicate suspensions for OD600 and 10 microscopic
field counts for TDC. concentration, especially at the start of reactions where the
quantifying densities of appropriately diluted pure cultures that
have been concentrated, post-staining, into a single focal plane
on a counterstained membrane filter. This method was initially
called the acridine orange direct counting (AODC) method.
Another method (Porter & Feig 1980) employs the less
toxic/mutagenic DAPI (4-6-diamidino-2-phenylindole).
Regardless of the fluorochrome used, the result is direct
microscopic counting of cell densities that can be used to
generate a standard curve relating TDC to OD600 values obtained
spectrophotometrically for the same cell suspensions. Access to
a high-quality epifluorescence microscope, equipped with one or
the other filter set, is likely available at most research
universities, and only periodic calibration of an OD600 versus cell
density curve is required.
4.1.1.4 Viable plate counts (VPC)
If the bacterium under study, when serially diluted in sterile
diluent and spread on plates of appropriate agar-gelled medium,
441
Figure 33. S. pasteurii quantification methods and agar efficiency. Three exponential growth curves of S. pasteurii in ATCC medium 1376 (28°C, 150
rpm) measured by OD600 and total direct counts in effective colony forming units (TDC ecfu; i.e., attached cells or cell clusters count for one ecfu) in (a)
and (b), total bacterial protein (TBP) in (a), viable plate counts on ATCC 1376 agar aged one week (aged agar) in (a) and (b), viable plate counts of two
separate batches of one day old ATCC 1376 agar (fresh agar) in (b), and viable plate counts on BHI agar in (c). The relationship between OD600 and cfu
from Ramachandran et al. (2001) is used to estimate culture density from the OD600 measures in all panels, while a power fit of the BHI data in (c) is to
estimate culture densities in (a) and (b).
relative change in product concentration is large compared to as might be found in a starving or toxified late
substrate. This is especially important for short duration assays exponential/stationary phase culture, could lead to substantial
required for accurate specific ureolytic activity determinations reduction in the cellular urease concentration. Similarly, the
but may be accordingly less important for longer duration soil average cell size of a culture elongates as it exits exponential
element measurements. growth, which can affect measures of per-cell, if not per-
biomass, activity. Thus, for the most repeatable results, cells
4.1.2.1 Specific ureolytic activity should be harvested during exponential growth after a minimum
The most sensitive and quantitative method for measuring of 4 to 6 doublings to ensure consistent, vigorous cell state, and
ureolytic activity in vitro uses the colorimetric Berthelot method used promptly. If late stationary or long-resting cells are to be
of ammonia determination (Weatherburn 1967, Krom 1980, used, more variability should be expected.
Section 3.4.1), though other ammonia methods, frequently Finally, caution is urged to identify the contribution to total
enzymatic, exist (Kaltwasser & Schlegel 1966) and are discussed ureolytic activity of free urease that may exist in the solution. As
elsewhere (Mobley & Hausinger 1989). The Berthelot method, cultures proceed, they can accumulate a population of dead cells.
however, is relatively slow and uses somewhat hazardous Due to the nature of exponential growth, this effect becomes
chemicals (e.g., phenol, nitroprusside, bleach), leading to other, much more significant during the later phases of growth, and is
especially prominent during the stationary and the appropriately
less involved measures of end-product accumulation being
named death phase. During this stage, dead cells’ membranes
preferred for some applications. The most prominent alternative
disintegrate and their cytoplasm is released into free solution,
utilizes the generation of charged end-products from the neutral
liberating urease. Free urease has five fold higher activity than
urea product, allowing a solution’s electrical conductivity to equivalent whole-cell urease which has an activity that is
serve as a measure of urea degradation (Lawrence 1971). seemingly limited by membrane diffusion (Larson & Kallio
Colorimetric change of pH indicator has also been suggested to 1954). Thus, it is the authors’ recommendation that cultures be
estimate urea degradation from medium alkalinization, though used soon after growth (ideally within a few hours) with storage
the method’s sensitivity, both of the method overall and the at 4°C. Longer storage may substantially increase variability and
buffering capacity of the chosen medium, may preclude accurate soluble urease concentration. In addition, the activity of media
quantification of specific activity (Cui et al. 2022). As discussed filtered and/or centrifuged to remove bacteria should be at a
below, accurate measures of specific activity must be done on minimum occasionally checked to quantify the impact of free
short timescales in which the chemical environment of the urease.
solutions remains effectively unchanged. Thus, direct
measurements of urea disappearance are typically not 4.1.2.2 Bulk ureolytic activity
appropriate for this application, where a detectable decrease in The determination of bulk ureolytic activity in treated soils can
the enzyme substrate is likely to affect the interpreted be achieved through pore fluid sampling and chemical analysis
instantaneous rate of hydrolysis. of the in situ conditions over time. Such measurements are
Measures of activity are best reported on a repeatable and typically obtained over a period of hours during which chemical
transferrable biomass basis. Optical density as a proxy for compositional changes are expected. The desired outcome
turbidity is not a quantitative, transferrable measure of cell primarily being to either describe the ureolytic behavior of some
density and should instead be calibrated internally and presented treated soil over a treatment residence period or merely to assess
on a cell or total protein basis, as discussed previously. urea concentrations at a particular point during treatments. Direct
Additionally, the specific ureolytic activity of a suspension of colorimetric urea assays on pore fluid samples (Knorst et al.
cells is subject to substantial change based on the growth phase 1997) are typically used for such measurements, when
in which they were harvested and the duration and conditions substantial urea is expected to degrade during the sampling time
under which they are held. Even for bacteria that express urease course. If instantaneous ureolytic rates are instead desired,
constitutively, stress-induced global metabolic regulation, such caution should be employed before assaying pore fluid samples
442
in vitro, as a substantial amount of ureolytic activity appears to
be unable to be sampled by simply extracting pore fluids and (12)
monitoring this fluid’s activity in isolation (Graddy et al. 2021).
In this case, assaying washed soil/cell pellets from soil samples
could better capture this activity and allow for more stringent At [S] >KM, the denominator approximates [S] and the
control over assay conditions, however, obtaining such samples reaction rate approaches Vmax. Since enzymes are catalytic and
not consumed in a reaction, Vmax remains a constant. In other
would likely require significant specimen disturbance.
words, at high substrate concentrations, the reaction rate is
4.1.2.3 Kinetic models for ureolysis independent of substrate concentration and is effectively
described by zero-order kinetics. At [S] <KM the denominator
For some applications, it can be desirable to model the kinetic
instead approximates KM, allowing Vmax and KM to be collected
process by which ureolytic activity degrades supplied urea in a
into a rate coefficient, k, yielding v = k[S]. At this point, the
given system. This can be accomplished by fitting appropriate reaction rate is linearly dependent on [S] and thus is best
kinetic models which can approximate experimentally observed described by first-order kinetics. The KM parameter describes the
urea degradation trends. Three kinetic models are commonly concentration at which the model transitions between kinetic
employed to fit observed urea degradation trends, and include regimes. From a biological perspective, KM describes the binding
zero-order, first-order, and Michaelis-Menten, each of which affinity of a given enzyme and its substrate. The more effective
have specific difficulties and limitations. the binding, the lower the KM, and the lower the concentration at
First-order reactions have reaction rates that change linearly which the enzyme becomes saturated and the kinetic behavior
as a function of a single reactant concentration (e.g., urea). This transitions.
is typical of the degradation of single species, with radioactive Michaelis-Menten kinetics are therefore applicable over a
decay being the canonical example. Trends are primarily large range of substrate concentrations and are uniquely suited to
captured by first-order rate coefficients which describe the describe kinetic behaviors around the transition from zero- to
proportion of a reactant that is degraded per unit time. The first-order kinetics. This utility, however, is somewhat offset by
integrated first-order rate equation is presented in Equation 10 the unwieldy integrated form (Equation 13, Beal 1982, Schnell
and is related to the current substrate concentration, [S]t, as well & Mendoza 1997), which requires the use of the Lambert W
as the initial substrate concentration, [S]0, rate coefficient, k, and function.
time, t.
(10) (13)
Zero-order kinetics describe reactions where the reaction rate 4.1.2.4 Ureolytic rate modeling
is constant, independent of reactant concentration. For enzyme In the context of describing the behavior of ureolytic MICP,
mediated reactions, if a substate is present at sufficient kinetic properties typically have two primary purposes: allowing
concentrations, the active sites of the enzymes that are present for ureolytic rate determinations of cell suspensions in vitro or
can remain occupied, and therefore the reaction can only proceed allowing for bulk ureolytic rate determinations and predictions in
as fast as the enzymes can complete these reactions, leaving the
situ. In both circumstances, rate determinations are being made
rate dependent upon the concentration and efficiency of the
for intact whole bacterial cells, and thus yield apparent whole cell
enzyme. The integrated zero-order rate equation (Equation 11) is
ureolytic rates. Nearly all of the pertinent organisms and
linear, with a slope of –k. As shown, the zero-order rate
coefficient is in units of concentration per time and the equation communities have apparent whole cell KM values in excess of 50
includes the initial substrate concentration, [S]0. mM, with the S. pasteurii type strain having a KM exceeding 300
mM (Lauchnor et al. 2015, Graddy et al. 2018). Thus, for most
(11) MICP applications and applied urea concentrations, the
concentration-dependence of ureolytic rates is an important
Most reactions only transiently operate under zero-order consideration.
conditions, as eventually one of the reactants is consumed to the In vitro measurements are typically relatively short, perhaps
point at which the catalyst begins to desaturate or become less several minutes, which satisfies the preconditions for the initial
than fully occupied. At this point, the behavior may transition rate approximation. The result being that zero-order fitting can
toward first- or higher-order kinetics as other reactants are do a reasonable job of describing the initial rate (Figure 34a);
subsequently consumed. This can limit the utility of zero-order however, the linearity of these urea concentration reductions will
fitting to relatively narrow ranges of concentrations and or times be dependent upon the activity and whole cell KM of the active
over which the reaction rate is approximately constant. To that cells. Similarly, it should be mentioned that the apparent rate will
end, zero-order kinetics are frequently used to estimate an be sensitive to the initial urea concentration when urea
instantaneous rate of a first or higher order reaction under a given concentrations are not in significant excess of the KM.
set of conditions by measuring the reaction rate for a very short For bulk activity measurements, relatively uniform sampling
period after initiation. In this initial rate approximation, the short over the entire retention period allows for characterization of
duration does not afford much deviation from the initial reaction behaviors over larger substrate concentration changes. For these
conditions, so the rate is largely constant and zero-order fitting instances, zero-order fits may not be appropriate unless the
of the pseudolinear portion of the data is typically a reasonable apparent KM is uncharacteristically small. Some researchers have
descriptor of the behavior. fit zero-order models to data observed at the start of in situ urea
Michaelis-Menten kinetics were developed to better describe degradation curves, with the result being an average rate over the
enzyme-substrate interactions and can better approximate the interval fit (Figure 34b). However, these estimates are sensitive
transition from a zero-order to first-order kinetic regime, as to the subjective definition of the linear range, which will vary
occurs upon catalyst desaturation. It takes the form seen in over the course of an experiment with the bulk ureolytic rate. For
Equation 12, where the reaction rate or velocity, V, is related to
the substrate concentration, [S], the Michaelis-Menten half-
saturation constant, KM, and the maximum reaction velocity,
Vmax, which in turn is the product of enzyme concentration, [E]
and the catalytic turnover number of the enzyme, k cat.
443
Figure 34. Examples of ureolytic rate analysis. Initial rates approximation of in vitro S. pasteurii ureolytic activity in (a), where the data are initially
reasonably linear, but quickly deviate based on changing reaction conditions. Comparison of the three primary urea degradation models in a single MICP
treatment (b) and estimating the initial rate for each treatment over the course of a whole MICP implementation (c). The shaded region in (c) represents
stimulation treatments. Error bars are triplicate standard deviations in (b) and treatments 1-5 in (c), with duplicates in treatments 6-8 (adapted from
Graddy et al. (2021).
these cases, fitting may be improved through more frequent incorporation of inhibition kinetics can be included should it be
sampling at the start of urea degradation time courses. In necessary (Qin & Cabral 1994, Fidaleo & Lavecchia 2003).
contrast, first-order fitting of these bulk ureolytic degradation
curves does not require subjective determination of data 4.2 Cementation process monitoring
inclusion and can reasonably apply over the entire times series.
Such fits may work best for degradation time courses where the During the MICP treatment process, various aspects of solution
apparent KM is high relative to the initial urea concentration, and geochemistry may be monitored to assess CaCO3 precipitate
therefore the degradation trend closely follows a first-order trend. formation. While a variety of different parameters may be
These conditions are generally encountered during ureolytic quantified, a summary of several of the most impactful and
MICP. The first-order rate coefficients determined, however, commonly used monitoring approaches for assessing CaCO3
have less direct, inherent meaning than an initial linear fit can precipitation are provided below.
provide, however, instantaneous rates can be calculated as the
product of the rate constant and substrate concentration. Lastly, 4.2.1 Aqueous calcium measurements
a Michaelis-Menten fit is the most physically appropriate model Measurement of aqueous calcium (Ca2+) concentrations provides
but is also the most computationally challenging. In particular, an opportunity to characterize cementation progression and
fitting can be difficult if the achieved data density is low at dissolution when such processes are of interest. Aqueous
concentrations near the KM. However, the model can well capture calcium, although somewhat infrequently measured in most
the behavior of ureolysis during MICP over larger time periods, geotechnical studies, can be easily performed using a variety of
affording useful kinetic parameters that represent the aggregate approaches including readily available colorimetric assays (e.g.,
behavior of all microorganisms present. QuantiChrom Calcium Assay; Bioassay Systems Inc.) and other
The performance of all three models is presented for a single high precision approaches including inductively coupled mass
urea degradation time course in Figure 34b, where all models do spectrometry (ICP-MS). During the cementation process, such
a reasonable job describing the urea degradation curve, though measurements may allow for characterization of CaCO3
the Michaelis-Menten fit seems to better maintain accuracy over precipitation kinetics, assessment of changes in saturation state,
an entire residence. The derived initial ureolytic rate for all three and the identification of reaction completion. Practically, such
models from simulated soil columns from Graddy et al. (2021)
aspects may be of importance for identifying optimal treatment
shows they all perform reasonably in describing relative initial
residence periods, enabling temporal and spatial modeling of the
rate behavior over an MICP experiment, though the first-order
precipitation process, and allowing for characterization of site-
model tends to overestimate initial ureolytic rates (Figure 34c).
Importantly, none of these models directly account for the specific conditions and their potential impact on long-term
impact of cell growth that is happening concurrent with urea material permanence, among other aspects. While such
degradation. The models reasonably approximate this behavior concentrations can be routinely measured in most basic
without the need to specifically attend to changes in cell growth. analytical laboratories, special care should be taken to ensure that
Additionally, a Monod term could have been included but would collected samples are indeed representative of pore fluids. In
require calibration of additional parameters which would be particular, it is recommended that aqueous samples be first
difficult to determine. Finally, product inhibition via generated filtered to remove any CaCO3 solids that may have been
total ammonium has largely been ignored in this discussion of collected during the sampling process and subsequently
ureolytic kinetics. However, it has been shown that whole cell stabilized samples in strong acid (i.e., hydrochloric acid) to
ureolytic activity is largely insensitive to ammonium and pH inhibit the potential for both further ureolytic activity and CaCO3
product inhibition in S. pasteurii (Lauchnor et al. 2015), likely precipitation (see example protocol in Burdalski & Gomez
due to the homeostatic efforts of the host cell, a feature thought 2020). An example of calcium and urea data obtained during
to be shared by other Sporosarcina. Nevertheless, the cementation injections involving three different equimolar urea
444
and calcium concentrations is provided in Figure 35. As shown, which must all be accounted for when completing measurements
urea and calcium trends were similar in time, suggesting that the for MICP solutions. When solution pH, total alkalinity, and
rate of CaCO3 precipitation in these experiments was governed ammonia (NH3) concentrations are known, Equations 14, 15, and
by urea hydrolysis rates. Calcium measurements offer the 16 can be used to determine CO32- and HCO3- concentrations
opportunity to monitor the progression of cementation process using the pKa2 for HCO3-/CO32- of 10.329. Similarly, total
directly, without need for further interpretation. inorganic carbon analyzers can be used to determine HCO3- and
CO32- concentrations. This process works instead by acidifying
4.2.2 Solution carbonate analyses specimens, therefore speciating carbonate species towards
Measurement of solution aqueous carbonate (CO32-) may be carbon dioxide (CO2) gas with subsequent capture and
equally important in assessing the potential for CaCO3 quantification of CO2 using various spectroscopic methods.
precipitation. Carbonate species are generated directly from urea
hydrolysis as a result of carbonic acid generation with pH rise (14)
from ammonia speciation driving the increased availability of
carbonate. CO32- concentrations play a key role in controlling
(15)
solution saturation states and can provide important
understandings regarding the thermodynamic favorability for
precipitation, equilibrium, or dissolution. Aqueous CO32- can be (16)
assessed through a variety of means including total alkalinity
(AT) measurements and total inorganic carbon analyses. Total
alkalinity measurements provide a means by which all proton
4.2.3 Saturation state determinations
acceptors in solution including carbonate species can be
Knowledge of carbonate and calcium concentrations can be used
quantified. Although an example method can be found in ASTM
to determine solution saturation state, however, such calculations
D1067-16, the basic premise of the measurement is the use of a
should consider other surrounding solution species that will
titration procedure wherein strong acid additions are performed
impact activities of carbonate and calcium. Thus, solution
in the presence of a pH indicator to determine concentrations of
activities are considered. Although many different models exist
proton acceptors. Proton acceptors are the conjugate bases of
for activity coefficients, model selection generally requires
knowledge of solution ionic strength (I) and temperature as well
as ion specific parameters including ion size. Under most cases
representative of MICP solutions, the Extended Debye-Hückel
equation can be used to determine activity coefficients for all
samples with I of less than 0.1 M and the Davies equation can be
used for samples with I values between 0.1 and 0.7 M. Once
concentrations, activity coefficients, and activities of Ca2+ and
CO32- have been determined, the solubility products (Ksp) of the
specific CaCO3 minerals of concern can be used to determine
saturation states. Under most conditions representative of MICP
solutions at both atmospheric pressure and near room
temperature (25°C), the solubility product of calcite mineral of
10-8.48
of solutions with respect to calcite, determined by dividing the
ion activity product (IAP), or product of Ca2+ and CO32- activities
(concentrations multiplied by activity coefficients), by the
solubility product (Ksp) for calcite mineral.
(17)
4.2.4 Solution electrical conductivity
Solution electrical conductivity (EC) can be used to assess
changes in conductive solution species during the MICP process.
Although commonly used to assess changes in ureolytic activity
due to increases in solution conductivity from the production of
polar ammonium and carbonate species from non-polar urea, the
measurement may also hold some promise for monitoring
solution changes during the precipitation process. The ability of
solution electrical conductivity to resolve important differences
in solution chemistry of this indirect measurement is dependent
on the urea to calcium ratios used and the pH of solutions. This
is due to the consumption of calcium and carbonate ions during
CaCO3 precipitation (expected to reduce EC) and the
Figure 35. Measurements of (a) calcium and (b) urea concentrations in
simultaneous release of ammonium and carbonate species due to
time during cementation treatments involving three different equimolar urea hydrolysis during the MICP process (expected to increase
urea and calcium concentrations (adapted from Burdalski & Gomez EC). As a result, solution conductivity changes are pH-dependent
2020). and will also depend on the balance between the concentrations
of ions being consumed and generated. While such
weak acids with pKa less than 4.5 in 25°C degree solutions with measurements can provide insights regarding urea hydrolysis
zero ionic strength. They include HCO3-, CO32-, OH-, and NH3, activity in the absence of added Ca2+ and mineral precipitation,
445
caution should be taken when attempting to use such inability for CaCO3 to precipitate cannot be inferred from
measurements to obtain meaningful insights during cementation solution pH measurements alone. Instead, such tendencies are
treatments. governed by saturation state and the activities of calcium and
carbonate species. From this perspective, solution pH may
4.2.5 Solution pH provide an indicator of when CaCO3 precipitation is either
Solution pH measurements provide only a fraction of the finished or unlikely, but conversely pH may convey very little
geochemical information needed to understand a solution’s information about when precipitation should be expected.
saturation state, though solution pH has commonly been used to
infer the potential for precipitation. This is because while urea 4.3 Byproduct process monitoring
hydrolysis in isolation can increase pH values by virtue of
ammonia generation and speciation, calcium consumption Following the MICP treatment process, the quantification and
during CaCO3 precipitation results in proton production and pH removal of generated ammonium byproducts may be of high
decreases when occurring in the absence of continued urea priority (Section 3.4). Such concentrations may be removed
hydrolysis. During MICP treatments, both processes often occur through post-treatment rinse injections and/or other remediation
simultaneously, and solution pH can appear to remain constant strategies, however, monitoring and characterization of
despite the presence of active CaCO3 precipitation. In fact, when ammonium concentrations both during and after removal will
urea is provided in excess, it may not be until near the end of likely remain needed. A brief summary of several approaches
reactions when calcium is finally fully consumed that any pH rise that can be used to monitor ammonium byproducts during MICP
is observed during cementation at all. Modeling and treatments and removal processes is provided below.
measurements of urea and solution pH changes during
4.3.1 Aqueous total ammonium measurements
stimulation and cementation are presented in Figure 36 and
Aqueous total ammonium concentrations can be measured using
further illustrate this point. As shown for cementation solutions
a variety of methods including colorimetric assays and titrations.
containing 350 mM urea and 250 mM calcium, solution pH
One of the most common methods employed is a colorimetric
increases rapidly once 250 mM calcium is consumed (~250 mM
assay that utilizes reactions between ammonium and salicylate
urea degradation) and the remaining 100 mM urea is hydrolyzed
(Krom 1980), wherein variations in aqueous ammonium
with minimal CaCO3 precipitation occurring. From this context,
manifest in differences in blue colorations that develop between
the increase in pH is not indicating a potential for precipitation,
aqueous samples once colorimetric reagents are supplied. Lee et
but rather is indicative of the end of precipitation due to a lack of
al. (2019a, b) used a method similar to Krom (1980) for
calcium. While the authors would not contend that solution pH
measurement of aqueous ammonium concentrations during post-
is entirely useless, as the speciation of
treatment rinsing of bio-cemented soils wherein two reagents
(Reagent A & B) were added to dilute sample volumes, samples
were allowed to sit for a consistent time period to develop
colorimetric differences, and sample absorbance were then
measured at 650 nm. Reagent A consisted of 1.9 mM sodium
nitroprusside, 811 mM sodium salicylate, 387 mM sodium
citrate, and 515 mM sodium tartrate in water. Reagent B
consisted of 1.32 mM sodium hypochlorite and 1.5 M sodium
hydroxide in water. The working concentration range for the
assay can be modified by changing supplied ratios of
colorimetric reagents to aqueous samples and/or sample dilution
factors depending on the magnitude of ammonium
concentrations. While the method quantifies total ammonium
(NH4+ & NH3), the specific concentrations of ammonium (NH4+)
and ammonia (NH3) may also be determined from such
measurements using the pKa of the ammonium/ammonia system
when solution pH is known. An example of aqueous total
ammonium concentrations measured during post-treatment
rinsing was shown earlier in Section 3.4 (Figure 27a).
4.3.2 Adsorbed ammonium measurements
Quantification of adsorbed ammonium masses may be of interest
in addition to pore fluid ammonium concentrations in many
cases. Adsorbed ammonium represents the fraction of positively
charged ammonium molecules that are bound to negatively-
Figure 36. Corresponding urea concentration and solution pH charged soil surfaces as a result of surface association. Such
measurements for standard stimulation (i.e., 350 mM urea, no calcium), concentrations although perhaps bound temporarily, merit post-
enhanced stimulation (i.e., 350 mM urea, no calcium), and cementation treatment characterization due to the potential for subsequent
treatments (i.e., 350 mM urea, 250 mM calcium chloride) with leaching in time and related impacts on pore fluid concentrations.
PHREEQC reaction modeling results. Trends show that solution pH
While a more complete method is described in Lee et al. (2019a),
remains relatively stable during cementation until supplied calcium is
depleted (~250 mM urea degradation) suggesting that solution pH a brief summary of the method is provided here. When
differences alone may provide very little insight about the status of an quantification of adsorbed ammonium masses are desired, moist
MICP system, but may be useful towards assessing reaction completion soil samples can be destructively sampled, frozen, and stored at
(adapted from Gomez et al. 2018b). -20°C until subsequent chemical analyses can be completed.
Once samples are ready for measurements, soils should be first
carbonate/bicarbonate/carbonic acid and ammonia/ammonium
thawed, and an extraction process should be used to extract and
are pH-dependent and pH may have important effects on
quantify NH4+ masses present in soils and pore fluids. First,
microbial enrichment, it is suggested that the possibility or
thawed soil samples are centrifuged to extract free soil solutions.
446
A small sample of the extracted free pore fluid is collected, 4.4 Precipitate characterization
frozen, and stored at -20° C until NH4+ concentrations are
measured. Next, a second soil sample of known water content is The impact of MICP precipitation on the improvement of soil
extracted using a KCl extraction process similar to Keeney & engineering behaviors is determined by the: (1) magnitude of
Nelson (1983). In this process, moist soil samples are mixed with precipitation, (2) mineralogy and morphology of the mineral
a known volume of 2 M KCl solution, allowed to equilibrate for solids generated, and (3) location of the precipitation including
at least 4 hours, and resulting mixtures are centrifuged to remove its association with particle surfaces and contacts.
KCl extracted soil solutions. A small sample of this KCl
extracted pore fluid is then collected, frozen, and stored at -20° 4.4.1 Magnitude of CaCO3
C until NH4+ concentrations are measured. Finally, absorbed The magnitude of CaCO3 precipitation achieved in a given soil
ammonium masses can be determined by accounting for NH4+ at the end of treatment can be evaluated by a number of different
present within free soil solutions as well as adsorbed NH4+ methods, including pressure chamber reactions (ASTM D4373-
masses obtained from the KCl extraction. Adsorbed NH4+ masses 21, Gomez et al. 2014, Choi et al. 2016), gravimetric acid
can then be estimated by subtracting NH4+ masses expected from washing (Rebata-Landa 2007, Mortensen et al. 2011, Burbank et
free solution from NH4+ measurements obtained following KCl al. 2013, Zhao et al. 2014, Li 2015), thermogravimetric analysis
extractions and adsorbed masses can be normalized per gram of (TGA) (van Tittelboom et al. 2010, Park et al. 2014a), x-ray
dry soil. While other approaches may be used, the key objective diffraction analysis (XRD) (Park et al. 2014b), and others (Chu
of such measurements is to quantify the presence of adsorbed et el. 2012, Li 2015, Lin et al. 2016a). None of these methods
ammonium masses rather than simply aqueous concentrations. measure soil CaCO3 contents. Rather, they induce some
As a soil’s cation exchange capacity (CEC) increases, larger measurable change (e.g., pressure chamber increase, mass
masses of adsorbed ammonium are possible and therefore reduction, diffraction of x-rays at specified peaks and intensities)
adsorbed ammonium quantification may be increasingly critical as a result of the procedure applied.
for sites containing soils with high fines contents. An example Pressure chamber reactions provide a reliable,
of sorbed ammonium masses measured in soil samples following straightforward, and low-cost method for analyzing relatively
post-treatment rinsing was shown earlier in Section 3.4 (Figure large soil samples (dry masses of 5 to 20 g). In this process, an
oven-dried soil sample is placed in a sealed vessel of constant
27b).
volume that is equipped with the capability to measure changes
4.3.3 Aqueous calcium measurements in pressure when CaCO3 is reacted with a relatively concentrated
strong acid (e.g., 1-5 M hydrochloric acid) (ASTM D4374-21).
Monitoring of cementation integrity during the ammonium
In this reaction, CaCO3 is dissolved, producing carbon dioxide
removal process may be desired when the potential for
gas and a corresponding increase in the pressure inside the vessel.
cementation degradation is identified. Solution calcium Prior to testing, known reagent-grade calcium carbonate masses
concentrations may provide an opportunity to detect CaCO3 are reacted in the same chamber to create a calibration curve.
dissolution directly, which would elevate solution calcium This relationship between CaCO3 masses and pressures
concentrations. Although calcium ions may be supplied in rinse generated in the vessel can then be used to interpret the unknown
solutions themselves, when other rinse solution chemistries are CaCO3 content of some soil sample. The method is non-specific,
used (e.g., KCl additions), calcium concentrations may be thus other carbonate and gas-generating minerals can also react
monitored during the rinsing process with integration of calcium with the provided acid and may confound results when more
concentrations over injected volumes allowing for determination complex soil mineral mixtures are present.
of CaCO3 mass losses. When completing such measurements, it TGA is similarly a relatively straight forward method. In this
may be important to determine baseline calcium concentrations process, a small dry soil sample of about 30 to 50 mg, is placed
expected in injected solutions and groundwater prior to in a TGA analyzer and the sample is heated at a constant rate
biocementation treatments and rinse injections in order to resolve while monitoring mass loss. In process, mineral solids can
dissolution induced changes. experience various decomposition reactions dependent on their
mineralogy. When calcium carbonate solids are heated to
4.3.4 Shear wave velocity measurements between 600 and 850°C, the mineral is decomposed, releasing
Shear wave velocity measurements can also be used to assess CO2 with only calcium oxide remaining. Ratios between the
cementation degradation during byproduct removal in addition molecular masses of calcium carbonate and carbon dioxide can
to during cementation treatment (see Section 3.4). Although be used in combination with the measured mass loss to estimate
previous studies have shown that when alkaline, cation-rich rinse the soil sample’s CaCO3 content. Due to the small sample size
injections are supplied to biocemented soils only minimal used in this process, obtaining a representative subsample is one
changes in Vs values are observed (Lee et al. 2019a), differences particular challenge.
in groundwater and soil chemistry may merit further monitoring XRD is primarily used as a qualitative tool to detect different
particularly when rinse injections are completed over prolonged crystalline mineral phases and the relative ratio of these minerals
time periods and injected volumes are large. Furthermore, when within some sample. The measurement exposes samples to x-
potential oxidation of residual ammonium is of concern, long- rays at varying inclinations and a detector is used to measure the
term monitoring of shear wave velocity may be used to assess intensity of the diffraction x-rays as a function of diffraction
potential impacts on cementation integrity following treatments. angles, commonly described by 2q. From this process, crystalline
Although such measurements do not provide more specific phases can be identified by comparing observed peaks from some
unknown sample to known diffraction angles for specific
insights regarding solution geochemical changes, such
minerals. This is an inverse problem and requires that the
measurements can provide important feedback during
considered mineral phases have at least a few distinct peaks that
implementation of ammonium removal strategies, which can be
do not appear in all samples. Calcite mineral has expected
used to further guide and modify field management efforts to diffraction peaks at 23 and 65° and its highest intensity peak at
minimize detrimental impacts on cementation integrity. approximately 30°. XRD measurements can be further analyzed
using other software packages such as TOPAS, JADE, or EVA,
to semi-quantitatively determine relative quantities of minerals
447
within a sample. Key to this measurement is the preparation of a
finely ground sample powder and the addition of some control
compound by which the semi-quantitative analysis can be further
refined.
ICP-AES, otherwise known as inductively coupled plasma
optical emission spectrometry (ICP-OES), can be used to detect
particular elements in a given sample. In this process, plasma is
used to excite atoms in some sample of interest, and elements can
be detected by the characteristic wavelengths that they emit
following excitation. For CaCO3 content measurements, a 2 to 15
g oven-dried sample is allowed to react with some acidic solution
to dissolve CaCO3 in the sample releasing Ca2+. The resulting
fluid can then be analyzed for total Ca2+. From the known added
fluid volume and the dry sample mass, the CaCO3 content of the
sample can be determined, assuming a 1:1 stochiometric ratio
between Ca2+and CaCO3 and complete dissolution of all
carbonate minerals.
Gravimetric acid washing is a rather crude method that has
been used by many researchers presumably due to its low cost,
minimal equipment requirements, ability to run many samples,
and need for minimal operator experience. In this process, an
oven-dried soil sample is weighed and subjected to repeated
washing with some strong acid to dissolve and remove carbonate
minerals. After, samples are fully reacted, soil solids can be Figure 37. Comparison of different calcite content methods against the
separated from solutions using a fine sieve (e.g., No. 200) to calcite content measured from the pressure chamber reactor (data from
remove all dissolved CaCO3 reaction byproducts. The retained Choi et al. 2017).
sample is then drained, oven dried, and a final dry mass is tomography (CT), and TGA, as discussed earlier in Sections 2
obtained. The difference between the initial and final dry mass is and 3. These methods require and analyze a small soil subsample
then assumed to be the result of CaCO3 masses lost in the and therefore evaluation of multiple samples may be needed to
specimen. In theory, this approach could account for all
ensure representativeness of observations. Application of these
carbonate minerals in a sample and thus could be used with
techniques is important for a variety of more fundamental
success for some soils. However, this method is also susceptible
reasons but can also be important for pilot studies when ensuring
to erroneous mass losses resulting from the rinsing out of fines
from a given sample. Thus, the method can frequently result in MICP treatment feasibility on sample soils confirming the end
overestimation of precipitated CaCO3 in natural and more well- state condition achieved in the field, and assessing material
graded soils, and would not be recommended for measurements permanence
in these materials. Combined SEM/EDS scans (e.g., Figure 13) provide visual
Choi et al. (2017) performed a study comparing the different SEM images of the precipitate microstructure enabling
methodologies described above, where 20 gram dry soil samples identification of mineral morphologies and qualitative evaluation
were taken at the center of a specimen after an unconfined of the spatial distribution of the precipitate while the EDS spatial
compression testing. Results from their study are reanalyzed in elemental mapping can be used to confirm the elemental
Figure 37 against the ASTM standard pressure chamber reaction composition of the imaged minerals. EDS line and planar scans
method. XRD, TGA, and the pressure chamber reaction method from can be used to spatially map expected elements such as
achieved similar values close to the average of all the methods calcium present in CaCO3 and silicon present in quartz and other
tested. In the authors’ collective experience, the pressure common parent soil minerals. The achieved elemental and visual
chamber reaction method is recommended due to its efficiency, information can be used to qualitatively assess mineral presence
increased sensitivity at low CaCO3 contents, and ease of use. but cannot be used to necessarily confirm mineralogy.
While many of the other methods can be similarly used with As discussed earlier, XRD analyses can be used to assess the
success, acid washing and ICP methods appear to be more mineralogy of soil samples identifying both the parent soil
susceptible to error and are therefore not recommended. minerals (e.g., quartz, feldspar) and precipitated cementation
For all measurement methods it is critical that a baseline (e.g., calcite, vaterite, aragonite). The measurement requires that
CaCO3 content measurement for an untreated soil be obtained the sample be ground into a fine powder (< ~100 m) and that the
and subtracted from post-treatment measurements to more minerals to be identified be crystalline with some definitive
accurately account for potential carbonate contributions from the material structure and thus distinct diffraction planes. Therefore,
parent soil material and the loss of fines, among other effects. for cases such as amorphous CaCO3, which lacks a definitive
Further care should be taken to rinse soils prior to measurements structure, this method is indeterminant. For such samples, other
to remove other soluble salts that may have resulted from the processes such as Fourier Transform Infrared Spectroscopy
drying process. For example, deionized water or ethanol can be (FTIR) may prove more useful. An example of XRD data
equilibrated with samples, and subsequently removed before obtained from biocemented Ottawa F-65 sand samples treated
drying and performing measurements. Additional controls are using augmented S. pasteurii at 3.4 x 108 cells/mL and a solution
also needed when significant carbonate minerals are present in of 250 mM urea and calcium is presented in Figure 38 (Burdalski
the parent soil material themselves such as the case for MICP & Gomez 2020). As shown, measured XRD diffraction patterns
treatment of biogenic carbonate sands. can be used to semiquantitatively fit known diffraction peaks for
different CaCO3 mineral polymorphs using various software
4.4.2 CaCO3 mineralogy, composition, & structure packages and diffraction databases thereby allowing for mineral
The mineralogy and composition of achieved precipitate can be quantities by mass to be estimated. In this particular sample, it
evaluated through a variety of different microscale techniques,
including scanning electron microscopy (SEM), energy-
dispersive spectroscopy (EDS), XRD, x-ray computed
448
Figure 38. X-ray diffraction measurements from a biocemented Ottawa F-65 test specimen and untreated Ottawa F-65 sand are compared to known
reference patterns for calcite, vaterite, and aragonite, with (a) patterns from measurements and calcite shown in an overview and (b) patterns for other
mineral polymorphs shown in a magnitude view. From such measurements, semiquantitative analyses can be used to determine compositional differences
with the test specimen shown containing ~83% calcite, ~9% vaterite, and ~8% aragonite (adapted from Burdalski & Gomez 2020).
was determined that the material had ~83% calcite, ~9% vaterite, While not yet demonstrated extensively, emerging methods
and ~8% aragonite. The observed diffraction patterns are shown including low-field nuclear magnetic resonance (NMR) may
in Figure 38a for the biocemented test specimen and the provide yet another method to examine soil microstructural
untreated sand along with known values for calcite with smaller changes during the MICP process. For example, Kirkland et al.
known peaks for vaterite and aragonite compared to the (2017) found that this method was sensitive to both physical and
biocemented test specimen in a magnitude view shown in Figure chemical changes caused by MICP including but not limited to
38b. As provided, such mineralogical compositional values may changes in soil pore size distributions, porosities, and
only be accurate to about 3 to 5%, however, such trends can still permeabilities.
provide important insights regarding mineralogical differences,
especially between similar test specimens wherein isolated 4.5 Non-destructive monitoring
factors are varied (Burdalski et al. 2022a). Lastly, when more
precise mineralogical compositional changes are desired, the Non-destructive monitoring methods can be particularly useful
authors recommend supplementing samples with some standard towards evaluating real-time improvements in engineering
material at known masses to act as a reference. properties during the biocementation process with no detrimental
Three-dimensional x-ray CT scans (Figure 16) can provide effects on further experimentation. Such measurements provide
some of the greatest insights regarding the spatial distribution of the opportunity to assess the current state of some material,
CaCO3 precipitation within a soil particle structure. With this thereby allowing for the number of cementation treatments
technology, scanning quality is governed by the maximum necessary to achieve target criterion to be adapted. Geophysical
achievable image resolution, which is typically about 0.01% of methods ,such as shear wave velocity (Vs), compression wave
the specimen dimension through which rays are penetrated. velocity (Vp), and resistivity measurements are commonly used
Hence, the detection of a CaCO3 coating with a thickness of 5 to monitor the mechanical improvement that can be realized with
m on a soil particle surface, would require a specimen size no MICP. Shear wave velocity and compression wave velocity
larger than 5 mm. Recent work by Roy (2021) has used CT data measurements induce essentially negligible strains (~10-6) within
to analyze the spatial distribution of CaCO3 as well as the effect a soil (Santamarina et al. 2001), while resistivity induces no
of this precipitation on soil pore space geometries and overall strains. In addition to these other methods, recently, spectrally-
connectivity. Although infrequently examined, such induced polarization (SIP) (Saneiyan et al. 2018) has been
characterizations can also allow for changes in soil particle investigated and shown to be sensitive to the presence of
microstructure and/or fabric to be quantified. This can include microbial cells (Ntarlagiannis et al. 2005), biofilm formation
the use of CT reconstructions to evaluate changes in soil particle (Davis et al. 2006), and other biogeochemical processes (Flores
and bond orientations as has been done by Terzis & Laloui Orozco et al. 2011).
(2018) and others. Observations from CT scans have been
instrumental in conceptually understanding how MICP treatment 4.5.1 Shear wave velocity measurements
can change the mesoscale soil properties and be used to guide Shear wave velocity (Vs) measurements have been extensively
particle-scale based investigations of biocemented soil used to monitor MICP progression due to their sensitivity to
engineering properties using discrete element codes (PFC3D) as changes in soil particle contact stiffnesses, their direct relation to
discussed further in Section 6.
449
soil small-strain shear modulii (Gmax = Vs2), and general m/s under fully saturated conditions and insensitive to changes
insensitivity to pore fluid saturation conditions. In the laboratory, in cementation. At higher cementation levels, the stiffening of
Vs measurements can be obtained by measuring the time required the soil skeleton can then be detected by Vp as compression wave
for a shear wave to propagate a known distance between a pair velocities become increasingly influenced by the cemented
of bender elements, which are typically constructed from particle matrix. This phenomenon was further explored by
piezoceramic materials (Stokoe et al. 1985, Lee & Santamarina cycling the saturation conditions of a cemented soil while
2005a). When completed appropriately, bender element measuring Vs and Vp concurrently. Figure 42a shows observed
measurements yield Vs values that are consistent with resonant Vs and Vp seismic wave velocities during dry and saturated
column measurements (Brignoli et al. 1996, Lee et al. 2005). conditions (Weil et ail. 2011). While the Vs measurements
Vs measurements have been widely used as an effective tool remain relatively constant, there is a noticeable difference in Vp,
to monitor the progressive improvement realized with each with the Vp of the dry specimen clearly increasing from about
MICP treatment and relationships between soil CaCO3 contents. 280 to 1400 m/s with increases in CaCO3 contents while the Vp
Figure 39 presents results from a range of studies wherein soil under saturated conditions begins at 1500 m/s and gradually
type, confining stresses, testing length scales, and biological and increases to about 1900 m/s.
cementation treatment schemes (open symbols are The evolution of Poisson’s ratio (n) with increasing
bioaugmentation, closed symbols are biostimulation) were all cementation, determined through measurements of Vs and Vp
varied. As shown, laboratory experiments include results from using Equation 18, can also be determined for both dry and
studies involving simple shear, triaxial, and 1D rigid column saturation conditions. Figure 42b shows the effect of saturation
specimens. Upscaling experiments include 1g models with on Poisson’s ratio, with the Poisson’s ratio being higher for the
length scales ranging from 0.5 to 3.7 m. System level saturated condition than the dry condition for all cases, but with
experiments include results from models tested on 1-m and 9-m values decreasing with cementation level increases. Specifically,
radius centrifuges. As evident, there is clear proportionality as CaCO3 content increased, the Poisson’s ratio for the saturated
between soil CaCO3 contents and shear wave velocities invariant condition decreased from 0.50 to 0.43. For the dry condition, the
of testing length scales, confining stresses, soil types, and Poisson’s ratio appeared to initially increase from 0.3 to 0.38
treatment schemes. However, the amount of Vs increase per with cementation, but then decreased to 0.33. A reduction in
percent soil CaCO3 by mass can vary considerably between Poisson’s ratio with increasing cementation is consistent with
studies. The source of this variability is difficult to isolate due to measurements by Gercek (2007) where the Poisson’s ratio for
the many inconsistencies between these studies, however, calcite was 0.31 but for intact sandstone, which can be thought
important differences affecting these relationships likely include of as an analog for biocemented sands, the ratio ranged from 0.05
differences in soil composition (classification, mineralogy, grain to 0.4.
size distribution, grain shape), prepared initial state (density,
coordination number), preparation method (moist tamped, air
pluviation), treatment scheme (formulations, stimulated versus (18)
augmented), and confining stress conditions during cementation
(confining stress, isotropic versus one-dimensional).
This same data set can be presented as a change in shear wave The ratio of Vp/Vs for both saturated and dry conditions versus
velocity ( Vs) versus soil CaCO3 by mass, as shown in Figure Vs is plotted in Figure 42c. As evident in the broader literature
40, which removes some dependence of these measurements on (Kramer 1996), this ratio increases as the Poisson’s ratio
the initial specimen conditions prior to treatment. Additional increases. As shown, data by Weil et al. (2011) for soils of
analysis of a data subset controlling for soil and experimental varying cementation agreed well with these expected trends
design showed that the Vs per percent calcite is not dependent under both saturated and dry conditions. Such trends also further
on whether stimulation or augmentation was used. confirmed that the observed Poisson’s ratio differences with
Figure 41 presents the same data set binned on the basis of higher cementation contents were also largely reasonable given
fines content and D10 values. Clean soils, defined as having less the related changes in Vp/Vs ratios.
than 3% fines, have a larger Vs increase per CaCO3 content than
sands with greater fines contents. This is reasonable as higher 4.5.3 Resistivity measurements
fines content can reduce precipitation efficiency due to sorption Direct and indirect resistivity measures are directly dependent on
of calcium ions to clay mineral surfaces (Gomez & DeJong 2017) how strongly a soil resists electric current, which in turn is
and the overall soil skeleton stiffness. The same trend is observed dependent on numerous factors including soil porosity, mineral
by separation according to the 10 values. As shown, the rate of composition, pore fluid composition, interactions at the soil
Vs increase per CaCO3 content increases as the 10 increases. particle surface, pore fluid saturation, and soil fabric. The
This apparent trend may result from the correlation between 10 biogeochemical and mechanical changes that occur during a
values and fines contents, however, it may also reflect other single MICP treatment cycle and over a treatment program can
physical mechanisms including but not limited to differences in make it difficult for resistivity measurements to reliably discern
soil particle coordination numbers and cementation magnitudes specific changes (e.g., chemical reaction progression versus
at each particle contact. amount of CaCO3 precipitated). Nonetheless, its potential for
tracking specific parameters has been demonstrated.
4.5.2 Compression wave velocity measurements Direct resistivity measurements can be readily implemented
Compression wave velocities are a function of the constrained by measuring the induced voltage between an electrode array
modulus (M = Vp2) with sensitivity to the degree of pore fluid (Rinaldi et al. 2006) and can be effective in monitoring isolated
saturation, soil Poisson's ratio, and Young’s modulus (Weil et al. soil changes when other conditions (including pore fluid
2011). Consequently, the utility in Vp for monitoring MICP
improvement can be obscured by saturation conditions.
Weil et al. (2011) observed that Vp only began to increase
significantly once the bulk modulus of the soil skeleton became
comparable to that of the pore fluid (i.e., governed by the
stiffness of water wherein Vp = ~1480 m/s, Lee and Santamarina
2005b). At low cementation, the compression waves primarily
travel through the pore fluid causing the Vp to be close to 1500
450
Figure 39. Shear wave velocity versus calcite content measurements from (a) laboratory experiments, (b) upscaling experiments, and (c) system level
testing in the centrifuge (closed symbol = stimulated, open symbol = augmented).
chemistry) remain constant. For MICP treatments, it can be resistivity) of a given treatment solution gradually increases as
difficult to monitor soil improvement since the chemical ureolysis progresses due changes in the presence of charged
composition of the pore fluid is continuously changing in a species (ammonium and carbonate). As an example, Whiffin
complex manner (DeJong et al. 2009). Weil et al. (2011) was (2004) used measured pore fluid conductivity rates to estimate
unable to detect significant and reliable changes that correlated urea hydrolysis rates. During cementation, however, monitoring
with CaCO3 precipitation during MICP treatments. However, with direct conductivity can be complicated by the continuous
changes that occur during the treatment enrichment stage, such precipitation of CaCO3 which can accordingly decrease
as during stimulation, may offer some possibility for improved conductivity. This is shown in Gomez (2017) wherein the ionic
monitoring. During stimulation, the conductivity (inverse of strength of a cementation solution was essentially constant
451
Figure 40. Increase in shear wave velocity versus calcite content for the
same set of experiments shown in Figure 39 differentiated by testing
length scales and reported enrichment method (closed symbol =
stimulated, open symbol = augmented).
during the cementation process due to the production of
ammonium and carbonate species and simultaneous
consumption of calcium and carbonate ions. However, if
employed urea to calcium concentration ratios significantly
exceed 1:1, it may be possible to detect when the calcium
Figure 41. Comparisons of increases in shear wave velocities versus
concentration has been depleted and excess urea is being calcite contents grouped by (a) fines contents and (b) D10 (mm).
consumed, thus generating excess ammonium ions in solution.
Indirect resistivity measurements can be obtained using hydraulic conductivity can be used as an indirect indicator of the
spectrally-induced polarization (SIP), which has been used in the magnitude of CaCO3 precipitation but can also indicate when
past for mineral exploration (Moon et al. 2006) and most recently CaCO3 precipitation is occurring non-uniformly, which can lead
in near surface environmental applications (Williams et al. 2009, to localized clogging and/or increased injection pressures during
Flores Orozco et al. 2012, Ntarlagiannis et al. 2016). In SIP, the treatments.
conductivity magnitude and phase are measured and analyzed to Baek et al. (2022) performed a series of tests in rigid columns
determine the real and imaginary components of complex equipped with differential pore pressure measurements from
conductivity, wherein the real component represents energy loss within the test specimen. Hydraulic conductivity was determined
while the imaginary component provides some insights through differential pressure transducer (DPT) measurements
regarding the energy stored. More specifically, complex obtained during steady state flow and during falling head tests,
conductivity measurements and polarization signals may provide with the latter expected to be adversely influenced by clogging
information about precipitation magnitudes and crystal sizes in the device plumbing and/or near the injection or extraction
which influence the polarization magnitude and relaxation times, specimen boundaries. MICP treatments were performed once
respectively (Wu et al. 2010). Saneiyan et al. (2018) further daily until a total CaCO3 content of about 20% was achieved.
suggested that SIP can allow for more long-term monitoring due Tests were performed on poorly graded soils with D50 values of
to its sensitivity to both carbonate mineral formation and 0.18, 0.51, 1.39, and 2.74 mm (100A, 100B, 100C, 100D), and
dissolution. A proof-of-concept experiment was conducted by final conditions at the end of treatment are shown in Figure 16.
Saneiyan et al. (2019) where an aquifer was stimulated in a 15- Figure 43 presents variations in hydraulic conductivity with
day field-scale MICP project. SIP was shown to have the respect to the CaCO3 contents, along with results from previous
potential to spatially and temporally track the progression of studies on clean sands (e.g., Al Qabany & Soga 2013, Martinez
MICP in the treated zone at this site. et al. 2013, Gomez & DeJong 2017). Continued CaCO3
precipitation causes gradual reductions in k in all sand columns
4.5.4 Hydraulic conductivity measurements up to a CaCO3 content of ~25% (Figure 43a). Figure 43b presents
Calcite precipitation can reduce the hydraulic conductivity (k) of variations in k normalized by its initial value, k/ko, with respect
a soil, with the magnitude being dependent on the amount and to CaCO3 content. CaCO3 precipitation of ~20% only causes a
spatial distribution of the precipitation. Consequently, changes in reduction in k of less than one order of magnitude (<70%). The
452
likely occurring near the specimen boundaries and/or in the
device plumbing.
Comparison of trends with previous studies (Al Qabany &
Soga 2013, Martinez et al. 2013, Gomez & DeJong 2017) show
that prior results indicate much more significant k reductions.
This significant difference could possibly be attributed to two
primary mechanisms. First, the treatment formulation could have
triggered initial precipitation in the pore space, which then
sedimented by gravity near pore throats forcing accelerated
reduction in k. Second, localized precipitation may have
occurred near the injection and/or extraction boundaries, or
within the device plumbing, resulting in a localized regions with
reduced k and misinterpretation of the actual global k values.
The observed trends can be reasonably modeled with a family
of Kozeny-Carman models. Specifically, the Kozeny grain-
coating model can be used which correlates k/ko with the pore
volume fraction of calcite Scc:
Figure 42. Changes in (a) seismic wave velocity and (b) Poisson’s ratio
with calcite content and (c) relationship between Vp/Vs ratio and
Poisson’s ratio (data from Weil et al. 2011.).
pore-scale imaging of the CaCO3 precipitation achieved also
further corroborated this minimal reduction in k. This is
consistent with the observed changes at the pore scale, as evident
in Figure 16, and in measured CaCO3 content distributions along
the length of the columns.
Changes in D50 showed an insignificant effect on the
Figure 43. Variations in (a) hydraulic conductivity and (b) normalized
normalized reduction rate in k for the tested sands that were hydraulic conductivity with an increase in calcite content for different
uniformly graded with Cu values of less than 1.8. Importantly, in soils. Note the open symbols are measured by using the differential
fine sands (100A and 100B) the k values measured by the falling pressure transducer or the piezometers connected to two ports at the
head method begin to deviate from the k values obtained by the column wall, and the solid symbols are measured by using the falling
DPT when CaCO3 contents >10%, indicating local clogging head method (adapted from Baek et al. 2022).
453
cementation gradient across the tested specimens. The average
(19) mid-depth CPT qc measurements respectively increased from a
baseline value near 3.6 MPa to a maximum value of 32.1 MPa.
where n is the power exponent (Kleinberg et al. 2003, Noh et al. Darby et al. (2019) (described in more detail in Section 7)
2016, Baek et al. 2019). This model is simple but robust, and the measured CPT qc pre-treatment, post-treatment, and following
model with n = 2 is identical to the original Kozeny-Carman select shaking events in the 1-m radius centrifuge at the Center
model. Figure 44 shows the model predictions while varying the for Geotechnical Modeling at the University of California, Davis.
empirical power exponent n, with n = 2 to 6 reasonably capturing Three different treatment levels were investigated: including
the observed k reductions. Previously published data show more lightly cemented ( Vs = 140 m/s), moderately cemented ( Vs =
significant k reductions, hence an n > 10 is required for
agreement with the previously reported data and may be again
somewhat confounded by the above testing limitations.
4.6 Cone penetration test measurements
Cone penetration (CPT) tip resistance, qc, is directly dependent
on soil stiffness and strength (Lunne et al. 1997), and is therefore
effective in estimating a wide spectrum of geotechnical
engineering parameters (e.g., relative density, friction angle,
stiffness, compressibility) capable of predicting engineering
performances, including liquefaction triggering (e.g., Idriss &
Boulanger 2008, Boulanger & Idriss 2014). For these reasons,
CPT soundings are frequently used to (1) evaluate initial baseline
site conditions and design the soil improvement programs,
including the treatment zone extent and desired improvement
magnitude, by performing soundings prior to an improvement
program and (2) verify improvement effectiveness by
performing additional soundings after the improvement program
has been completed. The efficacy of the CPT in evaluating
improvement maps directly to how MICP improves soils, with
CPT qc and Vs being the most effective measures of detecting
interparticle cementation and verifying treatment effectiveness.
Figure 44. Reduction in hydraulic conductivity (k/ko) with respect to pore Figure 45. qc profiles in untreated, lightly cemented, moderately
volume fraction of calcite (Scc). The curves are plotted with Kozeny grain cemented, and heavily cemented centrifuge models for the initial soil
model: k/ko = (1 – Scc)n+1. CC at the upper x-axis is calculated from the conditions as well as prior to shaking and after all shaking was completed
Scc assuming e of 0.667 and GS calcite of 2.71 (adapted from Baek et al. 2- (adapted from Darby et al. 2019).
22).
The effectiveness of CPT measurements to evaluate MICP 270 m/s), and heavily cemented ( Vs = 590 m/s). Figure 45
improvement has been primarily performed in large 1-g tank tests presents representative CPT qc profiles for untreated, after MICP
and in centrifuge tests. Gomez et al. (2018a) performed MICP treatment to various treatment levels, and following numerous
treatment on two large 1g tank tests, with CPT soundings shaking events that triggered liquefaction and degraded the
performed pre- and post-treatment in the sand layer to determine cementation. The CPT qc increased due to MICP treatments
how qc relates to CaCO3 content. CPT qc measurements were relative to the uncemented condition, and then degraded
obtained across a range of soil CaCO3 contents, from untreated following severe shaking. For example, at 4 m depth, q c for the
to 5.3%, as the overall test objective was to achieve a lightly, moderately, and heavily cemented models increased by a
454
Figure 46. qc, Vs, and calcite content relationships of data compiled from 1-g tank tests (Gomez et al. 2018a) and 1-m radius centrifuge tests (Darby et
al. 2019, San Pablo et al. 2022b).
factor of 2.2, 4.3, and 9.1, respectively. Following shaking,
however, the respective qc values largely decreased with
improvement factors of only 2.1, 2.6, and 4.8, respectively. (20)
Despite MICP degradation following numerous shaking events,
the qc at the end of the tests remained higher relative to the initial with the stress normalized penetration resistance, qc1N, defined as
uncemented condition, indicating that MICP at higher qc1N = (qc/pa)/( ’vo/pa)0.5 where pa is atmospheric pressure and
magnitudes could permit more permanent changes to qc. These ’vo is the vertical effective stress. For the centrifuge tests
results are consistent with that observed more recently by San
conducted by Darby et al. (2019), the corresponding average KG
Pablo et al. (2022b) in similar centrifuge tests.
factors for these soil profiles are 117, 1133, 2048, and 4117 for
Given the observed dependence of qc on cementation level
shown in Figure 46, and the dependence of Vs on cementation cementation levels of 0, 0.8, 1.33, and 2.18%. Figure 47b
level shown earlier (Figure 39), it is logical that qc and Vs are also presents results from 1-g and centrifuge tests for KG versus
correlated. Compiled results from 1-g and centrifuge tests where CaCO3 content. For reference, KG values between 110-330 are
qc, Vs, and CaCO3 content were measured are shown in Figure representative of Holocene and uncemented sands, whereas KG
46. As evident, both qc and Vs are correlated with CaCO3 content, values between 330 and 1100 represent aged, cemented, and
albeit it with some scatter. Nonetheless, trends show Vs to calcareous sands (Schneider & Moss 2011). Figure 48b shows
increase more rapidly than qc with CaCO3 content, particularly at uncemented tests to agree well with the range proposed by
low cementation levels. Direct comparison of qc versus Vs, Schneider & Moss (2011), while most of the MICP cemented
shown in Figure 46c, captures the relative rates of improvement. values are above a KG of 1100, indicating a strong microstructure
In aggregate, this data indicates that Vs is more sensitive than qc due to cementation and bonding that likely exceeded that
for cementation levels less than about 3% CaCO3 content by considered when developing the original KG framework.
mass, while at higher cementation levels both measures are Figure 49 presents the soil behavior type charts from
comparably sensitive to added cementation. The changes in q c Robertson et al. (1995) and Robertson (2016) where Go/qc is
and Vs relative to their respective uncemented baseline values, again used to differentiate the presence of cementation. Both
shown in Figure 47, further clarifies this relationship. Again, Vs charts successfully delineate uncemented soils from cemented
is more sensitive at low cementation levels than qc, while both soils, and furthermore the normalized parameters systematically
track added cementation effectively at higher cementation levels. track increasing cementation. These results are encouraging and
The influence of increased stiffness and strength for cemented suggest that existing CPT-based methods for identifying and
sands can be evaluated together considering the ratio of the quantifying cementation in sands can be directly applicable to
small-strain modulus, G0, to qc (Eslaamizaad & Robertson 1996, and effective for MICP improved sands.
Schnaid et al. 2004). As evident in Figure 48a, the ratio measured
in 1-g and centrifuge tests increases to a value of about 30 at 0.5%
CaCO3 content and asymptotically approaches a value of about 5 EXPERIMENTAL INVESTIGATIONS OF STRESS-
60 at 5% CaCO3 content. The small strain rigidity index (KG), STRAIN BEHAVIOR
adapted from Rix & Stokoe (1991) by Schneider & Moss (2011),
Biocementation of sands fundamentally changes conditions at
explicitly accounts for the impact of both relative density and
effective stress on qc, and is defined as: the particle scale, which map directly to distinct and significant
455
changes in element behavior when subjected to monotonic or
cyclic loading. As detailed in Section 2.4, precipitation within the
soil matrix on particle contacts and exposed particle surfaces
results in a variety of important microstructural changes
including: (1) cementation of particle-particle contacts, (2)
densification, and (3) increased particle angularity. The
following section is not an exhaustive review of past
experimentation, but rather provides a summary of the primary
mechanical behaviors of biocemented silica sands relative to
untreated silica sands.
Figure 48. (a) (Go/qc) and (b) KG versus calcite content of data
Figure 47. qc and Vs relationships of data compiled from 1-g tank
compiled from 1-g tank tests (Gomez et al. 2018a) and 1-m radius
tests (Gomez et al. 2018a) and 1-m radius centrifuge tests (Darby et al.
centrifuge tests (Darby et al. 2019, San Pablo et al. 2022b).
2019, San Pablo et al. 2021b).
50 mm diameter specimens that were augmented and received
cementation treatments under a constant confining stress in a
5.1 Small-strain dynamic behaviors
resonant column (RC) device (GDS Instruments, further
Cementation at particle contacts directly increases the soil matrix described in Na et al. 2022). It is noted that for specimens with
shear stiffness as evidenced by the linear increase in the shear higher cementation-induced initial stiffnesses, the maximum
wave velocity in proportion to the amount of CaCO3 precipitation strain levels achievable in the RC device became more limited
(up to ~10% CC) which indicates that non-destructive shear (i.e., elastic shear strain threshold decreased with increasing
wave velocity measurements are an effective real-time cementation).
cementation monitoring method (Section 4.4). The presence of The effect of cementation on the modulus reduction curves is
cementation also results in significant changes to the strain immediately evident in Figures 50a and 50d, with Gmax being
dependent shear modulus degradation and damping curves. proportional to the achieved CaCO3 contents, and secant G
However, the magnitude of these changes depends on progressively decreasing with increased strain. The threshold
cementation levels and confining stresses, among other shear strain at which the modulus reduction begins
variables. Although shear wave velocity measurements can (G=0.99*Gmax) decreases as cementation increases. This is more
effectively measure initial small-strain shear stiffness (Gmax) evident in the normalized G/Gmax curves in Figures 50b and 50e.
changes, resonant column testing can provide more The increase in Gmax follows a systematic trend with the relative
comprehensive insights into how cementation changes dynamic magnitude being proportional to the level of cementation
small-strain soil properties relevant to site response changes. similarly to what is seen for the Vs measurements. For example,
for the tests performed at 100 kPa confinement, the strain
Figure 50 presents results of resonant column tests
corresponding to G/Gmax = 0.5 decreases from 0.1 to 0.07 to 0.02
performed on Ottawa F-65 sand prepared to an initial Dr of 61%
to 0.01% as the cementation level is increased from 0 ( Vs = 0
and treated to various levels of cementation with confining
m/s) to 0.5 ( Vs = 91 m/s) to 2 ( Vs = 251 m/s) to 3.6 ( Vs =
stresses of 30 and 100 kPa. Varying cementation levels were 350 m/s) CaCO3 by mass. The threshold shear strain also
achieved through consecutive MICP treatment of untreated sand decreases systematically from roughly 0.002% for the
specimens, which resulted in shear wave velocity increases uncemented case to 0.001, 0.0006 and 0.0002% for cementation
between ~100 and 350 m/s, corresponding to CaCO3 contents levels of 0.5, 2 and 3.6% CaCO3 by mass. This is presumably
between 0.5 to 3.6%, respectively. Tests were performed using attributable to the brittleness of the CaCO3 cementing agent. The
456
broader understanding of dynamic soils properties for cemented
soils from past studies (e.g., Acar & El-Tahir 1986, Delfosse-
Ribay et al. 2004, Dupas & Pecker 1979, Haeri et al. 2005,
Saxena & Reddy 1987, Saxena et al. 1988, Simatupang et al.
2018, Tsai & Ni 2012).
The dependence of the small-strain shear modulus on
confining stress decreases with increasing cementation; at low
cementation levels, the shear modulus closely follows the stress
dependence documented for uncemented soils (Hardin & Black
1968, Stokoe et al. 1985), but becomes essentially independent
of confining stress for cementation levels above about 7% CaCO3
(or a Vs of about 650 m/s). Figure 52 presents results from
bender element measurements obtained in both triaxial and
centrifuge model experiments for a range of cementation levels
subjected to multiple confining stress unloading and reloading
cycles between about 1 and 80 kPa. As the cementation level
decreases, two observations are evident. First, stress cycling
gradually degrades the initial stiffness and integrity of the
cemented soils, resulting in G decreasing. Second, the
dependence on confining stress increases as the cementation
level decreases, as evidenced by the increasing slope of G with
confining stress.
These results show that, with increasing cementation, initial
shear moduli increase and threshold shear strains decrease,
collectively resulting in the G/Gmax and damping curves shifting
to lower strain levels. Further, the effect of confinement on
cementation integrity as well as modulus reduction and damping
curves is diminished as cementation levels increase and the
material approaches a stress-independent condition.
5.2 Monotonic stress-strain behaviors
5.2.1 Response to drained monotonic triaxial shearing
The effect of cementation on monotonic shearing behaviors,
and its dependence on the induced microstructural changes, are
evident in drained triaxial tests on miniature 1 cm diameter
specimens. Figure 53 presents shearing results for uncemented
and moderately cemented (CC ~5%) specimens. The details of
the first set of tests are detailed in Tagliaferri et al. (2011), with
a follow-up study across a range of cementation levels presented
in DeJong et al. (2017).
The triaxial tests were performed on Ottawa 50-70 sand (D50
= 0.21 mm, Cu= 1.0) under an isotropic confining stress of 100
Figure 49. (a) (Go/qc) and (b) KG versus calcite content of data compiled
kPa applied via vacuum. The specimen was sheared at a rate of
from 1-g tank tests (Gomez et al. 2018a) and 1-m radius centrifuge tests
(Darby et al. 2019, San Pablo et al. 2022b). 0.021 mm/min and paused at select global axial strain levels, at
which point a 3D X-ray computed tomography (CT) scan of the
effect of confining stress leads to an increase in Gmax and a shift specimen was obtained with a voxel resolution of about 15 mm
in G degradation to the higher strain levels, as shown in Figure by gradually rotating the specimen 360º while scanning
51a. The increase in confinement also extends the range the shear intermittently. Decreases in deviator stress shown in Figure 53a
strains that remain elastic, as shown by the threshold shear strain correspond to the stress relaxation that occurred while the
increasing from about 0.0002% for a confinement of 30 kPa to shearing process was paused for scanning.
0.0008% to a confinement of 200 kPa (Figure 51b). The stress strain response shown in Figure 53a for the
The corresponding effect of cementation on soil damping is cemented specimen exhibits the expected differences relative to
shown in Figures 50c and 50f. At very small strains, the the uncemented clean sand condition. Consistent with the
minimum damping increases with cementation when compared macroscopic observations described earlier and shown in Figures
to the uncemented soil. At strains greater than the threshold shear 50 – 53, added cementation induces a stiffer response at small
strain, damping begins to increase at lower strains in the most strains. Additionally, a higher peak strength is attained at a lower
highly cemented specimen followed by lower cementation levels strain level, followed by softening to a residual strength that
and the uncemented case. For example, for specimens at 100 kPa approaches a critical state condition comparable to that observed
confinement, the shear strain corresponding to 2% damping in the uncemented sand. This trend indicates that, as shearing
occurs at shear strains of 0.035, 0.01, 0.006, and 0.00025%, for progresses, the cementation bonds gradually break and therefore,
the 0 ( Vs = 0 m/s), 0.5 ( Vs = 91 m/s), ( Vs = 250 m/s), and the effects of the cementation on the small strain behavior
3.6% ( Vs = 350m/s) CaCO3 specimens, respectively. diminishes. At the tested cementation level (~3% CaCO3), this
Additionally, as confinement increases the degradation of moderate cementation magnitude does not appear to have a
cementation is suppressed until larger shear strains are achieved. significant effect on larger-strain behaviors after cohesive bonds
This is reflected by an increase in the threshold shear strain, as have been largely degraded.
shown in Figure 51b. However, above the respective threshold
shear strain the rate of modulus degradation appears to increase
with CaCO3 content. These trends are largely consistent with a
457
Figure 50. Modulus reductions and damping curves obtained from resonant column tests performed on Ottawa F-65 sand (Dr = 61%) at confining stresses
of 30 and 100 kPa at varying levels cementation ( Vs = 0 to 350 m/s) (adapted from Na et al. 2022).
Figure 51. Modulus reductions and damping curves obtained from resonant column tests at increasing confining stresses of 30, 100, and 200 kPa (adapted
from Na et al. 2022).
The presence of cementation forces the formation of shear the cementation level. Specifically, the local shear strains within
banding within the specimen where shearing and cementation the shear band corresponding to mobilization of the peak strength
degradation is localized, and critical state conditions may be was near 10% for the uncemented specimen, averaged about 4%
achieved. In contrast, cemented regions outside of the shear band for a specimen with a CaCO3 content of about 3.5%, and
remain largely intact with minimal deformations. This is evident decreased to about 1% for a specimen with a CaCO3 content of
in Figure 53b, where the mapped local porosity and shear strain 4.8%. The shear band for the uncemented case was rather diffuse
changes between select stages (beginning, middle, and end of the and developed gradually, eventually reaching a thickness in
test) are shown. Blue boxes indicate the progression in the excess of ~30 particle diameters after an axial global strain of
biocemented sand and the red boxes indicate the response of the about 5%. In contrast, the shear band became more localized with
uncemented sand. There was no observed evidence of increasing cementation level. The 3.5% CaCO3 specimen began
localization in the biocemented sand sample prior to the peak to localize after 3% shear strain and reached a shear band
strength, while in the uncemented specimen, significant dilation thickness of about 8 particle diameters at a shear strain of 5.8%,
is observed in a localized zone prior to peak. while the 4.5% CaCO3 specimen further localized within a shear
Further analyses of additional tests performed across a range band of about 6 particle diameters. Evidently, cementation
of cementation levels (DeJong et al. 2017) identified how the stabilizes the soil matrix, resulting in a more localized failure
progressive formation and thickness of the shear band depend on
458
mechanism within a shear band, with the shear band thickness (0% CaCO3) to values of 300 m/s (1.0% CaCO3), 650 m/s (1.3%
decreasing with increasing cementation. CaCO3), 1100 m/s (3.1% CaCO3), and 1400 m/s (5.3% CaCO3)
are shown in Figure 54.
The mean stress normalized deviatoric stress (q/p') versus
shear strain measurements shown in Figure 54a, highlight a
progressive increase in initial stiffness and peak strength with
cementation, followed by proportional softening towards a
constant residual normalized strength. Normalization of the
deviatoric stress with the initial isotropic consolidation stress
(q/p'c), shown in Figure 54b, captures the substantial absolute
increase in deviatoric stress with cementation, owing to the
significant increase in negative excess pore pressure generation
(Figure 54c).
The resulting q-p' stress paths for the tests, shown in Figure
54e, reveal several notable changes in behavior as the
cementation increases. Firstly, the initial stress path from the
isotropic consolidation state for the uncemented specimen
follows the expected vertical path consistent with initially
perfectly elastic behavior. This is immediately followed with a
reduction in p' due to positive pore pressure generation and
leftward curvature of the stress path indicative of contractive
volumetric tendencies. Once the CSL is reached, the specimen
follows the failure envelop typical of a clean sand with a cohesive
intercept of 0 kPa and a friction angle of 33º. In contrast, as
cementation increases, the increased dilatancy of the cemented
soil matrix affects the initial stress path at small strains, resulting
in the deviatoric load inducing an increase in the effective stress
as negative excess pore pressures are generated. The initial path
Figure 52. Shear stiffness results from triaxial and centrifuge model slope is 3.6:1 for the specimen treated to a Vs = 1400 m/s. It then
experiments at various cementation levels subjected to multiple stress follows that strain softening of q is not observed as strains
level cycles between 1 and 80 kPa (adapted from Montoya et al. 2013). progress in this particular test. Projecting the failure envelopes of
the treated specimens to the origin of the axes results in a
cohesion intercept and a friction angle which tend to increase
5.2.2 Response to undrained monotonic triaxial shearing with the level of cementation. This increase in strength
The increased stiffness, increased peak strength, and strong parameters (cohesion and friction angle) is more significant for
dilatancy observed at larger strains for biocemented sands under the cohesion but can also be observed, to a lesser degree, for the
drained conditions map directly to undrained conditions resulting friction angle.
in the generation of significant negative excess pore pressures Bender element measurements obtained throughout this
during the shearing process. Results from a series of 70 mm testing enabled the tracking of how the normalized shear
diameter undrained triaxial tests performed on Ottawa 50-70
sand prepared at a Dr of 40% under an isotropic confining stress
of 100 kPa following cementation treatments that increased the
Figure 53. Small-scale x-ray CT scanned triaxial test results including (a) stress-strain for cemented and uncemented specimen and (b) strains and
porosity values from image analysis of 3D CT scans with the upper red image series being for the uncemented specimen and the lower value image
series being for the cemented specimen (adapted from Tagliaferri et al. 2011).
shear wave velocity from the untreated baseline of Vs = 190 m/s
459
modulus degraded during the shearing process, defined as
G/(p/pa)0.5, to account for the changing effective stress conditions
due to excess pore pressure generation (Figure 54d). Due to the
placement of the bender elements in the top and bottom cap of
the apparatus, the shear wave velocity measurements can be
interpreted as being average values across the sheared sample,
meaning that the transmitted waves from the bender elements
traveled within an intact region (away from the shear band) as
well as an increasingly narrow sheared region (within the shear
zone) of the sample as the level of cementation increased.
Consequently, the magnitude of the reduction in the normalized
shear modulus (Figure 54d) is less than the actual decrease
expected within the shear band, which at high strains leads to
residual values that do not converge between conditions.
The breadth of monotonic drained and undrained triaxial tests
in the literature is extensive and has reaffirmed and expanded the
trends summarized here. Systematic, secondary dependencies
have been mapped for the effect of MICP cementation on clean
sand specimen behavior in triaxial tests across variations
including relative density (e.g., Tsukamoto et al. 2013, Xiao et
al. 2019), confining stress (e.g., Feng & Montoya 2015, Lin et al.
2016a, Nafisi et al. 2020b), stress paths (e.g., Montoya & DeJong
2015, Nafisi et al. 2021), and treatment solutions (e.g., Terzis et
al. 2016, Xiao et al. 2021a). The observations have also been
extended to variations in soil classification (size and gradation)
and mineral composition (e.g., Simatupang et al. 2018, Nafisi et
al. 2018, Nafisi et al. 2020b).
Aggregated data from the literature contains variability given
the (sometimes significant) differences in the soil characteristics,
MICP treatment procedure implemented, test device type and
compliance, and testing protocols and procedures. Nonetheless,
trends of properties dependent on CaCO3 content can be clearly
defined. For example, as detailed in a comprehensive synthesis
by El Kortbawi et al. (2022a) and presented in Figure 55, the
apparent cohesion increases with cementation at a proportion of
about 100 kPa increase in the cohesive intercept per 5 to 6%
increase in CaCO3 content and the peak friction angle on average
increases by about 5º per 5% increase in CaCO3 content
(observations consistent with those observed by Choi et al.
2020).
Degraded MICP can be readily re-established through re-
treatment as the open pore structure maintained after initial
MICP treatments allow for additional treatment injections as
desired. The effectiveness of retreatment following damage from
monotonic loading has been investigated using laboratory
triaxial tests, while the impacts of earthquake-induced damage
on cementation effectiveness have been investigated using
centrifuge model tests (Montoya & DeJong 2013). Figure 56
from (Montoya & DeJong 2013) shows the engineering behavior
to 10% axial strain in undrained triaxial testing following initial
treatment to a Vs = 650 m/s (1.3% CaCO3), and the subsequent
behavior following retreatment to an additional 10% axial strain
(strain redefined as 0% strain after retreatment). As evident, the
monotonic engineering behavior can be fully recovered through
retreatment.
5.2.3 Response to drained monotonic simple shearing
The previously described behaviors observed in drained and
undrained monotonic triaxial tests have been similarly observed
in direct simple shear tests on biocemented and uncemented
sands. A series of drained monotonic direct simple shear tests
were performed on loose (Dr = 30%) biocemented specimens of
varying cementation magnitudes ( Vs ~ 0 to 500 m/s) under a
vertical effective stress of 100 kPa (Lee et al. 2021, Lee & Gomez
2022). Figure 57 provides representative drained monotonic
responses for specimens of varying cementation including (a) Figure 54. (a) q/p' versus strain, (b) q/p'c versus strain, (c) u/p'c versus
shear stress versus shear strain, (b) shear stress versus vertical strain, (d) q versus p', and (e) G/(p'pa)1/2 versus strain for uncemented
effective stress, (c) shear stress versus log-scale shear strain, and ( Vs = 190 m/s) and cemented samples ( Vs = 300, 650, 1100, 1400
m/s) (adapted from Montoya & DeJong 2015).
(d) axial strain versus shear strain responses. Similar to drained
460
Figure 56. (a) stress ratios and (b) excess pore pressure generation for
undrained triaxial compression tests on untreated and biocemented
Ottawa 50-70 sand specimens, with the latter sheared after initial MICP
treatment and after MICP retreatment following shearing (adapted from
Montoya and DeJong 2013).
2021) and reflect the limited contribution of volumetric
behaviors to peak shear strengths observed in biocemented soils
with cementitious bonds instead providing the majority of the
observed peak strength enhancements. Mobilization of dilative
behaviors, although significantly impacted by the presence of
added cementation, instead occurs post-peak only after
interparticle shearing has been more fully mobilized. Progressive
increases in large strain shear strengths (strain > 5%) can also be
observed once cementation magnitudes exceed Vs values near
80 m/s. These large strain changes may be due to changes in
Figure 55. (a) cohesion vs CaCO3 content and (b) peak friction angle interparticle friction from the coating of soil particles surfaces
versus CaCO3 content trends summarized from multiple research studies with CaCO3 crystals, which could increase interparticle
(adapted from El Kortbawi et al. 2022a). frustration and constant volume friction angles.
triaxial tests, progressive increases in shear stiffnesses and peak
shear strengths can be observed with increases in cementation. 5.2.4 Response to undrained monotonic simple shearing
While an untreated specimen exhibited a peak shear strength near Undrained monotonic trends from DSS tests show similar shifts
in shear strength, stiffness, and volumetric behaviors as
75 kPa, the most highly cemented specimen ( Vs ~ 485 m/s)
undrained triaxial tests with increases in cementation
achieved a peak shear strength near twice as large near 140 kPa
magnitudes. A similar series of undrained monotonic direct
(Figure 57a). Furthermore, stress-strain responses show a
simple shear tests were performed on loose (Dr = 30%) bio-
progressive shift from strain-hardening to strain-softening
behaviors with increases in cementation. Significant cemented specimens of varying cementation magnitudes ( Vs ~
enhancements in initial shear stiffnesses can also be 0 to 500 m/s) subjected to an initial vertical effective stress of
distinguished from these responses when plotted versus shear 100 kPa (Lee et al. 2021, Lee & Gomez 2022). Figure 58
strains in log-scale (Figure 57c). When considering differences provides representative undrained monotonic responses for
in volumetric behaviors shown in Figure 57b and 57d, the specimens of varying cementation including (a) shear stress
uncemented and very lightly cemented samples were largely versus shear strain, (b) shear stress versus vertical effective
contractive as expected, while almost no volumetric changes stress, and (c) excess pore pressure versus shear strain responses.
were observed for the specimen treated to a light level of As shown in Figure 58b and Figure 58c, large differences in
volumetric tendencies were observed with even small
cementation near a Vs of 80 m/s. As cementation magnitudes
magnitudes of cementation. Although all samples exhibited
increased, however, strong dilation during shearing was
initially contractive behaviors at shear strains of less than 0.5%,
observed. In more heavily cemented specimens ( Vs > 300 m/s)
as expected, at larger shear strains, significant differences in
the mobilization of volumetric strains occurred only after the
peak strength was observed. Such behaviors are consistent with
other tests from the literature (e.g., Wu et al. 2021, Wang et al.
461
Figure 57. Summary of monotonic drained direct simple shear tests Figure 58. Summary of monotonic undrained direct simple shear tests
performed on specimens of varying biocementation ( Vs ~ 0 to 500 m/s) performed on specimens of varying biocementation ( Vs ~0 to 500 m/s)
including (a) shear stress versus shear strain, (b) axial strain versus shear including (a) shear stress versus shear strain, (b) excess pore pressure
strain, and (c) shear stress versus shear strain (in log-scale) (adapted from versus shear strain, and (c) shear stress versus vertical effective stress
Lee & Gomez 2022b). (adapted from Lee & Gomez 2022b).
excess pore pressure generation were observed. Uncemented ~ 100 m/s. Large negative excess pore pressures in more highly
specimens exhibited contractive excess pore pressure responses cemented samples resulted in greater effective stresses during
expected for loose of critical specimens. As cementation shearing and higher undrained shear strengths as expected. For
magnitudes increased, however, more limited volumetric the most highly cemented specimen ( Vs ~ 500 m/s), negative
behaviors were again observed at very light cementation ( Vs of excess pore pressures exceeded 600 kPa at shear strains near 6%
~ 44 m/s). Within increasing cementation, increasingly dilative and continued to increase with added strain. While such pore
behaviors were observed particularly once Vs values exceeded pressure values may be unrealistic for shallower soils relevant to
462
liquefaction mitigation applications, due to the cavitation of pore summary of how soil properties and monotonic behaviors are
fluids under suction pressures, the observed responses do provide expected to change with the addition of cementation. Table 3
insights regarding the progressive shift in undrained mechanical details how a range of properties and behaviors change as
behaviors with added cementation. Responses plotted in shear cementation increases for uncemented loose (Dr ~ 30%) poorly-
stress versus vertical effective stress space (Figure 58b) further graded silica sands when subjected to undrained and drained
suggest that while more lightly cemented specimens ( Vs < ~ monotonic shearing.
200 m/s) may follow a similar failure surface as uncemented
sands, this surface may shift with added cementation with greater 5.3 Cyclic stress-strain behaviors
differences observed once Vs values exceed ~300 m/s. The undrained cyclic behavior of biocemented soils has been
characterized at the element-scale using triaxial (Hayashi et al.
5.2.5 Summary of monotonic behaviors
2010, Burbank et al. 2013, Feng 2015, Sasaki & Kuwano 2016,
Collective observations from the above studies and broader
Simatupang & Okamura 2017, Simatupang et al. 2018, Xiao et
literature, in combination with fundamental understandings of
al. 2018, 2019) and direct simple shear (DSS) testing (Montoya
the MICP precipitation process, including the interparticle
et al. 2013, Zamani & Montoya 2018, 2019, Riveros &
distribution, morphology, mineralogy, and strength of the
Sadrekarimi 2020, Lee et al. 2020, 2021). Such tests have
achieved precipitates (Section 2 and 3), enables synthesis and
demonstrated the ability of biocementation to alter liquefaction
Table 3. Summary of the effect of biocementation on monotonic soil behaviors.
Effect as Cementation Increases on Uncemented Loose Poorly Graded Silica Sand
Property/Behavior
Subjected to Monotonic Undrained/Drained Shearing
Shear wave velocity increases at about 50-125 m/s per 1% CaCO3 content. Modulus
Small-strain Shear Modulus computed with G = r Vs2 . Density changes can be determined via CaCO3 while assuming
(Shear Wave Velocity) specific gravity of calcite comparable to silica (Gs ~ 2.71). When Gmax > ~400 MPa
(CaCO3 > 3%) modulus of biocemented soils is largely independent of confining stress.
Threshold shear strains decrease with cementation due to increased brittleness of
Threshold Shear Strain cementitious bonds. Threshold strains decrease from about 0.002% for uncemented sands
to 0.0002% at ~3.5% CaCO3 at 100 kPa confinement.
Shear modulus degradation occurs at lower strains with added cementation due to
Shear Modulus Degradation increasing brittleness. Strain to reach G50 at 100 kPa decreases by 10x relative to untreated
condition when CaCO3 is ~3.5%.
Damping Damping initiates at smaller strains due to lower threshold shear strain.
Shearing becomes increasingly localized with higher cementation due to increased
Shear Localization dilatancy. Effects are most pronounced at low confining stresses. Shearing zone decreases
down to about 8 particle diameters thick when CaCO3 is ~3.5%.
Dilation rate increases rapidly as CaCO3 increases due to cementation and densification.
Dilation Rate The strain level at which the maximum dilation rate occurs shifts increasingly beyond the
strain level corresponding to peak strength as the CaCO3 increases.
Peak shear strength increases regardless of drainage conditions or shearing mode, with
Peak Shear Strength characteristics similar to soft rock as CaCO3 increases (i.e., stress-invariant). Tensile
strength increases as CaCO3 increases.
Strain at which peak strength is mobilized decreases with CaCO3 due to increased initial
Strain at Peak Strength stiffness and failure localization. Strains at peak decrease from ~3.5% on average for the
uncemented condition, down to ~0.5% or less at 10% CaCO3 .
Apparent MC cohesive intercept increases due to cementation bonds, from 0 kPa for
MC Cohesive Intercept untreated sand up to about 350 kPa for 10% CaCO3 . Failure envelope shifts with the
accumulation of shear strain due to progressive damage.
Peak friction angle increases due to cementation bonds and increased interparticle friction,
Peak Friction Angle
from typical uncemented sand values of 30-34º to about ~43º for 10% CaCO3 .
Volumetric tendencies transition from contractive towards dilative due to added
Volumetric Tendencies cementation, increased soil density, and increased interparticle friction. Magnitude of
change depends primarily on cementation level and confining stress.
Critical state friction angle appears to change insignificantly at CaCO3 less than 3-5% . At
Critical State Friction Angle higher CaCO3 , the angle begins to increase slightly, however no significant changes are
evident at less than ~10% CaCO3 .
463
behaviors as well as considered a suite of other variables relatively minor for biocemented sands results (Lee et al. 2021),
including the presence of soil fines (Sasaki & Kuwano 2016, therefore liquefaction triggering resistances are discussed herein
Zamani & Montoya 2018, 2019), calcareous soils (Xiao et al. on the basis of shear strain development. Results from undrained
2018, 2019), and differences in specimen saturation during cyclic tests on loose biocemented Ottawa F-65 sand similar to
treatments (Simatupang & Okamura 2017). Collectively, these those shown earlier in Figure 59 can be further analyzed to assess
studies have clearly shown that biocementation can increase soil changes in liquefaction triggering resistance with changes in
liquefaction resistances, however, understanding of the effects of cementation magnitudes and applied loading demands. Figure 60
factors including confining stress changes, relative density presents (a) the number of cycles to 3% SASS for bio-cemented
differences, static shear stresses, cementation contents, and loose sand specimens versus Vs values reflective of
applied loading magnitudes as well as the post-triggering cementation-induced initial shear stiffness enhancements and (b)
consequences of such loading events have remained less the number of cycles to 3% SASS for bio-cemented loose sand
understood. Ongoing studies will continue to improve specimens of varying cementation and uncemented specimens
with varying initial relative densities (Dr = 30% to 85%) versus
understanding of the behaviors, particularly for biocemented
specimen dry densities as reflected by both dry density and
clean sands, however, the current understandings and more
porosity values. Although all bio-cemented specimens were
recent findings are summarized herein.
prepared initially loose (Dr = 30%) and cemented to varying
degrees, uncemented specimens were prepared to varying initial
5.3.1 Response to undrained cyclic simple shearing densities (Dr = 30% to 85%) to compare the effects of
Recent DSS testing has examined the undrained cyclic behavior biocementation to that of solely densification. As shown in
of loose (Dr initial = 30%) biocemented Ottawa F-65 Sand treated Figure 60a, near log-linear improvements in liquefaction
to varying cementation magnitudes ( Vs ~ 0 to 500 m/s) under triggering resistances can be observed with increases in Vs
an initial vertical effective stress of 100 kPa and subjected to values even at relatively low magnitudes of cementation ( Vs <
varying cyclic loading demands (Lee et al. 2021, Lee & Gomez 100 m/s, CaCO3 content < 0.9%). At higher cementation levels
2022b). All specimens were sheared under equivalent undrained (indicated by greater Vs increases), increases in applied CSR
conditions to achieve a double amplitude shear strain (DASS) of were necessary in order to trigger liquefaction over numbers of
24%, after which specimens were reconsolidated back to the loading cycles more representative of earthquake applications (<
initial effective stress of 100 kPa. 1,000 cycles). Such investigations preliminarily suggest that as
Representative responses for biocemented specimens treated applied loading magnitudes increase (i.e., CSR values increase)
to Vs values between 0 m/s (uncemented) and ~500 m/s, the effect of Vs or cementation increases may become less
subjected to Cyclic Stress Ratio (CSR = / 'v initial) values of pronounced, as indicated by trendlines with increasingly flatter
either 0.2 ( Vs = 0 m/s, 95 m/s), 0.75 ( Vs = 194 m/s, 294 m/s), slopes. At higher CSR values, a threshold amount of cementation
or 1.5 ( Vs = 476 m/s) are presented in Figure 59. Provided may also be required for specimens not to achieve 3% SASS
responses include measurements of shear stress versus effective monotonically.
vertical stress (column i), shear stress versus shear strain (column Figure 60b presents the number of cycles to 3% SASS for
ii), and excess pore pressure ratio (ru = u/ 'v initial) versus cycle biocemented and uncemented specimens versus dry density and
number (column iii). Modest increases in biocementation porosity values. As shown, cemented samples exhibited large
magnitudes can substantially decrease excess pore pressure increases in triggering resistances with minimal changes in dry
generation and shear strain development early on during densities (~ 0.43 to 0.93 %) under conditions where applied CSR
undrained cyclic loading for a given applied loading magnitude values were less than 0.3. When higher CSR magnitudes were
(i.e., CSR value). For example, for tests subjected to a CSR of applied, however, higher cementation magnitudes were needed
0.2, a small amount of cementation corresponding to to achieve similar numbers of cycles to 3% SASS and more
Vs = 95 m/s resulted in an increase in the number of cycles to significant cementation-induced increases in specimen dry
ru =0.95 from approximately 0.5 cycles to nearly 650 cycles densities were observed (~1.60 to 4.76%). Triggering resistance
when compared to the untreated condition. Similar enhancements for biocemented specimens, however, can be seen
improvements were observed at higher CSR values with the to exceed those achievable through dry density increases alone
number of cycles to ru =0.95 increasing from approximately 2.5 as shown by comparing biocemented specimens to uncemented
cycles to 575 cycles as Vs increased from 194 to 294 m/s. When specimens at similar dry densities and loading magnitudes.
considering shear stress versus vertical effective stress responses The relationships between applied CSRs versus number of
(column i), shifts in behaviors can occur, with significant cycles to liquefaction triggering (3% SASS) for specimens of
increases in vertical effective stresses and negative ru values varying cementation are presented in Figure 61. For the same
CSR value, the number of cycles to liquefaction triggering
observed early during cyclic loading with an abrupt transition to
increased by 2 to 3 orders of magnitude even for lightly cemented
increased contraction at some threshold cycle number after
specimens ( Vs < 100 m/s, CaCO3 < 0.9%). While extrapolated
which cemented bonds are degraded and the volumetric
trends are provided for higher cementation levels, ongoing
tendencies of the specimen become increasingly mobilized. At investigations will provide a more comprehensive assessment of
larger shear strains following liquefaction triggering, higher how such curves may vary with increasing cementation. These
magnitudes of cementation resulted in an increased number of liquefaction triggering curves can be described using the power-
cycles required to accumulate similar post-triggering shear law function CSR = a × , where Ncyc is number of cycles to
strains, however, such improvements were less pronounced than 3% SASS, and a and b values are parameters that capture the
triggering improvements. Failure envelopes observed during vertical position and slope of these curves, respectively. In past
tests exhibited minimal changes in apparent cohesion with studies, the slope of liquefaction triggering curves (b values)
increased cementation, however, clear shifts in such envelopes have been shown to decrease (Hernandez 2018, Xiao et al. 2018,
were evident during the loading process and were likely 2019), remain constant, and even increase (Riveros &
reflective of changes in interparticle friction and cohesive Sadrekarimi 2020) with increasing cementation. This functional
bonding as bonds degraded and interparticle shearing became
increasingly mobilized.
Although liquefaction triggering can be defined using either
shear strain (single amplitude shear strain [SASS] = 3%) or
excess pore pressure (ru= u/ 'v initial = 0.95) criteria, recent
tests have suggested that differences between criteria are likely
464
Figure 59. Responses of loose (Dr = 30%) biocemented sand specimens of varying cementation magnitudes ( Vs = 0 to ~500 m/s) subjected to differing
applied loading magnitudes including plots of (i) shear stress versus vertical effective stress, (ii) shear stress versus shear strain, and (iii) excess pore
pressure ratio versus cycle number (adapted from Lee & Gomez 2022b).
465
form did not appear to reasonably capture the trends observed for s values exceed ~250 m/s, after which more
bio-cemented specimens in Lee et al. (2021) and given the significant improvements may be observed with increases in
discrepancies in the literature further investigation is necessary cycles required to obtain the same shear strain increases
to assess the adequacy of power law trends to fit the data. approaching values near one order greater (Figure 62a). When
Figure 62 presents number of cycles from 3% to 24% SASS comparing trends instead on the basis of dry densities (Figure
s for loose specimens of varying cementation (Figure 62b), minimal differences in post-triggering strain accumulation
62a) as well as dry density and porosity differences for loose can be observed between uncemented and cemented specimens.
biocemented specimens and uncemented specimens of varying Although further investigations are needed, such trends suggest
initial relative densities (Figure 62b). More modest that cementation-induced density changes may be the primary
improvements in post-triggering strain accumulation can be contributor towards influencing post-triggering strain
observed, in contrast with earlier liquefaction triggering accumulation with the effects of cementitious bonding becoming
resistances that observed orders of magnitude increases with diminished at larger-strains once bonds become largely sheared
added cementation. Although for the tests presented in Figure 62 and cemented soil matrices become destructured.
both applied CSR values and cementation levels were varied, the The s for loose specimens
achieved trends collectively suggest that post-triggering strain of varying cementation and dry density and porosity differences
accumulation may be only minimally impacted by added for loose biocemented specimens and uncemented specimens of
varying relative densities are presented in Figures 63a and 63b,
respectively. Reconsolidation strains can be similar for all loose
specimens up to cementation s values of ~200 m/s,
above which more considerable reductions in reconsolidation
strains were observed. At Vs values exceeding ~ 300 m/s
reconsolidation strains of less than 0.5% were observed in
comparison to baseline values near 2.5% observed for loose
uncemented specimens (Figure 63a). When considering
reconsolidation strains as a function of dry density and porosity
values (Figure 63b), trends for cemented specimens appeared to
s values
of ~ 200 m/s, after which significantly lower reconsolidation
strains were observed for cemented specimens. These trends
were similar to those observed for post-triggering strain
accumulation and suggest that while higher cementation levels
may alter post-triggering reconsolidation at larger cementation
magnitudes, such enhancements were not as significant as pre-
triggering enhancements. That said, reconsolidation strains may
be altered at higher cementation levels than the post-triggering
strain accumulation behaviors discussed earlier, which may
reflect increased shear localization with added cementation and
the presence of largely intact cemented regions in more highly
Figure 61. Liquefaction triggering curves for biocemented specimens at
constant cementation levels (as expressed by Vs values) including
applied cyclic stress ratio (CSR) values versus number of cycles to
liquefaction triggering (as expressed by number of cycles to 3% SASS)
Figure 60. Number of cycles to liquefaction triggering (as expressed by (adapted from Lee et al. 2021).
number of cycles to 3% SASS) versus (a) Vs from cementation, and (b)
soil dry densities and porosities (adapted from Lee & Gomez 2022b).
466
s from
cementation for loose (Dr ~ 30%) specimens of varying bio-cementation s from cementation for
and (b) Dr for all bio-cemented specimens and uncemented specimens loose (Dr ~ 30%) specimens of varying bio-cementation levels and (b)
prepared at varying Dr (adapted from Lee & Gomez 2022b). total Dr for all bio-cemented loose specimens and uncemented specimens
prepared at varying initial Dr (adapted from Lee & Gomez 2022b).
cemented specimens following the cyclic loading and
reconsolidation process. and particle surface roughness changes. Current understanding
nonetheless allow for broader trends in undrained cyclic
5.3.2 Summary of cyclic behaviors behaviors with added cementation to be summarized. Table 4
Recent cyclic triaxial and direct simple shear testing has provided details how behaviors shift as cementation levels increase for
new insights regarding the liquefaction behavior of loose uncemented loose (Dr ~30%) poorly-graded silica sands when
biocemented sands. These results suggest that even small subjected to undrained cyclic shearing.
magnitudes of cementation can dramatically alter soil
liquefaction triggering resistances, however, improvement of 6 NUMERICAL INVESTIGATIONS OF STRESS-STRAIN
larger strain post-triggering behaviors including strain BEHAVIOR
accumulation and reconsolidation strains may require higher
magnitudes of cementation. The liquefaction behaviors of loose Changes induced to the mechanical behavior of sands through
clean sand s< MICP treatment evident in the experimental element test results
100 m/s) has been more extensively investigated recently at the (Section 5) have been studied numerically as well. Discrete
element scale, however, the impacts of higher cementation levels element modeling, wherein particle-scale properties and
for clean sands are still being explored through element- and mechanical contact laws are defined to provide insights into
systems-level testing (i.e., centrifuge testing). Such testing will micro- and element-scale behaviors, and constitutive models,
likely shed further light on potential improvements in post- wherein the element response are cast within a continuum scale
triggering behaviors due to cementation-induced densification
467
Table 4. Summary of the effect of biocementation on cyclic soil behaviors.
Effect as Cementation Increases on Uncemented Poorly Graded Silica Sand
Property/Behavior
Subjected to Cyclic Undrained Shearing
Number of cycles required to reach 3% SASS increases rapidly with cementation
primarily due to cementation induced initial shear stiffness improvements and to a lesser
Triggering
degree increased dilative tendencies. At CSR = 0.2, N increases from <1 cycle for
Resistance
uncemented sand to ~200 cycles for DVs = 100 m/s (@ 100kPa, Dr = 30%). Increase in
(3% SASS
cycles primarily occurs at very small shear strains. Once ~0.5-1% strain is exceeded, the
Criteria)
rate of strain accumulation is more similar to uncemented sands but depends on
cementation level.
Pore pressure generation up to r u = 0.95 largely maps with 3% single amplitude shear
Excess Pore strain accumulation for DVs < 100 m/s. Minimal excess pore pressure is generated per
Pressure cycle in cemented sands initially, followed by an abrupt increase at some threshold r u
Generation and shear strain whereas pore pressure in an uncemented sand is more gradual for all
cycles.
CSR vs. N "a" value increases with cementation. Reported changes in "b" values are not consistent
Fitting between past studies but have been shown to primarily decrease with added cementation
(CSR = a N-b ) for similar conditions. Functional form may require modification for cemented sands.
Post-triggering Number of cycles required to achieve an increase from 3% to 9% SASS increases with
Strain cementation level. Magnitude of decrease negligible at DVs < 100 m/s , but become
Accumulation significant at DVs > 250 m/s.
Post-triggering Post-liquefaction reconsolidation strains decrease with cementation level. Strain
Reconsolidation reductions are insignificant at DVs < 100 m/s due to limited cementation-induced
Strains densification, but become significant at DVs > 200 m/s.
constitutive model for upscaled system modeling, have both been defining the cementation level (tensile capacity) to have a
successfully implemented. This section reviews and highlights Weibull (or some other) distribution (Evans et al. 2015, de Bono
these current capabilities for simulating the behaviors of et al. 2015, Shen et al. 2016). This has allowed for more
biocemented sands. progressive failure of the cementation at contacts with loading,
resulting in a more gradual global specimen strain softening
6.1 Discrete element models & simulations response.
Inclusion of a physical representation of the CaCO3
Discrete element modeling (DEM) provides a platform to precipitate as a solid mass to investigate how its presence alters
interrogate at the particle-scale the hypotheses outlined in particle angularity and the total density (as well as the
Section 2 and 3 regarding how the cementation at particle- cementation) has been limited to one study (Evans et al. 2015).
contacts and on particle surfaces, the increased particle In this more realistic, but computationally expensive,
angularity, and the increased density drive the observed element implementation of the CaCO3 precipitate binding two sand
engineering behaviors detailed in Section 5, including altered particles together has been realized with a ring of miniature
stiffness, strength, dilative tendency, failure localization, and particles that are bound to the two sand particles. This approach
residual strength. has the potential to produce greater insights into load transfer,
failure progression, bond breakage, and coordination number
6.1.1 MICP representation in DEM evolutions with continued shearing.
The effect of MICP on the interparticle contact properties, the
increase in particle angularity, and the increase in total density 6.1.2 Monotonic stress-strain behavior – triaxial loading
have been studied to varying degrees. The inclusion of The DEM investigation by Feng et al. (2017) of the drained
cementation to the contact law can be incorporated in a triaxial behavior of MICP-cemented Ottawa 50-70 sand at
straightforward manner through parallel bonds, which follow a varying levels of cementation is representative of the agreement
Mohr-Coulomb failure criterion with a tension cut-off. The achievable and the insights produced when modeling the
inclusion of a bond model with a tensile capacity has been the cementation bond strength as being uniform with a parallel
primary modification in a majority of studies, both for MICP contact law. Their observations are generally consistent with
studies (Feng et al. 2017, Dong & Fatahi 2020) as well as for studies by Yang et al. (2017, 2019) and Abbas et al. (2020).
sands cemented by other means (Wang & Leung 2008, Obermayr Feng et al. (2017) performed a baseline calibration of the
et al. 2013, Jiang et al. 2015, Shen et al. 2016). particle properties using the measured experimental response of
The cementation strength (i.e., tension cut-off) has been loose, medium, and dense untreated Ottawa sand when subjected
typically defined as a constant value for all particle-particle to drained triaxial loading under 100 kPa confining stress.
contacts throughout a simulated specimen. However, this often Additional tests of loose sand MICP treated to light, moderate,
leads to the simulated responses being more brittle than MICP and heavy CaCO3 contents of 1, 3, and 4.3%, respectively, were
experimental data. More recent studies have defined the used for evaluation of the simulations. Calibration of the parallel
precipitation at particle contacts to have a variable distribution bond contact model was performed calibrating the properties to
throughout the simulated specimens, often modelled through obtain agreement with the specimen response of the heavily
cemented specimen, and the bond radii was subsequently
468
reduced proportionally to the cementation level to simulate the contacts per particle and is thus used to measure the
lightly and moderately cemented specimens. interconnectivity of an assembly of particles. During loading up
Figure 64 presents the DEM and experimental results for the to peak strength, which occurs at 0.3% axial strain, more than
lightly, moderately, and heavily cemented specimens in terms of 98% of bonds remain intact, which is consistent with the stiffness
stress-strain response and volumetric change. As evident, and remaining nearly linear up to the peak strength. As the peak
expected with direct calibration, good agreement was achieved strength is reached, the bond breakage increases and 20% of the
in the initial stiffness and peak strength. The rate of strain bonds have been broken between 0.3% and 0.5% strain. By 3%
softening post-peak is more rapid in the DEM simulation, which axial strain, about 50% of the bonds have been broken, which is
is attributed to the parallel bond contact model and the sufficient to generate a shear band across the specimen.
assignment of uniform cementation properties to all bonds in the Additional bond breakage with continued shearing to 10% axial
specimen. The residual strength decreases below the strain is relatively minimal. The coordination number evolves
experimental results at large strains, which is attributed to the with strain in a consistent manner. The initial average
cementation model not capturing the increase in density
coordination number of about 6 remains constant until the peak
associated with the presence of cementation. The volumetric
strength, after which it degrades rapidly to a value of about 5 by
response in the DEM simulations is initially more gradual than
0.5% axial strain and then remains at this value to 10% strain.
the experimental data, but then continues to increase at large
strains throughout the shearing process due to the simulated The average coordination number of the entire specimen reported
triaxial walls being more rigid than a latex membrane and does obscure the coordination number present within the shear
thereby amplifying the specimen volumetric change (see Yang et band that would be expectedly lower.
al. 2019 for implementation of a flexible membrane boundary). Improvements in DEM modeling of MICP treated soils with
The insights regarding how the cementation bonds break and incorporation of the MICP precipitated mass, have captured the
the coordination number changes with added strain are more gradual strain softening, residual strength values, and
provided greater insights into behavior of the contacts (Evans et
exemplified in Figure 65 using data from the heavily cemented
al. 2015, Khoubani et al. 2016, Khoubani 2018, Dong and Fatahi
2020). Khoubani (2018) implemented the CaCO3 precipitate
binding two sand particles using a ring of small CaCO3 particles
Figure 64. DEM numerical simulations and experimental specimen Figure 65. DEM numerical simulations for a heavily cemented specimen
data of (a) stress-strain and (b) volumetric behavior for varying exemplifying (a) bond breakage percentage and (b) coordination number
cementation levels (adapted from Feng et al. 2017). changes with axial strain level (adapted from Feng et al. 2017).
specimen. Coordination number refers to the average number of that are bound to the two sand particles and to each other (Figure
469
66). In this implementation, contact laws and associated rapidly, which is consistent with cementation bond breakage
properties are defined at sand-sand contacts, sand-calcite occurring primarily within 1.0% axial strain. The sand-sand
contacts, and calcite-calcite contacts. Calibration of the coordination numbers degrade to a value of about 5 during strain
parameters generally followed Feng et al. (2017), with additional softening and continue to decrease as straining continues. The
details required given the greater number of parameters. The sand contacts with both sand and calcite particles decreases to
parametric studies performed by Khoubani (2018) on two sand about 3.5 at 3% strain, and then remains fairly constant.
particles bound together showed more realistic behaviors with Finally, Figure 68c reveals how the global boundary stress is
respect to how a cementation bond degrades when subjected to distributed through the contact forces in the simulated specimens.
compression and tension, shear, bending, and twisting. The Prior to shearing, the isotopic global stress of 200 kPa is carried
cemented fragments that resulted from breakage occupy the pore solely through both the vertical and radial sand-sand contact
space and affect the system response as well. forces. The calcite-calcite and calcite-sand contact forces are
Figure 67 presents the DEM and experimental results for zero as the calcite bonds were formed after the specimen was
CaCO3 contents of 1.2, 2.4, and 5.3%, corresponding to lightly, subjected to the confining stress, consistent with experimental
moderately, and heavily cemented specimens, in terms of stress- processes. However, after only 0.5% axial strain, about 80% of
strain response and volumetric change for undrained drained the contact forces are transferred to calcite contacts. This occurs
triaxial loading at a confining stress of 200 kPa. As evident, good since the calcite-sand and calcite-calcite contacts have a tensile
agreement is achieved throughout the stress-strain curve,
including the initial stiffness, peak strength, rate of strain
softening, and residual strength. In particular, the more gradual
progression of cementation bond breakage this contact model
provides produces a more gradual softening of the stiffness as the
peak strength is approached as well as the subsequent non-linear
post-peak softening. The volumetric tendency at small strains is
also well captured, though the more rigid boundaries still
magnify the dilative response at higher strain levels.
Interrogation into particle scale trends provided in the DEM
simulation of the 5.3% CaCO3 specimen are shown in Figure 68.
Bond degradation with strain, separately for calcite-sand and
calcite-calcite bonds as well as in aggregate, are presented as a
percentage of respective total bonds in Figure 68a. The total bond
breakage percentage is generally consistent with Feng et al.
(2017), however the rate of change is more gradual at small strain
levels both before and after peak strength. About 50% of the
calcite-calcite bond failures occur by 0.5% axial strain, which
coincides with the peak strength. More than 80% have failed by
1% strain, and more than 90% fail by 10% strain. The percent of
sand-calcite failures are substantially less and occur more
gradually with further strain, with about 13, 30, and 60% reached
at strains of 0.5, 1.0, and 10%, respectively. This suggests that
the internal strength of CaCO3 is weaker than the calcite-sand
bond strength, an outcome which is also consistent with
experimental tests on two particles bonded together (Section
Figure 66. Image of spherical sand particles bound together with rings of Figure 67. DEM numerical simulations and experimental specimen data
small calcite particles binding sand particles together (adapted from of (a) q and (b) volumetric strain versus axial strain for varying
Evans et al. 2015). cementation levels (adapted from Khoubani 2018).
2.4).
capacity, while the sand-sand contacts do not. By 1.0% strain,
Figure 68b shows the coordination number of the sand
which corresponds to post-peak strain softening conditions, the
particle contacts with sand and CaCO3 particles, of sand particles
force distribution begins to rebalance, with sand-sand contacts
with other sand particles, and of CaCO3 particles with sand and
starting to carry a larger percentage of the forces. This trend
other calcite particles. For all metrics, the coordination numbers
are highest initially and the calcite particle contacts degrade most
470
continues, with the forces being approximately equally 6.1.3 Cyclic stress-strain behavior – simple shear loading
distributed at 3.8% strain, and after 10% strain, the sand-sand Very few DEM studies have been performed on the cyclic
contacts carry about 67% of the forces due to the breakage of the behavior of biocemented sands with Yang (2018) simulating
calcite bonds. cyclic simple shear loading and Zhang et al. (2022a) simulating
cyclic triaxial loading. Yang (2018) performed undrained cyclic
simple shear simulations of untreated and MICP treated sands.
Calibration of sand and cementation parameters was performed
using the consolidated undrained triaxial test simulation
approach detailed in Yang et al. (2019) and corresponding
experimental triaxial data of untreated and treated poorly graded
Ottawa 20-30 sand (O’Donnell et al. 2017a). The simple shear
device implemented was a square box where the leading and
trailing walls induced a uniform shear strain across the specimen
height during stress-controlled cycles at a frequency of 1 Hz.
Figure 69 shows images of a confined specimen within the rigid
walls and the boundary conditions applied, and the contact force
distribution with the specimen prior to cyclic shearing.
Figure 70 presents the DEM simulation results for untreated
and cemented specimens when subjected to a CSR = 0.30 under
a vertical confining stress of 100 kPa. As evident and consistent
with the experimental data shown in Section 5, cementation
results in a significantly stiffer initial response, which is largely
maintained through the first 10 cycles and will require 100s of
additional cycles before the liquefaction threshold criteria
(defined as 3.75% single amplitude shear strain). In contrast, the
uncemented specimen reaches the criteria after only 3 cycles.
Aggregated results from simulations of the untreated and
treated specimens in terms of cyclic strength and number of
cycles to trigger liquefaction (CSR versus N) curves are shown
in Figure 71. As evident, agreement between the numerical
simulations and the experimental data can be achieved, and the
presence of cementation significantly improves the cyclic
strength.
6.2 Continuum constitutive models & simulations
Constitutive models are stress-strain laws that describe the
element-level response of a target material under loading paths
of interest. Such laws for geomaterials have ranged from
relatively simple elastic or elastoplastic to more advanced
Figure 69. Images from DEM simulations performed by Yang (2018) of
the (a) initial confined specimen within the rigid walls and the boundary
conditions applied and (b) the contact force distribution within the
specimen prior to cyclic shearing.
Figure 68. Evolution of (a) bond breakage, (b) coordination number, and
(c) percentage of bonds active in force network from DEM numerical
simulations of a heavily cemented sand (CaCO3 = 5.3%) (adapted from
Khoubani 2018).
471
Figure 71. Compiled CSR versus Ncyc results from DEM cyclic simple
shearing simulations and experimental results for untreated and
biocemented sands (adapted from Yang 2018).
naturally cemented sands in the field and the establishment of
artificial bio-cementation as a feasible and sustainable soil
improvement method. Some models were developed specifically
for naturally cemented soils and later for artificially cemented
soils. Other models targeted soft rocks, structured, and
unsaturated soils recognizing that they all demonstrate
comparable macroscopic behaviors. For example, constitutively,
an increased tensile strength can be either due to matric suction
of an unsaturated soil or due to cementation bonds in an
artificially cemented soil. Developed models differ with respect
to their parent formulation, the constitutive components added or
modified to capture select mechanisms, their target material and
application (e.g., monotonic, cyclic, isotropic compression
loading), and the data used for their validation. Earlier models by
Hirai et al. (1989) and Reddy & Saxena (1992), and later by
Vatsala et al. (2001), Nova et al. (2003), Lee et al. (2004), Yu et
al. (2007), Arroyo et al. (2012), Gao & Zhao (2012), Gajo et al.
(2015, 2019), Rahimi et al. (2016, 2018), Rios et al. (2014),
Panico et al. (2018), Chong (2019), Zhang et al. (2021), and Fu
et al. (2022) have targeted bonded or cemented sands and clays
Figure 70. DEM cyclic simple shear simulations performed by Yang with a variety of artificial non-MICP treatment methods (tough-
(2018) on untreated and biocemented specimens.
lock, Portland cement, and lime). Select models have targeted
multisurface or bounding surface plasticity models. The soft rocks (e.g., Rumpelt 1990, Gens & Nova 1993, Vaunat &
formulation of a constitutive model typically depends on the Gens 2003, Ciantia & di Prisco 2016), structured soils (e.g.,
targeted behaviors as well as the availability of a validation basis Kavvadas & Amorosi 2000, Yang et al. 2008), and unsaturated
upon which comparisons can be performed, and conclusions can soils (e.g., Yang et al. 2008, Nguyen et al. 2014, 2017, Bruno et
be drawn regarding the model’s capabilities and limitations. al. 2020) recognizing that at certain high levels of cementations,
Constitutive models for cemented soils in a variety of the response of a treated soil macroscopically resembles that of a
applications must be able to approximate the broad mix of states soft rock or structured or unsaturated soil. Models specifically
of cementation and conditions in the field. Furthermore, the developed on the basis of bio-cementation have been those by
calibration of constitutive models in research and, even more Fauriel (2012) based on Laloui & François (2009), Nweke &
importantly, in practice must be feasible within the amount and Pestana (2017) on the basis of NorSand (Jefferies 1993), Gai &
extent of information available from characterizing and testing Sanchez (2018) on the basis of the Modified Cam Clay model,
these materials at various scales. The engineering effort is greatly and Xiao et al. (2021b). Out of all the mentioned models, few can
reduced if the constitutive model can already reasonably capture cyclic loading and specifically those of Xiao et al.
approximate the predicted stress-strain behaviors under a broad (2021b) and Zhang et al. (2021) (developed on the basis of
range of conditions. This can be realized only if enough Portland-cemented sands), the latter being a descendant of the
knowledge has been compiled from laboratory data that the SANISAND family of constitutive models and essentially of the
mechanical response of the material can be well established Dafalias & Manzari (2004) simple plasticity model accounting
within a framework. for fabric in the cyclic loading of sands.
Despite the progress made in the field of constitutive
6.2.1 Continuum modeling of cemented soils modeling for cemented soils, the developments have been
The development of several continuum models commenced in challenged by the lack of: (1) broader frameworks that describe
the late 1980s following the recognition and characterization of the mechanical response of these materials, (2) extensive
472
experimental datasets for a broad range of loading and boundary surface follows the Dafalias & Manzari (2004) model, the back-
conditions, and loading paths, and (3) most importantly large- or stress ratio defines its center and thus tracks its location, and
field-scale testing of treated soils such that the validity of a the parameter m defines the radius of the elastic cone in terms of
constitutive model could be examined more holistically towards stress ratio. The constitutive equations for the PM4Sand model
field upscaling and usage. Recognizing the need for usability and are presented in Ziotopoulou & Boulanger (2016) and Boulanger
calibration of a model with parameters commonly available in & Ziotopoulou (2017).
practice, the range and discrepancy of descriptors used to The elastic shear stiffness in PM4Sand is dependent on the
quantify cementation (e.g., CaCO3 content, increase in shear mean effective stress according to:
wave velocity, increase in cone penetration resistance) in the lab
and in the field further challenges the extensive adoption of a
constitutive model for (bio-)cemented soils. (23)
6.2.2 PM4SandC constitutive model Where G0 is the shear modulus coefficient, chosen to match
The PM4SandC constitutive model (El Kortbawi et al. 2022b) estimated or measured Vs, pA is the atmospheric pressure, p is the
for biocemented sands is a plasticity model aimed at capturing mean effective stress, and n is the exponent relating the
the mechanical response of a treated biocemented sand specimen. dependency of G on stress and is 0.5 for clean sands (Andrus &
As such, it does not simulate the process of MICP treatment Stokoe 2000). Accounting for the damage induced during cyclic
itself. The model inherits the formulation and input parameters loading by degrading G and subsequently the plastic modulus,
of PM4Sand (Boulanger & Ziotopoulou 2017) and extends it to has advanced the model’s ability to better approximate the stress-
bio-cemented sands through a targeted reformulation to strain response of a liquefying sand.
incorporate the initial contribution of the cementation bonds to Other constitutive modifications relative to the Dafalias and
the response and its subsequent degradation with damage Manzari (2004) have included: (1) revising the fabric
accumulation. Herein the baseline model PM4Sand is presented formation/destruction to depend on plastic shear rather than
followed by the modifications implemented to incorporate the plastic volumetric strains, (2) adding fabric history and
effects of cementation. cumulative fabric formation terms, (3) modifying the plastic
modulus relationship and making it dependent on fabric, (4)
6.2.2.1 PM4Sand model formulation modifying the dilatancy relationships to include dependence on
The PM4Sand (version 3.1, Boulanger & Ziotopoulou 2017) fabric and fabric history, and to provide more distinct control of
plasticity model for geotechnical earthquake engineering volumetric contraction versus expansion behavior, (5) providing
applications follows the basic framework of the stress-ratio a constraint on the dilatancy during volumetric expansion so that
controlled, critical state compatible, bounding surface plasticity it is consistent with Bolton’s (1986) dilatancy relationship, (6)
model for sand initially presented by Manzari & Dafalias (1997) modifying the elastic modulus relationship to include
and later extended by Dafalias & Manzari (2004). Modifications dependence on stress ratio and fabric history, (7) modifying the
to the Dafalias-Manzari model were developed and implemented logic for tracking previous initial back-stress ratios (i.e., loading
to improve its ability to approximate engineering design history effect), (8) incorporating a methodology for improved
relationships that are used to estimate the stress-strain behaviors modeling of post-liquefaction reconsolidation strains, and (9)
that are important to predicting liquefaction-induced ground providing default values for all but three primary input
deformations during earthquakes. The model was calibrated at parameters. The model is coded as a dynamic link library (DLL)
the equation level to provide for better approximation of the for use with the commercial program FLAC 8.1 (Itasca 2019).
trends observed across a set of experimentally- and case history-
based design correlations. 6.2.2.2 Modifications in PM4SandC model formulation
In PM4Sand, the mechanical response is dependent on the PM4SandC incorporates select extensions or reformulations of
functions controlling the bounding Mb and dilatancy Md surfaces, the PM4Sand equations in order to better approximate behaviors
both depending on the state of the sand and on the shear stiffness, observed in biocemented sands.
among other parameters. PM4Sand uses the relative state An equivalent tensile strength due to the cementation bonds,
parameter index R presented in Boulanger (2003) which in turn pcem,initial, is added to the effective mean stress p as shown in
uses Bolton’s (1986) dilatancy relationship to define the critical Figure 73. This approach has been used in a number of other
state line. The relative state parameter index, R, is defined as (all constitutive models for cemented soils (e.g., Nova et al. 2003,
stress terms are effective, and the prime symbol is dropped): Gao & Zhao 2012, Gai & Sanchez 2018, Rahimi et al. 2018). The
addition of pcem,initial represents an increase in the size of the
(21) bounding and dilatancy surfaces (Mb and Md, respectively)
related to the effective mean stress p through the relative state
(22) parameter index R. This modification is key to the behavior of
biocemented sands since the bounding and dilatancy surfaces
relate to the peak strength and the cyclic resistance. p cem,initial is
where DR,cs is the relative density at critical state for the current estimated internally as a function of the degree of cementation.
mean effective stress, p, DR is the apparent relative density, and As the magnitude of the cementation increases, the occurrence of
parameters Q and R are 10 and 1.5, respectively, for quartzitic cementation bonds at particle contacts increases, the cohesion
sands (Bolton 1986). The latter two parameters are secondary between particles increases and while assuming a slight increase
input parameters to the model and can be calibrated when data in the peak friction angle, pcem,initial increases accordingly. This is
are available. Together, the bounding, dilation, and critical state mathematically described by a failure envelope with an intercept
surfaces are shown in Figure 72, visualized as lines in q – p space similar to the one in the Mohr-Coulomb constitutive model for
cohesive soils:
(for a constant R) where the relative location of the dilatancy and
bounding lines are shown for a dense-of-critical state condition.
The model is simplified and restrained in plane strain conditions (24)
by removing the Lode angle dependency and the bounding and
dilatancy surfaces are scaled from the critical surface with
exponential functions of the relative state parameter. The yield
473
The elastic shear stiffness is increased with a larger G0, now but can also be assigned other values for a lab-specific
called Gcem, an enhanced p with the additional pcem,initial and a calibration. Therefore, the incremental plastic deviatoric strains
smaller n exponent (Equation 25), the latter being informed by d plq are factored using a multiplier Fs which increases from 0 to
the broader range of data collected on these materials. 1 and are accumulated into d plq,degrad (Equation 26). The
functional form for Fs (Equation 27) follows a zero plateau before
deg, picks up as the plastic deviatoric strains reach deg, and
(25) reaches a value of 1 as the plastic deviatoric strains reach or
exceed res:
Go can also be expressed as a function of in-situ data and
specifically of the shear wave velocity after cementation, VS,cem, (26)
with both Go and n being a function of the degree of cementation.
As cementation increases, Go also increases to reflect a stiffer
response and n decreases indicating that the stiffness becomes (27)
less depth (or effective stress) dependent, with values of n
ranging from 0 for heavily cemented sands (equivalent to rock)
to 0.5 for clean sands.
The accumulated plastic deviatoric strain d plq,degrad
Upon shearing, the initial contribution of the cementation
bonds to the stiffness and the strength (pcem,initial) degrades as a constitutes the damage tracked by the model to initiate the
degradation of the cementation pcem,initial to a “secant” or
function of damage tracked here through the accumulation of
degraded using a similar factor Fdeg. The factor Fdeg
strains. This approach of degrading cementation is consistent
with the approaches used in other constitutive models for follows a reverse trend compared to Fs, it starts and remains at a
value of 1 until the elastic threshold strain ( deg) is reached and
degrading cementation with strain accumulation or some damage
then it decays at a degradation rate (rate) and goes to 0 as the
index (e.g., Kavvadas & Amorosi 2000, Nova et al. 2003, Fauriel
& Laloui 2012, Gao & Zhao 2012, Gai & Sanchez 2018). The “residual” strain ( res) is attained. The degradation rate (rate) can
be a user input parameter or can default to a function of the range
accumulation of strains is initiated when the total accumulated
between deg and res. The rate is increased to have the
plastic deviatoric strain ( d plq,tot) reaches an elastic threshold
strain ( deg) and is fully contributing to the damage by the time it cementation degrade over a smaller res - deg range and vice
versa.
reaches a “residual” strain ( res). These parameters have default
values through the model’s generalized calibration targeted
towards capturing the broader set of data available (Section 5) (28)
Figure 72. Schematic of the PM4Sand yield, critical, dilatancy, and bounding lines in q-p space (Boulanger and Ziotopoulou 2017) shown for a
constant value of relative state parameter.
Figure 73. Schematic of PM4SandC model in q-p space at initialization shown for a constant value of relative state parameter. Following degradation,
it reverts to the PM4Sand space of Figure 72.
474
(29)
The equations from Equations 26 to 29 are essentially slowing
down the accumulation of plastic deviatoric strains at the onset
of degradation and then releasing it to progressively accumulate
damage and subsequently decay the cementation. This is found
to improve the slope of the relationship between the cyclic
resistance ratio (CRR) and the number of cycles required to
trigger liquefaction (Ncyc). From the evolution of the degradation
parameters and the progressive degradation of pcem the outcomes
include: (1) the shrinking of the bounding and dilatancy surfaces
to their original PM4Sand configuration (Figure 73) and (2) the
degradation of the stiffness and the value of n exponent from its
initial cemented value to 0.5, such that the model can default
back to a clean sand as is the case of a sheared and damaged
treated sand.
The initial contribution of cementation and its subsequent
degradation are controlled by five new secondary parameters.
The cohesion and peak friction angle enable the determination of
pcem,initial, while the strain at onset of degradation ( deg), the
“residual” strain ( res) and the degradation rate (rate) control the
degradation of the initial cementation. All these parameters have
a physical meaning and a straightforward interpretation and can
be calibrated based on laboratory test data which was is one of
the key goals of PM4SandC. Despite the common origin of their
proposed constitutive models, biocemented sands and clean
sands are different geomaterials with different mechanical
responses. As a result, the calibration of PM4SandC for a
biocemented sand should be independent from any prior
calibration of PM4Sand for its clean sand counterpart.
6.2.3. Monotonic stress-strain behavior – simple shear loading
Figure 74 illustrates PM4SandC simulation results for two
undrained monotonic DSS tests on loose Ottawa F-65 sand
treated at two different levels of cementation with DR = 28% and
initial shear wave velocity (Vs,o) 140 m/s, 'vc = 1 atm, and Vs
~44 m/s and Vs ~218 m/s respectively, corresponding to VsR (
= Vs,cem / Vs,o) ratios of 1.31 and 2.49 respectively. Figure 74
shows comparisons between experimental data and single
element monotonic undrained DSS numerical simulation results
in terms of shear stress versus shear strain deformations, stress
paths, and excess pore pressure build up as a function of
deformations. With the exception of some discrepancies, all
comparative results show a reasonable agreement after minimal
calibration, indicating that PM4SandC can capture such
responses of interest.
6.2.4. Cyclic stress-strain behavior – simple shear loading
Figure 75 illustrates PM4SandC simulation results for an
undrained cyclic DSS test on a MICP-treated Ottawa F-65 sand
with DR = 42%, 'vc = 1 atm, Vs,o ~146 m/s, Vs,cem ~182 m/s, and
thus a VSR of 1.25, subjected to a uniform CSR = 0.2. The
cemented specimen triggered at 8.5 cycles based on the two
triggering criteria: a single-amplitude strain of = 3% and an
excess pore water pressure ratio ru > 98%. While the experiment
triggered at 8.5 cycles, the simulation was continued to illustrate
the post-triggering strain accumulation and the damage to the
cementation bonds. The simulation results are consistent with the Figure 74. Comparison between experimental and simulations results for
experimental results with minimal calibration effort. The general monotonic undrained DSS testing on MICP-treated sands using
response was prioritized in this example calibration to illustrate PM4SandC (experimental data from Lee & Gomez et al. 2022b).
the behaviors of the biocemented sands instead of a test-specific
calibration.
475
Figure 75. Comparisons between experimental and simulation results for an undrained cyclic DSS test on an MICP-treated sand using PM4SandC
(experimental data from Lee et al. 2021).
Extending this calibration exercise to different CSRs defines 7 UPSCALING TO THE FIELD
the cyclic resistance ratio CRR versus number of cycles (Ncyc)
curves for bio-cemented sands, where CRR is the CSR required Realization of the MICP technology at field scale as well as
to reach liquefaction using some triggering criterion in a certain widescale acceptance as a viable and sustainable soil
number of cycles (Ncyc). In earthquake engineering, this improvement technique requires: (1) demonstration that the
relationship is often approximated with a power function CSR = technology can be reliably implemented at a cost, performance,
aNcyc-b, where the parameters and represent the CSR at one and feasibility level that is either competitive with or superior to
cycle and the slope of the curve, respectively. For clean sands, existing business-as-usual soil improvement techniques and (2)
the value of these parameters is a function of p and DR, with that engineering design guidance and numerical tools be
parameter b generally ranging from 0.25 to 0.37. For available such that treatment magnitudes and dimensions can be
biocemented sands, the same dependencies are expected to hold designed to mitigate site-specific liquefaction hazards, including
with an additional dependency on the level of cementation, those related to both liquefaction triggering resistance and the
however the applicability of the power law for treated sands is consequences when triggering does occur.
not fully established yet (Lee et al. 2021). While the value of Upscaling of the MICP technology from the research setting
parameter a should presumably be larger than sands due to the to the more practical field implementation setting will require
presence of the cementation, the value of parameter b is that challenges similar to those encountered by existing soil
hypothesized to be smaller due to the larger number of cycles improvement technologies (e.g., chemical grouting) and
needed to trigger liquefaction at low CSRs. Accordingly, geoenvironmental remediation techniques (e.g., dechlorination,
PM4SandC yields CRR curves that are visibly higher for contaminant immobilization) be overcome. Such challenges
biocemented sands than the curves for clean sands which, as include those related to treatment extent and uniformity,
shown in Figure 75. This is in agreement with the experimental environmental controls, material use, specialty equipment, and
data by Lee et al. (2021) on lightly cemented MICP-treated monitoring and QA/QC evaluation. Consequently, the
Ottawa F-65 sands under comparable conditions. similarities between MICP and some existing technologies will
enable use of equipment and materials processing methods that
are already established industry practices and readily available
from specialty contractors. Similarly, prior research on the effect
of natural and artificial cementation on the liquefaction
476
and will not be the most economical nor practical method for
every site. For instance, MICP will not likely be competitive
when addressing improvement of unoccupied project sites with
shallow liquefiable materials for which energy densification
methods (e.g., deep dynamic compaction, vibro-densification)
will remain most time, material, and cost efficient.
Significant effort has been placed on both upscaling of the
MICP treatment technology for field implementation and
determining the cementation magnitudes and extent required for
acceptable performance of supported geosystems, specifically
for instances when liquefaction susceptibility under seismic
loading is of concern. A generalized design process for MICP
treatment at the field scale is outlined first, followed by a section
focused on upscaling of the MICP treatment method to the field
and finally a section highlighting how MICP treatment can be
effective towards mitigating liquefaction hazards.
7.1 MICP design process for field implementation
The envisioned design process for MICP field implementation is
presented in Figure 77. The six stages, preliminary suitability
evaluation, technology compatibility with site conditions, site
characterization, design requirements, implementation plan, and
monitoring and QA/QC, each require that specific activities be
performed to achieve stage objectives. This process is expected
to be non-linear during process implementations, with several
iterations likely required in the latter stages (as indicated by the
feedback arrows in Figure 77). The process is also likely
Figure 76. Comparison between the experimental CSR versus Ncyc curves applicable to other emerging biomediated technologies,
from Lee et al. (2021) and the one from the calibration using PM4SandC including microbially induced desaturation and precipitation
for undrained cyclic DSS tests on MICP-treated sands (adapted from Lee
(MIDP) and enzymatically induced CaCO3 precipitation (EICP),
& Gomez et al. 2022b).
and is broadly consistent with the design processes used for
susceptibility of sands (e.g., Saxena et al. 1988. Clough et al. existing soil improvement technologies as detailed by Mitchell
1989) and the treatment extent required to provide structural et al. (1981, 2008), USACE (1999), NHI (2017a, 2017b) and
stability (Dashti et al. 2010a,b) provide a basis from which (Nicholson 2015). The following sections further detail each
MICP-specific guidelines can be further developed. Other stage of this design process.
aspects, including process monitoring and QA/QC evaluation,
will leverage both existing technologies and implementation 7.1.1 Preliminary suitability evaluation
methods including the CPT qc and Vs assessment approaches The initial screening stage evaluates the suitability of MICP for
described (Section 4), and new methods that may, in return, also some project by considering the anticipated site conditions,
provide benefits for existing soil improvement technologies.
earthquake loading, and performance requirements. Site
MICP’s unique attributes relative to other existing soil
conditions of specific interest to MICP implementations include
improvement technologies make it a promising alternative
information about groundwater chemistry, soil mineralogy,
technology specifically for liquefaction hazard mitigation. These
attributes as detailed in the prior sections include the ability to: classification, and state, obtained from prior site specific,
(1) reliably control reaction rates by using well understood regional, or geologic model information, which can be used to
biological catalysts (effectively controlling the “set time”), (2) further evaluate biogeochemical treatment compatibility and
potentially achieve greater uniformity and increased treatment expected success of MICP towards improving soil engineering
extent relative to existing grouting methods (thereby allowing for properties at the level needed for project requirements. The
reduced implementation costs), (3) treat soils with minimal site knowledge base regarding treatment feasibility, exemplified in
disruption requiring only fluid injections between installed wells, Figures 28 and 29, may be sufficient when initial screening
(4) maintain an open pore structure allowing for subsequent evaluations reflect advantageous conditions.
treatments and minimizing the impact on ambient groundwater The estimated earthquake loading, obtainable from regional
flow conditions, (5) re-treat and heal MICP treated soils during hazard maps or more site-specific studies, and the performance
the service life in the event of increased loading demands or requirements of the infrastructure system being
damage following earthquakes, and (6) verify and monitor stabilized/supported, such as deformation tolerances or system
improvement integrity using non-invasive and other more stability, may be used to estimate the magnitude and spatial
conventional methods following application and throughout the extent of improvement required. The magnitude of cementation
improvement’s service life. required can be approximated from laboratory derived cyclic
It is envisioned that these attributes will make MICP resistance curves (e.g., Figure 60), and given that MICP
particularly competitive and advantageous for liquefaction treatment can increase the CRR by an order of magnitude or more
mitigation applications near existing infrastructure, an for a given loading demand when sufficient cementation is
application for which arguably no satisfactory solution currently present (Figure 61), it is likely that the required improvement for
exists. Furthermore, MICP may also afford unique advantages some system could be realized. The treatment dimensions
with respect to the treatment of sands with high fines content required would be estimated at this stage primarily from
(e.g., silty sands) and a variety of other anthropogenically-
generated waste materials (e.g., dredge materials, hydraulic fill,
fly ash, mine tailings). Similar to current soil improvement
technologies, MICP will have a niche of effective applications
477
Preliminary Suitability Evaluation compromised (see Section 3.8.2). Thus, for sandy soils,
dedicated testing on reconstituted moist tamped samples where
treatments are applied to specimens under effective stress
conditions representative of in situ conditions and subsequently
Technology Compatibility w/ Site Conditions subjected to cyclic loading events are likely most appropriate.
7.1.3 Site characterization
A 3D subsurface model that is informed by geologic depositional
Site Characterization sequences and processes and contains comprehensive details
regarding the soil stratigraphy and layer connectivity,
mechanical properties, and hydraulic conditions is necessary to
Design Requirements (Performance Targets) characterize the site at a level that enables design requirements,
implementation plans, and monitoring and QA/QC methods to
be specified. This stage begins with the development of a
depositional process informed geologic model and follows with
Implementation Plan
in situ testing and laboratory characterization of the soil
stratigraphy and engineering properties of soil units, as is typical
for all geotechnical projects, with more detailed evaluation of site
Monitoring & QA/QC hydraulic conditions that will likely strongly influence the
implementation plan.
Identifying the treatment zone will require both simplified 1D
Figure 77. Overview of design process for MICP field implementation. analyses on individual soundings/borings and generation of
information obtained regarding stratigraphic conditions and site 2D/3D spatial realizations that are conditioned on the
footprints. soundings/borings. Initial evaluation of the engineering
properties and the target treatment zone for liquefaction
7.1.2 Technology compatibility with site conditions mitigation will include analysis of CPT/SPT profiles using 1D
Once identified as a potentially viable improvement alternative, liquefaction methods to estimate susceptibility to triggering
site-specific evaluations can commence to further verify specific (Boulanger & Idriss 2015) and potential volumetric and shear
applicability to and effectiveness for some targeted treatment strains (Iwasaki et al. 1978, Zhang et al. 2002, 2004, van
zone. The realization of and ability to control the MICP Ballegooy et al. 2014). Further insights may be obtained by
generating conditioned 2D/3D realizations of the subsurface,
biogeochemical reaction network is primarily dependent on site-
particularly for the identified liquefiable layers. Intra-spatial
specific environmental and geologic conditions, with an
modeling of the depositional variation within the potentially
understanding of groundwater chemistry and soil mineralogy
liquefiable layers can provide insights into density variations
being particularly important. Soil and groundwater samples may across the target treatment zone, while interspatial modeling of
be aseptically obtained during an investigation program and can the liquefiable and surrounding layers can better define the
be further analyzed to establish baseline in situ conditions, and bounds of treatment necessary. Stochastic conditional modeling
subsequently used to perform site-specific soil column (e.g., Fenton & Vanmarcke 1990, Griffiths et al. 2002, Jaksa et
laboratory tests to evaluate the efficacy of the MICP process, al. 2005, Montgomery & Boulanger 2017) and categorical
including the success of stimulation, cementation, and by- conditional modeling (e.g., Carle & Fogg 1996, DeJong et al.
product removal stages. A suite of columns tested in parallel may 2021, Bassal et al. 2021, 2022) approaches have been shown to
be desirable to parametrically evaluate treatment process be effective methods for generating realistic geostatistical
variables and identify the preferred treatment formulation. subsurface realizations that are conditioned on obtained in situ
Monitoring of pressures and flow rates during column tests can data. Obtained realizations of the subsurface conditions,
further constrain hydraulic conditions which may be desirable for including the engineering properties of each layer, can then be
field implementation as well as provide some understanding of directly incorporated into continuum models (e.g., FLAC) for
the potential for soil hydraulic conductivity reductions during the subsequent design purposes including non-linear dynamic
treatment process. Monitoring of shear wave velocity during analyses.
such tests will further indicate the rate of shear modulus Evaluation of site hydraulic conditions, with particular focus
improvement per treatment cycle, and when paired with post- on the potentially liquefiable layers, can begin using hydraulic
treatment soil CaCO3 content measurements can provide a properties estimated from in situ tests (e.g., CPT) and lab data,
measure of treatment efficiency that can be evaluated upon but may also include the use of additional dedicated in situ tests
comparison with the existing knowledge base for similar to better resolve hydraulic properties with depth (e.g., hydraulic
materials (e.g., Figure 40 & 41) and further used to constrain profiling tool). Such characterization will also likely include the
expected treatment solution concentrations and required numbers installation of a test well(s) through which both staged pumping
of treatment cycles. (lugeon) tests and long-term monitoring of ambient conditions
Evaluation of the improvement in soil undrained cyclic can be performed. The hydraulic model developed can be used
behaviors can either be estimated by comparing achieved metrics to assign hydraulic properties to the units in the conditioned
from columns tests, including shear wave velocities, CaCO3 stratigraphic model generated for liquefaction evaluation and
contents, and other soil strength parameters, to existing studies should resolve the locally defined values with the staged
for similar levels of improvement (Figures 39-41 & 60) or can hydraulic pumping test results, which provide a global spatially
be evaluated rigorously at a site-specific level through treatment averaged hydraulic system response. Reconciliation of the fine
and testing of simple shear or triaxial specimens using the stratigraphically linked hydraulic conditions against the global
preferred treatment formulation identified in the first portion of pumping response can be achieved using a variety of hydraulic
this stage. Due to the small levels of cementation necessary to modeling software programs (e.g., ModFlow, PhT3D, etc.).
produce a significant improvement in cyclic resistance (e.g., Once resolved, these 2D/3D hydraulic models afford subsequent
Figure 61), specimens obtained from rigid column tests should investigation and development of specific treatment
not be extruded and subsequently tested to obtain estimates of implementation plans.
cyclic resistances, as material integrity will almost certainly be
478
7.1.4 Design requirements & performance targets 7.1.5 Implementation plan
The MICP treatment program design requires specification of the The implementation plan to deploy MICP at some project site
targeted improvement magnitudes and spatial extent. The requires details regarding the achieved microbial and related
improvement magnitude can be estimated based upon the precipitation reaction rates, solution transport rates and
expected seismic demand throughout the soil profile as distributions during applied treatment injections, specific
calculated by 1D liquefaction triggering evaluation methods and dimensions regarding well spacings, layouts, and their relation to
the level of cementation required to increase the CRR in the the intended treatment zone, as well as injection details such as
susceptible soil layers such that a factor of safety of one or higher treatment volumes and the associated required infrastructure
exists throughout the soil profile. The cementation level required, needed for preparation, mixing, and application of MICP
as identified by either a soil CaCO3 content or shear wave treatment solutions.
velocity increase (Figure 40), can be estimated from general Treatment spatial distribution and uniformity is governed
relationships such as those shown in Figure 61. However, it may largely by the advective-dispersive injection regime as well as
be preferrable if a site-specific CSR versus Ncyc curve is the reactive transport during injections which is directly
developed using simple shear or triaxial test results performed on dependent on the achieved ureolytic reaction rates relative to
soil specimens sampled from the liquefiable layer and applied solution transport rates. Ideally, every treatment solution
reconstituted to conditions similar to that in situ. The would be injected with minimal reactions occurring during
improvement magnitude estimated with this approach is the injections, thereby achieving a near uniform distribution of
improvement required for free field conditions only and ignores reactants after injections that is governed only by the advective-
three dimensional aspects of treatment non-uniformity, dispersive flow characteristics of the application system and
subsurface non-uniformity, infrastructure load distributions and surrounding soil. In this case, a stop flow retention period could
other specifics, as well as site-specific dynamic wave be provided and near uniform improvement could be achieved.
propagation, among other factors. Consequently, it is likely that Instead for actual field applications, some fraction of reactants
the improvement level specified during this phase will will be consumed during injections thereby precipitating more
CaCO3 closer to injection sources, with dispersion generating
underestimate the improvement level required for some
further gradients in transported concentrations, and surrounding
acceptable performance.
groundwater gradients perhaps altering the distribution of
MICP improvements as realized in the subsurface will
generated CaCO3 during the stop-flow retention and reaction
inevitably have some degree of inherent non-uniformity despite
period. The practical objective to achieve greater cementation
the design measures taken to minimize such variations.
uniformity is to ensure that reactions occurring during injections
Therefore, it is likely that some zones will be improved below
are minimized. This is accomplished when the microbial reaction
the target improvement level with others being at or above the
rates are much slower than transport rates, resulting in only a
desired improvements. From this perspective, the improved soil
small percentage of reactants being consumed during the
and related system will likely behave as a composite of both
injection of solutions. Non-uniformity is inevitable, however,
weaker zones, which would trigger at a lower cyclic demand if
practical tolerances for desired reaction rates and flow rates can
considered in isolation, and stronger, stiffer, and more dilative
be determined from passive tracer simulations and more refined
zones, which may act to dampen excess pore pressure generation
reactive transport simulations.
and limit shear strains in surrounding weaker soil regions. Such
Microbial reaction rates and thus CaCO3 precipitation rates
systems-level behaviors have yet to be fully explored and
can be controlled by varying the applied treatment solution
understood, however, it is clear that treatment uniformity and its
formulation during both the stimulation (Section 3.2) and
post-treatment characterization will play an important role in this
cementation phases (see Section 3.3). As shown previously,
response and thus continues to be a focus of ongoing research.
variations in reaction rates can be largely accomplished by
The treatment extent and spatial distribution required depends
varying supplied growth factors and nutrients as shown by
directly on whether or not the site supports a structure, and if so,
varying yeast extract in the authors’ experiments (Figure 20).
what structural characteristics are critically important. When
The transport rate is primarily controlled by the hydraulic
infrastructure is already present at some site to be improved, the
conductivity, hydrostatic pressure, and hydraulic fracturing
desired treatment zone can be estimated following related
pressure and the distribution of flow is primary controlled by soil
earthquake engineering and soil improvement studies which
porosity, dispersive properties, and fluid density gradients, when
have characterized the effect of discrete treatment zone volumes
present. Conceptually, when low ureolytic and precipitation rates
(i.e., depths and widths) and normalized such dimensions in
are achieved, reaction completion for a single treatment could
accordance with the surrounding infrastructure dimensions and
occur over 50 to 100 hours providing the opportunity for longer
knowledge of stratigraphic layers with simplified geometries in
injection durations (10 to 20 hours) to still achieve significant
order to enable more general use (e.g., Hausler 2002, Dashti et
treatment uniformity and a large extent of improvement. The
al. 2010a, b). MICP specific guidelines for treatment zone
authors’ experiments, while not exhaustive, have shown that
extents and distributions, however, have not yet been developed
ureolytic activities can be slowed to achieve reaction completion
to a level where these guidelines can be detailed.
over at least 48 hours (Lee et al. 2022). Maximizing treatment
Furthermore, non-linear dynamic analyses using FLAC or
extent is key towards enabling larger spacings between treatment
other numerical platforms provide the opportunity to develop
wells, thereby reducing implementation costs. In reasonably
site-specific treatment magnitude and extent designs that
permeable sands with low dispersivities, well spacings near 5 to
incorporate the three-dimensional aspects of subsurface non-
10 m may be achievable and practical.
uniformity, infrastructural load distributions, and dynamic wave
The well layout pattern at the project site is driven by the: (1)
propagation. Implementation of this complexity requires
geometry of the desired treatment zone, (2) level of uniformity
accurate calibration of appropriate constitutive models for
required, (3) flow rate and dispersion constraints, and (4) need
cemented soils (such as PM4SandC described earlier in Section
for post-treatment byproduct removal. Pilot field trials may
6.2) and verification and validation of such models against
involve simplistic injection layouts such as well dipoles or larger
detailed case histories (e.g., centrifuge model data). At this time,
well networks consisting of some center injection/extraction well
geomechanical numerical modeling of a project site with an
surrounded by a set of injection/extraction wells (Sorenson 2021)
MICP improved zone has not yet been performed. Instead, only
and can be used to evaluate soil advective-dispersive flow
the initial study shown subsequently in Section 7.3 has been
characteristics, potential treatment distributions, as well as guide
performed using a simplified 2D analysis.
optimization of well patterns, treatment control, and uniformity.
479
When upscaled it is expected that some variation of a repeating some extreme loading event (e.g., earthquake loading). The array
triangular or square well pattern with both injection and of monitoring tools available, as described in Section 4, are
extraction wells can form a near-closed loop treatment system extensive and are capable of discerning specific biogeochemical
that can be implemented. It may be advantageous to extract and geotechnical changes that occur throughout all treatment
solutions at higher volumes than those injected to better control stages (i.e., stimulation, cementation, byproduct removal).
treatment extent and contain byproducts. It may also be However, only a subset of such measures are well suited for real-
advantageous to reverse flow directions on occasion in order to time monitoring of the process in situ during treatments.
improve treatment uniformity. This strategy is consistent with Prior to injections, solutions should be subsampled and
that used for some chemical grouting and bioremediation chemical analyses should be performed to verify that intended
technologies; hence, the application system infrastructure solution chemical compositions were achieved. Similar
required and associated operations are already known in industry. subsamples can be retained from production well locations
To ensure ease of monitoring, monitoring wells and or other during injections to examine changes in effluent solution
sampling devices should be installed prior to treatments, with chemistry in time during solution replacement. During stop-flow
locations determined by non-reactive and reactive transport residence periods, aqueous samples can be periodically obtained
simulations as discussed later. from monitoring wells and embedded non-destructive probes
Design of the desired biomediated reaction rate, injection and and/or sampling devices should be present to capture reaction
extraction rates and volumes, and well spacings, layouts, and dynamics. While in-situ chemical specific probes can be
other geometries, have interdependencies that are best optimized suspended within injection, production, and monitoring wells at
by performing numerical advective-dispersive flow simulations some treatment depth, it may be more practical to use simple
that are anchored to site characteristics determined during in probes capable of continuous fluid conductivity and pH
prior stages and calibrated to achieved field data. Although bulk measurements, if solution chemistry allows for some indirect
and or localized permeabilities can be crudely determined using measure of reaction progression in time. For example, as shown
a single injection well test (with packers used to potentially in Figure 3 and 36, solution pH can provide a useful method for
isolate some treatment zone), calibration of soil advective- indirect monitoring of cementation completion, as a large pH
dispersive properties requires at least two well tests wherein a increase can be observed once calcium is fully consumed,
conservative tracer is injected at some known concentration and provided that urea is provided in excess of calcium (urea to
observed at surrounding outlet wells. This process, referred to as calcium ratio > 1.0).
a breakthrough curve or passive tracer testing, will yield data Shear wave velocity is the preferred practical approach for
similar to that shown in Figure 26 from a simple 1D soil column tracking mechanical improvements in soils during treatments
test receiving injected bromide ions, however, similar processes given its direct proportionality to soil CaCO3 content (Figure 40),
and other conservative tracers, such as chloride ions, can and the breadth of established methods for measuring such velocities
have been used with equal success in the field. Such data can in situ (e.g., SCPT, cross-hole), and its non-destructive nature.
allow for refinement of the hydrogeological model in order to With a sufficiently dense sensor array tomographic images of the
account for bulk properties of the porous media in the treatment spatial changes in shear wave velocity could be tracked during
zone including soil porosities, dispersion coefficients, and other the treatment phase (e.g., Lee et al. 2005, Kim & Kim 2010).
physical flow non-uniformities resulting from containment loss Other measures, such as compression wave velocity, are either
and or inherent soil hydraulic heterogeneity, among other less sensitive to such mechanical enhancements, and/or may only
features. Hydrogeological modeling software capable of 3D non- be practically useful at existing well/borehole locations.
reactive transport modeling include FEFFLOW, MODFLOW, Preferred post-treatment QA/QC processes include many of
and Hydrus, among others. the same methods used to determine the initial site baseline
To account for reactive transport during simulations, conditions. Such methods may include drilling and destructive
however, more complicated means are necessary. Reaction rates sampling (e.g., split spoon, sonic coring), seismic CPT
can be designed initially leveraging soil column pilot test data, soundings, and other geophysical measurements for
however, ultimately, monitoring of reaction rates as observed in
biogeochemical and mechanical engineering analysis. Data
the field condition will be necessary. Soil column data, therefore,
obtained during assessment of the final improved state can be
can be used to guide initial implementation plans with collected
compared against the targeted designed improvement levels
real-time site data allowing for modifications in applied
treatments and updating of reactive transport simulations, as (e.g., soil CaCO3 content, Vs increases) and against existing
necessary. While a variety of software packages can be used for broader databases of such values (e.g., Figure 46) to ensure that
simple 1D reactive transport simulations including PHREEQC, the intended treatment program was realized and that the process
TOUGHREACT, and others, field scale modeling will require followed conventional expectations for improvement
more complicated 3D reactive transport platforms such as progression. Collection of such data provides an opportunity to
PHT3D, OPENFOAM, COMSOL, and TOUGHREACT V3.3- contribute to refinement of such relationships and our broader
OMP, among others. Ultimately, such models require a non- understanding of these behaviors.
trivial effort, thus, considerations should be made regarding the Finally, continued monitoring over the service life of the
value of such reactive transport simulations over the use of more improvement may be necessary, particularly when MICP
simplified non-reactive simulations. The need for refinement improved zones are subject to somewhat undersaturated
may be largely determined by considering the effect of groundwater and extreme loading events that might compromise
uncertainty on such simulations and the related impacts cementation integrity. For these instances, depending on the
anticipated for predicted improvement distributions. mechanisms (i.e., mechanical versus chemical degradation), both
soil CaCO3 content and shear wave velocity measurements may
7.1.6 Monitoring & QA/QC be useful towards characterizing degradation (e.g., Figure 19 &
Monitoring during the treatment implementation period is 54). Furthermore, additional evaluation could be performed by
necessary to ensure that the project is being realized as predicted SCPT or cross-hole seismic methods (particularly if cased holes
and designed, while post-treatment QA/QC is necessary to verify remain accessible), and in some cases surface wave methods
that the desired design was realized in situ. It may also be (e.g., SASW) when the treatment zone is relatively shallow.
desirable to continue to monitor the integrity of the cementation
over the service life of some project to assess the potential for
long-term chemical damage and or mechanical degradation after
7.1.7 Summary
480
The design process recommended and outlined above has been how different stages of the treatment design and implementation
informed by past upscaling and liquefaction mitigation process can be addressed.
experiments performed using MICP to date as well as by
established practices developed by others performing 7.2.2 Test design
conventional soil improvement and geoenvironmental The test design leveraged insights from small-scale laboratory
remediation methods. This envisioned process articulates the soil column tests involving the same alluvial sand which
authors’ vision regarding how MICP can be successfully and demonstrated advances regarding: (1) the ability of modified
systematically designed and implemented in the field. It is stimulation techniques to reliably control achieved bulk
important to reiterate that this design process is not a linear series ureolytic rates and related cementation rates thereby minimizing
of events and stages. Instead, this process may be frequently reactions during solution transport (Lee et al. 2022), (2) the
revisited during the implementation of MICP at some site to optimization of the treatment solution compositions used during
incorporate new information and understandings that will guide stimulation and cementation which afforded reduced material
final application strategies. It should also be mentioned that field usage and environmental impacts (San Pablo et al. 2022a), (3)
applications of MICP for liquefaction mitigation purposes have monitoring of biogeochemical and geomechanical aspects of the
not yet been realized to the authors’ knowledge. With this in MICP treatment process in near-real time (Gomez et al. 2019),
mind, the following sections focus on specific studies performed and (4) techniques for effective removal of ammonium
in a research context wherein understanding of how MICP can byproducts following cementation (Lee et al. 2019a).
be upscaled to the field and effectively mitigate liquefaction Prior to treatments, reactive transport pre-modeling was
hazards in sandy soils has been both demonstrated and expanded. performed involving similar 3.7 m long 1D columns in
PHREEQC. In order to ensure representative transport, modelled
7.2 Upscaling of MICP treatment methods column advective-dispersive properties and porosities were
calibrated to observed passive tracer test data obtained from the
Following its initial conception, research efforts regarding MICP physical column experiments immediately prior to treatments
have been largely divided between exploring the more (Lee et al. 2019a, b). Leveraging measured bulk ureolytic
fundamental science governing the biogeochemical process and reaction rates from the earlier laboratory experiments, a
developing the engineering capacity needed to implement the Michaelis-Menten ureolytic kinetic model similar to that
technology at field scale and achieve practical industry described in Section 4.1.2 was calibrated to the experimental
acceptance and adoption. This dual approach has proven very data. Column reactive transport models then assumed these
useful, with improved scientific understandings guiding similar rates applied at all locations along columns and reaction
upscaling efforts, and upscaling efforts in turn allowing for the rates were varied to explore the sensitivity of predicted
identification of critical development challenges and knowledge cementation distributions in columns to these factors while
gaps requiring improved fundamental understandings. assuming a constant injection flow rate that would be used for all
Table 5 presents a summary of the upscaling projects to date columns. From these results, it was determined that one column
that have been performed at length scales exceeding ~0.5 m, with would receive modified stimulation and cementation treatments
target applications including slope stabilization, fractured rock intended to achieve “slow” ureolytic reaction rates which were
sealing, strontium immobilization, and grouted piles, among expected to minimize cementation non-uniformity and increase
other uses. Projects are listed in Table 5 in order of publication cementation extent. These “slow” reaction rate treatments
year, with information regarding the length scales, flow and targeted ~90% degradation of the supplied 250 mM urea over a
geometric conditions, treatment methods, targeted applications, 48 hour period, with the injection period (3.1 hrs.) consisting of
and primary reference(s) also provided. As shown, numerous only ~6% of this primary reaction time. A second column
upscaling studies have been performed by the broader MICP received unmodified stimulation and cementation treatments
research community, each of which have provided new insights used frequently in past stimulation experiments performed on
and address aspects of practical upscaling challenges. San Pablo cm-scale laboratory columns and achieved a relatively “fast”
et al. (2020) presents the most recent upscaling experiment by bulk ureolytic rate. In this case, “fast” reaction rate treatments
the authors and is described in greater detail as it addressed targeted ~90% degradation of the supplied 250 mM urea over an
aspects of all six stages of the proposed design process presented 8 hour period, with the injection period now consisting of ~40%
in Figure 77. Although each of the above studies in Table 5 merit of the primary reaction time. While these two columns provided
an extended discussion, this particular study is the focus of the the opportunity to contrast differences in reactive transport
next section as it highlights current state of the art processes for conditions afforded by differences ureolytic rates, a third column
technology upscaling as recommended by the authors. was augmented with S. pasteurii to compare the success of
stimulation approaches to that afforded by augmentation. This
7.2.1 MICP upscaling in ~4 m long 1D columns augmentation approach did not provide additional nutrients
San Pablo et al. (2020), with particular aspects detailed further in (yeast extract) following introduction of non-native cells, thus
Lee et al. (2019b), performed MICP treatments in five 3.7-m long the potential for unintended stimulation was minimized. The
horizontal columns (0.15 m3 total soil volume) containing one of provided cell densities in the augmented column were intended
three different natural sands. The upscaling study was performed to enable a similar bulk ureolytic rate as the “high” rate
specifically to address outstanding field upscaling issues
including the: (1) success of pure augmentation approaches
versus different stimulation strategies, (2) improvement of
treatment spatial uniformity, (3) optimization of stimulation and
cementation treatment formulations, (4) evaluation of practical
treatment extents, and (5) removal of generated treatment
byproducts. A subset of these test results from three columns
containing the same alluvial poorly graded silica sand (Concrete
Sand) is presented herein to highlight the study’s findings with
respect to the above objectives, highlight the interdependencies
between different aspects of the treatment process, and exemplify
481
Table 5. Summary of MICP upscaling projects performed.
Model vs. Flow geometry (1D,
Year Dimensions 1 (m) Soil Type Treatment Method Application(s) References
In situ 2D, 3D), Shape
2007 In situ 1D, column 5.0 long x 0.06 diameter SP (Itterbeck Sand) Augmentation Soil improvement Whiffin et al. 2007
2009 Model 3D, box 0.9 x 1.1 x 1.0 SP (Maas River Sand) Augmentation Soil improvement van Paassen et al. 2009
Soil stabilization for horizontal
2009 Model 3D, box 8.0 x 5.6 x 2.5 SP (Itterbeck Sand) Augmentation van Paassen et al. 2009, 2010
drilling, QA/QC methods
Stimulation, soil improvement,
2011 In situ 3D, zone 1.25 deep x 0.9 diameter SP (alluvial sand) Stimulation Burbank et al. 2011
QA/QC methods
Soil stabilization for horizontal
2011 In situ 3D, zone 1000 m 3 (limits unknown) GP Augmentation van Paassen 2011
drilling, QA/QC methods
2012 Model 3D, box 6.0 x 4.5 x 1.5 SP (Itterbeck Sand) Augmentation Soil improvement Esnault Filet et al. 2012
2012 In situ 3D, zone 4.7 x 4.3 x 1.5 GP (river alluvium) Stimulation Strontium immobilization Smith et al. 2012
dacite volcanic rock
2013 In situ 2D, planar 25 m deep x 4 m 2 area Augmentation Fracture sealing Cuthbert et al. 2013
formation
2014 Model 2D, 1/4 well pattern 0.5 x 0.5 x 0.15 SP (Ottawa 50-70) Augmentation Soil improvement DeJong et al. 2014
2015 In situ 3D, test plots 4.9 x 2.4 x 0.3 SP Augmentation Surface erosion Gomez et al. 2015
2016 Model 2D, rock slabs 1.75 x 1.75 granite Augmentation Fracture sealing Minto et al. 2016
2016 In situ 3D, zone 341 m deep injection fractured sandstone Augmentation Fracture sealing Phillips et al. 2016
Augmentation & Augmentation vs. stimulation,
2017 Model 2D, circular tanks 1.7 diameter x 0.5 long SP w/ fines (Concrete Sand) Gomez et al. 2016, 2018, Graddy et al. 2018
Stimulation treatment monitoring
482
2018 In situ 3D, zone 310 m deep injection wellbore cement Augmentation Wellbore sealing Phillips et al. 2018
2019 Model 3D, box 2.0 x 2.0 x 2.5 m various sands Augmentation Erosion control Beguin et al. 2019, Esnault Filet et al. 2020
Augmentation vs. stimulation,
SP w/ fines (Concrete, Augmentation &
2019 Model 1D, square column 3.7 x 0.2 x 0.2 treatment monitoring, QA/QC Lee et al. 2019, San Pablo et al. 2020
Delta, & Covelo Sands) Stimulation
methods, byproduct removal
Biogeochemical processes,
2019 In situ 3D, zone 4.7 x 4.3 x 1.5 GP (river alluvium) Stimulation Saneiyan et al. 2019, Ohan et al. 2020
geophysical measurements
2020 Model 3D, box 0.91 x 0.91 x 0.91 SP (silica sand) Augmentation Scour mitigation around piles Do et al. 2020, Montoya et al. 2021
2020 In situ 3D, test plot 1.0 x 8.0 x 0.2 SP Augmentation Slope stability Ghasemi and Montoya 2020, 2022
2020 In situ 3D, zone 698 m deep injection wellbore cement Augmentation Wellbore sealing Kirkland et al. 2020
2020 In situ 3D, slope 100 x 10 x 4.5 silty clay Augmentation Slope stability Terzis et al. 2020
2020 Model 3D, box 1.0 x 0.6 x 0.6 SP calcareous sand Augmentation Pile capacity Xiao et al. 2020
2021 In situ 3D, zone 310 m deep injection wellbore cement annulus Augmentation Wellbore sealing Kirkland et al. 2021a
2021 In situ 3D, zone 697 m deep injection sandstone Augmentation Wellbore sealing Kirkland et al. 2021b
2021 Model 3D, zone 1.0 x 1.0 x 1.6 SP (Bar Sand) Augmentation Pile capacity Lin et al. 2021
2021 Model 3D, zone 15 x 1.2 x 0.5 SP Augmentation Erosion control Montoya et al. 2021
2021 Model 3D, box 1.1 x 1.0 x 0.9 SP (Ottawa sand) Augmentation Soil improvement Wu et al. 2021
2021 In situ 3D, zone 5.0 deep x 5.0 diameter fine grained soils Augmentation Soil improvement Zeng et al. 2021
2022 Model 3D, box 1.0 x 0.4 x 0.3 SP (silica, calcareous) Augmentation Road base improvement Xiao et al. 2022
Figure 78. Overview of (a) 3.7 m column set-up and (b) treatment program (adapted from San Pablo et al. 2020).
stimulated column (at the time of the first cementation treatment) detailed in San Pablo et al. (2020) as well as provided in Table 2
across the entire column length. Herein these three treatment in Section 3.3. Prior to and at select times following each
approaches are referred to as SL for “stim low”, SH for “stim treatment injection, aqueous samples were collected for chemical
high”, and AH for “aug high”. and biological measurements, bender element measurements
The test configuration, designed to simulate a single stream were performed to examine shear wave velocity progression, and
tube within a well-to-well half space, is shown in Figure 78a. flow rate and fluid pressure measurements were obtained during
Each setup consisted of a horizontal 3.7-m long steel column injections for hydraulic conductivity determination. Following
with a square cross-section (0.2 x 0.2 m) that contained 0.15 m3 test completion, all columns were destructively sampled along
of soil. The natural soil was moist tamped in place to a relative their length for further evaluation, including measurements of
density of about 62% (and a corresponding porosity of about soil CaCO3 contents and residual sorbed ammonium masses as
0.35), resulting in the total pore volume (PV) being about 50 L. well as SEM imaging of select soil subsamples.
Pairs of bender element sensors were placed at equally spaced
intervals along the column to measure soil shear wave velocities 7.2.3 Bacterial distributions and progression
at various times during treatments. Aqueous sampling ports Aqueous total cell densities along column lengths were measured
enabled direct sampling from the middle of column cross- using total direct counts of aqueous subsamples at various times
sections and were placed at five offset locations along the column including after initial saturation and after the 1st, 4th, and 9th
lengths. Solutions injections during all treatment stages except cementation treatments are presented in Figure 79a. While
for the byproduct removal stage was performed at a constant rate aqueous counts do not account for cell densities attached to soil
of about 0.6 L/min and involved a treatment volume of near 1.5 particles, it is reasonable to assume that there is some equilibrium
PV (76 L). Injection durations were ~3.1 hours for all treatments. between suspended and attached cells and therefore the relative
The pumping system involved a closed circulation system at the abundance of suspended cells may be proportional to total cell
influent location in order to control injection flow rates and densities. It should also be mentioned that total aqueous cells in
pressures and a static tailwater head was used to maintain full stimulated columns do not solely reflect ureolytic cells and
saturation. Pressure sensors were used to measure fluid pressures cannot distinguish between cell densities facilitating ureolysis
during injections in order to track soil hydraulic conductivity and those that may not be active in this process. Nevertheless,
changes in time. such measurements can be useful towards verifying that
The 26 day treatment program implemented, as summarized microbial growth is occurring during stimulation and/or the
in Figure 78b, consisted of the following steps. First, columns distribution of augmented cells. As shown in Figure 79a, such
were saturated with artificial groundwater and passive tracer
measurements demonstrate total cell growth within the treatment
testing was performed by injecting 1.5 PV of a 15 mM NaBr
zone during the stimulation phase with aqueous cell densities
solution followed by 1.5 PV of deionized water in all columns
largely maintained throughout subsequent cementation
while monitoring effluent concentrations in time to establish
passive tracer breakthrough curves. Soil columns were then treatments. Initial suspended total cell densities were initially
subjected to varying biological treatments (either stimulation or uniform along column lengths with values near 1 x 106 cells/mL.
augmentation), followed by cementation and ammonium Measurements after the 1st cementation treatment, which
byproduct removal phases, with the treatment timeline and followed all stimulation and augmentation treatments, however,
injection frequency of each phase indicated in Figure 78b. showed a near 350% increase in the suspended cell populations
Importantly, the number of stimulation treatments were identical for the SL column and a greater than 600% increase in the SH
between the two stimulation columns (SL and SH) and the and AH columns. Increased cell densities in SH in comparison to
number of cementation solution injections were identical SL were consistent with the rate altering stimulation processes
between all three columns. As shown, augmentation treatment employed and the similarity between the SH and AH values
injections were performed twice due to the poor performance
observed during the initial cementation treatments following the
first augmentation. Particular treatment formulations used are
483
Figure 79. Comparison of SL, SH, and AH columns for (a) aqueous cell densities obtained initially, during the 1st (C1), 4th (C4), and 9th (C9) cementation
treatment, (b) urea concentrations in time during the 9th cementation treatment sampled at 0, 1, 2, 4, 8, 24, and 48 hours after injection, (c) V s measured
initially, during the 1st (C1), 5th (C5), and 9th (C9) cementation treatment, (d) end condition improvement through calcite content and V s, (e) aqueous
NH4+ concentrations during by-product rinsing with initial (0 PV), after the full injection (10.5 PV), and 12 hours after injection (12.5 hr after 10.5 PV),
and (f) sorbed NH4+ concentrations from soil samples (adapted from San Pablo et al. 2020).
484
indicated that comparable cell densities that were targeted were injection source, leaving much of the column length with more
indeed realized. Further measurements after the 4th and 9th minimal cementation. Urea degradation trends in the AH column
cementation treatments demonstrated that aqueous cell indicated that less than 50% of the injected urea concentrations
populations remained relatively constant along column lengths were consumed along the column length, resulting in a much
in all columns, with the largest changes occurring closer to lower precipitation yield.
injection sources. This can be at least partially attributed to the
region closer to the injection location receiving higher 7.2.5 Shear wave velocities
concentrations of treatment solution constituents (including Shear wave velocity measurements provide the most direct
organics) as well as dissolved oxygen concentrations during the indication of CaCO3 precipitation occurring along the column
treatment injection process, which likely yielded higher bacterial lengths due to the direct and robust correlation between
cell growth in stimulated columns. In contrast, trends in the biocementation magnitudes and soil shear stiffnesses (Figure
augmented column, which received no nutrients, suggested that 41). Figure 79c presents shear wave velocity profiles measured
cells may have been eluted near the injection source during along column lengths after initial saturation and at the end of the
repeated treatments. retention period for the 1st, 5th, and 9th cementation treatments.
As shown, differences in the magnitude and uniformity of
7.2.4 Urea degradation rates improvement between columns is clearly evident. In the SL
Urea degradation was further monitored in all columns to column, Vs trends are relatively flat along the column length,
characterize bulk reaction rates during injections. Measured urea indicating a relatively uniform improvement that was consistent
concentrations along column lengths are shown in Figure 79b for with the treatment objective for this column. After the 9 th and
various times after the 9th cementation treatment in all three final cementation treatment, the shear wave velocity ratio
columns. Such measurements confirmed that urea degradation between the furthest (Vs = 546 m/s) and closest (Vs = 1186 m/s)
and related CaCO3 precipitation reactions occurred along the bender element measurement locations from the injection source
column length as designed for different treatment variations. The was ~0.46. The rate of Vs increase per cementation treatment was
distribution of urea concentrations across the column lengths also relatively constant with treatment number, such that
measured immediately following injections, indicated as 0 hrs. projections to higher treatment levels could be made. The
after injection in Figure 79b, reveal the concentrations of urea increase in shear wave velocity closest to the injection source
that were delivered versus consumed during the injection period. was near 1000 m/s and exceeded many of the improvement levels
For the SL column, the urea concentration near the injection discussed earlier in Section 5. It is also evident that an increase
source was near 230 mM, only a slight decrease from the injected in Vs would have been realized in SL at distances further than the
concentration of 250 mM, while the urea concentration at the 3.7 m column length if the test set up was different, indicating
most distant sampling location ( > 3 m) was near 160 mM (based on extrapolation and conservation of added reagent
(reflective of both dispersion and reactions during transport). In masses) that measurable improvements at distances greater than
SL, the urea concentrations measured from all ports right after 5 m likely could have been realized.
injection were 190 mM on average, about 76% of the injected In contrast with trends observed in the SL column, SH column
concentration of 250 mM suggesting that more minimal urea shear wave velocity trends are higher near the injection source
degradation had occurred during injection. In contrast, urea but decrease significantly with injection distance and particularly
concentrations in the SH column immediately after injections after a distance of 2.3 m. This results in a lower Vs ratio between
decreased rapidly with increases in column length such that the the furthest and closest measurement locations of only ~0.25.
concentration at the most distant sampling location ( > 3 meters) These Vs trends are consistent with the earlier urea degradation
was near 13 mM which was only 6% of the 220 mM urea trends shown in Figure 79b and suggest that a majority of the
reactions and precipitation observed in this column occurred
concentration measured near the injection source. Consequently,
during the treatment injection period. However, surprisingly the
this more rapid urea degradation during the injection in SH
magnitude of Vs near the middle of the column is higher than
resulted in a minimal amount of cementation occurring near the
what the urea trends might have suggested. In this column, the
end of the column. Interestingly, urea concentrations in the AH rate of Vs increase per treatment was more rapid initially, but
column immediately after the treatment injection indicated that a began to decrease with additional treatments, particularly at
minimal amount of urea was consumed during treatments. This further treatment distances, indicating that additional treatment
was unexpected, and indicated that the augmented bacteria injections (if applied) would have likely only magnified the
population present in AH was not distributed in a manner that treatment non-uniformity observed. The AH column only
achieved bulk rates that were similar and spatially comparable to showed detectable Vs increases close to the injection source at
SH. the first measurement location, with negligible Vs increases
Examination of the additional trends of urea concentrations further along the column length. This further confirms the
along column lengths over the retention time, which was ~24 ineffectiveness of the employed non-nutritive augmentation
hours for SH and AH and ~48 hours for SL, provide further strategy used, particularly for applications where achieving
details regarding the spatiotemporal reduction of urea. In the SL improvement over large treatment distances is of high priority.
column, the rate of urea reduction was slowest near the injection However, other augmentation protocols could have been
source, with a 90% reduction of the initial urea concentration designed and employed to likely improve performance. Such
requiring more than 30 hrs., and was most rapid at the largest strategies could have included additional/more frequent
injection distance, where a similar 90% reduction required only augmentation treatments and a mixture of augmentation and
about 4 hours. Nevertheless, the ability to control the treatment stimulation through the inclusion of organics in treatments, such
process such that the majority of urea degradation and associated as yeast extract.
CaCO3 precipitation occurred during the retention period instead
of during the injection period resulted in more uniform CaCO3 7.2.6 Soil CaCO3 contents
precipitation along the column length. In contrast, the urea trends Final soil CaCO3 contents, measurable through destructive
in time for the SH column show that urea is consumed along testing after completion of the entire treatment program
almost the entire column length within about 1 hour (near (including the byproduct removal phase discussed subsequently),
injection source). This high reaction rate was expected to provide the first direct measure of treatment success and directly
produce non-uniformity in delivered reagents and more localized map to the targeted changes in engineering properties (Sections
CaCO3 precipitation at relatively short distances from the
485
4 & 5). Soil CaCO3 content measurements in samples obtained conclusively demonstrated the feasibility of removing generated
from the middle of column cross-sections with column length ammonium byproducts, subsequent studies (Lee & Gomez
(including between bender elements) are plotted in Figure 79d. 2022a) (Section 3.3) have shown that even more effective
The increase in Vs at the end of the cementation treatment is also removal of the remaining low concentrations can occur through
plotted for comparison. The final measured soil CaCO3 contents staged injections and the use of alternative cation formulations
were about 5 to 8% immediately adjacent to the injection source (i.e., K+ ions).
in all columns, but the differences in uniformity and magnitude
of improvement are apparent at injection distances of 1 m and 7.2.8 Summary
beyond. A comparison of the theoretical soil CaCO3 contents that The 3.7 meter long soil column experiments performed in San
could have been precipitated (considering treatment Pablo et al. (2020) (as well as the complimentary works
concentrations and volumes) against the measured soil CaCO3 summarized in Table 5) represents the integrated depth of
contents showed that in the SL column appreciable aqueous understanding and control of the enabling biogeochemical
calcium and urea concentrations exited the column and remained processes and associated monitoring methods that can be used
available for improvements at a distances beyond the 3.7 m throughout the treatment process and shows that the technology
column length. In the SH column, the amount of precipitation has matured to a point where deployment at field sites may be
near the injection source was larger, but reduced significantly performed reliably in a controlled manner.
with distance due to reactions during injections, which reduced
transported reagent concentrations. 7.3 MICP mitigation of liquefaction hazards
Soil CaCO3 content trends agreed very well with observations Significant research effort has focused on understanding the
made on the basis of Vs; the final trends of SL were the most relationships between engineering behaviors afforded by MICP
uniform and indicated that improvement could have been in parallel with the previously described upscaling efforts
realized at even further injection distances, the SH trends
considering that the magnitude and extent of MICP improvement
indicated that improvement was concentrated within 2.3 m of the
required at a given site will be application specific. This has been
injection source, and the AH trends indicated that improvement
done more extensively at the laboratory element scale, as detailed
was only realized within about 0.5 m of the injection source. The
robust correlation between Vs and soil CaCO3 contents allows for in Sections 5 and 6, than at the systems level, however, such
treatment progression and uniformity to be assessed in real time, efforts have significantly advanced understandings of the
is critical to ensuring that targeted improvements in engineering technology for liquefaction hazard mitigation applications.
properties can be realized, and affords the opportunity to adjust Table 6 presents a summary of the system level studies that
the treatment program during implementation, if/as necessary. have been performed which have evaluated the effect of MICP
improvement on liquefaction triggering susceptibility and the
7.2.7 Byproduct removal consequences of liquefaction once triggered. As evident, the
Following the completion of all cementation treatments, majority of these systems level evaluations have been
ammonium byproduct concentrations residing within columns investigated using geotechnical centrifuge modeling.
were targeted for removal by pumping ~10.5 PV of an alkaline One of these more recent studies, Darby et al. (2019), used
calcium rich rinse solution over continuously over 700 min in a centrifuge modeling to examine how MICP treatment could
single injection. Following this injection, rinse solutions resided increase soil liquefaction triggering resistance and alter the
within columns under stop flow conditions for 12 hours, which consequences of liquefaction once triggered using models
small scale laboratory test results showed was effective in consisting of a uniform free-field liquefiable loose sand deposit.
displacing sorbed NH4+ ions from particle surfaces (Lee et al. This study is highlighted in the following section to describe the
2019b). Figure 79e presents aqueous ammonium concentrations primary insights gained, and is further supplemented by recent
along soil column lengths at various times during rinsing. As 1D numerical simulations, which demonstrate the agreement that
shown, gradients in aqueous ammonium were observed within can be realized between the physical experiments and responses
columns during rinsing with greater ammonium removed closer predicted using the recently developed PM4SandC model
to the injection source (distance of 0 m) and continued reductions described in Section 6.2.
in concentrations at all locations with increases in injected
volumes. Initial ammonium concentrations near 500 mM were 7.3.1 Centrifuge modeling of MICP soil liquefaction
observed, which was expected given the applied urea behaviors
concentrations of 250 mM, and were reduced to values between Darby et al. (2019) prepared saturated Ottawa F-65 sand models,
0.01 and 10 mM after rinsing. Final aqueous ammonium treated to varying degrees of MICP improvement, and subjected
concentrations were less than 10 mM throughout the SL column models to repeated shaking events using the 1m radius centrifuge
and within the first 3 meters of the SH and AH columns, with all at the UC Davis Center for Geotechnical Modeling in order to
columns achieving near 99.7% aqueous ammonium removal. evaluate how (1) MICP improvement changes the free-field
Although not shown, the rate of aqueous ammonium removal system response as indicated by measured accelerations, induced
was much more rapid during the first few pore volumes of shear strains, and pore pressure generation during shaking, (2)
pumping, with outlet well ammonium concentrations decreasing cementation degrades spatially and temporally with increases in
from 500 mM to about 10 mM after 3 PVs of continuous rinse the applied loading magnitudes, (3) cone penetration resistances
injections and further reductions occurring more gradually. increase following treatments and degrade with progressive
In addition to aqueous ammonium concentrations, sorbed loading events, (4) existing liquefaction triggering curves may
ammonium masses were also quantified after rinsing. Figure 79f shift as a function of MICP cementation, and (5) resulting
presents sorbed ammonium masses for all columns with length. displacements following triggering events can be reduced by
These results suggested that not insignificant sorbed ammonium MICP improvement.
may persist after rinse injections and may require
characterization following field-scale applications (Lee et al.
2019b). Such results also demonstrated that an equilibrium may
exist between sorbed and aqueous ammonium concentrations
with a near linear relationship observed between both at low
ammonium concentrations (< 10 mM). While these results
486
Table 6. Summary of MICP upscaling projects performed focused on liquefaction hazard mitigation.
Model & [Prototype] Shake Table (1g) vs.
Year Container Soil Type Application(s) References
Dimensions (m) Centrifuge (radius @ Ng)
0.50 x 0.24 x 0.11 Flexible Shear Triggering, system reponse, Vs
2013 Centrifuge (1m @ 50g) SP (Ottawa 50-70) Montoya et al. 2013
[25 x 12 x 6] Beam monitoring, settlement
0.50 x 0.24 x 0.11 Flexible Shear Triggering relationships, Vs & qc
2019 Centrifuge (1m @ 80g) SP (Ottawa F-65) Darby et al. 2019
[40 x 19 x 9] Beam monitoring, settlement
Triggering, spectral response,
2020 0.7 x 0.7 x 0.5 Laminar Shear Shake Table (1g) SP (Calcareous Sand) Zhang et al. 2020
settlement
Triggering, foundation
1.65 x 0.80 x 0.53 Flexible Shear
2021 Centrifuge (9m @ 60g) SP (Ottawa F-65) interactions, settlement, teatment Zamani et al. 2021
[99 x 48 x 32] Beam
zone, stratigraphic effects
0.50 x 0.24 x 0.11 Flexible Shear Triggering, settlement, treatment
2022 Centrifuge (1m @ 80g) SP (Ottawa F-65) San Pablo et al. 2022b
[40 x 19 x 9] Beam zone
Following model construction, the moderately cemented
7.3.1.1 Test design specimen was treated using three cementation injections to
Darby et al. (2019) leveraged previous advances and findings achieve an average shear wave velocity of about 400 m/s, which
from an earlier centrifuge study by Montoya et al. (2013) which corresponded to a soil CaCO3 content near 1.4% by mass. In
showed how MICP treatment can alter systems level responses, order to establish ureolytic activity, augmentation with S.
a previous upscaling study by Gomez et al. (2018a) that indicated pasteurii was used to minimize the potential for injection related
how CPT qc can progressively track increases in MICP model disturbances and to maximize time efficiency. As shown
cementation, and laboratory studies (Section 5) that earlier in Section 4.5, regardless of biological approach
characterized how the monotonic and cyclic behaviors of sandy (stimulation versus augmentation), comparable improvements in
soils can be modified as a function of MICP treatment engineering properties can be observed on the basis of similar
magnitudes and other factors. The resulting test design sought to soil CaCO3 contents and thus comparable results would have
build upon these earlier findings and extend them to realistic field been expected if stimulation had been used instead.
stress conditions to examine achieved performances at a more Augmentation treatments applied to the model were also
global systems level. nutritive and may also have allowed for unintended stimulation
A set of centrifuge models comprised of a uniform saturated as described in Section 2.3. Following augmentation, models
sand profile representing a simple free-field condition were received a flush injection to remove generated carbonate species
treated to varying levels of MICP cementation and were and cementation injections proceeded. Specific solution
subjected to repeated shaking events, which progressively formulations for each treatment stage are detailed in Darby et al.
increased the applied loading magnitudes. Separate models with (2019). After preparation, both the cemented and uncemented
an initial relative density (Dr) of 38% were treated to “light”, models were saturated under vacuum with a solution of
“moderate”, and “heavy” cementation levels, and two additional
untreated models prepared to 38% and 53% relative density were
performed and allowed for characterization of baseline reference
conditions. For simplicity, only the untreated and “moderately”
treated specimens are discussed herein, with details regarding all
other tests available in Darby et al. (2019).
The models consisted of Ottawa F-65 sand placed at an initial
relative density (Dr) of 38% by dry pluviation into a flexible
shear beam container, thereby producing a level ground free-field
profile. The constructed model dimensions as well as the
instrumentation and cone penetration sounding locations are
shown in Figure 80. Instrumentation included: (1) matching
arrays of accelerometers and pore pressure transducers which
allowed for monitoring of the system’s dynamic response during
shaking, (2) two arrays of bender elements that enabled shear
wave monitoring of cementation formation during MICP
treatments as well as the gradual mechanical degradation of
cementation during progressive shaking events, and (3) a
displacement sensor that allowed for measurement of ground
surface settlements. Furthermore, cone penetration tests were
performed using a 6 mm penetrometer at four locations across
the model, at various times including: (1) after model saturation
(and before MICP treatment), (2) after all MICP treatments
(when performed), (3) after the first shaking event triggering soil
liquefaction, and (4) following the completion of all shaking
Figure 80. Design of centrifuge test model (adapted from Darby et al.
events. 2019).
487
methylcellulose and DI water prepared to a viscosity of which is attributed to the deformations required to degrade the
approximately 5x10 m2/s, near 50 times the viscosity of water, cementation to level at which liquefaction could be induced.
to improve scaling of dynamic pore pressure generation and
related dissipation during shaking events. 7.3.1.3 Cone penetration resistances
After saturation, models were mounted on the centrifuge and The cone penetration resistance profiles obtained in the loose
subjected to an 80g acceleration field for testing, with spin- uncemented and moderately cemented models are shown in
up/spin-down cycles occurring as necessary to mount/unmount Figure 83. Profiles from the uncemented model show a steady
the cone penetrometer between particular shaking events. The increase in tip resistances with continued shaking throughout the
applied shaking event consisted of 15 cycles of a uniform sounding depth, which can be attributed to gradual densification.
amplitude 1.25-Hz frequency sine wave, with the acceleration The profiles obtained in the moderately cemented model
magnitude varying from 0.02 g to 0.2 g, as required to trigger indicated that pre-treatment penetration resistances were similar
liquefaction in the model. In total, the uncemented model was to the untreated model, but then increased by more than 100%
subjected to 11 shaking events that progressed from 0.02 g to following MICP cementation. For example, the qc at mid-depth
0.25 g while the moderately cemented model was subjected to 9 was 2.3 MPa prior to treatment (similar to the uncemented
shaking events that progressed from 0.02 g to 0.20 g.
7.3.1.2 Dynamic responses
Shaking sequences for the loose and moderately cemented
models are shown in Figure 81, with the progression of peak base
accelerations (PBA) and mid-depth cone tip resistances (qc),
shear wave velocities (Vs), maximum shear strains ( max), and
excess pore pressure ratios (ru) plotted versus shaking event
number. As shown, the shaking magnitude gradually intensified
until the model liquefied at mid-depth, which was defined to have
occurred when excess pore pressure ratios (ru = u/ ’v) exceeded
0.95. Model responses were evaluated immediately after each
shaking event. Initial triggering of liquefaction in the
uncemented model occurred at a PBA of 0.06 g during shaking
event 2 and occurred in the moderately cemented model at 0.17
g during shaking event 5. Shear strains, computed following the
procedure by Brandenberg et al. (2010), remained relatively
small throughout all shaking events, an outcome which was
primarily attributed to the boundary conditions induced by the
flexible shear beam container. Vs measurements obtained
between every shaking event increased slightly with progressive
shaking in the uncemented model due to gradual soil
densification; however, Vs remained nearly constant and then
decreased rapidly upon liquefaction triggering in the moderately
cemented model due to cementation degradation. CPT qc values
measured at mid-depth correspondingly increased slightly with
progressive densification in the uncemented model, while such
values decreased dramatically in the moderately cemented model
following liquefaction triggering, with more minor increases
than observed during subsequent shaking events, which
gradually densified the model.
The substantial increase in liquefaction triggering resistances
provided by the moderate level of MICP treatment is particularly
evident in Figure 82, which presents time histories for the 0.06 g
shaking event 2 applied to the uncemented sand and the 0.17 g
shaking event 5 applied to the moderately cemented sand.
Provided are excess pore pressures measured at mid-depth,
surface accelerations (approximately 2.0 m from the surface),
CSR values at mid-depth (determined following Kamai &
Boulanger 2010), shear strains at mid-depth, and base
accelerations in time. As evident, the CSR sustained throughout
the shaking time history reached a maximum value of about 0.8
and remained above about 0.3 in the moderately cemented
specimen until the end of shaking, whereas the uncemented soil
had an initial CSR at mid-depth of 0.15 and then attenuated to
insignificant values after the onset of liquefaction.
Correspondingly, the measured excess pore pressure and surface
acceleration time histories track well with the applied shaking
frequency. The cemented model demonstrated clear
amplification of the input motion, with the input base
acceleration of 0.17 g amplifying to 0.49 g near the ground
surface, whereas the uncemented model exhibited Figure 81. Summary of dynamic response at mid-depth in the
deamplification of the input motion. Shear strains were higher in uncemented and moderately cemented models (adapted from Darby et al.
the moderately cemented model than in the uncemented model, 2019).
488
Figure 83. Progression of cone penetration resistances for (a)
Figure 82. Dynamic response comparison of uncemented and uncemented and (b) moderately cemented model (where both models
moderately cemented of (a) u at mid-depth, (b) surface acceleration, initially were prepared at a relative density of 38%) (adapted from Darby
(c) CSR at mid-depth, (d) shear strain at mid depth, and (e) base et al. 2019).
acceleration versus time (adapted from Darby et al. 2019).
were present and compensated for one another (Darby et al.
model) and increased to 9.9 MPa following MICP. Following the 2019).
0.17 g shaking event 5 in the cemented model, the qc at mid-depth
decreased to 4.5 MPa, and then gradually increased to a final 7.3.1.4 Liquefaction triggering
value near 6.3 MPa after all shaking events were completed. This Cone penetration measurements in combination with the
degradation after initial liquefaction triggering followed by embedded instrumentation provided an opportunity to
progressive densification was consistent with Vs trends. investigate how MICP improvement may affect the CPT case
Furthermore, this moderate cementation level appeared to be history-based liquefaction triggering curves for clean sands used
unique in the sense that the treated soil degraded towards similar routinely in practice (i.e., Boulanger and Idriss 2015). Further
conditions as observed in the uncemented model, albeit with processing of the obtained centrifuge data was necessary for such
some clear residual benefits resulting from cementation. This comparisons to be made. First, equivalent uniform time series
outcome can be attributed to the interaction of three hypothesized were produced from the irregular CSR time series induced in the
mechanisms: (1) densification of the soil skeleton, (2) damage to centrifuge models following the procedure by Seed et al. (1975)
generated cemented bonds, and (3) the reduction of initial void as detailed in (Idriss & Boulanger 2008), with an assumed b-
ratios by the addition of precipitate solids. It is noted that the value of 0.20, which was deemed reasonable following review of
additional cementation levels tested did confirm these processes cyclic laboratory data. Measured cone tip resistances and
and indicated that the moderately cemented model existed at a developed CSRs at mid-depth in all models were subjected to a
unique cementation threshold at which all three mechanisms vertical effective stress of 35 kPa. Typically, measured qc values
would be normalized to 1 atm overburden stress using a CN
489
relationship, however, similar stress-normalization relationships this new triggering curve would indicate that MICP treatment can
for measurements in cemented soils do not currently exist. increase the resistance to initial triggering by up to 400% when
Therefore, the CPT based liquefaction triggering curve for clean compared to uncemented soils at similar qcN values. Furthermore,
sands by Boulanger & Idriss (2015), typically presented for 1 atm a trendline drawn between points observed during the final
(100 kPa) stress normalized conditions, was instead shifted to an shaking event in all models would indicate a residual benefit
overburden effective stress of 35 kPa to allow for comparisons to across all cementation levels of 100% or more.
be made.
The results from all shaking events applied to all models is 7.3.1.5 Liquefaction consequences
shown in Figure 84, along with the liquefaction triggering curve The consequences of liquefaction for the described free-field
for uncemented sands. All model shaking events that did not level ground models can be captured by evaluating how
trigger liquefaction are shown with open symbols, while all estimated maximum shear strains induced in the model at mid-
model shaking events that did trigger liquefaction, regardless of depth and measures of incremental surface settlements varied
prior shaking history, are shown as closed symbols. Provided with the applied equivalent CSR magnitudes in the presence and
arrows indicate the progression of shaking events for the absence of triggering. Figure 85 presents this data for the
moderately cemented specimen only. From this comparison, uncemented and moderately cemented models for triggering and
several insights can be made. First, the progression of the non-triggering events. As shown, a distinct separation between
moderately cemented model data points clearly illustrates the trends exists for events where liquefaction was and was not
increased resistance that MICP can provide with respect to triggered. The maximum shear strain induced during non-
liquefaction triggering relative to an uncemented soil as well as triggering events remained below 0.05% for the uncemented
how gradual cementation degradation over multiple shaking model and below 0.5% for the moderately cemented model.
events eventually decreases qc and then further decreases the However, such strains increased to 1.7% for the uncemented
CRR. For the moderately cemented soil, liquefaction triggering model at the maximum applied CSR of about 0.11, and
occurs at an equivalent CRR of 0.56, which is significantly maximum strains in the moderately cemented model were near
higher than the expected value for uncemented soil of about 0.13. 1% when subjected to a CSR of 0.56. Incremental settlements
In addition, the initial normalized cone tip resistance of 88 were negligible for the uncemented soils for non-triggering
decreases to 60 due to the liquefaction triggering event, and events, however, settlements increased between 35 and 75 mm
further decreases to a value of about 37 before it then increases
for all liquefaction events, regardless of the CSR loading
due to densification, ultimately reaching a final value near 55.
magnitude or shaking sequence. The moderately cemented
Even near the end of the applied shaking sequence it is evident
model similarly did not have any significant settlements during
that the degraded cementation still increases the CRR by 100%
or more relative to the uncemented triggering curve. When non-triggering shaking events prior to the initial liquefaction
considering data from all three cemented models, it is clear that event. However, the liquefaction event and subsequent strong
MICP treatments provides significant benefits with respect to shaking events resulted in incremental settlements between 40
liquefaction triggering resistances. The Boulanger & Idriss and 80 mm. Some additional settlement also was observed in this
(2015) triggering curve for uncemented sands is based upon the model near the end of the shaking sequence even though the
strong correlation between CPT, qcN, and Dr; with higher stress- liquefaction triggering criteria of ru > 0.95 was not formally met.
normalized qcN values corresponding to denser (higher Dr) sands.
However, all of the centrifuge models were prepared to the same
initial density, and therefore increases in qcN are solely the result
Figure 84. Comparison to existing liquefaction triggering correlation showing progression of shaking events for moderately cemented model (adapted
from Darby et al. 2019).
of cementation increases instead. If a trendline were drawn
between the initial triggering points for each of the three models,
490
cemented Ottawa F-65 sand as described earlier. A 1D site
response analysis was performed using the explicit finite
difference program FLAC (Fast Lagrangian Analysis of
Continua, Itasca 2016) and the user-defined plasticity model for
bio-cemented sands PM4SandC (Section 6.2) was used to predict
the response of biocemented sands when subjected to cyclic
loading. The two analyses presented here represent a 1D soil
column model in FLAC where model boundary conditions,
constraints, and soil properties were prescribed to best
approximate the conditions of the reference centrifuge model
tests. Initially, a grid of 11.4 m prototype-scale is defined with
two soil layers: (1) the liquefiable Ottawa sand and (2) a mix of
Monterey sand and fine gravel that was beneath the Ottawa F-65
sand and used for drainage. The base layer was 2.4 m thick and
was discretized into 0.6 m elements, and the liquefiable sand was
9 m thick and was discretized into 0.5 m elements (Figure 86).
The liquefiable Ottawa sand was further divided into three
sublayers: (1) bottom, (2) middle, and (3) top, which was based
upon the location of the included instrumentation and
specifically the bender element sensors. This was performed to
account for the modest nonuniformity observed in MICP
treatments within the centrifuge models as discerned by bender
element Vs measurements. As such, the liquefiable Ottawa sand
is assigned three different sets of parameters to better
approximate the realized soil conditions which were present in
the centrifuge model. The soil column was fixed at the base in
both the x- and y- directions prior to initializing stresses. Stresses
were then initialized in the model due to gravity and the model
was run to establish static equilibrium with no water present.
After static equilibrium was confirmed, pore pressures were
initialized by fully saturating the model. Pore pressures increased
from zero at the top of the model to hydrostatic pressures at the
11.4 m depth. The cyclic loadings applied at the base of the
models were comparable to the shaking events which initially
triggered liquefaction in these models were prescribed at the base
of the models. As described earlier, for the uncemented model,
liquefaction was initially triggered during shaking event 2 which
applied a PBA of 0.06 g, whereas for the moderately cemented
model liquefaction was initially triggered during shaking event 5
Figure 85. (a) Maximum shear strain and CSR15cyc and (b) incremental
settlement and CSR15cyc for non-triggering events (open symbol) and
triggering events (closed symbol) (adapted from Darby et al. 2019).
7.3.2 Numerical modeling of centrifuge tests
Nonlinear effective stress 1D and 2D site response analyses with
an appropriately calibrated constitutive model are routinely
performed to predict the response of uniform and layered soils
during shaking and in turn capture the generation of excess pore
pressures along with stress strain responses. While these analyses
have been performed for uniform and nonuniform soils with
stochastic Vs layering (e.g., Montgomery and Boulanger 2017)
or conditioned upon site investigation data (e.g., Pretell et al.
2022) and sloping ground (e.g., Montgomery and Ziotopoulou
2020), no site response analyses have been performed for
deposits that have been either partially or fully treated using
biocementation. The following subsection presents a numerical
simulation of the Darby et al. (2019) centrifuge test described in
the previous Section 7.3.1. This analysis serves as a validation of
the constitutive model for biocemented sands PM4SandC
(Section 6.2) at the system level as well as under free-field
conditions.
7.3.2.1 Model design
The centrifuge model considered involved an instrumented level- Figure 86. 1D site responses analysis in FLAC represented by a 1D soil
ground free-field centrifuge test on uncemented and moderately column with different soil layers in prototype scale.
491
which applied a PBA of 0.16 g. Due to the use of a flexible shear liquefaction at 15 cycles. For this analysis, a liquefaction
beam container in the centrifuge, the lateral nodes of the column criterion of an ru larger than 95% was selected in agreement with
at each elevation are connected to each other and, as such, forced the centrifuge test. The ratios between the cemented and
to move together as would be the case in the free field. uncemented shear wave velocities (VsR) were further specified.
For the uncemented model, VsR was prescribed to be 1.0 and the
7.3.2.2 Model parameters constitutive model therefore reverts to PM4Sand. For the
Soil properties for each layer were assigned using values cemented model, VsR was assigned as reported in Table 8,
determined during the experimental centrifuge tests as well as following the measured Vs from the experiment. For the base
from other values available in the literature. The minimum and layer, the dynamic behavior was represented using the hysteretic
maximum void ratios and specific gravity of Ottawa F-65 sand model (sig3 model) with selected parameters of a = 1.014, b = -
were assumed using reported values from previous tests (Carey 0.4792 and xo= -1.249 as recommended by Seed & Idriss (1970).
et al. 2020a). The relative density (Dr) of the base drainage layer
was estimated from previous experiments (Dr ~70 %) as direct 7.3.2.3 Dynamic responses
Dr measurements were not available. The relative density of the Validation of numerical results against the centrifuge
Ottawa F-65 sand layer was adjusted from the reported 38% measurements necessitates the selection of recording points on
based on the measured cone penetration resistances at 2 m depth the soil column, which closely approximate the location of the
(Carey et al. 2020b). The permeability of the base layer actual instrumentation used in the centrifuge tests. At about 2 m
(Monterey sand and fine gravel) and the Ottawa F-65 sand were (i.e., TOP), 4 m (i.e., MID), and 5.5 m (i.e., BOT) from the soil
set to 2 x10-3 and 1.3 x 10-5 m/s, respectively, to better surface, numerical time histories were extracted. Figure 87
approximate pore pressure generation. Table 7 summarizes the presents a comparison of the 1D site response simulation and
soil properties for both the uncemented and cemented models. centrifuge experimental results obtained at mid-depth in the
Soil shear wave velocities were determined from bender Ottawa F-65 sand layer. For the uncemented sand, the dynamic
element measurements obtained near the bottom, middle, and top response from the simulation agrees well with the centrifuge test.
of the model as reported in Darby et al. (2019). The observed Vs As shown, simulated shear strain, CSR, surface acceleration, and
nonuniformity warranted further sublayering of the untreated and pore pressure generation time histories are comparable to the
treated Ottawa F-65 sand layers in the model. While Vs values in experimental recordings. The simulated maximum strain (Figure
the uncemented Ottawa sand model were more uniform than in 87d) and CSR (Figure 87c) are 0.11% and 0.26, respectively, for
the moderately cemented sand model, both were subdivided into the uncemented model. These values are only slightly different
three layers for consistency between analyses. Since the base from those measured in the centrifuge experiment (1.7% shear
layer was not instrumented during the experiment, the Vs of this strain; 0.15 CSR) and were therefore deemed acceptable.
layer was estimated to be 204 m/s using available relationships In addition, measured and simulated surface accelerations
between Vs, Dr, and effective vertical stresses and the assumption (Figure 87b) confirm that the applied base accelerations were de-
that this compacted base layer was stiffer than the liquefiable amplified for the uncemented case (from PBA of 0.06 g to a peak
layer (e.g., Andrus & Stokoe 2000, Biryaltseva et al. 2016). surface acceleration of about 0.03 g). Good agreement between
The parameters requiring calibration in the PM4SandC the centrifuge measurements and the simulation results are also
constitutive model were determined through a variety of means. observed for the moderately cemented sand model. The
Table 8 summarizes these calibrated parameters for both the maximum strain achieved in the simulation (Figure 87d) was
uncemented and cemented models. In both models, mass-based 0.37% and was consistent with the experimentally observed
Dr values were adjusted to be consistent with the measured cone behavior from the standpoint that it increased when compared to
penetration resistances and stresses. Small-strain shear moduli the uncemented sand. However, the experimental centrifuge
(Go) were computed using both the measured Ottawa F-65 sand results for both the uncemented and moderately cemented cases
dry densities and the square of the measured shear wave velocity show a higher maximum shear strain than simulations, which
at each layer, which were first stress-normalized using may be attributable to the reconsolidation processes that occur
atmospheric pressure and an assumed coefficient of lateral earth simultaneously during shaking in the experiments that are not
pressure at-rest (Ko) value (Allmond & Kutter 2012, CGM considered in the simulations. The maximum achieved CSR in
Report 12-01). The contraction rate parameter (hpo) was the simulation (Figure 87c) is 0.81, which is identical to the CSR
iteratively estimated using single element direct simple shear Table 8. Soil properties and constitutive models selected for the
(DSS) calibrations where a target CRR, initial hpo, Dr, and Go uncemented and moderately cemented model simulations.
were specified until the single element driver triggered
Bottom Mid Top
Soil Base Ottawa Ottawa Ottawa
Table 7. Initial soil properties selected for uncemented and moderately
layer
cemented models. sand layer sand layer sand layer
Uncemented Model
Bottom Mid Top
Base dynamic
Soil Ottawa sand Ottawa Ottawa hysteretic PM4SandC PM4SandC PM4SandC
layer model
layer sand layer sand layer
hpo - 0.2 0.44 0.44
Uncemented Model
Go - 547.8 513 513
emin 0.536 0.51 0.51 0.51
VsR (Vs ratio) - 1.0 1.0 1.0
emax 0.843 0.78 0.78 0.78
Gs 2.64 2.67 2.67 2.67 Moderately Cemented Model
Dr (%) 70 41 41 41 dynamic
hysteretic PM4SandC PM4SandC PM4SandC
model
Vs (m/s) 300 155 150 150
hpo - 5 2 4
Moderately Cemented Model Go - 1074 1206 912
Vs (m/s) 300 155 150 150 VsR (Vs ratio) - 1.40 1.53 1.33
Vs cem edeg - 0.002 0.002 0.002
300 217 230 200 eres - 0.01 0.01 0.01
(m/s)
492
observed from the centrifuge experiment. The measured surface ground surface; Figure 88c), and (4) a location within the Ottawa
acceleration (Figure 87b) shows a similar amplification of the F-65 sand layer towards the top of the layer (2 m from ground
base acceleration as the shaking propagates to the ground surface surface; Figure 88d). It can be observed in the spectral
with a maximum achieved surface acceleration of 0.4 g accelerations shown that the uncemented sand deamplified the
(compared to a PBA of 0.16 g). The pore pressure generation acceleration achieved near the ground surface from a peak
simulated and measured at mid-depth in the Ottawa F-65 sand ground acceleration (PGA) of 0.06 g to ~0.03 g at period of 0
layer (Figure 87a) shows similar pore pressure generation seconds, whereas the moderately cemented sand tended to
behaviors for both uncemented and moderately cemented cases. amplify the acceleration near the ground surface from a PGA of
Figure 88 presents the response spectra computed at various 0.16 g to 0.4 g at a period of 0 seconds. This behavior is seen in
locations within the liquefiable Ottawa F-65 sand layer both the centrifuge measurements and simulation results. The
consistent with the experimental accelerometer locations. These predominant period (~0.75 seconds) experimentally observed for
locations include: (1) the base of the Monterey sand and fine both the uncemented and cemented models are largely consistent,
gravel mix (11.4 m from the ground surface; Figure 88a), (2) a however, the spectral accelerations observed at this predominant
location within the Ottawa F-65 sand layer towards the bottom period for the moderately cemented sand case are significantly
of the layer (5.5 m from the ground surface; Figure 88b), (3) a
Figure 87. Dynamic responses from centrifuge experiments and simulations compared for the uncemented and moderately cemented cases at the middle
of Ottawa F-65 sand layer (centrifuge data from Darby et al. 2019).
location at the middle of the Ottawa F-65 sand layer (4 m from
493
Figure 88. Response spectra obtained for the (a) uncemented and (b) moderately cemented centrifuge experiments and simulations at various depths
within the Ottawa F-65 sand layer (centrifuge data from Darby et al. 2019).
amplified when compared to the uncemented case. While this the presented results are an encouraging validation of the
amplification behavior is generally also captured by the formulation and implementation of the model and highlight its
simulation, simulated results do not perfectly match those from ability to capture the undrained monotonic and cyclic behaviors
the centrifuge experiments because pore pressure generation was of biocemented sands at both the element and system levels.
prioritized during this calibration, as prediction of liquefaction Furthermore, it is important to note that the presented simulated
triggering was of primary interest (ru > 95%). responses were obtained through minimal calibration effort (i.e.,
by prioritizing only primary input variables). This modeling
7.3.3 Summary decision was intentional to demonstrate the utility of the model
Geotechnical centrifuge experiments have and will continue to for future practical field-scale implementations of MICP and
improve understanding of the liquefaction behaviors of potentially other similar cementitious technologies.
biocemented sands at the systems level. Recent centrifuge tests
have demonstrated that biocementation can afford clear changes 8 CONCLUSIONS
in dynamic wave propagation behaviors, excess pore pressure
generation, cone penetration responses, and related shear strains, This paper presents current state of the art understandings
with large improvements in liquefaction triggering resistances. regarding MICP soil improvement and its applicability to
Following triggering, such tests suggest that the consequences of liquefaction mitigation. The emerging field of biogeotechnics
liquefaction can also be reduced using MICP, although to a lesser presents a breadth of opportunities to develop bioinspired and
extent than triggering behaviors, consistent with earlier results biomediated technologies. Of these solutions, biomediated soil
described in Section 5. Leveraging these results, the PM4SandC improvement processes have advanced rapidly over the past 25
constitutive model (Section 6) was used to model both the years and been shown to be capable of modifying the soil
uncemented and moderately cemented centrifuge experiments engineering behaviors through unconventional mechanisms
and showed good agreement between experimental responses including the generation of inorganic minerals, gas phases, and
and 1D site response simulations following initial model biopolymers within the soil matrix. Such technologies have the
calibration. While future analyses will continue to expand potential to realize transformative engineering performance and
understanding of the utility of this model for MICP treated soils, sustainability gains relative to current soil improvement
494
technologies, and will benefit from continued evaluation during outlined. The range of soils and environmental conditions that
their development using frameworks such as life cycle are physically and practically amendable to biocementation
sustainability assessment to maximize their environmental, include an extensive range of quartz-based soils, carbonate
financial, and societal benefits. Of the biomediated soil sands, silty materials, and anthropogenically-produced waste
improvement technologies, MICP is the most mature, with materials. The capability of stimulated MICP to improve soils
envisioned applications spanning across civil engineering using solely the transport of soluble reactants may extend the
subdisciplines and its application to liquefaction mitigation range of treatable soils into smaller particle size ranges relative
arguably at the forefront of geotechnical development efforts. to that which has already been established.
The MICP process wherein carbonate generation occurs via An extensive range of monitoring processes exist to track
microbial ureolysis enables production rates that are nearly 100 microbiological, geochemical, and mechanical changes during
times greater than any known alternative microbial pathway. MICP implementation. Microbiological methods provide the
Augmentation with non-native Sporosarcina pasteurii, a ability to quantify microbial growth, activities, strain purity, and
ureolytic bacterium, has been used to complete the MICP soil community changes, and can provide the fundamental basis
process, but has associated practical challenges and clear for reactive transport models needed for upscaling. Geochemical
disadvantages related to cell injection processes and treatment measurements can be used to interrogate the status of
uniformity. Stimulation is the alternative and preferred biological precipitation processes, quantify byproducts produced and their
approach for MICP and involves the enrichment of indigenous removal, as well as indirectly assess microbially-mediated
ureolytic bacteria through the application of solutions containing reaction rates. Cementation magnitudes can be effectively
nutrients and environmental factors that effectively select for quantified using many different methods and metrics, with soil
ureolytic microorganisms. Use of the stimulation process calcite contents and shear wave velocity changes being the
provides important practical implementation and performance authors’ preferred metrics through which expected behavioral
benefits, including improved treatment control and uniformity, improvements can be inferred. Advanced material analyses and
reduced soil ecological and environmental impacts, and non-destructive imaging techniques can further characterize
reductions in implementation costs and field requirements. The other important aspects regarding the generated cementation,
stimulation process is robust and soils incapable of stimulation including its mineralogy, morphology, and interparticle spatial
efforts have yet to be encountered. Although the presented distribution. Monitoring and assessment methods also include
treatment formulations have been shown to select for the genus more traditional geotechnical testing techniques including
Sporosarcina effectively, there is also a significant amount of hydraulic conductivity and cone penetration testing. Existing
diversity between the microbial species identified in stimulated
frameworks that leverage information at both large (CPT) and
soils, suggesting that ureolytic bacteria may be ubiquitous in
small (Vs) strains are effective in identifying cemented materials
geotechnically-relevant soils and the success of stimulated MICP
in the presence of other complicating factors expected at field-
is not dependent on the presence of a small subset of bacterial
strains. scale including soil and confining stress variations.
Differences in the treatment processes employed in The mechanical behaviors of MICP treated soils have been
combination with surrounding environmental conditions have comprehensively characterized for a variety for different strain
been shown to impact the mineralogy of the precipitation levels, loading types, and drainage conditions. Low-strain
generated during MICP, which may have important implications dynamic behaviors shift with added cementation and include
with respect to the realized engineering benefits. Furthermore, increases in initial shear stiffnesses, decreases in the threshold
microstructural changes resulting from MICP have been shear strains at which shear stiffness degradation occurs, and
captured through a diverse range of direct imaging approaches increases in damping ratios at lower strain levels. Monotonic
and have yielded new understandings regarding interparticle undrained and drained shearing behaviors are established for
precipitation distributions and their effects on engineering biocemented sands, with significant improvements in strengths,
behaviors. Recent studies have further advanced understandings stiffnesses, and volumetric behaviors possible. Undrained cyclic
regarding the long-term geochemical permanence of MICP by testing has shown that biocementation can produce orders of
characterizing the abundance of CaCO3 polymorphs present in magnitude improvements in liquefaction triggering resistances
precipitates as well as their dissolution kinetics. Existing even at low cementation levels ( Vs <100 m/s), though
dissolution kinetic models have been incorporated in reactive significant improvements in post-triggering strain accumulation
transport models and calibrated to physical experiments, thus and reconsolidation behaviors require greater cementation
providing a path for assessment of in situ material longevity and magnitudes to achieve more significant precipitation-induced
lifespan prior to and following treatment implementation. densification.
Preferred treatment techniques and associated influencing Numerical models of the mechanical behaviors of
factors have been identified through the practical application of biocemented soils have leveraged the wealth of experimental
this realized fundamental understanding. The preferred treatment data on MICP treated soils and provide the opportunity to
process presented uses stimulation of indigenous ureolytic advance understandings regarding how treated soils will behave
bacteria and represents a culmination of understandings with when subjected to conditions that have not yet been
respect to process geochemistry, microbiological phenomena, experimentally assessed. Discrete element models are able to
and engineering property enhancements, all within the context of appropriately account for experimental behaviors for both
optimizing sustainability metrics. Specific recommendations for monotonic and cyclic loading cases. Such models can also yield
stimulation and cementation treatment phases have been insights that cannot be achieved in physical experiments, such as
discussed, including the impact of treatment formulations and the effect of global shearing conditions on microscale calcite
injection processes on reactive transport conditions and their bond integrity. At the continuum scale a number of constitutive
impact on the realized biocementation spatial distribution. The models have attempted to capture the impact of cementation on
presented techniques may merit modification for specific soil mechanical behaviors. The fundamental basis of the recently
applications, with such decisions ultimately requiring that the developed PM4SandC constitutive model was presented, with
fundamentals described in the earlier paper sections be simulations providing validation that the model can accurately
understood so that such treatment decisions can be informed and capture both monotonic and cyclic behaviors for biocemented
carefully considered. Recent advances with respect to byproduct sands.
management including how generated ammonium can be An envisioned design process was presented for future MICP
removed and effectively monitored after treatments are also field-scale applications and encompasses six stages including:
(1) preliminary suitability evaluation, (2) technology
495
compatibility with site conditions, (3) site characterization, (4) 10 REFERENCES
design requirements, (5) implementation plan, and (6)
monitoring and QA/QC. Considerations and activities specific to Abbas M., Kunhappan D., Dadda A., Geindreau C., Emeriault F.,
each design stage were presented and discussed in detail. Current Esnault-Filet A., and Garandet A. 2020. Discrete element modelling
for biocemented sand: Effect of calcite distribution at the
understandings regarding process upscaling were evident microscopic scaleCardoso, R., Jommi, C., and Romero, E. (Eds.),
through a review of nearly 30 meter-scale experiments, with 4th European Conference on Unsaturated Soils. 6.
detailed review of one project clearly demonstrating the https://doi.org/10.1051/e3sconf/202019505005
advances that have been realized with respect to process
efficiency, cementation spatial uniformity and control, and artificially cemented sand. Journal of Geotechnical Engineering 112
byproduct management, among other aspects. (11), 15. https://doi.org/10.1061/(ASCE)0733-
The effect of MICP on liquefaction behaviors at the systems 9410(1986)112:11(1001)
level are highlighted through review of a recent centrifuge Al Qabany A., Mortensen B., Martinez B., Soga K., and DeJong J. 2011.
experimental program which improved understandings of both Microbial Carbonate Precipitation: Correlation of S-Wave Velocity
with Calcite Precipitation Geo-Frontiers 2011. American Society of
liquefaction triggering and post-triggering behaviors for Civil Engineers, Dallas, Texas, United States, 3993–4001.
biocemented sands as well as reexamined how well-established https://doi.org/10.1061/41165(397)408
liquefaction susceptibility assessment frameworks may need to Al Qabany A., and Soga K. 2013. Effect of chemical treatment used in
be modified to incorporate the benefits of biocementation. The micp on engineering properties of cemented soils. Geotechnique;
PM4SandC constitutive model has been further shown to London 63 (4), 331–339. https://doi.org/10.1680/geot.SIP13.P.022
accurately simulate the responses of past centrifuge experiments, Al Qabany A., Soga K., and Santamarina C. 2012. Factors affecting
including those related to liquefaction triggering, post-triggering, efficiency of microbially induced calcite precipitation. J. Geotech.
and dynamic wave propagation behaviors. These simulations Geoenviron. Eng. 138 (8), 992–1001.
provide an important validation of the model’s formulation, and https://doi.org/10.1061/(ASCE)GT.1943-5606.0000666
Allmond J., and Kutter B. 2012. Centrifuge testing of rocking
clearly demonstrate how behaviors key to liquefaction mitigation foundations on saturated and submerged sand: centrifuge data report
applications can be effectively captured. for jda01 (No. UCD/CGM report12/01).
The path forward for MICP is exciting. The collective Almajed A., Lateef M.A., Moghal A.A.B., and Lemboye K. 2021. State-
advancements to date have matured the technology to a level of-the-art review of the applicability and challenges of microbial-
where it is ready for more comprehensive field-scale trials, which induced calcite precipitation (MICP) and enzyme-induced calcite
will hopefully lead to widespread practical adoption and precipitation (EICP) techniques for geotechnical and
implementation in time for liquefaction mitigation as well as a geoenvironmental applications. Crystals 11 (4), 370.
host of other broader applications. Continued advancements will https://doi.org/10.3390/cryst11040370
present both new challenges and opportunities as MICP Andrus R.D., and Stokoe II K.H. 2000. Liquefaction Resistance of Soils
from Shear-Wave Velocity. Journal of Geotechnical and
transitions to practical field implementation, and the state of the Geoenvironmental Engineering 126 (11), 1015–1025.
art understandings presented herein may provide a knowledge https://doi.org/10.1061/(ASCE)1090-0241(2000)126:11(1015)
base from which these challenges can be addressed and APHA 2005. Standard methods for the examination of water and
overcome. wastewater, 21st ed. American Public Health Association/American
Water Works Association/Water Environment Federation,
9 ACKNOWLEDGEMENTS Washington DC.
Arroyo M., Ciantia M., Castellanza R., Gens A., and Nova R. 2012.
The work of Kyunguk (Thomas) Na in generating the resonant Simulation of cement-improved clay structures with a bonded elasto-
column data and of Seung-Hun (Phillip) Baek in generating the plastic model: A practical approach. Computers and Geotechnics 45,
140–150. https://doi.org/10.1016/j.compgeo.2012.05.008
X-ray CT images and the hydraulic conductivity data is greatly
Asghari S., Chittoori B.C.S., Burbank M., and Hudyma N. 2021.
appreciated. The contributions of the following individuals in Studying the relationship between indigenous microbial
providing data, references, and/or reviewing the manuscript is communities, urease activity, and calcite precipitation in artificial
also appreciated: Dellen Behrend, Ben Blair, Liam Bradshaw, mixes of clay and sand IFCEE 2021. American Society of Civil
Scott Braswell, Bob Burdalski, Susan Burns, Jian Chu, Grainne Engineers, Dallas, Texas, 130–138.
El Mountassir, Annette Esnault Filet, Matthew Evans, David https://doi.org/10.1061/9780784483428.014
Frost, Soo-Min Ham, Hai Lin, Nick Gorski, Alejandro Martinez, Assadi- .,
Ekinci A., Garcia F.E., Harbottle M., Tagarelli V., Jefferson I.,
Ashly Cabas Mijares, James Minto, Narayanan Neithalath, Maghoul P., Masoero E., El Mountassir G., Muhunthan B., Geng X.,
Adrienne Phillips, Bruna G.O. Ribeiro, Michael Tsesarsky, Leon Ghadr S., Mirzababaei M., Mitrani H., and van Paassen L. 2022.
van Paassen, Yuze Wang, Matthew Weil, Pu Yang, Xiao Yang, Recent advances in nature-inspired solutions for ground engineering
and Samantha Young. (NiSE). Int. J. of Geosynth. and Ground Eng. 8 (1), 3.
The manuscript development has been supported by the https://doi.org/10.1007/s40891-021-00349-9
Engineering Research Center Program of the National Science ASTM International 2021. Standard test method for rapid determination
Foundation under NSF Cooperative Agreement No. EEC-
1449501 as well as the National Science Foundation grant ECI- https://doi.org/10.1520/D4373-21
ASTM International 2016. Standard test methods for acidity or alkalinity
1824647. Operation of the centrifuge facility at the University of
California, Davis was supported by the NSF as part of the Natural https://doi.org/10.1520/D1067-16
Hazards and Engineering Research Infrastructure (NHERI) Bachmeier K.L., Williams A.E., Warmington J.R., and Bang S.S. 2002.
network under award CMMI-1520581. Some SEM images Urease activity in microbiologically-induced calcite precipitation.
presented were made possible by the Molecular Analysis Journal of Biotechnology 93 (2), 171–181.
Facility, a National Nanotechnology Coordinated Infrastructure https://doi.org/10.1016/S0168-1656(01)00393-5
site at the University of Washington, which is supported in part Baek S.-H., Hong J.-W., Kim K.Y., Yeom S., and Kwon T.-H. 2019. X-
by the NSF Grant NNCI-1542101, the University of Washington, ray computed microtomography imaging of abiotic carbonate
precipitation in porous media from a supersaturated solution:
the Molecular Engineering & Science Institute, and the Clean
Insights into effect of CO2 mineral trapping on permeability. Water
Energy Institute. Any opinions, findings and conclusions or Resources Research 55 (5), 3835–3855.
recommendations expressed in this manuscript are those of the https://doi.org/10.1029/2018WR023578
authors and do not necessarily reflect the views of the National Baek S.-H., Kwon T.-H., and DeJong J.T. 2022. Impact of microbially
Science Foundation or other supporting agencies. induced calcite precipitation (MICP) on hydraulic conductivity of
496
coarse sands 20th International Conference on Soil Mechanics and Boulanger R.W., and Idriss I.M. 2015. CPT-based liquefaction triggering
Geotechnical Engineering. Sydney, Australia. procedure. J. Geotech. Geoenviron. Eng. 142 (2), 04015065.
Bakken L.R., and Olsen R.A. 1983. Buoyant densities and dry-matter https://doi.org/10.1061/(ASCE)GT.1943-5606.0001388
contents of microorganisms: conversion of a measured biovolume Boulanger R.W., and Idriss I.M. 2014. CPT and SPT based liquefaction
into biomass. Appl Environ Microbiol 45 (4), 1188–1195. triggering procedures (UCD/CGM-14/01).
https://doi.org/10.1128/aem.45.4.1188-1195.1983 Boulanger R.W., and Ziotopoulou K. 2017. PM4Sand (Version 3.1): A
Bang S.S., Galinat J.K., and Ramakrishnan V. 2001. Calcite precipitation sand plasticity model for earthquake engineering applications
induced by polyurethane-immobilized Bacillus pasteurii. Enzyme (Center for Geotechnical Modeling, Department of Civil and
and Microbial Technology 28 (4), 404–409. Environmental Engineering No. UCD/CGM-17/01). University of
https://doi.org/10.1016/S0141-0229(00)00348-3 California, Davis, Davis, CA.
Barkouki T., Martinez B., Mortensen B., Weathers T., Dejong J., Spycher Bradford M.M. 1976. A rapid and sensitive method for the quantitation
N., Ginn T., Fujita Y., Smith R., Barkouki T., Martinez B., of microgram quantities of protein utilizing the principle of protein-
Mortensen B., Weathers T., Dejong J., Spycher N., Ginn T., Fujita dye binding. Analytical Biochemistry 78, 248–254.
Y., and Smith R. 2011. Forward and inverse bio-geochemical https://doi.org/10.1016/0003-2697(76)90527-3
modeling of microbially induced calcite precipitation in half-meter Brandenberg S.J., Wilson D.W., and Rashid M.M. 2010. Weighted
column experiments. Transp. Porous Media 23–39. Residual Numerical Differentiation Algorithm Applied to
https://doi.org/10.1007/s11242-011-9804-z Experimental Bending Moment Data. Journal of Geotechnical and
Bassal P.C., Boulanger R.W., and DeJong J.T. 2022. System response of Geoenvironmental Engineering 136 (6), 854–863.
an interlayered deposit with spatially distributed ground https://doi.org/10.1061/(ASCE)GT.1943-5606.0000277
deformations in the Chi-Chi earthquake.To Be Submitted. Brignoli E.G., Gotti M., and Stokoe K.H. 1996. Measurement of shear
Bassal P.C., Boulanger R.W., and DeJong J.T. 2021. Dynamic analyses waves in laboratory specimens by means of piezoelectric
of liquefaction and lateral spreading for an interlayered deposit in the transducers. Geotech. Test. J. 19 (4), 384.
Chi-Chi earthquake 426–441. https://doi.org/10.1520/GTJ10716J
https://doi.org/10.1061/9780784483695.042 British Columbia Contaminated Sites Regulation 2020. British columbia
Baum M.J., Kovalev A.E., Michels J., and Gorb S.N. 2014. Anisotropic environmental management act contaminated sites regulation
friction of the ventral scales in the snake lampropeltis getula [WWW Document]. Contaminated Sites Regulation.
californiae. Tribol Lett 54 (2), 139–150. Bruno A.W., Gallipoli D., Rouainia M., and Lloret-Cabot M. 2020. A
https://doi.org/10.1007/s11249-014-0319-y bounding surface mechanical model for unsaturated cemented soils
Baveye P.C., Otten W., Kravchenko A., Balseiro-Romero M., Beckers under isotropic stresses. Computers and Geotechnics 125, 103673.
É., Chalhoub M., Darnault C., Eickhorst T., Garnier P., Hapca S., https://doi.org/10.1016/j.compgeo.2020.103673
Kiranyaz S., Monga O., Mueller C.W., Nunan N., Pot V., Schlüter Burbank M., Weaver T., Lewis R., Williams T., Williams B., and
S., Schmidt H., and Vogel H.-J. 2018. Emergent properties of Crawford R. 2013. Geotechnical tests of sands following bioinduced
microbial activity in heterogeneous soil microenvironments: calcite precipitation catalyzed by indigenous bacteria. J. Geotech.
Different research approaches are slowly converging, yet major Geoenviron. Eng. 139 (6), 928–936.
challenges remain. Front. Microbiol. 9, 1929. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000781
https://doi.org/10.3389/fmicb.2018.01929 Burbank M.B., Weaver T.J., Green T.L., Williams B.C., and Crawford
Beal S.L. 1982. On the solution to the Michaelis-Menten equation. R.L. 2011. Precipitation of calcite by indigenous microorganisms to
Journal of Pharmacokinetics and Biopharmaceutics 10 (1), 109– strengthen liquefiable soils. Geomicrobiology Journal 28 (4), 301–
119. https://doi.org/10.1007/BF01059186 312. https://doi.org/10.1080/01490451.2010.499929
Béguin R., Oxarango L., Sapin L., Garandet A., Viglino A., François E., Burbank M.B., Weaver T.J., Williams B.C., and Crawford R.L. 2012.
Mora H., Martins F., Duchesne L., Albrecht D., Esnault-Filet A., Urease activity of ureolytic bacteria isolated from six soils in which
Gutjahr I., and Lépine L. 2019. Experimental tests of soil calcite was precipitated by indigenous bacteria. Geomicrobiology
reinforcement against erosion and liquefaction by microbially Journal 29 (4), 389–395.
induced carbonate precipitation Internal Erosion in Earthdams, https://doi.org/10.1080/01490451.2011.575913
Dikes and Levees, Lecture Notes in Civil Engineering. Springer Burdalski R.J. 2020. Investigating the effect of biological and chemical
International Publishing, Cham, 16–24. https://doi.org/10.1007/978- factors on the reaction kinetics and mineralogy of ureolytic bio-
3-319-99423-9_2 cementation (Master of Science in Civil & Environmental
Benini S., Rypniewski W.R., Wilson K.S., Miletti S., Ciurli S., and Engineering). University of Washington, Seattle, Washington.
Mangani S. 1999. A new proposal for urease mechanism based on Burdalski R.J., and Gomez M.G. 2020. Investigating the effect of
the crystal structures of the native and inhibited enzyme from microbial activity and chemical concentrations on the mineralogy
Bacillus pasteurii: why urea hydrolysis costs two nickels. Structure and morphology of ureolytic bio-cementation Geo-Congress 2020,
7 (2), 205–216. https://doi.org/10.1016/S0969-2126(99)80026-4 Proceedings. American Society of Civil Engineers, Minneapolis,
Bhutange S.P., and Latkar M.V. 2020. Microbially induced calcium Minnesota, 83–95. https://doi.org/10.1061/9780784482834.010
carbonate precipitation in construction materials. J. Mater. Civ. Eng. Burdalski R.J., Gomez M.G., and Gorman-Lewis D. 2022a. Effect of
32 (5), 03120001. https://doi.org/10.1061/(ASCE)MT.1943- environmental factors on the mineralogy and morphology of
5533.0003141 ureolytic bio-cementation.To Be Submitted.
Biryaltseva T., Lunne T., Kreiter S., and al et 2016. Relative density Burdalski R.J., Ribeiro B.G.O., Gomez M.G., and Gorman-Lewis D.
prediction based on in-situ and laboratory measurements of shear 2022b. Effect of ureolytic rates and treatment additions on the
wave velocity 5th International Conference on Geotechnical and mineralogy and morphology of ureolytic bio-cementation.To Be
Geophysical Site Characterisation, ISC 2016. Proceedings. Vol.1. Submitted.
Australian Geomechanics Society, Australia, 389–394. Burdalski R.J., Ribeiro B.G.O., Gomez M.G., and Gorman-Lewis D.
Blackwell P.G., Rennolls K., and Coutts M.P. 1990. A root anchorage 2022c. Effect of biological factors on the mineralogy and
model for shallowly rooted Sitka spruce. Forestry 63 (1), 73–91. morphology of ureolytic bio-cementation.To Be Submitted.
https://doi.org/10.1093/forestry/63.1.73 Burrall M., DeJong J.T., Martinez A., and Wilson D.W. 2020. Vertical
Bladh K.W., Bideaux R.A., Anthony-Morton E., and Nichols B.G. 2001. pullout tests of orchard trees for bio-inspired engineering of
Handbook of mineralogy. Mineralogical Society of America, anchorage and foundation systems. Bioinspir. Biomim. 16 (1),
Berkeley. 016009. https://doi.org/10.1088/1748-3190/abb414
Bolton M.D. 1986. The strength and dilatancy of sands. Géotechnique 36 Carey T.J., Gavras A., and Kutter B.L. 2020a. Grain size analysis and
(1), 65–78. https://doi.org/10.1680/geot.1986.36.1.65 maximum and minimum dry density testing of ottawa f-65 sand for
Borela R., Frost J.D., Viggiani G., and Anselmucci F. 2021. Earthworm- leap-ucd-2017Kutter, B.L., Manzari, M.T., and Zeghal, M. (Eds.),
inspired robotic locomotion in sand: An experimental study using x- Model Tests and Numerical Simulations of Liquefaction and Lateral
ray tomography. Géotechnique Letters 11 (1), 1–22. Spreading. Springer International Publishing, Cham.
https://doi.org/10.1680/jgele.20.00085 Carey T.J., Stone N., and Kutter B.L. 2020b. Comparison of leap-ucd-
Boulanger R.W. 2003. High overburden stress effects in liquefaction 2017 cpt resultsKutter, B.L., Manzari, M.T., and Zeghal, M. (Eds.),
analyses. Journal of Geotechnical and Geoenvironmental Model Tests and Numerical Simulations of Liquefaction and Lateral
Engineering 129 (12), 1071–1082. Spreading. Springer International Publishing, Cham, 117–129.
https://doi.org/10.1061/(ASCE)1090-0241(2003)129:12(1071) https://doi.org/10.1007/978-3-030-22818-7_6
497
Carle S., and Fogg G. 1996. Transition probability-based indicator Ciurli S., Benini S., Rypniewski W.R., Wilson K.S., Miletti S., and
geostatistics. Mathematical Geology 28, 453–476. Mangani S. 1999. Structural properties of the nickel ions in urease:
https://doi.org/10.1007/BF02083656 novel insights into the catalytic and inhibition mechanisms.
Castanier S., Le Métayer-Levrel G., and Perthuisot J.-P. 1999. Ca- Coordination Chemistry Reviews 190–192, 331–355.
carbonates precipitation and limestone genesis — the https://doi.org/10.1016/S0010-8545(99)00093-4
microbiogeologist point of view. Sedimentary Geology 126 (1–4), Clarke M.A. 2003. SyrupsCaballero, B. (Ed.), Encyclopedia of Food
9–23. https://doi.org/10.1016/S0037-0738(99)00028-7 Sciences and Nutrition (Second Edition). Academic Press, Oxford,
Chen X., and Achal V. 2020. Effect of simulated acid rain on the stability 5711–5717. https://doi.org/10.1016/B0-12-227055-X/01175-5
of calcium carbonate immobilized by microbial carbonate Clough G.W., Iwabuchi J., Rad N.S., and Kuppusamy T. 1989. Influence
precipitation. Journal of Environmental Management 264, 110419. of cementation on liquefaction of sands. Journal of Geotechnical
https://doi.org/10.1016/j.jenvman.2020.110419 Engineering 115 (8), 1102–1117.
Chen Y., Khosravi A., Martinez A., DeJong J., and Wilson D. 2020. https://doi.org/10.1061/(ASCE)0733-9410(1989)115:8(1102)
Analysis of the self-penetration process of a bio-inspired in situ Cohen D., Schwarz M., and Or D. 2011. An analytical fiber bundle model
testing probe Geo-Congress 2020. American Society of Civil for pullout mechanics of root bundles. J. Geophys. Res. 116 (F3),
Engineers, Minneapolis, Minnesota, 224–232. F03010. https://doi.org/10.1029/2010JF001886
https://doi.org/10.1061/9780784482834.025 Collingridge D. 1980. The social control of technology. Frances Pinter,
Cheng L., and Cord-Ruwisch R. 2014. Upscaling effects of soil London, UK.
improvement by microbially induced calcite precipitation by surface Colombani J. 2016. The alkaline dissolution rate of calcite. J. Phys.
percolation. Geomicrobiology Journal 31 (5), 396–406. Chem. Lett. 7 (13), 2376–2380.
https://doi.org/10.1080/01490451.2013.836579 https://doi.org/10.1021/acs.jpclett.6b01055
Cheng L., and Cord-Ruwisch R. 2012. In situ soil cementation with Cortes D.D., and John S. 2018. Earthworm-inspired soil penetration
ureolytic bacteria by surface percolation. Ecological Engineering Proceedings of Biomediated and Bioinspired Geotechnics (B2G)
42, 64–72. https://doi.org/10.1016/j.ecoleng.2012.01.013 Conference. Atlanta, GA, 7.
Cheng L., Cord-Ruwisch R., and Shahin M.A. 2013. Cementation of sand Cubillas P., Köhler S., Prieto M., Chaïrat C., and Oelkers E.H. 2005.
soil by microbially induced calcite precipitation at various degrees Experimental determination of the dissolution rates of calcite,
of saturation. Can. Geotech. J. 50 (1), 81–90. aragonite, and bivalves. Chemical Geology 216 (1–2), 59–77.
https://doi.org/10.1139/cgj-2012-0023 https://doi.org/10.1016/j.chemgeo.2004.11.009
Cheng L., Shahin M., Cord-Ruwisch R., Addis M., Hartanto T., and Elms Cui M.-J., Teng A., Chu J., and Cao B. 2022. A quantitative, high-
C. 2014. Soil stabilisation by microbial-induced calcite precipitation throughput urease activity assay for comparison and rapid screening
(MICP): Investigation into some physical and environmental of ureolytic bacteria. Environmental Research 208, 112738.
aspects.Soil stabilisation by microbial-induced calcite precipitation https://doi.org/10.1016/j.envres.2022.112738
(micp) Cunningham A.B., Characklis W.G., Abedeen F., and Crawford D. 1991.
Cheng L., and Shahin M.A. 2015. Assessment of different treatment Influence of biofilm accumulation on porous media hydrodynamics.
methods by microbial-induced calcite precipitation for clayey soil Environ. Sci. Technol. 25 (7), 1305–1311.
improvement 7. https://doi.org/10.1021/es00019a013
Cheng L., Shahin M.A., and Cord-Ruwisch R. 2017a. Surface Cuthbert M.O., McMillan L.A., Handley-Sidhu S., Riley Michael.S.,
percolation for soil improvement by biocementation utilizing in situ Tobler D.J., and Phoenix Vernon.R. 2013. A field and modeling
enriched indigenous aerobic and anaerobic ureolytic soil study of fractured rock permeability reduction using microbially
microorganisms. Geomicrobiology Journal 34 (6), 546–556. induced calcite precipitation. Environ. Sci. Technol. 47 (23), 13637–
https://doi.org/10.1080/01490451.2016.1232766 13643. https://doi.org/10.1021/es402601g
Cheng L., Shahin M.A., and Mujah D. 2017b. Influence of key -Seget Z. 2017. Bioaugmentation
environmental conditions on microbially induced cementation for as a strategy for the remediation of pesticide-polluted soil: A review.
soil stabilization. Journal of Geotechnical and Geoenvironmental Chemosphere 172, 52–71.
Engineering 143 (1), 04016083. https://doi.org/10.1016/j.chemosphere.2016.12.129
https://doi.org/10.1061/(ASCE)GT.1943-5606.0001586 Dadda A., Geindreau C., Emeriault F., du Roscoat S.R., Garandet A.,
Cheshomi A., and Mansouri S. 2019. Study the grain size and infiltration Sapin L., and Filet A.E. 2017. Characterization of microstructural
method effects for sand soil improvement using the microbial and physical properties changes in biocemented sand using 3D x-ray
method. Geomicrobiology Journal 37 (4), 355–365. microtomography. Acta Geotech. 12 (5), 955–970.
https://doi.org/10.1080/01490451.2019.1705437 https://doi.org/10.1007/s11440-017-0578-5
Choi S.-G., Chang I., Lee M., Lee J.-H., Han J.-T., and Kwon T.-H. 2020. Dafalias Y.F., and Manzari M.T. 2004. Simple plasticity sand model
Review on geotechnical engineering properties of sands treated by accounting for fabric change effects. Journal of Engineering
microbially induced calcium carbonate precipitation (MICP) and Mechanics 130 (6), 622–634. https://doi.org/10.1061/(ASCE)0733-
biopolymers. Construction and Building Materials 246, 118415. 9399(2004)130:6(622)
https://doi.org/10.1016/j.conbuildmat.2020.118415 Darby K.M., Hernandez G.L., DeJong J.T., Boulanger R.W., Gomez
Choi S.-G., Park S.-S., Wu S., and Chu J. 2017. Methods for calcium M.G., and Wilson D.W. 2019. Centrifuge model testing of
carbonate content measurement of biocemented soils. J. Mater. Civ. liquefaction mitigation via microbially induced calcite precipitation.
Eng. 29 (11), 06017015. https://doi.org/10.1061/(ASCE)MT.1943- J. Geotech. Geoenviron. Eng. 145 (10), 04019084.
5533.0002064 https://doi.org/10.1061/(ASCE)GT.1943-5606.0002122
Choi S.-G., Wu S., and Chu J. 2016. Biocementation for sand using an Dashti S., Bray J.D., Pestana J.M., Riemer M., and Wilson D. 2010a.
eggshell as calcium source. J. Geotech. Geoenviron. Eng. 142 (10), Mechanisms of seismically induced settlement of buildings with
06016010. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001534 shallow foundations on liquefiable soil. Journal of Geotechnical and
Chong S.-H. 2019. Development of constitutive model for simulation of Geoenvironmental Engineering 136 (1), 151–164.
cemented soil. Geotech Geol Eng 37 (5), 4635–4641. https://doi.org/10.1061/(ASCE)GT.1943-5606.0000179
https://doi.org/10.1007/s10706-019-00903-3 Dashti S., Bray J.D., Pestana J.M., Riemer M., and Wilson D. 2010b.
Chu J., Ivanov V., He J., Naeimi M., Li B., and Stabnikov V. 2011. Centrifuge testing to evaluate and mitigate liquefaction-induced
Development of microbial geotechnology in singapore Geo- building settlement mechanisms. J. Geotech. Geoenviron. Eng. 136
Frontiers 2011. American Society of Civil Engineers, Dallas, Texas, (7), 918–929. https://doi.org/10.1061/(ASCE)GT.1943-
United States, 4070–4078. https://doi.org/10.1061/41165(397)416 5606.0000306
Chu J., Stabnikov V., and Ivanov V. 2012. Microbially induced calcium Davis C.A., Atekwana Estella, Atekwana Eliot, Slater L.D., Rossbach S.,
carbonate precipitation on surface or in the bulk of soil. and Mormile M.R. 2006. Microbial growth and biofilm formation in
Geomicrobiology Journal 29 (6), 544–549. geologic media is detected with complex conductivity
https://doi.org/10.1080/01490451.2011.592929 measurements: microbial growth and geophysics. Geophys. Res.
Ciantia M.O., and di Prisco C. 2016. Extension of plasticity theory to Lett. 33 (18), n/a-n/a. https://doi.org/10.1029/2006GL027312
debonding, grain dissolution, and chemical damage of calcarenites. de Bono J., McDowell G., and Wanatowski D. 2015. Investigating the
Int. J. Numer. Anal. Meth. Geomech. 40 (3), 315–343. micro mechanics of cemented sand using DEM. Int. J. Numer. Anal.
https://doi.org/10.1002/nag.2397 Meth. Geomech. 39 (6), 655–675. https://doi.org/10.1002/nag.2340
498
De Muynck W., Verbeken K., De Belie N., and Verstraete W. 2013. Duggan A.R., Goggins J., Clifford E., and McCabe B.A. 2017. The use
Influence of temperature on the effectiveness of a biogenic carbonate of carbonation depth techniques on stabilized peat. Geotech. Test. J.
surface treatment for limestone conservation. Appl Microbiol 40 (6), 20160223. https://doi.org/10.1520/GTJ20160223
Biotechnol 97 (3), 1335–1347. https://doi.org/10.1007/s00253-012- Dupas J.-M., and Pecker A. 1979. Static and dynamic properties of sand-
3997-0 cement. Journal of the Geotechnical Engineering Division 105
DeJong J.T. 2020. Biogeotechnics: Bio-mediated processes and bio- (GT3), 18. https://doi.org/10.1061/AJGEB6.0000778
inspired ideas for geotechnical engineering innovation. Dupuy L.X., Fourcaud T., Lac P., and Stokes A. 2007. A generic 3D
DeJong J.T., Burrall M., Wilson D.W., and Frost J.D. 2017. A bio- finite element model of tree anchorage integrating soil mechanics
inspired perspective for geotechnical engineering innovation and real root system architecture. American Journal of Botany 94
Geotechnical Frontiers 2017. American Society of Civil Engineers, (9), 1506–1514. https://doi.org/10.3732/ajb.94.9.1506
Orlando, Florida, 862–870. Dyson A.S., and Rognon P.G. 2014. Pull-out capacity of tree root
https://doi.org/10.1061/9780784480472.092 inspired anchors in shallow granular soils. Géotechnique Letters 4
DeJong J.T., Fritzges M.B., and Nüsslein K. 2006. Microbially induced (4), 301–305. https://doi.org/10.1680/geolett.14.00061
cementation to control sand response to undrained shear. J. Geotech. El Kortbawi M., Ziotopoulou K., Gomez M.G., and Lee M. 2022a.
Geoenviron. Eng. 132 (11), 1381–1392. Mechanical behavior of bio-cemented sands: State review of
https://doi.org/10.1061/(ASCE)1090-0241(2006)132:11(1381) experimental and numerical developments.To Be Submitted.
DeJong J.T., and Kavazanjian E. 2019. Bio-mediated and bio-inspired El Kortbawi M., Ziotopoulou K., and Boulanger R.W. 2022b. A
geotechnicsLu, N., and Mitchell, J.K. (Eds.), Geotechnical constitutive model for bio-cemented sands for plane-strain
Fundamentals for Addressing New World Challenges, Springer earthquake engineering applications.To Be Submitted.
Series in Geomechanics and Geoengineering. Springer International El Mountassir G., Minto J.M., van Paassen L.A., Salifu E., and Lunn R.J.
Publishing, Cham, 193–207. https://doi.org/10.1007/978-3-030- 2018. Applications of microbial processes in geotechnical
06249-1_7 engineering Advances in Applied Microbiology. Elsevier, 39–91.
DeJong J.T., Krage C.P., Boulanger R.W., and DeGroot D.J. 2021. https://doi.org/10.1016/bs.aambs.2018.05.001
Optimization of CPT soundings to reduce uncertainty in Ellis D.E., Lutz E.J., Odom J.M., Buchanan R.J., Bartlett C.L., Lee M.D.,
interpretation of subsurface stratigraphy 6th International Harkness M.R., and DeWeerd K.A. 2000. Bioaugmentation for
Conference on Geotechnical and Geophysical Site Characterization. accelerated in situ anaerobic bioremediation. Environ. Sci. Technol.
Budapest, Hungary, 20. 34 (11), 2254–2260. https://doi.org/10.1021/es990638e
DeJong J.T., Martinez B.C., Ginn T.R., Hunt C., Major D., and Tanyu B. Eslaamizaad S., and Robertson P. 1996. Seismic cone penetration test to
2014. Development of a scaled repeated five-spot treatment model identify cemented sands In Proceedings of the 49th Canadian
for examining microbial induced calcite precipitation feasibility in Geotechnical Conference. St. John’s, NL, 23–25.
field applications. Geotech. Test. J. 37 (3), 20130089. Esnault Filet A., Gadret J.-P., Loygue M., and Borel S. 2012. Biocalcis
https://doi.org/10.1520/GTJ20130089 and its applications for the consolidation of sands 1767–1780.
DeJong J.T., Martinez B.C., Mortensen B.M., Nelson D.C., Waller J.T., https://doi.org/10.1061/9780784412350.0152
Weil M.H., Ginn T.R., Weathers T., Barkouki T., Fujita Y., Redden Esnault Filet A., Gutjahr I., Garandet A., Viglino A., Béguin R., Monier
G., Hunt C., Major D., and Tanyu B. 2009. Upscaling of bio- J.-M., Oxarango L., Emeriault F., and Perthuisot S.C. 2020.
mediated soil improvement. Proceedings of the 17th International BOREAL, Bio-reinforcement of embankments by biocalcification.
Conference on Soil Mechanics and Geotechnical Engineering E3S Web Conf. 195, 05001.
(Volumes 1, 2, 3 and 4) 2300–2303. https://doi.org/10.3233/978-1- https://doi.org/10.1051/e3sconf/202019505001
60750-031-5-2300 European Union 2020. Directive (EU) 2020/2184 of the European
DeJong J.T., Mortensen B.M., Martinez B.C., and Nelson D.C. 2010. parliament and of the Council of 16 December 2020 On the quality
Bio-mediated soil improvement. Ecological Engineering 36 (2), of water intended for human consumption (No. 435). Off. J. Eur.
197–210. https://doi.org/10.1016/j.ecoleng.2008.12.029 Union.
DeJong J.T., Soga K., Kavazanjian E., Burns S., Van Paassen L.A., Al Evans T., Khoubani A., and Montoya B. 2015. Simulating mechanical
Qabany A., Aydilek A., Bang S.S., Burbank M., Caslake L.F., Chen response in bio-cemented sandsOka, F., Murakami, A., Uzuoka, R.,
C.Y., Cheng X., Chu J., Ciurli S., Esnault-Filet A., Fauriel S., and Kimoto, S. (Eds.), Computer Methods and Recent Advances in
Hamdan N., Hata T., Inagaki Y., Jefferis S., Kuo M., Laloui L., Geomechanics. CRC Press, 1569–1574.
Larrahondo J., Manning D.A.C., Martinez B., Montoya B.M., https://doi.org/10.1201/b17435-277
Nelson D.C., Palomino A., Renforth P., Santamarina J.C., Seagren Ezzat S.M., and Ewida A.Y.I. 2021. Smart soil grouting using innovative
E.A., Tanyu B., Tsesarsky M., and Weaver T. 2013. Biogeochemical J Appl Microbiol
processes and geotechnical applications: progress, opportunities and 131 (5), 2294–2307. https://doi.org/10.1111/jam.15117
challenges. Géotechnique 63 (4), 287–301. Fauriel S., and Laloui L. 2012. A bio-chemo-hydro-mechanical model
https://doi.org/10.1680/geot.SIP13.P.017 for microbially induced calcite precipitation in soils. Computers and
Delfosse-Ribay E., Djeran-Maigre I., Cabrillac R., and Gouvenot D. Geotechnics 46, 104–120.
2004. Shear modulus and damping ratio of grouted sand. Soil https://doi.org/10.1016/j.compgeo.2012.05.017
Dynamics and Earthquake Engineering 24 (6), 461–471. Feng K. 2015. Constitutive response of microbial induced calcite
https://doi.org/10.1016/j.soildyn.2004.02.004 precipitation cemented sands (PhD Dissertation). North Carolina
Dhami N.K., Alsubhi W.R., Watkin E., and Mukherjee A. 2017. State University, Raleigh, NC.
Bacterial community dynamics and biocement formation during Feng K., and Montoya B.M. 2015. Drained shear strength of MICP sand
stimulation and augmentation: implications for soil consolidation. at varying cementation levels IFCEE 2015. American Society of
Front. Microbiol. 8, 1267. Civil Engineers, San Antonio, Texas, 2242–2251.
https://doi.org/10.3389/fmicb.2017.01267 https://doi.org/10.1061/9780784479087.208
Dickson A.G., and Goyet C. 1994. Handbook of methods for the analysis Feng K., Montoya B.M., and Evans T.M. 2017. Discrete element method
of the various parameters of the carbon dioxide system in sea water. simulations of bio-cemented sands. Computers and Geotechnics 85,
version 2 (No. ORNL/CDIAC-74). Oak Ridge National Lab., TN 139–150. https://doi.org/10.1016/j.compgeo.2016.12.028
(United States). https://doi.org/10.2172/10107773 Fenton G.A., and Vanmarcke E.H. 1990. Simulation of random fields via
Dikshit R., Jain A., Dey A., and Kumar A. 2020. Microbially induced local average subdivision. Journal of Engineering Mechanics 116
calcite precipitation using Bacillus velezensis with guar gum. PLoS (8), 1733–1749. https://doi.org/10.1061/(ASCE)0733-
ONE 15 (8), e0236745. 9399(1990)116:8(1733)
https://doi.org/10.1371/journal.pone.0236745 Ferris F.G., Phoenix V., Fujita Y., and Smith R.W. 2004. Kinetics of
Do J., Montoya B.M., and Gabr M.A. 2021. Scour mitigation and calcite precipitation induced by ureolytic bacteria at 10 to 20°C in
erodibility improvement using microbially induced carbonate artificial groundwater. Geochimica et Cosmochimica Acta 68 (8),
precipitation. Geotech. Test. J. 44 (5), 20190478. 1701–1710. https://doi.org/10.1016/S0016-7037(03)00503-9
https://doi.org/10.1520/GTJ20190478 Ferris F.G., Stehmeier L.G., Kantzas A., and Mourits F.M. 1997.
Dong Y., and Fatahi B. 2020. Discrete element simulation of cavity Bacteriogenic mineral plugging. Journal of Canadian Petroleum
expansion in lightly cemented sands considering cementation Technology 36 (09). https://doi.org/10.2118/97-09-07
degradation. Computers and Geotechnics 124, 103628. Fidaleo M., and Lavecchia R. 2003. Kinetic study of enzymatic urea
https://doi.org/10.1016/j.compgeo.2020.103628 hydrolysis in the ph range 4–9. Chem. Biochem. Eng. Q. 8.
499
Finnerty W.R., and Singer M.E. 1983. Microbial enhancement of oil Ghasemi P., and Montoya B.M. 2022. Field implementation of
recovery. Nat Biotechnol 1 (1), 47–54. microbially induced calcium carbonate precipitation for surface
https://doi.org/10.1038/nbt0383-47 erosion reduction of a coastal plain sandy slope.In Review.
Flores Orozco A., Kemna A., Oberdörster C., Zschornack L., Leven C., Ghasemi P., and Montoya B.M. 2020. Field application of the
Dietrich P., and Weiss H. 2012. Delineation of subsurface microbially induced calcium carbonate precipitation on a coastal
hydrocarbon contamination at a former hydrogenation plant using sandy slope Geo-Congress 2020. American Society of Civil
spectral induced polarization imaging. Journal of Contaminant Engineers, Minneapolis, Minnesota, 141–149.
Hydrology 136–137, 131–144. https://doi.org/10.1061/9780784482834.016
https://doi.org/10.1016/j.jconhyd.2012.06.001 Giri R.K., and Reddy K.R. 2014. LCA and sustainability assessment for
Flores Orozco A., Williams K.H., Long P.E., Hubbard S.S., and Kemna selecting deep foundation system for high-rise buildings. American
A. 2011. Using complex resistivity imaging to infer biogeochemical Society of Civil Engineers, Reston, VA, 621–630.
processes associated with bioremediation of an uranium- https://doi.org/10.1061/9780784478745.057
contaminated aquifer. J. Geophys. Res. 116 (G3), G03001. Gollapudi U.K., Knutson C.L., and Bang S.S. 1995. A new method for
https://doi.org/10.1029/2010JG001591 controlling leaching through permeable channels. Chemosphere 30
Fu Z., Chen S., Zhong Q., and Ji E. 2022. A damage hypoplasticity (4), 695–705. https://doi.org/10.1016/0045-6535(94) 00435-W.
constitutive model for cemented sand and gravel materials. Acta Gomez M.G. 2017. Up-scaling of bio-cementation soil improvement
Geotechnica 17. https://doi.org/10.1007/s11440-021-01206-9 using native soil microorganisms (Ph.D.). ProQuest Dissertations
Fujita Y., Ferris F.G., Lawson R.D., Colwell F.S., and Smith R.W. 2000. and Theses. University of California, Davis, Davis, CA.
Subscribed content calcium carbonate precipitation by ureolytic Gomez M.G., Anderson C.M., DeJong J.T., Nelson D.C., and Lau X.H.
subsurface bacteria. Geomicrobiology Journal 17 (4), 305–318. 2014. Stimulating in situ soil bacteria for bio-cementation of sands
https://doi.org/10.1080/782198884 Geo-Congress 2014 Technical Papers, Proceedings. American
Fujita Y., Taylor J.L., Gresham T.L.T., Delwiche M.E., Colwell F.S., Society of Civil Engineers, Atlanta, Georgia, 1674–1682.
McLing T.L., Petzke L.M., and Smith R.W. 2008. Stimulation of https://doi.org/10.1061/9780784413272.164
microbial urea hydrolysis in groundwater to enhance calcite Gomez M.G., Anderson C.M., Graddy C.M.R., DeJong J.T., Nelson
precipitation. Environ. Sci. Technol. 42 (8), 3025–3032. D.C., and Ginn T.R. 2017. Large-scale comparison of
https://doi.org/10.1021/es702643g bioaugmentation and biostimulation approaches for biocementation
Gai X., and Sánchez M. 2018. An elastoplastic mechanical constitutive of sands. J. Geotech. Geoenviron. Eng. 143 (5), 04016124.
model for microbially mediated cemented soils. Acta Geotech. 14 https://doi.org/10.1061/(ASCE)GT.1943-5606.0001640
(3), 709–726. https://doi.org/10.1007/s11440-018-0721-y Gomez M.G., and DeJong J.T. 2017. Engineering properties of bio-
Gajo A., Cecinato F., and Hueckel T. 2019. Chemo-mechanical modeling cementation improved sandy soils Grouting 2017. American Society
of artificially and naturally bonded soils. Geomechanics for Energy of Civil Engineers, Honolulu, Hawaii, 23–33.
and the Environment 18, 13–29. https://doi.org/10.1061/9780784480793.003
https://doi.org/10.1016/j.gete.2018.11.005 Gomez M.G., DeJong J.T., and Anderson C.M. 2018a. Effect of bio-
Gajo A., Cecinato F., and Hueckel T. 2015. A micro-scale inspired cementation on geophysical and cone penetration measurements in
chemo-mechanical model of bonded geomaterials. International sands. Can. Geotech. J. 55 (11), 1632–1646.
Journal of Rock Mechanics and Mining Sciences 80, 425–438. https://doi.org/10.1139/cgj-2017-0253
https://doi.org/10.1016/j.ijrmms.2015.10.001 Gomez M.G., DeJong J.T., Anderson C.M., Nelson D.C., and Graddy
Gal J.Y., Bollinger J.C., Tolosa H., and Gache N. 1996. Calcium C.M. 2016. Large-scale bio-cementation improvement of sands
carbonate solubility: a reappraisal of scale formation and inhibition. Geotechnical and Structural Engineering Congress 2016. American
Talanta 43 (9), 1497–1509. https://doi.org/10.1016/0039- Society of Civil Engineers, Phoenix, Arizona, 941–949.
9140(96)01925-X https://doi.org/10.1061/9780784479742.079
Ganendra G., De Muynck W., Ho A., Arvaniti E.C., Hosseinkhani B., Gomez M.G., DeJong J.T., Martinez B.C., Hunt C.E., deVlaming L.A.,
Ramos J.A., Rahier H., and Boon N. 2014. Formate oxidation-driven Major D.W., and Dworatzek S.M. 2013. Bio-mediated Soil
calcium carbonate precipitation by Methylocystis parvus OBBP. Improvement Field Study to Stabilize Mine Sands GeoMontreal
Appl Environ Microbiol 80 (15), 4659–4667. 2013. Montreal, Canada.
https://doi.org/10.1128/AEM.01349-14 Gomez M.G., Graddy C.M.R., DeJong J.T., and Nelson D.C. 2019.
Gao Z., and Zhao J. 2012. Constitutive modeling of artificially cemented Biogeochemical changes during bio-cementation mediated by
sand by considering fabric anisotropy. Computers and Geotechnics stimulated and augmented ureolytic microorganisms. Sci Rep 9 (1),
41, 57–69. https://doi.org/10.1016/j.compgeo.2011.10.007 11517. https://doi.org/10.1038/s41598-019-47973-0
Gartner 2018. Understanding Gartner’s hype cycles [WWW Document]. Gomez M.G., Graddy C.M.R., DeJong J.T., Nelson D.C., and Tsesarsky
URL www.gartner.com M. 2018b. Stimulation of native microorganisms for biocementation
Gat D., Ronen Z., and Tsesarsky M. 2017. Long-term sustainability of in samples recovered from field-scale treatment depths. J. Geotech.
microbial-induced CaCO3 precipitation in aqueous media. Geoenviron. Eng. 144 (1), 04017098.
Chemosphere 184, 524–531. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001804
https://doi.org/10.1016/j.chemosphere.2017.06.015 Gomez M.G., Martinez B.C., DeJong J.T., Hunt C.E., deVlaming L.A.,
Gat D., Ronen Z., and Tsesarsky M. 2016. Soil bacteria population Major D.W., and Dworatzek S.M. 2015. Field-scale bio-cementation
dynamics following stimulation for ureolytic microbial-induced tests to improve sands. Proceedings of the Institution of Civil
CaCO3 precipitation. Environ. Sci. Technol. 50 (2), 616–624. Engineers - Ground Improvement 168 (3), 206–216.
https://doi.org/10.1021/acs.est.5b04033 https://doi.org/10.1680/grim.13.00052
Gat D., Tsesarsky M., and Shamir D. 2011. Ureolytic calcium carbonate Gowthaman S., Iki T., Nakashima K., Ebina K., and Kawasaki S. 2019.
precipitation in the presence of non-ureolytic competing bacteria Feasibility study for slope soil stabilization by microbial induced
Geo-Frontiers 2011. American Society of Civil Engineers, Dallas, carbonate precipitation (MICP) using indigenous bacteria isolated
Texas, United States, 3966–3974. from cold subarctic region. SN Appl. Sci. 1 (11), 1480.
https://doi.org/10.1061/41165(397)405 https://doi.org/10.1007/s42452-019-1508-y
Gehlen M., Bassinot F.C., Chou L., and McCorkle D. 2005. Reassessing Graddy C.M.R., Gomez M.G., DeJong J.T., and Nelson D.C. 2021.
the dissolution of marine carbonates: II. Reaction kinetics. Deep Sea Native bacterial community convergence in augmented and
Research Part I: Oceanographic Research Papers 52 (8), 1461– stimulated ureolytic MICP biocementation. Environ. Sci. Technol.
1476. https://doi.org/10.1016/j.dsr.2005.03.011 55 (15), 10784–10793. https://doi.org/10.1021/acs.est.1c01520
Gens A., and Nova R. 1993. Conceptual bases for a constitutive model Graddy C.M.R., Gomez M.G., Kline L.M., Morrill S.R., DeJong J.T., and
for bonded soils and weak rocks Proc. of Geotech. Eng. of Hard Nelson D.C. 2018. Diversity of Sporosarcina-like bacterial strains
Soils-Soft Rocks. Athens, Greece, 485–494. obtained from meter-scale augmented and stimulated
Gercek H. 2007. Poisson’s ratio values for rocks. International Journal biocementation experiments. Environ. Sci. Technol. 52 (7), 3997–
of Rock Mechanics and Mining Sciences 44 (1), 1–13. 4005. https://doi.org/10.1021/acs.est.7b04271
https://doi.org/10.1016/j.ijrmms.2006.04.011 Gray J., and Lissmann H.W. 1950. The kinetics of locomotion of the
Gerhardt P. 1994. Methods for general and molecular bacteriology. grass-snake. J Exp Biol 26 (4), 24.
American Society for Microbiology, Washington, D.C. https://doi.org/10.1242/jeb.26.4.354
500
Griffiths D.V., Fenton G.A., and Manoharan N. 2002. Bearing capacity Ingraham J.L., Maaløe O., Neidhardt F.C., and others 1983. Growth of
of rough rigid strip footing on cohesive soil: Probabilistic study. the bacterial cell. Sinauer Associates.
Journal of Geotechnical and Geoenvironmental Engineering 128 Islam M.T., Chittoori B.C.S., and Burbank M. 2020. Evaluating the
(9). https://doi.org/10.1061/(ASCE)1090-0241(2002)128:9(743) applicability of biostimulated calcium carbonate precipitation to
Gui R., Pan Y., Ding D., Liu Y., and Zhang Z. 2018. Experimental study stabilize clayey soils. J. Mater. Civ. Eng. 32 (3), 04019369.
on the fine-grained uranium tailings reinforced by MICP. Advances https://doi.org/10.1061/(ASCE)MT.1943-5533.0003036
in Civil Engineering 2018, 1–10. Itasca 2019. FLAC (version 8.1) – fast lagrangian analysis of continua.
https://doi.org/10.1155/2018/2928985 Itasca Consulting Group, Inc, Minneapolis, MN.
Guinée J. 2016. Life cycle sustainability assessment: what is it and what Itasca 2016. FLAC – fast lagrangian analysis of continua, version 8.0.
are its challenges?Clift, R., and Druckman, A. (Eds.), Taking Stock Itasca Consulting Group, Inc, Minneapolis, MN.
of Industrial Ecology. Springer, Cham, 45–68. Iwasaki T., Tatsuoka F., Tokida K., and Yasuda S. 1978. A practical
https://doi.org/10.1007/978-3-319-20571-7_3 method for assessing soil liquefaction potential based on case studies
Haeri S.M., Hosseini S.M., Toll D.G., and Yasrebi S.S. 2005. The at various sites in Japan Proc. Second Int. Conf. Microzonation Safer
behaviour of an artificially cemented sandy gravel. Geotech Geol Construction Research Application, 1978. Alexandria, VA:, 885–
Eng 23 (5), 537–560. https://doi.org/10.1007/s10706-004-5110-7 896.
Hall C.A., Hernandez G., Darby K.M., van Paassen L., Jr. E.K., DeJong Jahns T. 1996. Ammonium/urea-dependent generation of a proton
J., and Wilson D. 2018. Centrifuge model testing of liquefaction electrochemical potential and synthesis of ATP in Bacillus pasteurii.
mitigation via denitrification-induced desaturation 117–126. Journal of Bacteriology 178 (2), 403–409.
https://doi.org/10.1061/9780784481455.011 https://doi.org/10.1128/jb.178.2.403-409.1996
Ham S.-Mi., Kwon T.-H., and DeJong J.T. 2022a. Mechanical testing of Jain S., and Arnepalli D.N. 2019. Biochemically induced carbonate
single particle contacts.To Be Submitted. precipitation in aerobic and anaerobic environments by Sporosarcina
Ham S.-Mi., Martinez A., Han G., and Kwon T.-H. 2022b. Grain-scale pasteurii. Geomicrobiology Journal 36 (5), 443–451.
tensile and shear strengths of glass beads cemented by microbially https://doi.org/10.1080/01490451.2019.1569180
induced calcite precipitation.In Review. Jaksa M.B., Goldsworthy J.S., Fenton G.A., Kaggwa W.S., Griffiths
Hamdan N., Kavazanjian E., Rittmann B.E., and Karatas I. 2011. D.V., Kuo Y.L., and Poulos H.G. 2005. Towards reliable and
Carbonate mineral precipitation for soil improvement through effective site investigations 109–121.
microbial denitrification Geo-Frontiers 2011: Advances in https://doi.org/10.1680/geot.2005.55.2.109
Geotechnical Engineering. Dallas, Texas, 10. Jefferies M.G. 1993. Nor-Sand: a simple critical state model for sand.
Hardin B.O., and Black W.L. 1968. Vibration modulus of normally Géotechnique 43 (1), 91–103.
consolidated clay. J. Soil Mech. and Found. Div. 94 (2), 353–369. https://doi.org/10.1680/geot.1993.43.1.91
https://doi.org/10.1061/JSFEAQ.0001100 Jefferis S.A. 2008. Moving towards sustainability in geotechnical
Harkes M.P., van Paassen L.A., Booster J.L., Whiffin V.S., and van engineering. ASCE, New Orleans, LA, 844–851.
Loosdrecht M.C.M. 2010. Fixation and distribution of bacterial https://doi.org/10.1061/40971(310)105
activity in sand to induce carbonate precipitation for ground Jenneman G.E., Knapp R.M., McInerney M.J., Menzie D.E., and Revus
reinforcement. Ecological Engineering, Special Issue: BioGeoCivil D.E. 1984. Experimental studies of in-situ microbial enhanced oil
Engineering 36 (2), 112–117. recovery. Society of Petroleum Engineers Journal 24 (01), 33–37.
https://doi.org/10.1016/j.ecoleng.2009.01.004 https://doi.org/10.2118/10789-PA
Hausler E.A. 2002. Influence of ground improvement on settlement and Jiang M., Zhang F., and Thornton C. 2015. A simple three-dimensional
liquefaction: A study based on field case history evidence and distinct element modeling of the mechanical behavior of bonded
dynamic geotechnical centrifuge tests (PhD Dissertation). Ph.D. sands. Int. J. Numer. Anal. Meth. Geomech., 39 (16), 1791–1820.
Thesis. University of California, Berkeley, Berkeley. Jiang N.-J., and Soga K. 2019. Erosional behavior of gravel-sand
Hayashi K., Okamura M., and Yasuhara H. 2010. Identification of mixtures stabilized by microbially induced calcite precipitation
liquefaction resistances in sand improved by calcium carbonate (MICP). Soils and Foundations 59 (3), 699–709.
precipitated. Jiban Kogaku Janaru (Japanese Geotechnical Journal) https://doi.org/10.1016/j.sandf.2019.02.003
5 (2), 391–400. https://doi.org/10.3208/jgs.5.391 Jiang N.-J., and Soga K. 2017. The applicability of microbially induced
He J., and Chu J. 2014. Undrained responses of microbially desaturated calcite precipitation (MICP) for internal erosion control in gravel–
sand under monotonic loading. J. Geotech. Geoenviron. Eng. 140 sand mixtures. Géotechnique 67 (1), 42–55.
(5), 04014003. https://doi.org/10.1061/(ASCE)GT.1943- https://doi.org/10.1680/jgeot.15.P.182
5606.0001082 Jiang N.-J., Soga K., and Kuo M. 2017. Microbially induced carbonate
Hermansson M. 1999. The DLVO theory in microbial adhesion. Colloids precipitation for seepage-induced internal erosion control in sand–
and Surfaces B: Biointerfaces 14 (1), 105–119. clay mixtures. J. Geotech. Geoenviron. Eng. 143 (3), 04016100.
https://doi.org/10.1016/S0927-7765(99)00029-6 https://doi.org/10.1061/(ASCE)GT.1943-5606.0001559
Hernandez G.L. 2018. Centrifuge model testing of liquefaction Jiang N.-J., Tang C.-S., Yin L.-Y., Xie Y.-H., and Shi B. 2019.
mitigation via microbially induced calcite precipitation (Master’s Applicability of microbial calcification method for sandy-slope
Thesis). University of California, Davis, Davis, CA. surface erosion control. Journal of Materials in Civil Engineering 31
Hirai H., Takahashi M., and Yamada M. 1989. An elastic-plastic (11), 04019250. https://doi.org/10.1061/(ASCE)MT.1943-
constitutive model for the behavior of improved sandy soils. Soils 5533.0002897
and Foundations 29 (2), 69–84. Kaltwasser H., and Schlegel H.G. 1966. NADH-dependent coupled
https://doi.org/10.3208/sandf1972.29.2_69 enzyme assay for urease and other ammonia-producing systems.
Hobbie J.E., Daley R.J., and Jasper S. 1977. Use of nuclepore filters for Analytical Biochemistry 16 (1), 132–138.
counting bacteria by fluorescence microscopy. Appl. Environ. https://doi.org/10.1016/0003-2697(66)90088-1
Microbiol. 33 (5), 1225–1228. Kamai R., and Boulanger R. 2010. Characterizing localization processes
Huang S., Tang Y., Bagheri H., Li D., Ardente A., Aukes D., Marvi H., during liquefaction using inverse analyses of instrumentation arrays
and Tao J. (Julian) 2020. Effects of friction anisotropy on upward 219–238. https://doi.org/10.1201/b10826-23
burrowing behavior of soft robots in granular materials. Advanced Kang C.-H., Kwon Y.-J., and So J.-S. 2016. Soil bioconsolidation
Intelligent Systems 2 (6), 1900183. through microbially induced calcite precipitation by Lysinibacillus
https://doi.org/10.1002/aisy.201900183 sphaericus WJ-8. Geomicrobiology Journal 33 (6), 473–478.
Huang S., and Tao J. 2020. Modeling clam-inspired burrowing in dry https://doi.org/10.1080/01490451.2015.1053581
sand using cavity expansion theory and DEM. Acta Geotech. 15 (8), Kantzas A., Stehmeier L., Marentette D.F., Ferris F.G., Jha K.N., and
2305–2326. https://doi.org/10.1007/s11440-020-00918-8 Maurits F.M. 1992. A novel method of sand consolidation through
Huang S., and Tao J. 2018. Modeling of the burrowing mechanism by bacteriogenic mineral plugging Annual Technical Meeting.
razor clam: Role of penetration kinematics IFCEE 2018. American Petroleum Society of Canada, Calgary, Alberta.
Society of Civil Engineers, Orlando, Florida, 547–556. https://doi.org/10.2118/92-46
https://doi.org/10.1061/9780784481585.053 Karatas I., Kavazanjian E.Jr., and Rittmann B.E. 2008. Microbially
Idriss I.M., and Boulanger R.W. 2008. Soil liquefaction during induced precipitation of calcite using Pseudomonas denitrificans.
earthquakes. Earthquake Engineering Research Institute (EERI), Proc. 1st Int. Conf. Biogeotech. Engng, Delft. Delft, Netherlands.
Oakland, California, United States.
501
Kavazanjian E., and Hamdan N. 2015. Enzyme induced carbonate Larson A.D., and Kallio R.E. 1954. Purification and properties of
precipitation (EICP) columns for ground improvement IFCEE 2015. bacterial urease. J Bacteriol 68 (1), 67–73.
American Society of Civil Engineers, San Antonio, Texas, 2252– https://doi.org/10.1128/jb.68.1.67-73.1954
2261. https://doi.org/10.1061/9780784479087.209 Lauchnor E.G., Topp D.M., Parker A.E., and Gerlach R. 2015. Whole
Kavvadas M., and Amorosi A. 2000. A constitutive model for structured cell kinetics of ureolysis by Sporosarcina pasteurii. Journal of
soils. Géotechnique 50 (3), 263–273. Applied Microbiology 118 (6), 1321–1332.
https://doi.org/10.1680/geot.2000.50.3.263 https://doi.org/10.1111/jam.12804
Keeney D.R., and Nelson D.W. 1983. Nitrogen—Inorganic Forms Lawrence A.J. 1971. Conductimetric enzyme assays. European Journal
Methods of Soil Analysis. John Wiley & Sons, Ltd, 643–698. of Biochemistry 18 (2), 221–225. https://doi.org/10.1111/j.1432-
https://doi.org/10.2134/agronmonogr9.2.2ed.c33 1033.1971.tb01234.x
Kendall A., Raymond A.J., Tipton J., and DeJong J.T. 2018. Review of Lee J.-S., Fernandez A.L., and Santamarina J.C. 2005. S-wave velocity
life-cycle-based environmental assessments of geotechnical tomography: Small-scale laboratory application. Geotech. Test. J. 28
systems. Proceedings of the Institution of Civil Engineers - (4), 12638. https://doi.org/10.1520/GTJ12638
Engineering Sustainability 171 (2), 57–67. Lee J.-S., and Santamarina J.C. 2005a. P-wave reflection imaging.
https://doi.org/10.1680/jensu.16.00073 Geotechnical Testing Journal 28 (2).
Khodadadi Tirkolaei H., Javadi N., Krishnan V., Hamdan N., and Lee J.-S., and Santamarina J.C. 2005b. Bender elements: Performance
Kavazanjian E. 2020. Crude urease extract for biocementation. and signal interpretation. J. Geotech. Geoenviron. Eng. 131 (9),
Journal of Materials in Civil Engineering 32 (12), 04020374. 1063–1070. https://doi.org/10.1061/(ASCE)1090-
https://doi.org/10.1061/(ASCE)MT.1943-5533.0003466 0241(2005)131:9(1063)
Khosravi A., Martinez A., DeJong J., and Wilson D. 2018. Discrete Lee K., Chan D., and Lam K. 2004. Constitutive model for cement
element simulations of bio-inspired self-burrowing probes in sands treated clay in a critical state frame work. Soils and Foundations 44
of varying density Proceedings of Biomediated and Bioinspired (3), 69–77. https://doi.org/10.3208/sandf.44.3_69
Geotechnics (B2G) Conference. Atlanta, GA, 6. Lee M., and Basu D. 2018. An integrated approach for resilience and
Khoubani A. 2018. A new bonding model for the particulate simulation sustainability in geotechnical engineering. Indian Geotechnical
of bio-cemented sand (with a side excursion on percolation in Journal 48 (2), 207–234. https://doi.org/10.1007/s40098-018-0297-
granular mixtures (PhD Dissertation). Oregon State University. 3
Khoubani A., Evans T.M., and Montoya B.M. 2016. Particulate Lee M., and Gomez M.G. 2022a. Examination of ammonium by-product
simulations of triaxial tests on bio-cemented sand using a new removal following ureolytic bio-cementation using post-treatment
cementation model Geo-Chicago 2016. American Society of Civil rinsing.To Be Submitted.
Engineers, Chicago, Illinois, 84–93. Lee M., and Gomez M.G. 2022b. Effect of field-representative factors on
https://doi.org/10.1061/9780784480120.010 the liquefaction triggering and post-triggering behavior of
Kim N.-R., and Kim D.-S. 2010. A shear wave velocity tomography biocemented sands.To Be Submitted.
system for geotechnical centrifuge testing. Geotechnical Testing Lee M., Gomez M.G., El Kortbawi M., and Ziotopoulou K. 2021. Effect
Journal 33 (6), 434–444. of light biocementation on the liquefaction triggering and post-
Kirkland C.M., Akyel A., Hiebert R., McCloskey J., Kirksey J., triggering behavior of loose sands. J. Geotech. Geoenviron. Eng. 148
Cunningham A.B., Gerlach R., Spangler L., and Phillips A.J. 2021a. (1), 04021170. https://doi.org/10.1061/(ASCE)GT.1943-
Ureolysis-induced calcium carbonate precipitation (UICP) in the 5606.0002707
presence of CO2-affected brine: A field demonstration. Lee M., Gomez M.G., El Kortbawi M., and Ziotopoulou K. 2020.
International Journal of Greenhouse Gas Control 109, 103391. Examining the liquefaction resistance of lightly cemented sands
https://doi.org/10.1016/j.ijggc.2021.103391 using microbially induced calcite precipitation (MICP) Geo-
Kirkland C.M., Hiebert R., Hyatt R., McCloskey J., Kirksey J., Thane A., Congress 2020. American Society of Civil Engineers, Minneapolis,
Cunningham A.B., Gerlach R., Spangler L., and Phillips A.J. 2021b. Minnesota, 53–64. https://doi.org/10.1061/9780784482834.007
Direct injection of biomineralizing agents to restore injectivity and Lee M., Gomez M.G., Graddy C.M.R., San Pablo A.C.M., DeJong J.T.,
wellbore integrity. SPE Production & Operations 36 (01), 216–223. and Nelson D.C. 2022. Improving the spatial control of bio-
https://doi.org/10.2118/203845-PA cementation soil improvement using stimulated indigenous
Kirkland C.M., Thane A., Hiebert R., Hyatt R., Kirksey J., Cunningham microorganisms.To Be Submitted.
A.B., Gerlach R., Spangler L., and Phillips A.J. 2020. Addressing Lee M., Kolbus C.M., Yepez A.D., and Gomez M.G. 2019a.
wellbore integrity and thief zone permeability using microbially- Investigating ammonium by-product removal following stimulated
induced calcium carbonate precipitation (micp): a field ureolytic microbially-induced calcite precipitation. 260–270.
demonstration. Journal of Petroleum Science and Engineering 190, Lee M., Gomez M.G., San Pablo A.C.M., Kolbus C.M., Graddy C.M.R.,
107060. https://doi.org/10.1016/j.petrol.2020.107060 DeJong J.T., and Nelson D.C. 2019b. Investigating ammonium by-
Kirkland C.M., Zanetti S., Grunewald E., Walsh D.O., Codd S.L., and product removal for ureolytic bio-cementation using meter-scale
Phillips A.J. 2017. Detecting microbially induced calcite experiments. Sci Rep 9 (1), 1–15. https://doi.org/10.1038/s41598-
precipitation in a model well-bore using downhole low-field NMR. 019-54666-1
Environ. Sci. Technol. 51 (3), 1537–1543. Lee Y.S., Kim H.J., and Park W. 2017. Non-ureolytic calcium carbonate
https://doi.org/10.1021/acs.est.6b04833 precipitation by Lysinibacillus sp. YS11 isolated from the
Kleinberg R.L., Flaum C., Straley C., Brewer P.G., Malby G.E., Peltzer rhizosphere of Miscanthus sacchariflorus. J Microbiol. 55 (6), 440–
E.T., Friederich G., and Yesinowski J.P. 2003. Seafloor nuclear 447. https://doi.org/10.1007/s12275-017-7086-z
magnetic resonance assay of methane hydrate in sediment and rock. Leejeerajumnean A., Ames J.M., and Owens J.D. 2000. Effect of
J. Geophys. Res. 108 (B3). https://doi.org/10.1029/2001JB000919 ammonia on the growth of bacillus species and some other bacteria.
Knorst M.T., Neubert R., and Wohlrab W. 1997. Analytical methods for Letters in Applied Microbiology 30 (5), 385–389.
measuring urea in pharmaceutical formulations. Journal of https://doi.org/10.1046/j.1472-765x.2000.00734.x
Pharmaceutical and Biomedical Analysis 15 (11), 1627–1632. Li B. 2015. Geotechnical properties of biocement treated sand and clay.
https://doi.org/10.1016/S0731-7085(96)01978-4 Nanyang Technological University, Singapore.
Kramer S.L. 1996. Geotechnical earthquake engineering. Prentice-Hall https://doi.org/10.32657/10356/62560
Inc. Li X., Scheibe T.D., and Johnson W.P. 2004. Apparent decreases in
Krom M.D. 1980. Spectrophotometric determination of ammonia: a colloid deposition rate coefficients with distance of transport under
study of a modified Berthelot reaction using salicylate and unfavorable deposition conditions: a general phenomenon. Environ.
dichloroisocyanurate. Analyst 105 (1249), 305. Sci. Technol. 38 (21), 5616–5625.
https://doi.org/10.1039/an9800500305 https://doi.org/10.1021/es049154v
Lai H., Wu S., Cui M., and Chu J. 2021. Recent development in Liang J., Guo Z., Deng L., and Liu Y. 2015. Mature fine tailings
biogeotechnology and its engineering applications. Front. Struct. consolidation through microbial induced calcium carbonate
Civ. Eng. 15 (5), 1073–1096. https://doi.org/10.1007/s11709-021- precipitation. Can. J. Civ. Eng. 42 (11), 975–978.
0758-0 https://doi.org/10.1139/cjce-2015-0069
Laloui L., and François B. 2009. ACMEG-T: Soil thermoplasticity Liang J., Guo Z., Deng L., and Liu Y. 2014. Oil sands mature fine tailings
model. Journal of Engineering Mechanics 135 (9), 932–944. consolidation through microbial induced calcite precipitation.
https://doi.org/10.1061/(ASCE)EM.1943-7889.0000011 OSRIN (Report No, TR-66), 39.
502
Liang T., Bengough A.G., Knappett J.A., MuirWood D., Loades K.W., ground improvement 321–329.
Hallett P.D., Boldrin D., Leung A.K., and Meijer G.J. 2017. Scaling https://doi.org/10.1061/9780784482834.035
of the reinforcement of soil slopes by living plants in a geotechnical Martinez A., DeJong J., Akin I., Aleali A., Arson C., Atkinson J., Bandini
centrifuge. Ecological Engineering 109, 207–227. P., Baser T., Borela R., Boulanger R., Burrall M., Chen Y., Collins
https://doi.org/10.1016/j.ecoleng.2017.06.067 C., Cortes D., Dai S., DeJong T., Del Dottore E., Dorgan K.,
Lin H., O’Donnell S.T., Suleiman M.T., Kavazanjian E., and Brown D.G. Fragaszy R., Frost J.D., Full R., Ghayoomi M., Goldman D.I.,
2021. Effects of enzyme and microbially induced carbonate Gravish N., Guzman I.L., Hambleton J., Hawkes E., Helms M., Hu
precipitation treatments on the response of axially loaded pervious D., Huang L., Huang S., Hunt C., Irschick D., Lin H.T., Lingwall B.,
concrete piles. J. Geotech. Geoenviron. Eng. 147 (8), 04021057. Marr A., Mazzolai B., McInroe B., Murthy T., O’Hara K., Porter M.,
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002565 Sadek S., Sanchez M., Santamarina C., Shao L., Sharp J., Stuart H.,
Lin H., Suleiman M.T., and Brown D.G. 2020. Investigation of pore- Stutz H.H., Summers A., Tao J., Tolley M., Treers L., Turnbull K.,
scale CaCO3 distributions and their effects on stiffness and Valdes R., van Paassen L., Viggiani G., Wilson D., Wu W., Yu X.,
permeability of sands treated by microbially induced carbonate and Zheng J. 2021. Bio-inspired geotechnical engineering:
precipitation (MICP). Soils and Foundations 60 (4), 944–961. principles, current work, opportunities and challenges.
https://doi.org/10.1016/j.sandf.2020.07.003 Géotechnique 1–19. https://doi.org/10.1680/jgeot.20.P.170
Lin H., Suleiman M.T., Brown D.G., and Kavazanjian E. 2016a. Martinez A., DeJong J.T., Jaeger R.A., and Khosravi A. 2020. Evaluation
Mechanical behavior of sands treated by microbially induced of self-penetration potential of a bio-inspired site characterization
carbonate precipitation. J. Geotech. Geoenviron. Eng. 142 (2), probe by cavity expansion analysis. Can. Geotech. J. 57 (5), 706–
04015066. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001383 716. https://doi.org/10.1139/cgj-2018-0864
Lin H., Suleiman M.T., Jabbour H.M., Brown D.G., and Kavazanjian E. Martinez A., and O’Hara K.B. 2021. Skin friction directionality in
2016b. Enhancing the axial compression response of pervious monotonically- and cyclically-loaded bio-inspired piles in sand.
concrete ground improvement piles using biogrouting. J. Geotech. DFIJ 15 (1). https://doi.org/10.37308/DFIJnl.20200831.222
Geoenviron. Eng. 142 (10), 04016045. Martinez A., Palumbo S., and Todd B.D. 2019. Bioinspiration for
https://doi.org/10.1061/(ASCE)GT.1943-5606.0001515 anisotropic load transfer at soil–structure interfaces. Journal of
Lin H. (Thomas), Suleiman M., Jabbour H., and Brown D. 2017. Bio- Geotechnical and Geoenvironmental Engineering 145 (10),
grouting to enhance axial pull-out response of pervious concrete 04019074-undefined. https://doi.org/10.1061/(ASCE)GT.1943-
ground improvement piles. Canadian Geotechnical Journal 55, 1– 5606.0002138
12. https://doi.org/10.1139/cgj-2016-0438 Martinez B.C., DeJong J.T., Ginn T.R., Montoya B.M., Barkouki T.H.,
Liu Q. 2021. Experimental investigation of microbially induced Hunt C., Tanyu B., and Major D. 2013. Experimental optimization
carbonate precipitation on fine mineral tailings (Ph.D.). North of microbial-induced carbonate precipitation for soil improvement.
Carolina State University, United States -- North Carolina. J. Geotech. Geoenviron. Eng. 139 (4), 587–598.
Liu Q., and Montoya B.M. 2021. Microbial-induced calcium carbonate https://doi.org/10.1061/(ASCE)GT.1943-5606.0000787
precipitation to accelerate sedimentation of fine tailings. J. Geotech. Marvi H., Bridges J., and Hu D.L. 2013. Snakes mimic earthworms:
Geoenviron. Eng. 147 (10), 02821001. propulsion using rectilinear travelling waves. J. R. Soc. Interface. 10
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002651 (84), 20130188. https://doi.org/10.1098/rsif.2013.0188
Liu Q., and Montoya B.M. 2020. Experimental study of consolidation Mickovski S.B., Bengough A.G., Bransby M.F., Davies M.C.R., Hallett
behavior of mature fine tailings treated with microbial induced P.D., and Sonnenberg R. 2007. Material stiffness, branching pattern
calcium carbonate precipitation Geo-Congress 2020. American and soil matric potential affect the pullout resistance of model root
Society of Civil Engineers, Minneapolis, Minnesota, 196–204. systems. Eur J Soil Science 58 (6), 1471–1481.
https://doi.org/10.1061/9780784482834.022 https://doi.org/10.1111/j.1365-2389.2007.00953.x
Liu S., Wen K., Armwood C., Bu C., Li C., Amini F., and Li L. 2019. Minto J.M., MacLachlan E., El Mountassir G., and Lunn R.J. 2016. Rock
Enhancement of MICP-treated sandy soils against environmental fracture grouting with microbially induced carbonate precipitation.
deterioration. Journal of Materials in Civil Engineering 31 (12), Water Resour. Res. 52 (11), 8827–8844.
04019294. https://doi.org/10.1061/(ASCE)MT.1943-5533.0002959 https://doi.org/10.1002/2016WR018884
Lunne T., Powell J.J.M., and Robertson P.K. 1997. Cone penetration Misra A. 2010. A multicriteria based quantitative framework for
testing in geotechnical practice, Reprinted. ed. E & FN Spon, assessing sustainability of pile foundations. University of
London. Connecticut, Mansfield, Connecticut.
Luria S. 1960. The bacterial protoplasm: composition and Mitchell J.K. 2008. Recent developments in ground improvement for
organizationGunsalus, I.C., and Stanier, R.Y. (Eds.), The Bacteria. mitigation of seismic risk to existing embankment dams 1–20.
Academic Press, New York, NY, 1–34. https://doi.org/10.1061/40975(318)98
Mahawish A., Bouazza A., and Gates W.P. 2019a. Factors affecting the Mitchell J.K. 1981. Soil improvement – state of the art report
bio-cementing process of coarse sand. Proceedings of the Institution Proceedings of 10th ICSMFE. Balkema, Stockholm, Sweden, 509–
of Civil Engineers - Ground Improvement 172 (1), 25–36. 565.
https://doi.org/10.1680/jgrim.17.00039 Mitchell J.K., and Santamarina J.C. 2005. Biological considerations in
Mahawish A., Bouazza A., and Gates W.P. 2019b. Unconfined geotechnical engineering. J. Geotech. Geoenviron. Eng. 131 (10),
compressive strength and visualization of the microstructure of 1222–1233. https://doi.org/10.1061/(ASCE)1090-
coarse sand subjected to different biocementation levels. Journal of 0241(2005)131:10(1222)
Geotechnical and Geoenvironmental Engineering 145 (8), Mobley H.L., and Hausinger R.P. 1989. Microbial ureases: significance,
04019033. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002066 regulation, and molecular characterization. Microbiol. Rev.
Mahawish A., Bouazza A., and Gates W.P. 2018. Effect of particle size https://doi.org/10.1128/mr.53.1.85-108.1989
distribution on the bio-cementation of coarse aggregates. Acta Montgomery J., and Boulanger R.W. 2017. Effects of spatial variability
Geotech. 13 (4), 1019–1025. https://doi.org/10.1007/s11440-017- on liquefaction-induced settlement and lateral spreading. Journal of
0604-7 Geotechnical and Geoenvironmental Engineering 143 (1),
Major D.W., McMaster M.L., Cox E.E., Edwards E.A., Dworatzek S.M., 04016086. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001584
Hendrickson E.R., Starr M.G., Payne J.A., and Buonamici L.W. Montgomery J., and Ziotopoulou K. 2020. Numerical Simulations of
2002. Field demonstration of successful bioaugmentation to achieve Selected LEAP Centrifuge Experiments with PM4Sand in
dechlorination of tetrachloroethene to ethene. Environ. Sci. Technol. FLACKutter, B.L., Manzari, M.T., and Zeghal, M. (Eds.), Model
36 (23), 5106–5116. https://doi.org/10.1021/es0255711 Tests and Numerical Simulations of Liquefaction and Lateral
Mallett S.D. 2019. Mechanical behavior of fibrous root-inspired Spreading. Springer International Publishing, Cham, 481–497.
anchorage systems (PhD Dissertation). Georgia Institute of Montoya B.M., and DeJong J.T. 2015. Stress-strain behavior of sands
Technology, Atlanta, GA. cemented by microbially induced calcite precipitation. Journal of
Manzari M.T., and Dafalias Y.F. 1997. A critical state two-surface Geotechnical and Geoenvironmental Engineering 141 (6),
plasticity model for sands. Géotechnique 47 (2), 255–272. 04015019. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001302
https://doi.org/10.1680/geot.1997.47.2.255 Montoya B.M., and Dejong J.T. 2013. Healing of biologically induced
Martin K.K., Khodadadi T.H., Chester M., and Kavazanjian E. 2020. cemented sands. Géotechnique Letters 3 (3), 147–151.
Hotspot life cycle assessment for environmental impacts of EICP for https://doi.org/10.1680/geolett.13.00044
503
Montoya B.M., Dejong J.T., and Boulanger R.W. 2013. Dynamic Geomechanics 17 (7), 04017010.
response of liquefiable sand improved by microbial-induced calcite https://doi.org/10.1061/(ASCE)GM.1943-5622.0000863
precipitation. Géotechnique 63 (4), 302–312. Nguyen L.D., Fatahi B., and Khabbaz H. 2014. A constitutive model for
https://doi.org/10.1680/geot.SIP13.P.019 cemented clays capturing cementation degradation. International
Montoya Brina M., Do J., and Gabr M.A. 2021. Distribution and Journal of Plasticity 56, 1–18.
properties of microbially induced carbonate precipitation in https://doi.org/10.1016/j.ijplas.2014.01.007
underwater sand bed. J. Geotech. Geoenviron. Eng. 147 (10), Nicholson P.G. 2015. Soil improvement and ground modification
04021098. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002607 methods. Butterworth-Heinemann.
Montoya Brina M, Evans T.M., Wengrove M.E., Bond H., Yazdani E., Nielsen S.D., Koren K., Löbmann K., Hinge M., Scoma A., Kjeldsen
Dadashiserej A., and Liu Q. 2021. Resisting dune erosion with bio- K.U., and Røy H. 2020. Constraints on CaCO3 precipitation in
cementation 11. superabsorbent polymer by aerobic bacteria. Appl Microbiol
Montoya B.M., Safavizadeh S., and Gabr M.A. 2019. Enhancement of Biotechnol 104 (1), 365–375. https://doi.org/10.1007/s00253-019-
coal ash compressibility parameters using microbial-induced 10215-4
carbonate precipitation. J. Geotech. Geoenviron. Eng. 145 (5), Noh D.-H., Ajo-Franklin J.B., Kwon T.-H., and Muhunthan B. 2016. P
04019018. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002036 and S wave responses of bacterial biopolymer formation in
Moon C.J., Whateley M.E.G., and Evans A.M. 2006. Introduction to unconsolidated porous media. Journal of Geophysical Research:
mineral exploration. Blackwell Publishing, Sydney. Biogeosciences 121 (4), 1158–1177.
Mörsdorf G., and Kaltwasser H. 1989. Ammonium assimilation in https://doi.org/10.1002/2015JG003118
Proteus vulgaris, Bacillus pasteurii, and Sporosarcina ureae. Arch. Nova R., Castellanza R., and Tamagnini C. 2003. A constitutive model
Microbiol. 152 (2), 125–131. https://doi.org/10.1007/BF00456089 for bonded geomaterials subject to mechanical and/or chemical
Morse J.W., and Arvidson R.S. 2002. The dissolution kinetics of major degradation. Int. J. Numer. Anal. Meth. Geomech. 27 (9), 705–732.
sedimentary carbonate minerals. Earth-Science Reviews 58 (1–2), https://doi.org/10.1002/nag.294
51–84. https://doi.org/10.1016/S0012-8252(01)00083-6 Ntarlagiannis D., Robinson J., Soupios P., and Slater L. 2016. Field-scale
Morse J.W., and Berner R.A. 1972. Dissolution kinetics of calcium electrical geophysics over an olive oil mill waste deposition site:
carbonate in sea water: II. A kinetic origin for the lysocline. Evaluating the information content of resistivity versus induced
American Journal of Science 272, 840–851. polarization (IP) images for delineating the spatial extent of organic
https://doi.org/10.2475/ajs.272.9.840 contamination. Journal of Applied Geophysics 135, 418–426.
Mortensen B.M., Haber M.J., DeJong J.T., Caslake L.F., and Nelson https://doi.org/10.1016/j.jappgeo.2016.01.017
D.C. 2011. Effects of environmental factors on microbial induced Ntarlagiannis D., Yee N., and Slater L. 2005. On the low-frequency
calcium carbonate precipitation: Environmental factors on MICP. electrical polarization of bacterial cells in sands. Geophys. Res. Lett.
Journal of Applied Microbiology 111 (2), 338–349. 32 (24), L24402. https://doi.org/10.1029/2005GL024751
https://doi.org/10.1111/j.1365-2672.2011.05065.x Nweke C.C., and Pestana J.M. 2017. Modeling bio-cemented sands:
Mujah D., Shahin M.A., and Cheng L. 2016. State-of-the-art review of Shear strength and stiffness with degradation 34–45.
biocementation by microbially induced calcite precipitation (MICP) https://doi.org/10.1061/9780784480793.004
for soil stabilization. Geomicrobiology Journal 34 (6), 524–537. Nzila A., Razzak S.A., and Zhu J. 2016. Bioaugmentation: an emerging
https://doi.org/10.1080/01490451.2016.1225866 strategy of industrial wastewater treatment for reuse and discharge.
Mukherjee S., Sahu R.B., and Mukherjee J. 2021. Effect of biologically International Journal of Environmental Research and Public Health
induced cementation via ureolysis in stabilization of silty soil. 13 (9), 846. https://doi.org/10.3390/ijerph13090846
Geomicrobiology Journal 39 (1), 66–82. Obermayr M., Dressler K., Vrettos C., and Eberhard P. 2013. A bonded-
https://doi.org/10.1080/01490451.2021.2005188 particle model for cemented sand. Computers and Geotechnics 49,
Multiphysical modelling of soils with a focus on microbially induced 299–313. https://doi.org/10.1016/j.compgeo.2012.09.001
calcite precipitation (PhD Dissertation) 2012. . EPFL, Lausanne, O’Donnell S.T., Kavazanjian E., and Rittmann B.E. 2017a. MIDP:
Switzerland. https://doi.org/10.5075/epfl-thesis-5413 Liquefaction mitigation via microbial denitrification as a two-stage
Na T.K., Cabas A., and Montoya B.M. 2022. Resonant column testing process. II: MICP. J. Geotech. Geoenviron. Eng. 143 (12),
procedure for microbial induced carbonate precipitated sands. 04017095. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001806
Geotechnical Testing Journal. O’Donnell S.T., Rittmann B.E., and Kavazanjian E. 2017b. MIDP:
Nafisi A., Khoubani A., Montoya B.M., and Evans T.M. 2018. The effect Liquefaction mitigation via microbial denitrification as a two-stage
of grain size and shape on mechanical behavior of MICP sand I: process. i: desaturation. J. Geotech. Geoenviron. Eng. 143 (12),
Experimental study 9. 04017094. https://doi.org/10.1061/(ASCE)GT.1943-5606.0001818
Nafisi A., Liu Q., and Montoya B.M. 2021. Effect of stress path on the Ohan J.A., Saneiyan S., Lee J., Bartlow A.W., Ntarlagiannis D., Burns
shear response of bio-cemented sands. Acta Geotech. 16 (10), 3239– S.E., and Colwell F.S. 2020. Microbial and geochemical dynamics
3251. https://doi.org/10.1007/s11440-021-01286-7 of an aquifer stimulated for microbial induced calcite precipitation
Nafisi A., Mocelin D., Montoya B.M., and Underwood S. 2020a. Tensile (MICP). Front. Microbiol. 11, 1327.
strength of sands treated with microbially induced carbonate https://doi.org/10.3389/fmicb.2020.01327
precipitation. Can. Geotech. J. 57 (10), 1611–1616. O’Hara K.B., and Martinez A. 2020. Monotonic and cyclic frictional
https://doi.org/10.1139/cgj-2019-0230 resistance directionality in snakeskin-inspired surfaces and piles.
Nafisi A., Montoya B.M., and Evans T.M. 2020b. Shear strength Journal of Geotechnical and Geoenvironmental Engineering 146
envelopes of biocemented sands with varying particle size and (11), 04020116-undefined. https://doi.org/10.1061/(asce)gt.1943-
cementation level. J. Geotech. Geoenviron. Eng. 146 (3), 04020002. 5606.0002368
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002201 Ortiz D., Gravish N., and Tolley M.T. 2019. Soft robot actuation
Nafisi A., Safavizadeh S., and Montoya B.M. 2019. Influence of microbe strategies for locomotion in granular substrates. IEEE Robot. Autom.
and enzyme-induced treatments on cemented sand shear response. J. Lett. 4 (3), 2630–2636. https://doi.org/10.1109/LRA.2019.2911844
Geotech. Geoenviron. Eng. 145 (9), 06019008. O’Toole C.O., Liu Q., Montoya B.M., Kananizadeh N., and Odle W.
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002111 2022. The effect of microbial induced carbonate precipitation on
Nassar M.K., Gurung D., Bastani M., Ginn T.R., Shafei B., Gomez M.G., fine-grained mine tailings GeoCongress 2022. Charlotte, North
Graddy C.M.R., Nelson D.C., and DeJong J.T. 2018. Large-scale Carolina.
experiments in microbially induced calcite precipitation (micp): Pacheco-Torgal F., and Labrincha J.A. 2013. Biotechnologies and
reactive transport model development and prediction. Water bioinspired materials for the construction industry: an overview.
Resources Research 54 (1), 480–500. International Journal of Sustainable Engineering 7 (3), 235–244.
https://doi.org/10.1002/2017WR021488 https://doi.org/10.1080/19397038.2013.844741
Nelson D.C., and Jannasch H.W. 1983. Chemoautotrophic growth of a Panico F., Fonseca A.V. da, and Vaunat J. 2018. Modelling the small
marine beggiatoa in sulfide-gradient cultures. Arch. Microbiol. 136 strain behaviour of a cemented silty sand with bounding plasticity
(4), 262–269. https://doi.org/10.1007/BF00425214 Numerical Methods in Geotechnical Engineering IX. CRC Press.
Nguyen L., Fatahi B., and Khabbaz H. 2017. Development of a Park S.-S., Choi S.-G., Kim W.-J., and Lee J.-C. 2014a. Effect of
constitutive model to predict the behavior of cement-treated clay microbially induced calcite precipitation on the strength of cemented
during cementation degradation: C3 model. International Journal of sand New Frontiers in Geotechnical Engineering. American Society
504
of Civil Engineers, Shanghai, China, 47–56. precipitation and its sustainability in engineering applications.
https://doi.org/10.1061/9780784413456.006 Sustainability 12 (15), 6281. https://doi.org/10.3390/su12156281
Park S.-S., Choi S.-G., and Nam I.-H. 2014b. Effect of plant-induced Ramachandran S.K., Ramakrishnan V., and Bang S.S. 2001.
calcite precipitation on the strength of sand. J. Mater. Civ. Eng. 26 Remediation of concrete using micro-organisms. ACI Materials
(8), 06014017. https://doi.org/10.1061/(ASCE)MT.1943- Journal 98 (1), 3–9. https://doi.org/10.14359/10154
5533.0001029 Raymond A.J. 2020. Development and implementation of a life cycle
Peng J., and Liu Z. 2019. Influence of temperature on microbially sustainability assessment framework for geotechnical systems (PhD
induced calcium carbonate precipitation for soil treatment. PLoS Dissertation). University of California, Davis, Davis, CA.
ONE 14 (6), e0218396. Raymond A.J., DeJong J.T., Kendall A., Blackburn J.T., and Deschamps
https://doi.org/10.1371/journal.pone.0218396 R. 2021. Life cycle sustainability assessment of geotechnical ground
Pham V.P., van Paassen L.A., van der Star W.R.L., and Heimovaara T.J. improvement methods. Journal of Geotechnical and
2018. Evaluating strategies to improve process efficiency of Geoenvironmental Engineering.
denitrification-based MICP. Journal of Geotechnical and https://doi.org/10.1061/(ASCE)GT.1943-5606.0002646
Geoenvironmental Engineering 144 (8), 04018049. Raymond A.J., Kendall A., and DeJong J.T. 2020a. Life cycle
https://doi.org/10.1061/(ASCE)GT.1943-5606.0001909 sustainability assessment (LCSA): a research evaluation tool for
Phang I.R.K., Wong K.S., Chan Y.S., and Lau S.Y. 2019. Effect of emerging geotechnologies. American Society of Civil Engineers,
surcharge load on microbial-induced calcite precipitation (MICP) Reston, VA, 330–339. https://doi.org/10.1061/9780784482834.036
treatment of tropical peat. IOP Conf. Ser.: Mater. Sci. Eng. 495, Raymond A.J., Kendall A., DeJong J.T., Gomez M.G., San Pablo
012068. https://doi.org/10.1088/1757-899X/495/1/012068 A.C.M., Lee M., Graddy C.M.R., and Nelson D.C. 2022. Life cycle
Phillips A.J., Cunningham A.B., Gerlach R., Hiebert R., Hwang C., sustainability assessment of microbially induced calcite precipitation
Lomans B.P., Westrich J., Mantilla C., Kirksey J., Esposito R., and (MICP) bio-cementation techniques.To Be Submitted.
Spangler L. 2016. Fracture sealing with microbially-induced Raymond A.J., Kendall A., DeJong J.T., Kavazanjian E., Woolley M.A.,
calcium carbonate precipitation: a field study. Environ. Sci. Technol. and Martin K.K. 2020b. Life cycle sustainability assessment of
50 (7), 4111–4117. https://doi.org/10.1021/acs.est.5b05559 fugitive dust control methods. J. Constr. Eng. Manage. 147 (3),
Phillips A.J., Troyer E., Hiebert R., Kirkland C., Gerlach R., 04020181. https://doi.org/10.1061/(ASCE)CO.1943-7862.0001993
Cunningham A.B., Spangler L., Kirksey J., Rowe W., and Esposito Rebata-Landa V. 2007. Microbial activity in sediments: Effects on soil
R. 2018. Enhancing wellbore cement integrity with microbially behavior (PhD Dissertation). Georgia Institue of Technology.
induced calcite precipitation (micp): a field scale demonstration. Rebata-Landa V., and Santamarina J.C. 2006. Mechanical limits to
Journal of Petroleum Science and Engineering 171, 1141–1148. microbial activity in deep sediments: microbes. Geochem. Geophys.
https://doi.org/10.1016/j.petrol.2018.08.012 Geosyst. 7 (11). https://doi.org/10.1029/2006GC001355
Phradichith C. 2020. Control of biofilm permeability reduction in sand Reddy K.R., and Saxena S.K. 1992. Constitutive modelling of cemented
columns by varying the treatment delivery system (Master’s Thesis). sand. Mechanics of Materials 14, 155–178.
University of California, Davis. https://doi.org/10.1016/0167-6636(92)90012-3
Pinske M.A. 2011. Life cycle assessment of ground improvement Ribeiro B.G.O., and Gomez M.G. 2022a. Investigating the dissolution
methods (Master of Science in Civil & Environmental behavior of calcium carbonate based bio-cemented sands.
Engineering). University of California, Davis, Davis, CA. GeoCongress 2022 Technical Papers.Accepted.
Plummer L.N., and Busenberg E. 1982. The solubilities of calcite, Ribeiro B.G.O., and Gomez M.G. 2022b. Dissolution behavior of
aragonite and vaterite in CO2-H2O solutions between 0 and 90°C, ureolytic bio-cementation: physical experiments and reactive
and an evaluation of the aqueous model for the system CaCO3-CO2- transport modeling.To Be Submitted.
H2O. Geochim. 46, 1011–1040. https://doi.org/10.1016/0016- Rinaldi V., Guichon M., Ferrero V., Serrano C., and Ponti N. 2006.
7037(82)90056-4 Resistivity survey of the subsurface conditions in the estuary of the
Porter H., Dhami N.K., and Mukherjee A. 2018. Sustainable road bases rio de la plata. J. Geotech. Geoenviron. Eng. 132 (1), 72–79.
with microbial precipitation. Proceedings of the Institution of Civil https://doi.org/10.1061/(ASCE)1090-0241(2006)132:1(72)
Engineers - Construction Materials 171 (3), 95–108. Rios S., Da Fonseca A., and Baudet B. 2014. On the shearing behaviour
https://doi.org/10.1680/jcoma.16.00075 of an artificially cemented soil. Acta Geotechnica 9.
Porter K.G., and Feig Y.S. 1980. The use of DAPI for identifying and https://doi.org/10.1007/s11440-013-0242-7
counting aquatic microflora. Limnol. Oceanogr. 25 (5), 943–948. Riveros G.A., and Sadrekarimi A. 2020. Liquefaction resistance of Fraser
https://doi.org/10.4319/lo.1980.25.5.0943 River sand improved by a microbially-induced cementation. Soil
Pretell R., Ziotopoulou K., and Abrahamson N.A. 2022. Conducting 1D Dynamics and Earthquake Engineering 131, 106034.
site response analyses to capture 2D VS spatial variability effects. https://doi.org/10.1016/j.soildyn.2020.106034
Earthquake Spectra 87552930211069400. Rix G.J., and Stokoe K.H. 1991. Correlation of initial tangent modulus
https://doi.org/10.1177/87552930211069400 and cone penetration resistance. Calibration Chamber Testing 351–
Proto C.J., DeJong J.T., and Nelson D.C. 2016. Biomediated 361.
permeability reduction of saturated sands. J. Geotech. Geoenviron. Robertson P., Fear C., Woeller D., and Weemees I. 1995. Estimation of
Eng. 142 (12), 04016073. https://doi.org/10.1061/(ASCE)GT.1943- sand compressibility from seismic CPT Proceedings of the 48th
5606.0001558 Canadian Geotechnical Conference. Vancouver, BC, 25–27.
Pungrasmi W., Intarasoontron J., Jongvivatsakul P., and Likitlersuang S. Robertson P.K. 2016. Cone penetration test (CPT)-based soil behaviour
2019. Evaluation of microencapsulation techniques for MICP type (SBT) classification system — an update. Can. Geotech. J. 53
bacterial spores applied in self-healing concrete. Sci Rep 9 (1), (12), 1910–1927. https://doi.org/10.1139/cgj-2016-0044
12484. https://doi.org/10.1038/s41598-019-49002-6 Roden E.E., Leonardo M.R., and Ferris F.G. 2002. Immobilization of
Purdy C.M., Raymond A.J., DeJong J.T., Kendall A., and Krage C. 2021. strontium during iron biomineralization coupled to dissimilatory
Life cycle sustainability assessment of geotechnical site hydrous ferric oxide reduction. Geochimica et Cosmochimica Acta
investigation. Canadian Geotechnical Journal. 66 (16), 2823–2839. https://doi.org/10.1016/S0016-7037(02)00878-
Qin Y., and Cabral J.M.S. 1994. Kinetic studies of the urease-catalyzed 5
hydrolysis of urea in a buffer-free system. Appl Biochem Biotechnol Roy N. 2021. Pore space architecture of particulate materials:
49 (3), 217–240. https://doi.org/10.1007/BF02783059 Characterization & applications (PhD Dissertation). Georgia
Rahimi M., Chan D., and Nouri A. 2018. Constitutive model for Institute of Technology, Atlanta, GA.
monotonic and cyclic responses of loosely cemented sand Rumpelt T.K. 1990. Constitutive modeling of weak porous rocks (Ph.D.).
formations. Journal of Rock Mechanics and Geotechnical University of California, Berkeley, United States -- California.
Engineering 10 (4), 740–752. Rütting T., Boeckx P., Müller C., and Klemedtsson L. 2011. Assessment
https://doi.org/10.1016/j.jrmge.2018.01.010 of the importance of dissimilatory nitrate reduction to ammonium for
Rahimi M., Chan D., and Nouri A. 2016. Bounding surface constitutive the terrestrial nitrogen cycle. Biogeosciences 8 (7), 1779–1791.
model for cemented sand under monotonic loading. Int. J. Geomech. https://doi.org/10.5194/bg-8-1779-2011
16 (2), 04015049. https://doi.org/10.1061/(ASCE)GM.1943- Safavizadeh S., Montoya B.M., and Gabr M.A. 2019. Microbial induced
5622.0000534 calcium carbonate precipitation in coal ash. Géotechnique 69 (8),
Rahman M.M., Hora R.N., Ahenkorah I., Beecham S., Karim M.R., and 727–740. https://doi.org/10.1680/jgeot.18.P.062
Iqbal A. 2020. State-of-the-art review of microbial-induced calcite
505
Safdar M.U., Mavroulidou M., Gunn M.J., Purchase D., Gray C., Payne carbonate precipitation. Soil Dynamics and Earthquake Engineering
I., and Garelick J. 2021. Towards the development of sustainable 115, 710–718. https://doi.org/10.1016/j.soildyn.2018.09.027
ground improvement techniques—biocementation study of an Sjöberg E.L. 1978. Kinetics and mechanism of calcite dissolution in
organic soil. Circ.Econ.Sust. https://doi.org/10.1007/s43615-021- aqueous solutions at low temperatures. Stockholm Contrib. Geol. 32,
00071-8 1–96.
San Pablo A.C.M., Graddy C.M.R., DeJong J.T., Gomez M.G., and Sjöberg E.L., and Rickard D.T. 1984. Temperature dependence of calcite
Nelson D.C. 2022a. Solution optimization of stimulated microbially dissolution kinetics between 1 and 62°C at pH 2.7 to 8.4 in aqueous
induced calcite precipitation (MICP) for up-scaling by reduction of solutions. Geochimica et Cosmochimica Acta 48 (3), 485–493.
urea inputs.To Be Submitted. https://doi.org/10.1016/0016-7037(84)90276-X
San Pablo A.C.M., DeJong J.T., and Carey T.J. 2022b. Effect of MICP Smith D.G., Russell W.C., Ingledew W.J., and Thirkell D. 1993.
treatment zones on mitigating liquefaction investigated via Hydrolysis of urea by Ureaplasma urealyticum generates a
centrifuge modeling.To Be Submitted. transmembrane potential with resultant ATP synthesis. J Bacteriol
San Pablo A.C.M., Lee M., Graddy C.M.R., Kolbus C.M., Khan M., 175 (11), 3253–3258. https://doi.org/10.1128/jb.175.11.3253-
Zamani A., Martin N., Acuff C., DeJong J.T., Gomez M.G., and 3258.1993
Nelson D.C. 2020. Meter-scale biocementation experiments to Smith P.K., Krohn R.I., Hermanson G.T., Mallia A.K., Gartner F.H.,
advance process control and reduce impacts: Examining spatial Provenzano M.D., Fujimoto E.K., Goeke N.M., Olson B.J., and
control, ammonium by-product removal, and chemical reductions. J. Klenk D.C. 1985. Measurement of protein using bicinchoninic acid.
Geotech. Geoenviron. Eng. 146 (11), 04020125. Analytical Biochemistry 150 (1), 76–85.
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002377 https://doi.org/10.1016/0003-2697(85)90442-7
Saneiyan S., Ntarlagiannis D., Ohan J., Lee J., Colwell F., and Burns S. Smith R.W., Fujita Y., Ginn T.R., and Hubbard S.S. 2012. Field
2019. Induced polarization as a monitoring tool for in-situ microbial investigations of microbially facilitated calcite precipitation for
induced carbonate precipitation (MICP) processes. Ecological immobilization of strontium-90 and other trace metals in the
Engineering 127, 36–47. subsurface (No. DOE/ER64404-01). Univ. of Iowa, Iowa City, IA
https://doi.org/10.1016/j.ecoleng.2018.11.010 (United States). https://doi.org/10.2172/1052856
Saneiyan S., Ntarlagiannis D., Werkema D.D., and Ustra A. 2018. Sneath P.H.A. 1962. Longevity of micro-organisms. Nature 195 (4842),
Geophysical methods for monitoring soil stabilization processes. 643–646. https://doi.org/10.1038/195643a0
Journal of Applied Geophysics 148, 234–244. Song C., and Liu S. 2020. A novel approach of bulk strength
https://doi.org/10.1016/j.jappgeo.2017.12.008 enhancement through microbially-mediated carbonate cementation
Santamarina J.C., Klein K.A., and Fam M.A. 2001. Soils and waves: for mylonitic coal. Geomicrobiology Journal 37 (8), 1–12.
particulate materials behavior, characterization and process https://doi.org/10.1080/01490451.2020.1764675
monitoring. J. Wiley & Sons, Chichester, New York. Sorenson K. 2021. Field trials and long-term monitoring of microbially-
Sasaki T., and Kuwano R. 2016. Undrained cyclic triaxial testing on sand induced desaturation for the treatment of liquefiable silty soils (PhD
with non-plastic fines content cemented with microbially induced Dissertation). Portland State University, Portland.
CaCO3. Soils and Foundations 56 (3), 485–495. Spaulding C., Masse F., and LaBrozzi J. 2008. Ground improvement
https://doi.org/10.1016/j.sandf.2016.04.014 technologies for a sustainable world. ASCE, New Orleans, LA, 891–
Saxena S.K., and Reddy K.R. 1987. Mechanical behavior of cemented 898. https://doi.org/10.1061/40971(310)111
sands (No. IIT-CE-8701), National Science Foundation Under Grant Stocks-Fischer S., Galinat J.K., and Bang S.S. 1999. Microbiological
No.CEE 8313935. Illinois Institute of Technology, Chicago, IL. precipitation of CaCO3. Soil Biology and Biochemistry 31 (11),
Saxena S.K., Reddy K.R., and Avramidis A.S. 1988. Liquefaction 1563–1571. https://doi.org/10.1016/S0038-0717(99)00082-6
resistance of artificially cemented sand 19. Stokoe K.H., Lee S.H.H., and Knox D.P. 1985. Shear moduli
https://doi.org/10.1061/(ASCE)0733-9410(1988)114:12(1395) measurements under true triaxial stresses. Advances in the Art of
Schnaid F., Lehane B., and Fahey M. 2004. In situ test characterization Testing Soils Under Cyclic Conditions. Detroit, Michigan, 166–185.
of unusual geomaterials In Proceedings of the 2nd International Su F., and Yang Y.Y. 2020. Microbially induced carbonate precipitation
Conference on Site Characterization. Millpress, Porto, Portugal, 49– via methanogenesis pathway by a microbial consortium enriched
74. from activated anaerobic sludge. J. Appl. Microbiol. 131 (1), 236–
Schneider J.A., and Moss R.E.S. 2011. Linking cyclic stress and cyclic 256. https://doi.org/10.1111/jam.14930
strain based methods for assessment of cyclic liquefaction triggering Tagliaferri F., Waller J., Andò E., Hall S.A., Viggiani G., Bésuelle P.,
in sands. Géotechnique Letters 1 (2), 31–36. and DeJong J.T. 2011. Observing strain localisation processes in bio-
https://doi.org/10.1680/geolett.11.00021 cemented sand using x-ray imaging. Granular Matter 13 (3), 247–
Schnell S., and Mendoza C. 1997. Closed form solution for time- 250. https://doi.org/10.1007/s10035-011-0257-4
dependent enzyme kinetics. Journal of Theoretical Biology 187 (2), Tang C.-S., Yin L., Jiang N., Zhu C., Zeng H., Li H., and Shi B. 2020.
207–212. https://doi.org/10.1006/jtbi.1997.0425 Factors affecting the performance of microbial-induced carbonate
Schwarz M., Cohen D., and Or D. 2011. Pullout tests of root analogs and precipitation (MICP) treated soil: a review. Environ Earth Sci 79 (5),
natural root bundles in soil: experiments and modeling: tensile forces 94. https://doi.org/10.1007/s12665-020-8840-9
and failure of root bundles. J. Geophys. Res. 116 (F2). -
https://doi.org/10.1029/2010JF001753 burrowing-out robot in dry sand. J. Geotech. Geoenviron. Eng. 145
Seed H.B., and Idriss I.M. 1970. Soil Moduli and Damping Factors for (12), 02819002. https://doi.org/10.1061/(ASCE)GT.1943-
Dynamic Response Analyses. (No. EERC 70-10). California Univ., 5606.0002177
Berkeley. Earthquake Engineering Research Center. Terzis D., Bernier-Latmani R., and Laloui L. 2016. Fabric characteristics
Seed H.B., Idriss I.M., Makdisi F., and Banerjee N. 1975. Representation and mechanical response of bio-improved sand to various treatment
of irregular stress time histories by equivalent uniform stress series conditions. Géotechnique Letters 6 (1), 50–57.
in liquefaction analyses. College of Engineering, University of https://doi.org/10.1680/jgele.15.00134
Terzis D., and Laloui L. 2019. A decade of progress and turning points
Shen Z., Jiang M., and Thornton C. 2016. DEM simulation of bonded in the understanding of bio-improved soils: a review. Geomechanics
granular material. Part I: Contact model and application to cemented for Energy and the Environment 19, 100116.
sand. Computers and Geotechnics 75, 192–209. https://doi.org/10.1016/j.gete.2019.03.001
https://doi.org/10.1016/j.compgeo.2016.02.007 Terzis D., and Laloui L. 2018. 3-D micro-architecture and mechanical
Shillaber C.M., Mitchell J.K., and Dove J.E. 2016. Energy and carbon response of soil cemented via microbial-induced calcite
assessment of ground improvement works. II: Working model and precipitation. Sci Rep 8 (1), 1416. https://doi.org/10.1038/s41598-
example. J. Geotech. Geoenviron. Eng. 142 (3), 04015084. 018-19895-w
https://doi.org/10.1061/(ASCE)GT.1943-5606.0001411 Terzis D., Laloui L., Dornberger S., and Harran R. 2020. A Full-Scale
Simatupang M., and Okamura M. 2017. Liquefaction resistance of sand Application of Slope Stabilization via Calcite Bio-Mineralization
remediated with carbonate precipitation at different degrees of Followed by Long-Term GIS Surveillance Geo-Congress 2020.
saturation during curing. Soils and Foundations 57 (4), 619–631. American Society of Civil Engineers, Minneapolis, Minnesota, 65–
https://doi.org/10.1016/j.sandf.2017.04.003 73. https://doi.org/10.1061/9780784482834.008
Simatupang M., Okamura M., Hayashi K., and Yasuhara H. 2018. Small- Tian Z.-F., Tang X., Xiu Z.-L., and Xue Z.-J. 2020. Effect of different
strain shear modulus and liquefaction resistance of sand with biological solutions on microbially induced carbonate precipitation
506
and reinforcement of sand. Marine Georesources & Geotechnology Engineers, Oakland, California, United States, 968–977.
38 (4), 450–460. https://doi.org/10.1080/1064119X.2019.1595229 https://doi.org/10.1061/9780784412121.100
Tobler D.J., Cuthbert M.O., Greswell R.B., Riley M.S., Renshaw J.C., van Paassen L.A., Pham V., Mahabadi N., Hall C., Stallings E., and
Handley-Sidhu S., and Phoenix V.R. 2011. Comparison of rates of Kavazanjian E. 2018. Desaturation via biogenic gas formation as a
ureolysis between Sporosarcina pasteurii and an indigenous ground improvement technique 244–256.
groundwater community under conditions required to precipitate https://doi.org/10.1061/9780784481677.013
large volumes of calcite. Geochimica et Cosmochimica Acta 75 (11), van Tittelboom K., De Belie N., De Muynck W., and Verstraete W. 2010.
3290–3301. https://doi.org/10.1016/j.gca.2011.03.023 Use of bacteria to repair cracks in concrete. Cement and Concrete
Tobler D.J., Minto J.M., El Mountassir G., Lunn R.J., and Phoenix V.R. Research 40 (1), 157–166.
2018. Microscale analysis of fractured rock sealed with microbially https://doi.org/10.1016/j.cemconres.2009.08.025
induced CaCO3 precipitation: Influence on hydraulic and Vatsala A., Nova R., and Murthy B.R.S. 2001. Elastoplastic model for
mechanical performance. Water Resour. Res. 54 (10), 8295–8308. cemented soils. Journal of Geotechnical and Geoenvironmental
https://doi.org/10.1029/2018WR023032 Engineering 127 (8), 679–687.
Tominaga T., An S.-Y., Oyaizu H., and Yokota A. 2009. Sporosarcina https://doi.org/10.1061/(ASCE)1090-0241(2001)127:8(679)
luteola sp. nov. isolated from soy sauce production equipment in Vaunat J., and Gens A. 2003. Bond degradation and irreversible strains
Japan. The Journal of General and Applied Microbiology 55 (3), in soft argillaceous rock Soil mechanics and geotechnical
217–223. https://doi.org/10.2323/jgam.55.217 engineering. 479–484.
Trujillo M.E., Dedysh S., DeVos P., Hedlund B., Kämpfer P., Rainey F., Vego I., Ceccato F., Simonini P., Frost J.D., Mallett S.D., and Cola S.
and Whitman W.B. 2015. Sporosarcina Bergey’s Manual of 2021. Numerical investigation of failure mechanism during pullout
Systematics of Archaea and Bacteria. John Wiley & Sons, Ltd, 1–7. of root inspired anchorages IACMAG 2021: Challenges and
https://doi.org/10.1002/9781118960608.gbm00563 Innovations in Geomechanics. Springer, Cham, 111–118.
Tsai P.-H., and Ni S.-H. 2012. Effects of types of additives on dynamic https://doi.org/10.1007/978-3-030-64518-2_14
properties of cement stabilized soils. IJASE 10 (2), 131–144. Wadell H. 1935. Volume, shape, and roundness of quartz particles. The
Tsukamoto M., Inagaki T., Sasaki Y., and Oda K. 2013. Influence of Journal of Geology 43 (3), 250–280. https://doi.org/10.1086/624298
relative density on microbial carbonate precipitation and mechanical Wadell H. 1933. Sphericity and roundness of rock particles. The Journal
properties of sand Proceedings of the 18th International Conference of Geology 41 (3), 310–331. https://doi.org/10.1086/624040
on Soil Mechanics and Geotechnical Engineering. Paris, 2613– Wadell H. 1932. Volume, shape, and roundness of rock particles. The
2616. Journal of Geology 40 (5), 443–451. https://doi.org/10.1086/623964
United Nations 2019. The sustainable development goals report. New Wang L., Chu J., Wu S., and Wang H. 2021. Stress–dilatancy behavior
York. of cemented sand: comparison between bonding provided by cement
Updegraff D.M. 1983. Plugging and penetration of petroleum reservoir and biocement. Acta Geotech. 16 (5), 1441–1456.
rock by microorganisms Proceedings of the 1982 International https://doi.org/10.1007/s11440-021-01146-4
Conference on Microbial Enhancement of Oil Recovery. Shangri-La, Wang Y., Soga K., DeJong J.T., and Kabla A.J. 2019. Microscale
Afton, Oklahoma. B. Linville, Bartlesville Energy Technology …, visualization of microbial-induced calcium carbonate precipitation
85. processes. J. Geotech. Geoenviron. Eng. 145 (9), 04019045.
U.S. Army Corps of Engineers 1999. Guidelines on ground improvement https://doi.org/10.1061/(ASCE)GT.1943-5606.0002079
for structures and facilities (1110-1-185 No. Engineer Technical Wang Y.-H., and Leung S.-C. 2008. A particulate-scale investigation of
Letter). cemented sand behavior. Can. Geotech. J. 45 (1), 29–44.
U.S. Department of Agriculture, Agricultural Research Service 2021. https://doi.org/10.1139/T07-070
USDA food and nutrient database for dietary studies 2017-2018 Warthmann R., van Lith Y., Vasconcelos C., McKenzie J.A., and Karpoff
[WWW Document]. Food Surveys Research Group Home Page. A.M. 2000. Bacterially induced dolomite precipitation in anoxic
URL http://www.ars.usda.gov/nea/bhnrc/fsrg culture experiments 4.
US EPA 2013. Aquatic life ambient water quality criteria for ammonia – Weatherburn M.W. 1967. Phenol-hypochlorite reaction for
freshwater. Washington, DC. determination of ammonia. Anal. Chem. 39 (8), 971–974.
U.S. Federal Highway Association 2017. Ground modification methods https://doi.org/10.1021/ac60252a045
reference manual – Volume I (No. FHWA-NHI-16-027). Weaver T.J., Burbank M., Lewis A., Lewis R., Crawford R., and
U.S. Federal Highway Association, Berg R.R., Collin J.G., DiMaggio Williams B. 2011. Bio-induced calcite, iron, and manganese
J.A., Filz G.M., Bruce D.A., and Ayala D. 2017. Ground precipitation for geotechnical engineering applications Geo-
modification methods reference manual – Volume II (No. FHWA- Frontiers 2011. American Society of Civil Engineers, Dallas, Texas,
NHI-16-028). United States, 3975–3983. https://doi.org/10.1061/41165(397)406
van Ballegooy S., Malan P., Lacrosse V., Jacka M.E., Cubrinovski M., Weil M.H., DeJong J.T., Martinez B.C., and Mortensen B.M. 2011.
Bray J.D., O’Rourke T.D., Crawford S.A., and Cowan H. 2014. Seismic and resistivity measurements for real-time monitoring of
Assessment of liquefaction-induced land damage for residential microbially induced calcite precipitation in sand. Geotechnical
Christchurch. Earthquake Spectra 30 (1), 31–55. testing journal 35 (2), 1–12. https://doi.org/10.1520/GTJ103365
https://doi.org/10.1193/031813EQS070M Whiffin V.S. 2004. Microbial caco3 precipitation for the production of
van Paassen L.A. 2011. Bio-mediated ground improvement: from biocement (phd). Whiffin, Victoria S.
laboratory experiment to pilot applications Geo-Frontiers 2011, <https://researchrepository.murdoch.edu.au/view/author/Whiffin,
Proceedings. American Society of Civil Engineers, Dallas, Texas, Victoria.html> (2004) Microbial CaCO3 precipitation for the
United States, 4099–4108. https://doi.org/10.1061/41165(397)419 production of biocement. PhD thesis, Murdoch University.
van Paassen L.A. 2009. Biogrout: ground improvement by microbially Murdoch University, Perth, Australia.
induced carbonate precipitation. (PhD Dissertation). Delft Whiffin V.S., Paassen L.A. van, and Harkes M.P. 2007. Microbial
University of Technology, Delft, Netherlands. carbonate precipitation as a soil improvement technique.
van Paassen L.A., Daza C.M., Staal M., Sorokin D.Y., van der Zon W., Geomicrobiology Journal 24 (5), 417–423.
and van Loosdrecht Mark.C.M. 2010a. Potential soil reinforcement https://doi.org/10.1080/01490450701436505
by biological denitrification. Ecological Engineering, Special Issue: Williams K.H., Kemna A., Wilkins M.J., Druhan J., Arntzen E.,
BioGeoCivil Engineering 36 (2), 168–175. N’Guessan A.L., Long P.E., Hubbard S.S., and Banfield J.F. 2009.
https://doi.org/10.1016/j.ecoleng.2009.03.026 Geophysical monitoring of coupled microbial and geochemical
van Paassen L.A., Ghose R., van der Linden T.J.M., van der Star W.R.L., processes during stimulated subsurface bioremediation. Environ.
and van Loosdrecht M.C.M. 2010b. Quantifying biomediated Sci. Technol. 43 (17), 6717–6723. https://doi.org/10.1021/es900855j
ground improvement by ureolysis: large-scale biogrout experiment. Wilson W.W., Wade M.M., Holman S.C., and Champlin F.R. 2001.
J. Geotech. Geoenviron. Eng. 136 (12), 1721–1728. Status of methods for assessing bacterial cell surface charge
https://doi.org/10.1061/(ASCE)GT.1943-5606.0000382 properties based on zeta potential measurements. Journal of
van Paassen L.A., Harkes M.P., and van Zwieten G.A. 2009. Scale up of Microbiological Methods 12. https://doi.org/10.1016/S0167-
biogrout: a biological ground reinforcement method 6. 7012(00)00224-4
van Paassen L.A., Hemert W.J. van, Star W.R.L. van der, Zwieten G. Wu S., Li B., and Chu J. 2021. Stress-dilatancy behavior of MICP-treated
van, and Baalen L. van 2012. Direct shear strength of biologically sand. Int. J. Geomech. 21 (3), 04020264.
cemented gravel GeoCongress 2012. American Society of Civil https://doi.org/10.1061/(ASCE)GM.1943-5622.0001923
507
Wu S., Li B., and Chu J. 2019. Large-scale model tests of biogrouting for Can. Geotech. J. 56 (12), 1889–1900. https://doi.org/10.1139/cgj-
sand and rock. Proceedings of the Institution of Civil Engineers - 2018-0551
Ground Improvement 1–10. https://doi.org/10.1680/jgrim.18.00074 Zamani A., Phradichith C., Graddy C.M.R., DeJong J.T., Nelson D.C.,
Wu Y., Hubbard S., Williams K.H., and Ajo-Franklin J. 2010. On the and Parales R.E. 2022. Control of biofilm permeability reduction in
complex conductivity signatures of calcite precipitation. J. Geophys. sand columns by varying the treatment delivery system.To Be
Res. 115 (G2), n/a-n/a. https://doi.org/10.1029/2009JG001129 Submitted.
Xiao P., Liu H., Stuedlein A.W., Evans T.M., and Xiao Y. 2019. Effect Zamani A., Xiao P., Baumer T., Carey T.J., Sawyer B., DeJong J.T., and
of relative density and biocementation on cyclic response of Boulanger R.W. 2021. Mitigation of liquefaction triggering and
calcareous sand. Can. Geotech. J. 56 (12), 1849–1862. foundation settlement by MICP treatment. J. Geotech. Geoenviron.
https://doi.org/10.1139/cgj-2018-0573 Eng. 147 (10), 04021099. https://doi.org/10.1061/(ASCE)GT.1943-
Xiao P., Liu H., Xiao Y., Stuedlein A.W., and Evans T.M. 2018. 5606.0002596
Liquefaction resistance of bio-cemented calcareous sand. Soil Zambare N.M., Naser N.Y., Gerlach R., and Chang C.B. 2020.
Dynamics and Earthquake Engineering 107, 9–19. Mineralogy of microbially induced calcium carbonate precipitates
https://doi.org/10.1016/j.soildyn.2018.01.008 formed using single cell drop-based microfluidics. Sci Rep 10 (1),
Xiao Y., Stuedlein A.W., Pan Z., Liu H., Matthew Evans T., He X., Lin 17535. https://doi.org/10.1038/s41598-020-73870-y
H., Chu J., and van Paassen L.A. 2020. Toe-bearing capacity of Zeng C., Veenis Y., Hall C.A., Young E.S., van der Star W.R.L., Zheng
precast concrete piles through biogrouting improvement. J. Geotech. J., and van Paassen L.A. 2021. Experimental and numerical analysis
Geoenviron. Eng. 146 (12), 06020026. of a field trial application of microbially induced calcite precipitation
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002404 for ground stabilization. J. Geotech. Geoenviron. Eng. 147 (7),
Xiao Y., Wang Y., Wang S., Evans T.M., Stuedlein A.W., Chu J., Zhao 05021003. https://doi.org/10.1061/(ASCE)GT.1943-5606.0002545
C., Wu H., and Liu H. 2021a. Homogeneity and mechanical Zhang A., Dafalias Y.F., and Jiang M. 2021. A bounding surface
behaviors of sands improved by a temperature-controlled one-phase plasticity model for cemented sand under monotonic and cyclic
MICP method. Acta Geotech. 16 (5), 1417–1427. loading. Géotechnique 1–18. https://doi.org/10.1680/jgeot.20.P.275
https://doi.org/10.1007/s11440-020-01122-4 Zhang F., Wang C., Chang J., and Feng H. 2022. DEM analysis of cyclic
Xiao Y., Xiao W., Ma G., He X., Wu H., and Shi J. 2022. Mechanical liquefaction behaviour of cemented sand. Computers and
performance of biotreated sandy road bases. J. Perform. Constr. Geotechnics 142, 104572.
Facil. 36 (1), 04021111. https://doi.org/10.1061/(ASCE)CF.1943- https://doi.org/10.1016/j.compgeo.2021.104572
5509.0001671 Zhang G., Robertson P.K., and Brachman R.W.I. 2004. Estimating
Xiao Y., Zhang Z., Stuedlein A.W., and Evans T.M. 2021b. Liquefaction liquefaction-induced lateral displacements using the standard
modeling for biocemented calcareous sand. Journal of Geotechnical penetration test or cone penetration test. J. Geotech. Geoenviron.
and Geoenvironmental Engineering 147 (12), 04021149. Eng. 130 (8), 861–871. https://doi.org/10.1061/(ASCE)1090-
https://doi.org/10.1061/(ASCE)GT.1943-5606.0002666 0241(2004)130:8(861)
Yanez V.R. 2021. Effect of microbially induced calcite precipitation in Zhang G., Robertson P.K., and Brachman R.W.I. 2002. Estimating
carbonate rich soils (Master’s Thesis). University of California, liquefaction-induced ground settlements from CPT for level ground.
Davis, Davis, CA. Can. Geotech. J. 39 (5), 1168–1180. https://doi.org/10.1139/t02-047
Yanez V.R., San Pablo A.C.M., Graddy C.M.R., DeJong J.T., Gomez Zhang X., Chen Y., Liu H., Zhang Z., and Ding X. 2020. Performance
M.G., and Nelson D.C. 2022. Bio-stimulated MICP soil evaluation of a MICP-treated calcareous sandy foundation using
improvement in marine and brackish environments.To Be Submitted. shake table tests. Soil Dynamics and Earthquake Engineering 129,
Yang C., Cui Y.J., Pereira J.M., and Huang M.S. 2008. A constitutive 105959. https://doi.org/10.1016/j.soildyn.2019.105959
model for unsaturated cemented soils under cyclic loading. Zhang X., Knappett J.A., Leung A.K., Ciantia M.O., Liang T., and Nicoll
Computers and Geotechnics 35 (6), 853–859. B.C. 2022. Centrifuge modelling of root-soil interaction of laterally
https://doi.org/10.1016/j.compgeo.2008.08.005 loaded trees under different loading conditions. Géotechnique 1–39.
Yang M., Défossez P., Danjon F., and Fourcaud T. 2018. Analyzing key https://doi.org/10.1680/jgeot.21.00088
factors of roots and soil contributing to tree anchorage of Pinus Zhao Q., Li L., Li C., Li M., Amini F., and Zhang H. 2014. Factors
species. Trees 32 (3), 703–712. https://doi.org/10.1007/s00468-018- affecting improvement of engineering properties of MICP-treated
1665-4 soil catalyzed by bacteria and urease. Journal of Materials in Civil
Yang P. 2018. Discrete element modeling of the mechanical response of Engineering 26 (12), 04014094.
cemented granular materials (PhD Dissertation). Arizona State https://doi.org/10.1061/(ASCE)MT.1943-5533.0001013
University, Tempe, Arizona. Zheng J., and Hryciw R.D. 2015. Traditional soil particle sphericity,
Yang P., Kavazanjian E., and Neithalath N. 2019. Particle-scale roundness and surface roughness by computational geometry.
mechanisms in undrained triaxial compression of biocemented Géotechnique 65 (6), 494–506.
sands: Insights from 3D DEM simulations with flexible boundary. https://doi.org/10.1680/geot.14.P.192
Int. J. Geomech. 19 (4), 04019009. Ziotopoulou K., and Boulanger R.W. 2016. Plasticity modeling of
https://doi.org/10.1061/(ASCE)GM.1943-5622.0001346 liquefaction effects under sloping ground and irregular cyclic
Yang P., O’Donnell S., Hamdan N., Kavazanjian E., and Neithalath N. loading conditions. Soil Dynamics and Earthquake Engineering 84,
2017. 3D DEM simulations of drained triaxial compression of sand 269–283. https://doi.org/10.1016/j.soildyn.2016.02.013
strengthened using microbially induced carbonate precipitation. Int.
J. Geomech. 17 (6), 04016143.
https://doi.org/10.1061/(ASCE)GM.1943-5622.0000848
Yu H.S., Tan S.M., and Schnaid F. 2007. A critical state framework for
modelling bonded geomaterials. Geomechanics and Geoengineering
2 (1), 61–74. https://doi.org/10.1080/17486020601164275
Yu T., Souli H., Pechaud Y., and Fleureau J.-M. 2021. Review on
engineering properties of MICP-treated soils. Geomechanics and
Engineering 27 (1), 13–30.
https://doi.org/10.12989/GAE.2021.27.1.013
Zamani A., and Montoya B.M. 2019. Undrained cyclic response of silty
sands improved by microbial induced calcium carbonate
precipitation. Soil Dynamics and Earthquake Engineering 120, 436–
448. https://doi.org/10.1016/j.soildyn.2019.01.010
Zamani A., and Montoya B.M. 2018. Undrained monotonic shear
response of MICP-treated silty sands. J. Geotech. Geoenviron. Eng.
144 (6), 04018029. https://doi.org/10.1061/(ASCE)GT.1943-
5606.0001861
Zamani A., Montoya B.M., and Gabr M.A. 2019. Investigating
challenges of in situ delivery of microbial-induced calcium
carbonate precipitation (MICP) in fine-grain sands and silty sand.
508
Corporate Associates Presidential Group
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Are we overdesigning: Results of ISSMGE survey
Sommes-nous en surconception : résultats de l'enquête ISSMGE
Peter Day
University of Stellenbosch, Stellenbosch, South Africa
Jean-Louis Briaud
Texas A&M University, College Station, Texas, USA
ABSTRACT: In June 2018, geotechnical engineers from the Asian Region of the ISSMGE asked: “Are we overdesigning?” In
response, the ISSMGE initiated a survey involving ten simple geotechnical problems on clays or sands, based on real-life
geotechnical data from the two National Geotechnical Experimentation Sites at Texas A&M University. With the assistance of the
various regions of the ISSMGE, the survey was taken online in 2020 and 2021, and close to 240 responses were received. These
responses are analysed and presented in this paper and compared with results of full-scale tests conducted on the Texas A&M sites,
model calculations, and reliability-based solutions by TC 304. In almost all cases, the average of the responses gave a reasonable
answer to the problem. However, the range of responses was unacceptably large and contained many highly improbable outcomes.
Only two of the ten problems considered indicated a tendency towards overdesign. The remainder of the problems included
evidence of both over- and underdesign. There is a clear need for the ISSMGE to focus on the implementation of existing knowledge
and on improving the state of practice in the industry.
RÉSUMÉ : En Juin 2018, les ingénieurs géotechniciens de la région asiatique de la SIMSG ont demandé "est-ce que nous
surconcevons ?". En réponse, la SIMSG a lancé une enquête impliquant dix problèmes géotechniques simples sur des argiles ou des
sables, basés sur des données géotechniques réelles provenant des deux sites d'essai, Sites Nationaux d'Expérimentation Géotechnique
a Texas A&M University. Avec l'aide des différentes régions de la SIMSG, l'enquête a été mise en ligne en 2020 et 2021 et près de 240
réponses ont été reçues. Ces réponses sont analysées et présentées dans cet article et comparées aux résultats des tests à grande échelle
menés sur les sites Texas A&M, aux calculs de modèles et aux solutions basées sur la fiabilité par le TC 304. Dans presque tous les cas,
la moyenne des réponses a donné un résultat raisonnable en réponse aux problems poses. Cependant, l'éventail des réponses était
inacceptablement large et contenait de nombreux résultats hautement improbables. Seuls deux des 10 problèmes examinés indiquaient
une tendance à la surconception. Le reste des problèmes comprenait des preuves de sur- et de sous-conception. Il est clairement nécessaire
que la SIMSG se concentre sur la mise en œuvre des connaissances existantes et sur l'amélioration de l'état des pratiques dans l'industrie.
KEYWORDS: geotechnical design, overdesign, design calibration, practice survey.
selection of parameters and calculation models. The design
1 INTRODUCTION problems were aimed at assessing the provisions made for safety
and serviceability in geotechnical design. Respondents were
At the XVI Danube European Conference (Skopje Macedonia, asked to answer as many of the problems as they wished,
June 2018), the question was asked: “Are we overdesigning?” preferably based on the work they do on a day-to-day basis.
Although the question originated from the Asian Region, it is Partial solutions were also accepted.
valid internationally where different countries and users of Initially, respondents were requested to submit their responses
various testing techniques or design codes tend to follow local by email in .pdf format. Later, due largely to the assistance
practice for geotechnical design. received from the Young Members Presidential Group (YMPG),
In response, the Corporate Associates Presidential Group the survey was taken online.
(CAPG) of the ISSMGE conducted a survey of current
geotechnical design practices worldwide. The survey was set up
to assess the consistency of calculation models and design 2 GEOTECHNICAL DATA
methods for a variety of geotechnical structures and, where
possible, to compare the results with full-scale tests, model 2.1 Data provided
solutions and reliability analyses.
The survey was based on two soil profiles one in clay and Although not disclosed at the time, the geotechnical data was
the other in sand. Soil test results, typical of those one would find from the National Geotechnical Experimentation Sites at Texas
in a geotechnical investigation report, were given for each soil A&M University. The data was extracted from Briaud (1997),
profile. Excel spreadsheets were also provided, with numerical was digitised, and provided in both graphical and numerical
data to assist with data analysis. Unbeknown to the respondents, format.
the data came from investigations conducted on two test sites on The information included moisture content, grading, SPT,
the campus of Texas A&M University at College Station, Texas, CPT/CPTu and Pressuremeter test data at both sites. In addition,
USA (Briaud 1997; Briaud 2021a; Briaud 2021b.) bulk density, dry density, over-consolidation and undrained
Ten specimen problems were posed, including concentrically shear strength data was provided for the clay site, and dilatometer
and eccentrically loaded spread footings, axially and laterally and cross-hole seismic data for the sand site. The graphical
loaded piles, slopes and retaining structures. The idea was to representation of this data is given in Annexures A and B at the
keep the problems easy to analyse and representative of everyday end of this paper.
geotechnical structures. The location of the sites, their geological history, site
Some problems required prediction of the performance of the stratigraphy and soil properties are presented in Briaud (1997).
geotechnical structure (prediction problems) while others called The original survey and the numerical data for both sites can be
for the design of the structure as it would be constructed (design found at original survey and soils data (scroll down to the
problems). The prediction problems were aimed at assessing the overdesign survey files).
511
2.2 Clay site Figure 2. Occupation of respondents.
The soils at the clay site are predominantly alluvial in origin and
0 20 40 60 80 100 120
are over-consolidated. The profile can be summarised as follows:
0 – 5.5m Very stiff clay Consultant / designer 104
5.5 – 6.5m Medium dense sand parting (intermittent)
6.5 – 12.0m Very stiff clay Student 70
12.0m + Highly weathered shale.
Academic 24
The water table is at a depth of about 6m below ground level. Contractor 17
2.3 Sand site Graduate in training 6
The soils at the sand site are also primarily of alluvial origin and Other 15
the profile is as follows:
0 – 4.0m Silty sand
4.0 – 8.0m Clean sand 0 25 50 75 100 125
8.0 – 12.5m Clayey sand
12.5m + Highly weathered shale. Student 63
The water table is at a depth of about 5m below ground level. 0 - 10 years 107
10 - 20 years 37
3 ANALYSIS OF DATA
As will be seen from the discussion of the ten sample problems, 30 years + 14
the answers varied over a wide range. Some answers were
clearly incorrect, and some were frivolous. These answers were Figure 3. Respondents’ years of experience.
omitted from the analysis of the data. However, outliers (very
high or very low values) still existed within the remaining data.
These outliers distort the average values and skew the analysis.
Extreme outliers were also omitted, but values that had the 0 20 40 60 80 100
appearance of a considered response (e.g. accompanied by input
parameters, calculation methods, etc.) were retained in the data PF LSD 86
set. WLD 78
The elimination of data is a subjective process. To make clear LRFD 31
where high or low values have been removed, the range of values
considered in each analysis is shown on the figures where WLD & LRFD 7
appropriate. Anyone interested in re-analysing the complete WLD & PF LSD 6
data set can retrieve the original data from overdesign survey
Other 1
combined responses. This Excel file includes all individual
responses to the problems, complete with parameters used, No answer 27
comments and references.
Figure 4. Design methods used in respondent’s country.
4 DETAILS OF SURVEY RESPONDENTS In addition to the submissions by survey respondents,
ISSMGE Technical Committee 304 (TC 304: Engineering
The first section of the survey gathered general information about Practice of Risk Assessment and Management) compiled a report
the respondents. providing probabilistic analyses of the survey questions (Ching
Two hundred and thirty-six responses were received from 39 & Zhang 2020). This report can be downloaded from TC 304
countries. The number of responses received from the top ten report.
countries is shown in Figure 1. The occupation and experience of
the respondents are shown in Figures 2 and 3. The design
methods used (WLD working load design, LRFD load and 5 CLAY 1: VERTICALLY LOADED STRIP
resistance factor design, PF LSD partial factor limit states FOOTING
design, or Other) are shown in Figure 4.
Two additional questions were asked regarding the design 5.1 The problem (Figure 5)
codes used in the respondent’s country and the respondent’s
favourite textbooks. These responses are not analysed in this
paper. They can, however, be viewed by downloading the P
original data from the website given above. • 2.0m wide concrete strip
footing founded on clay at
0 10 20 30 40 50 60 n.g.l. 0.75m below ground level
• Concentric line load P
South Africa 46 excludes weight of footing
0.75m
Argentina 33
1.2m
France 32 2.0m width
Canada 18 Figure 5. CLAY 1 – strip footing on clay.
Australia 15
Morocco 13
Singapore 9
CLAY 1 is a prediction problem in which respondents were asked
to predict the applied load that will cause bearing capacity failure
Chile 7 of the footing (Pult) and the applied load that will cause the
Indonesia 7 footing to settle 25mm in the long term (P25mm).
Bolivia 6
Others 50
Figure 1. Number of responses from top ten countries.
512
5.2 Responses 3 000 kN/m) were from drained analyses. Direct correlations
with in situ tests gave an average prediction of 1 027 kN/m.
The predicted bearing capacity (Pult) is given in Figure 6 and the
load that will cause 25mm settlement (P25mm) is given in Figure
7.
Exclusive median
150 - 7000 kN/m
TC304:
1 154 (mean)
TC304
N = 155
1 154 (mean)
Average 1 227 kN/m
(43 - 8 872 kN/m)
Figure 8. Bearing capacity (Pult) v. experience.
Figure 9 was compiled to check if the wide range of
Figure 6. Predicted bearing capacity (Pult). predictions is attributable to the estimation of shear strength from
the data provided or the calculation of bearing capacity using the
estimated shear strength. This figure represents the results from
the undrained (total stress) analyses. The blue dots are the survey
TC304 responses. The solid black line in Figure 9 is the result of a
650 (mean) Skempton-type bearing capacity calculation (q = c Nc + p0) with
a bearing capacity factor Nc = 5.7 (for z/B = 0.375), a bulk density
of 21.5 kN/m3 for the clay and a density of 24 kN/m3 for concrete.
The red cross is the result of the reliability analysis by TC 304
N = 150 (Yang et al. 2020) based on an average undrained shear strength
Average 538 kN/m of 111.3 kPa, a coefficient of variation of 26.6% and a log-normal
(18 - 2 431 kN/m) distribution.
Two conclusions can be drawn from Figure 9. Firstly, there is
a large range of inferred undrained shear strengths for the clay
(50 kPa 200 kPa). The low values are probably a reflection of
the lower bound to the results from triaxial tests on the clay. The
mid-range values are a reasonable interpretation of all the triaxial
test results and of the in situ test data. The justification for the
high values (approaching 200 kPa) is unclear. The second
conclusion is that the bearing capacities calculated for any given
undrained shear strength show a large variation and deviate
significantly from the results of traditional bearing capacity
theory. The low values may be the result of the inclusion of a
margin of safety, but the reason for the high values is unclear.
There is insufficient data to permit a similar analysis of the
Figure 7. Predicted load for 25mm settlement (P25mm).
drained (effective stress) results.
5000
5.3 Analysis of responses Skempton
Although the mean values may be reasonably acceptable, it is Nc = 5.7
clear from Figures 6 and 7 that there is a considerable range of 4000
Bearing capacity (kN/m)
results. The probabilistic solutions prepared by TC 304 (Yang et
TC304 (mean)
al. 2020) show that it is highly improbable that the bearing
capacity of the footing will exceed 3 000 kN/m or the load for 3000
25mm settlement will exceed 2 000 kN/m.
To assess whether the wide range of answers can be attributed
to inexperience of some respondents, 30% of whom were still
students, the results were sorted according to the experience of 2000
the respondents as shown in Figure 8. The average values for
each group are not unreasonable, but there were more outliers in
the less experienced categories. This is not necessarily so for the 1000
remaining problems.
Of the respondents who stated which method of analysis was
used, approximately 70% used an undrained analysis (total 0
stress), 26% used a drained analysis (effective stress) and 5% 0 50 100 150 200 250
used direct correlations with in situ tests. The average prediction Undrained shear strength cu (kPa)
of the drained bearing capacity (2 081 kN/m) was more than
double the average prediction of undrained bearing capacity Figure 9. Comparison of results for varying values of shear strength.
(968 kN/m). Note that most of the high predictions (above
513
No direct measurement of the compressibility of the clay
profile is available. The stiffness of the clay must therefore be
determined from in situ test results. For a drained elastic modulus TC304
profile varying from 10 MPa at founding level to 24 MPa at a mean
depth of 12m, as may be inferred from the SPT N-values 40 % Avearge 40%
interpreted using Stroud’s method (Stroud 1989), a simple elastic TC304 N = 82
settlement analysis gives the applied load under which 25mm deterministic
settlement will occur as 210 kN/m. This agrees reasonably with 25 %
the deterministic value of 261 kN/m calculated by Yang et al.
(2020) based on CPT and Pressuremeter test results. The survey
responses shown in Figure 7 (average value 538 kN/m) are closer
to the result of the probabilistic analysis by Yang et al. (2020)
which gave a mean value of 650 kN/m.
6 CLAY 2: AXIALLY LOADED PILE
6.1 The problem (Figure 10)
P
Figure 12. Predicted base capacity as a percentage of total capacity.
n.g.l.
• 0.76m diameter, bored, cast in Load (kN)
0.76m situ pile in clay 0 1000 2000 3000 4000
10.0m
• Depth 10.0m below natural 0
ground level.
20
Settlement (mm)
Figure 10. CLAY 2 – bored pile in clay. 40
CLAY 2 is a prediction problem in which respondents were asked
to predict the ultimate capacity of the pile, the end bearing and
shaft resistance (if calculated separately), and the load-settlement 60
curve up to 80% of the failure load. N = 63
Flemming
6.2 Responses 80
The predicted pile capacity (Pult) is given in Figure 11, the ratio Measured
of end bearing to total pile capacity (Pb/Pult) in Figure 12, and the 100
load-settlement curves in Figure 13.
The TC 304 probabilistic assessment is given in Shuku and Figure 13. Predicted load settlement curves.
Hachich (2020a).
Exclusive median
Measured (75 - 12 000 kN)
N = 86
2 033 kN Avearge 2 460 kN
(75 - 12 000 kN) TC 304
TC304 (probabilistic
deterministic TC304 mean) TC 304
3 186 kN (deterministic)
probabilistic
mean
4 066 kN
Measured
Figure 14. Pile capacity v. experience.
Figure 11. Predicted pile capacity.
6.3 Analysis of responses
Once again, the average values are reasonable but there is a wide
range of responses. Figure 14 shows the range of results v.
experience of the respondents. This time, the most significant
outlier was not from the least experienced respondents. It is
514
clear from this graph how outliers can affect the interpretation of
results. TC304 Measured
This problem is based on a full-scale pile load test performed mean 409 kN
at the Texas A&M National Geotechnical Experimentation Site 279 kN
(Briaud 2021a; King et al. 2009), the site from which the test TC304
results given in Section 2 were obtained. The result of this pile deterministic
252 kN
load test is shown in black solid lines on Figures 11, 13 and 14.
N = 53
The blue dashed line in Figure 13 is a prediction by the first Avearge 446kN
author using the method proposed by Fleming (1992). In this case, (25 - 1 700 kN)
the pile capacity was based on empirical correlations with SPT
results (Byrne et al. 2019). The pile capacity was assessed as
2 400 kN, the contribution of the base capacity (end bearing) as
25%, and the elastic modulus of the soil below the tip of the pile
as 40 MPa. Changing the contribution of the base capacity to 40%
(based on Figure 12) gives an even closer alignment between the
measured load settlement curve and the results given by
Fleming’s method.
It is interesting to note that many respondents opted for a
straight-line load-settlement response. Some respondents even
opted for a curve where the axial stiffness of the pile increased
with load. Neither of these options accords with the typical load- Figure 16. Predicted horizontal load capacity.
settlement response shown by the two curves superimposed on
Figure 13.
Horizontal Load (kN)
0.0 200.0 400.0 600.0 800.0 1000.0
7 CLAY 3: LATERALLY LOADED PILE 0
7.1 The problem (Figure 15) 10
20
H
Lateral deflection (mm)
• 0.76m diameter, bored, cast in 30
0.8m
n.g.l. situ pile in clay
• Founding depth 4.6m below
40
natural ground level 50
0.76m • Load applied at 0.8m above
60
4.6m
ground level
• Pile reinforcement sufficient to 70
N = 44
prevent bending or shear
80
failure of pile shaft. TC 304
Figure 15. CLAY 3 – laterally loaded pile in clay. 90
Measured
CLAY 3 is a prediction problem in which respondents were asked 100
to predict the ultimate lateral load capacity of the pile, and the Figure 17. Load-deflection curves.
load-deflection curve up to 80% of the failure load. Like problem
CLAY 2, this problem is based on tests conducted on the Texas
A&M National Geotechnical Experimentation Site. 1800
1600
Horizontal load capacity (kN)
7.2 Responses
1400
The predicted lateral load capacity is given in Figure 16 and the Broms
load-deflection curves in Figure 17. The TC 304 analysis was 1200
undertaken by Shuku and Hachich (2020b). TC 304
1000 deterministic
7.3 Analysis of responses
800
In the Figures 16 and 17, two very optimistic responses were
omitted as they were so high that they would have distorted the 600
analysis.
The measured values (black solid lines) in Figures 16 and 17 400
come from a test on a pile of the same dimensions at the clay site 200
at Texas A&M in 1979 as reported by Bierschwale et al. (1981).
It is pleasing to note that the measured lateral load capacity of 0
the pile corresponds well with both the mean and the median 40 80 120 160
values from the responses received.
The analysis by TC 304 (Shuku and Hachich 2020b) was Undrained shear strength cu (kPa)
based on an approximation of the shape of the load-deflection
curve using a bi-linear function shown by the dashed red line in Figure 18. Comparison of results for varying values of shear strength.
Figure 17. The load-deflection curve represented by the bi-linear
approximation has a shape similar to the measured results in Figure 19 shows the range of results v. experience of the
Figure 17. respondents.
Figure 18 shows the predicted lateral load capacity of the pile
plotted against the undrained shear strength, where this was
stated by the respondents. The solid black line in this figure is the
lateral resistance from the Broms (1964) method. The
deterministic value from the TC 304 calculation is shown by the
red cross.
515
When analysing the slope, the respondent must decide whether
to use an effective stress analysis (drained analysis, based on c’
and ’) or a total stress analysis (undrained analysis, based on cu).
Exclusive median
Of the 61 respondents who indicated which analysis was used,
36% used an effective stress analysis and 56% a total stress
analysis, and 8% used both. The results from the effective and
total stress analyses are given in Figures 22 and 23.
The other decision to be made is whether to include a tension
crack and whether the tension crack can fill with water. Only 13%
Measured of respondents made provision for a tension crack and only 7%
TC304 allowed the crack to fill with water.
mean
TC 304 Deterministic
N = 26
Average 29.1o
(4o - 75o)
Figure 19. Predicted horizontal load capacity v. experience.
8 CLAY 4: PERMANENT SLOPE
8.1 The problem (Figure 20)
n.g.l.
• Permanent cutting for
6m cutting Figure 22. Design angle of cutting (effective stress analysis).
a public road.
• No surcharges.
TC 304 N = 40
1m deep toe drain Pf = 1.10-4 Average 42.3o
Figure 20. CLAY 4: Permanent slope in clay. 34o (14o - 90o)
Unlike the previous problems which involved the prediction of
the capacity of piles and footings, this is a design problem. There TC 304 TC 304
is therefore an added component, namely allowance for a margin Pf = 1.10-3 Pf = 1.10-2
of safety. In this case, respondents were asked to specify the 56o 83o
maximum slope angle for a permanent road cutting in clay.
8.2 Responses
Figure 21 summarises the responses received.
Figure 23. Design angle of cutting (total stress analysis).
N = 74 8.3 Analysis of responses
Average 36.5o
(4o - 90o) Given that the clay is overconsolidated and that the slope is to be
permanent, there is a strong possibility that failure will occur in
the drained state after a prolonged period of time. This seems to
be borne out by Figures 22 and 23 which show a lower design
angle for the effective stress (drained) analysis than for the total
stress (undrained) analysis.
In the undrained condition, it is likely that the slope can be cut
vertically. This was confirmed by the analysis by TC 304 (Jiang
et al. 2020a) which showed a factor of safety exceeding 2.0 for
an undrained, uncracked, vertical cut. However, such a slope is
unlikely to remain stable in the long term due to the formation of
tension cracks and desiccation of the clays leading to ravelling
and sloughing.
The reliability analysis by TC 304 (Jiang et al. 2020b) showed
that the probability of failure of an undrained, uncracked slope
Figure 21. Design angle of cutting (all responses). cut at 83o is 1.10-2, at 56o is 1.10-3 and at 34o is 1.10-4.
There are too many different combinations of parameters to
permit a more theoretical analysis of this problem.
516
TC 304
Exclusive median TC 304
Measured
N = 74 SPT/CPT
Mean 4 838 kN
Mean 17 097 kN
N=147
Average = 4 704 kN
(200 - 23 800 kN)
Figure 26. Bearing capacity of footing.
Figure 24. Design angle of cutting v. experience of respondents.
TC 304
9 SAND 1: VERTICALLY LOADED SPREAD Measured
FOOTING Mean 2 427 kN
9.1 The problem (Figure 25) Derived from
Briaud & Gibbens (1999)
3 000 kN
P
N=140
• 2.25m square concrete footing Average = 2 313 kN
founded on sand at 0.75m 200 - 23 270 kN)
n.g.l. below ground level
• Concentric applied load P
excludes weight of footing
0.75m
1.2m
2.25m square
Figure 25. SAND 1: Vertically loaded square footing.
Respondents were asked to predict the maximum load that the
footing could carry before failing in bearing and the load that
would induce 25mm settlement.
This problem is based on full-scale load tests carried out on Figure 27. Load for 25 mm settlement.
five spread footings on sand at the Texas A&M National
Geotechnical Experimentation Site in 1993 (Briaud 2021b;
Briaud & Gibbens 1997 & 1999). These tests were carried out on Exclusive median
square footings with side lengths of 1.0m, 1.5m, 2.5m, and 3.0m N = 147
(×2). TC 304, SPT / CPT
9.2 Responses
TC 304, Measured
The predicted bearing capacity is shown in Figure 26, and the Measured
predicted load for 25mm settlement in Figure 27. Figure 28
shows the bearing capacity predictions v. experience.
9.3 Analysis of responses
Most respondents assumed a frictional material with no cohesion.
Where cohesion was assumed, it was generally small (<5 kPa).
Figure 29 shows predicted bearing capacity plotted against the
assumed friction angle of the sand. The two curves plotted on this
figure are the results of bearing capacity calculations using
Meyerhof (1951, 1963 & 1965) and EN1997-1:2004, Annex D.
The difference between these curves is due mainly to differences
in the shape and depth factors used by the two methods. Figure 28. Bearing capacity v. experience.
TC 304 (Yang & Ching 2020) derived two sets of results, one From Figure 29, most respondents have underestimated the
based on measured values of and the other with values
derived from SPT and CPT data. These two values are indicated bearing capacity of the footing, some seriously so. The isolated
by the two crosses in Figure 29. Both agree reasonably well with points that plot above the “Meyerhof” line are probably due to
the Meyerhof and EC 1997 calculations. the inclusion of cohesion in the calculation. The load that will
cause 25mm settlement (P25mm) given in Figure 27 also appears
to be underestimated by many respondents. To check whether
there is a correlation between low Pult and low P25mm values, the
ratio P25mm / Pult was calculated for each response. This ratio
varied from 0.001 to 3.15 with an average of 0.59. If the bearing
517
capacity is taken as the maximum load that the footing can carry, method of Archer and Heymann (2015) and the stiffness
values of P25mm / Pult greater than 1.0 do not make sense. Values degradation curve proposed by Rollins et al. (1998). The result is
at the low end of the range should also have raised concern in the shown by the solid blue line in Figure 31. It compares well with
the mean of results of the full-scale load tests in Figure 30 for a
mind of the respondent. footing width of 2.25m.
20000 Figure 31. Settlement calculated from G0 values using Archer and
Heymann (2015) and Rollings et al. (1996).
18000
Meyerhof
16000
Bearing capacity (kN)
EN 1997-1 0.8
14000
TC 304
12000
0.6
Pressure (MPa)
10000
8000
0.4
6000 Briaud & Gibbens
inferred from Fig. 30
4000
2000
0.2 Archer & Heymann
/ Rollings et al.
0
20 25 30 35 40 45 0
Friction angle (deg) 0 10 20 30 40 50
Settlement (mm)
Figure 29. Bearing capacity v. friction angle ( ).
Briaud and Gibbens (1999) normalised the load-settlement 10 SAND 2: AXIALLY LOADED PILE
results from the five full-scale load tests on the square footings
referred to in Section 9.1 by dividing the measured settlement by
the width of the foundation as shown in Figure 30. All five sets 10.1 The problem (Figure 32)
of normalised results plotted in a narrow band. Although the
footing size used in the SAND 1 problem (2.25m × 2.25m)
differed from the size of the footings tested in 1993, Figure 30 n.g.l.
can be used to derive the load-settlement curve for the SAND 1
footing (Figure 31). The bearing pressure causing 25mm • 0.76m diameter, bored,
settlement (600 kPa) corresponds to an applied load of 3 000 kN cast in situ pile in sand.
which is about 30% higher than the average value from the Installed under slurry.
responses in Figure 26 and the TC 304 value based on measured 0.76m
10.0m
values. • Depth 10.0m below
natural ground level.
1.8
1.5 1m
2.5m
Pressure (MPa)
Figure 32. SAND 2: Axially loaded pile.
1.2
3m 1.5m Respondents were asked to predict the ultimate load capacity of
0.9 the pile, the base and shaft resistance, and the load-settlement
3m curve up to 80% of the pile capacity.
This problem is based on full-scale tests carried out on drilled
0.6 shafts in sand at the Texas A&M National Geotechnical
Experimentation Site in 2006 (King et al. 2009; Briaud 2021b).
0.3
10.2 Responses
0
0 0.02 0.04 0.06 0.08 0.10 The pile capacity, base capacity and load-settlement curves are
Settlement / Width, S/B shown in Figures 33, 34 and 35. The pile capacity responses for
the various experience categories are given in Figure 36.
Figure 30. Normalised pressure-settlement for five footings on sand
(Briaud & Gibbens 1999).
10.3 Analysis of responses
When performing a routine settlement calculation, such as for
the footing in question, many geotechnical engineers would The average pile capacity predicted by the respondents
select an appropriate elastic modulus based on the available (3 039 kN) agreed well with the measured pile capacity from the
information and perform a simple elastic settlement calculation. Texas A&M National Geotechnical Experimentation Site
Most of these solutions give a linear relationship between load (2 937 kN). However, the range of predicted capacities is
and settlement which, as illustrated by Figure 30, is clearly not extremely high, anywhere from 0.14 to 2.9 times the measured
the case for the entire curve, but reasonable up to 50% of the value. Forty six out of 82 respondents underpredicted the pile
ultimate load. For more complex problems, finite element load capacity. It is interesting to note that the group that predicted
analysis may be carried out incorporating a constitutive model the highest pile capacity on average was the group with the most
for which the necessary parameters can be determined or experience. The average prediction for this group was over
estimated. However, for this problem (SAND 1) and the data 40% too high.
available, a simple calculation based on small-strain shear
stiffness (G0) provided the closest approximation to the test
results shown in Figure 30.
The small-strain shear stiffness (G0) data gives an average
value of G0 100 MPa dropping to 60 MPa between 7m and
9m. This small-strain stiffness profile was analysed using the
518
Measured TC 304 Deterministic Exclusive median
2 937 kN 3 644 kN N = 82
TC 304 Probabilistic
3 965 kN
N = 82
Avearge 3 039 kN
(429 - 8450 kN)
Measured
Figure 33. Predicted ultimate capacity of pile. Figure 36. Predicted pile capacity v. experience.
The deterministic and probabilistic predictions by TC 304
(Shuku & Hachich 2020c) in Figure 33 were respectively 24% and
35% higher than the measured resistance.
The load-settlement curves also showed significant variation,
TC 304 with many respondents predicting a linear load-settlement
41% response.
N = 78
Avearge 58%
11 SAND 3: SPREAD FOOTING WITH
HORIZONTAL LOAD
11.1 The problem (Figure 37)
H Qwind 50kN
V G Qwater
320kN 100kN
4.5m
n.g.l.
Figure 34. Ratio of base capacity to pile capacity.
0.75m
Load (kN)
L=B
0 1000 2000 3000 4000
0
• Foundation is for a water tank stand
20 • Vertical load V includes dead load G excluding
weight of footing) and weight of water in tank Qwater.
Settlement (mm)
40 • Horizontal load H is due to wind only Qwind. Can act
in any horizontal direction.
60 • Loads are characteristic values.
80
Figure 37. SAND 3: Square footing with horizontal load.
100
Measured This is a design problem in which the respondents were asked to
N = 59 determine the required size of a base for a water tank stand.
120
Figure 35. Load-settlement predictions.
519
11.2 Responses From the 71 respondents who stated the friction angle assumed
in their calculations, the average friction angle was 34o, varying
Ninety-nine responses were received. The size of the footing from 26o to 44o. Unfortunately, it is not possible to determine
required by the respondents is given in Figure 38, the variation from the responses whether these are average values or
in required size with friction angle in Figure 39 and the characteristic values.
experience of the respondents in Figure 40. Forty-two respondents determined the footing size for the
“tank empty” condition and 49 for the “tank full” condition.
EN 1997-1 11.3 Analysis of responses
DA1 ALL RESULTS
' = 34o N = 99 TC 304 did not submit an analysis of this problem, as current
Average = 2.66m knowledge about the model uncertainty for the design method of
(0.8 - 8.2 m) spread footings with horizontal load is limited (Ching & Zhang
2020). Furthermore, there is no corresponding load test. The
EN 1997-1 responses received were therefore compared to an analysis
EQU performed using Design Approach 1 from EN1997-1:2004,
together with the default values of partial load and material
factors and load combination factors from EN1990:2002 and
EN1997-1:2004, following the model solution for a similar
problem by Orr (2005). This comparison is shown in Figure 39.
The above analysis shows that the STR (DA1 C1) and GEO
(DA1 C2) design approaches yield very similar results. For all
but the lowest friction angle considered (25o), the tank empty
condition dictates the required size of the base. The relatively
small variation in base size with friction angle is due to the
dominating effect of eccentricity of load on the calculation. The
minimum base size required to satisfy the EQU limit state is
1.91m. If unfactored values are used, the base will overturn if the
base is smaller than 1.24m.
Figure 38. Required footing size (all results – tank full and tank empty). Based on Figure 39, most of the respondents overestimated the
required size of the base.
10.0
EN 1997-1 EQU n=71
EN 1997-1 DA1 12 SAND 4: PERMANENT ROAD CUTTING
8.0
Footing size L=B (m)
12.1 The problem (Figure 41)
n.g.l.
6.0
• Permanent
4.0 6m cutting cutting for a
public road.
• No surcharges.
2.0
1m deep toe drain
0.0 Figure 41. SAND 4: Permanent road cutting.
25 30 35 40 45 This is a design problem in which respondents were requested to
Friction angle (deg) provide the angle at which the slope should be constructed.
Figure 39. Variation in required size with friction angle. 12.2 Responses
Sixty-nine responses were received. The required slope angles
are shown in Figure 42, the slope angle required for various
Exclusive median friction angles in Figure 43 and the slope angle versus experience
N = 99
of the respondents in Figure 44.
Where working load design methods were used, the average
value of the factor of safety used by the respondents was 1.5,
ranging from 1.2 to 2.0.
EN 1997-1: DA1 The average friction angle was 33.8o varying from 25o to 41o.
' = 34o
12.3 Analysis of responses
The solid red line in Figure 43 is from a Morgenstern-Price slope
stability analysis (Morgenstern & Price 1965) performed using
commercial software (Geostudio SLOPE/W, version 2021 R2)
with a factor of safety of 1.5. This line is virtually
indistinguishable from the result of an infinite slope analysis with
EN 1997-1: EQU a factor of safety of 1.5, i.e. = arctan (tan /1.5).
For the average friction angle used by the respondents of 33.8o,
the average slope angle of 26o (from Figure 42) corresponds to a
factor of safety of 1.37. This is a reasonable value for a slope
where the failure surface is likely to be shallow and the quantity
of material involved in the failure will be small.
Figure 40. Required footing size v. experience.
520
a slope angle of 17.2o resulted in a probability of failure of 1.10-3
and that the slope would have to be flattened to less than 14 o to
TC 304 TC 304 achieve a failure probability of 1.10-4.
17.2o 28o
Pf = 0.001 FoS = 1.5
13 SAND 5: PROPPED EMBEDDED RETAINING
TC 304 WALL
21.6o N = 69
Pf = 0.01 Average = 26o 13.1 The problem (Figure 45)
(16 - 45o)
This is a design problem in which respondents were requested to
determine the required depth of embedment (d), the propping
force (P) and the maximum bending moment in the wall (Mmax).
n.g.l. P 20kPa traffic loading
6m • Permanent retaining
Figure 42. Required slope angle. wall propped at top.
50
• 6.0m deep excavation.
excavation level
N = 53 • 20kPa variable loading
45 (surcharge) due to
d traffic.
40
Slope Angle (deg)
35
30 Figure 45. SAND 5: Propped embedded retaining wall.
25 13.2 Responses
Based on the responses received, the required embedment depth
20 is shown in Figure 46, the propping force in Figure 47 and the
minimum required bending resistance of the wall in Figure 48.
15 The average friction angle used by the respondents in their
20 30 40 50 calculations was 34o. Again, there is insufficient information to
ascertain if this is an average value or a characteristic value.
Friction angle (deg)
Figure 43. Variation in required slope angle with friction angle.
TC 304 TC 304
2.6 m 3.3 m
EN 1997-1: DA1 GEO
TC 304 4.0m
FoS = 1.5 Free earth support
'k = 34o
N=42
Average 4.1 m
(0.45 - 11.2 m)
Morgenstern-Price
FoS = 1.5
Figure 46. Required embedment depth.
Figure 44. Experience of designers.
It is interesting to note that three respondents assumed a slope
angle equal to the friction angle (i.e. = ). This implies a factor
of safety of 1.0. Two respondents selected slope angles above the
friction angle. For such slopes to be stable, some cohesion must
be present.
Although a slope angle of 16o (minimum submitted value) may
seem low, the analyses by TC 304 (Jiang et al. 2020b) found that
521
TC 304 EN 1997-1: 2004 n.g.l. 10kPa construction traffic
34/38 kN/m DA1 GEO • Temporary soil nailed wall.
-1.4m
106 kN/m • 10kPa variable loading
Free earth support -2.9m
(construction traffic).
'k = 34o 6m • 10o nail inclination.
-4.4m • 20mm nails: Fy = 140kN.
• All nails L m long.
N=46
Average 109 kN/m
(23 - 415 kN/m) Figure 49. SAND 5: Propped embedded retaining wall.
14.2 Responses
From the replies of 38 respondents, the required nail length, the
horizontal spacing of the nails and the total length of nails per
metre length of face are given in Figures 50 to 52.
TC 304
Figure 47. Required propping force. SAICE 1989 N = 38
Pf =10-4
5.5 m Average 7.1 m
4.9 m
(3.7 - 12.0 m)
EN 1997-1: 2004
DA1 GEO
324 kNm/m
Free earth support
'k = 34o
N=35
Average 449 kNm/m
(110 - 2100 kNm/m)
Figure 50. Required length of nails.
TC 304
Figure 48. Required minimum moment capacity of wall. Pf =10-3 SAICE 1989
1.4 m 1.5 m
13.3 Analysis of responses
N = 38
No full-scale load test on a similar structure has been carried out. TC 304
TC 304 (Shuku et al. 2020) performed a deterministic and Average 1.47 m
Pf =10-4
probabilistic analysis of the depth of embedment and the (0.8 -3.0 m)
1.3 m
propping force required. Their responses are shown by the
dashed red lines in Figures 46 and 47.
For a further comparison, a limit states design was carried out
based on Design Approach 1, Combination 2 (GEO) from
EN1997-1:2004. The average characteristic value of the friction
angle ( ’k) was taken as 34o, equal to the average friction angle
used by the respondents. This calculation, which made use of the
free-earth support method, was based on the model solution for
a propped cantilever wall prepared by Orr (2005). The results are
shown by the solid black lines in Figures 46 to 48.
There is insufficient data to allow meaningful comparison of
the responses in terms of experience.
Figure 51. Horizontal spacing of nails in each row.
14 SAND 6: SOIL NAILED WALL
14.3 Analysis of responses
14.1 The problem (Figure 49) The construction of soil-nailed walls is a highly competitive
market in South Africa and full advantage is taken of the typically
This is a design problem in which respondents were requested to partially saturated, relatively stiff soil profile. Design is carried
determine the required length of soil nails (L), and the maximum out in accordance with a local lateral support code (SAICE 1989)
horizontal spacing (SH). using a single- or double-wedge analysis with a FoS of 1.5 on
wedge stability and a FoS of 2.0 on the bond stress. This method
of analysis has been used for comparison purposes for this
problem as shown by the solid black lines in Figures 50 to 52.
The analysis again used the average friction angle of 34o obtained
522
from the responses received. In terms of total meterage of soil serviceability in geotechnical design. There are thus three things
nails, the results compare favourably with those from the that are being assessed: (1) the selection of design parameters
analyses by TC 304 (Cao et al. 2020). from test results, (2) the methods of analysis (calculation models)
As correctly pointed out by TC 304, the nail length and the nail used, and (3) the provision made for safety/serviceability. The
spacing are not independent. Wider spacings require longer nails. results of the survey would have been easier to analyse had the
However, the nail spacing and length of nails can be combined survey been split into three sections, dealing with each of the
into a single variable for comparison purposes, namely the total above issues. However, that would have reduced the survey to
length of nails per metre of face as shown in Figure 52. The the level of an undergraduate assignment and would not have
average length of nails per metre from the survey (15.9m) is assessed the full range of design solutions put forward by the
greater than required. respondents to this survey.
15.2 Assessment of responses
SAICE 1989
11.0 m/m TC 304 Probably the most surprising result of this survey is the wide
Pf =10-4 range of responses to almost every problem. The obvious starting
TC 304 11.3 m/m point for any analysis would be to compare the predictions and
Pf =10-3 N=38 the designs to known outcomes based on either measurement or
10.5 m/m Average 15.9 m the application of established theory. However, as stated above,
(5.8 - 20 m) geotechnical engineering is not an exact science. Even where we
have measured values from full-scale tests on the sites from
which the data originated, these are typically single test results
which may either be below or above the mean of a larger number
of test results. Nevertheless, a best estimate of the result is a
better starting point than no estimate at all.
To this end, the measured values from full-scale tests on the
National Geotechnical Experimentation Site have been adopted
as best estimates. Where no test results are available, the results
from calculations based on widely accepted codes or calculation
models have been used. The input values to these calculations
have been taken as the mean value of the parameters given in the
responses received. The authors do not contend that the best-
estimate values are “correct”. The complexity of geotechnical
design and the uncertainties involved make it impossible to say
Figure 52. Total length of nails per metre length of face. what is correct or incorrect. These values are simply a starting
point for the rational assessment of the results.
For this specific problem, an increase in nail spacing results in Table 1 summarises the responses received, grouping them
an increase in the length of nails but an overall decrease in the into predicted values (load capacity, load for 25mm settlement,
length of nails per metre length of face. However, there is a etc.), design dimensions (size of footing, slope angle, etc.) and
practical maximum spacing of nails, typically 1.5m in sandy soils parameter values (cohesion and friction angle). In the final three
such as present on the sand site. Spacings of 3m (maximum from columns of Table 1, the range (minimum and maximum) and
responses received) may result in less total meterage of nails but mean (average) values from the responses have been normalised
is likely to give problems during construction. The average by dividing by the best-estimate value.
spacing of 1.47m is closer to the norm.
The rule of thumb for the length of soil nails is 0.7× – 0.9× the Table 1. DEM input parameters
height of the face. Designs using shorter soil nails are often Result Best Estimate Responses Normalised
governed by limiting bond stresses, while designs with longer Problem Parameter
Prediction
Estimate Source Min Mean Max Min Mean Max
nails may be governed by the yield strength of the nail. In the CLAY 1 Pult 1154 kN/m Mean TC 304 43 1127 8872 0.04 0.98 7.69
SAICE and TC 304 designs, the bar lengths are 0.92 and 0.82 CLAY 1 P25 650 kN/m TC 304 18 538 2413 0.03 0.83 3.71
times the height of the face respectively. The average length of CLAY 2
CLAY 3
Pult
Pult
2033 kN
409 kN
Measured
Measured
75
25
2460
446
12000
1700
0.04
0.06
1.21
1.09
5.90
4.16
nails from the responses received (7.1m) is higher than is SAND 1 Pult 6235 kN EN1997 ' = 34 o
200 4704 23800 0.03 0.75 3.82
normally expected. SAND 1 P25 2900 kN Measured 200 2313 23270 0.07 0.80 8.02
SAND 2 Pult 2937 kN Measured 429 3039 8450 0.15 1.03 2.88
Design dimension/slope
15 DISCUSSION CLAY 4 undrained 34 deg TC 304 P f = 10-4 14 40.9 90 - 1.20 -
SAND 3 L=B 2.07 m EN1997 ' = 34o 0.8 2.66 8.2 0.39 1.29 3.96
o
15.1 Correctness and adequacy of geotechnical data SAND 4
SAND 5 d
24 deg
4m
Inf. slope
EN1997
' = 34
' = 34o
16
0.45
26
4.1
45
11.2
-
0.11
1.08
1.03
-
2.80
' = 34o
Some respondents questioned the adequacy and correctness of SAND 5
SAND 5
P
Mreq
106 kN/m EN1997
324 kNm/m EN1997 ' = 34o
23
110
109
449
415
2100
0.22
0.34
1.03
1.39
3.92
6.48
the data. In particular, the CPT data was subject to much scrutiny, SAND 6 m/m 11 m/m SAICE ' = 34o 5.8 15.9 20 0.53 1.45 1.82
and inconsistencies in the data were pointed out.
It should be borne in mind that some of the data comes from Material property
CLAY 1 c u 111 kPa TC 304 CLAY 1&4 51 98 199.3 0.46 0.88 1.80
the 1980s (Briaud 1997) when testing methods were less CLAY 3 cu 111 kPa TC 304 CLAY 1&4 65 98 150 0.59 0.88 1.35
advanced than they are today, and digital recording of data was CLAY 4 cu 111 kPa TC 304 CLAY 1&4 41 77 150 0.37 0.69 1.35
non-existent. As a result, the data was digitised from the plotted SAND 1
SAND 3
'
'
35.6
35.6
deg
deg
TC 304 mean
TC 304 mean
23.4
26
33.8
34
44
41
0.66
0.73
0.95
0.96
1.24
1.15
soil test results extracted from published papers, as the original SAND 4 ' 35.6 deg TC 304 mean 25 33.8 41 0.70 0.95 1.15
data was not available in modern electronic form.
The compilers of the survey acknowledge discrepancies in the The most significant outcome is that the average value of the
data. However, geotechnical engineering is not an exact science. responses typically falls within a range of 30% above or below
In real life, geotechnical designers seldom have the quantity and the best estimate value, i.e. the normalised mean in the second
quality of data they would like, and they have to make do with last column typically lies between 0.7 and 1.3.
what is available. This survey is thus a better reflection of As far as the range of responses is concerned, the smallest
conditions in geotechnical engineering practice than it would range was in the material properties assessed from the available
have been with a well-manicured set of test results. geotechnical data (cu for clay and ’ for sand). The likely reason
Having said that, it is worth reflecting on the effect on the for this small range is that most experienced respondents will
outcome of the survey of both the data itself and interpretation of have a feel for the values of these parameters. Some results are
the data. As stated in the introduction, the survey concerned two regarded as highly optimistic or pessimistic, but very few are
types of problems prediction and design. The prediction highly improbable. It is noted that the material properties
problems were aimed at assessing the selection of parameters and submitted by the respondents include both best-estimate values
calculation models. The design problems were aimed and characteristic values. In the context of the problems included
additionally at assessing the provision made for safety and in this survey, characteristic values are typically lower than best-
523
estimate values. This will tend to skew the average and the range Confidence Tshilande in the African Region; Chriswini Tanaka,
of the responses towards the lower end of the scale. This is a Aswin Lim, Rara Noviarti, Chung Siung Choo, Dongsheng
shortcoming of the survey, but one which could be addressed by Chang, Baby Lavanya and Pan How Yen in South-East Asia;
further analysis of the results. Dena Gabbassova in Australasia; Leanne McLaren and Kshama
The next lowest range was for design dimensions. Once again,
most experienced respondents will have a feel for the correctness Roy in Canada; Alexandre Lopes, Ahmed Hamza Mridakh, Jean
of a result. Nevertheless, the design dimensions submitted de Sauvage and Minh Tuan Hoang in the Francophone countries;
include values that are highly improbable, being either unsafe or and Mario Terceros Arce in South America. These are the names
excessively conservative. we are aware of, but there are sure to have been many more.
The highest range of responses was for the prediction Thank you to all of you.
problems, all of which relate to load capacity. This is somewhat Thanks to Bengt Fellenius and Sergio Augustin for their
surprising as prediction of load capacity includes only two main detailed comment and analysis of the soil test data.
categories of uncertainty, namely parameter uncertainty and Thanks to the eminent persons who participated in the online
calculation model uncertainty, while the assessment of design sessions, including Harry Poulos, John Carter, Paulus Pramono
dimensions includes a third category of uncertainty, namely
provision for safety. Despite the average values of the predictions Rahardjo, Ng Tiong Guan, Alejo Sfriso, Roger Frank, Fahd
being reasonable estimates, the extremes in virtually all cases are Cuira, Jér me Racinais and Julien Habert, Derek Martin, Dennis
highly improbable. Becker, Nicol Chang, and Hany El-Naggar.
This effort was also supported by numerous member societies
including the Canadian Geotechnical Society, Geotechnical
Society of Singapore, Indonesian Society for Geotechnical
16 CONCLUSIONS Engineering, Hong Kong Geotechnical Society, Malaysian
Geotechnical Society, French Society for Soil Mechanics and
As a result of the very wide range of responses to almost every Geotechnical Engineering, Moroccan Society for Soil Mechanics
problem in the survey, it is difficult to provide a clear answer to
the question, “Are we overdesigning?” The only problem where and Geotechnical Engineering, Australian Geomechanics
there appeared to be a clear case of overdesign (underprediction Society and the Geotechnical Division of SAICE (South Africa).
of capacity) was the prediction of the failure load for problem Sponsorship was received from Thurber Engineering,
SAND 1 (vertically loaded spread footing). This is surprising, as Northern Crescent, Tencate Geosynthetics, Terrasol, Menard,
the calculations involved are straight-forward and the calculation Coffey (now Tetra Tech Coffey), Gamuda Australia, Franki
models are well established. There was also a tendency towards Africa (now Keller) and Jones & Wagener. Thank you.
overdesign of the base size required for SAND 3 (footing on sand
with horizontal load). This is more understandable, as the
problem is more complex and the methods of calculation less
clearly defined. In all other cases there was evidence of both 18 REFERENCES
overdesign and underdesign.
A point of concern is the apparent failure, even among Archer, A., and Heymann, G. 2015. Using small strain stiffness to predict
experienced designers, to consider potentially critical load cases the load settlement behaviour of shallow foundations on sand.
or recognise clearly incorrect predictions. Examples are the Journal of the South African Institution of Civil Engineering, 57(2),
failure to consider the possibility of drained slope failure in over- 28 35. DOI: 10.17159/2309-8775/2015/v57n2a4.
consolidated clays (CLAY 4, permanent slope cut), critical load Bierschwale, M.W., Coyle, H.M., and Bartoskewitz, R.E. 1981. Field
conditions (e.g. tank empty in problem SAND 3) and critical tests and new design procedure for laterally loaded drilled shafts in
limit states (e.g. the EQU limit state in SAND 3). clay. Report FHWA/TX-81/09+211-3F. Federal Highway
In all cases where measured values are available, accepted Administration, USA.
methods of analysis yielded similar values. The wide range of Briaud, J-L. 1997. The national geotechnical experimentation sites at
responses received points to two possibilities. The first is that Texas A&M University: clay and sand – a summary. Report NGES-
established methods of analysis are not being applied or are being TAMU-07-September 1997.
incorrectly applied. The second is that less attention was paid to Briaud, J-L. 2021a. 40 Years of full-scale infrastructure testing at a
the analysis of the survey problems than would be paid to national geotechnical experimentation site: clay site. International
geotechnical design in practice, with less detailed checking of Journal of Geoengineering Case Histories, 6(3), 1 24, doi:
calculations, and less critical scrutiny of the final result. It is 10.4417/IJGCH-06-03-01.
concerning that simple checks could have ruled out many of the Briaud, J-L. 2021b. 40 Years of full-scale infrastructure testing at a
improbable responses. national geotechnical experimentation site: sand site. International
Much more can be done with the data collected during this Journal of Geoengineering Case Histories, 6(3), 25 53, doi:
survey. This is why this article was written and why the data is 10.4417/IJGCH-06-03-02.
available on the websites cited in the article. Briaud, J-L., and Gibbens, R. 1997. Large scale load tests and database
The results of this survey clearly indicate that the ISSMGE of spread footings on sand. Report FHWA-RD-97-068, Federal
should pay as much attention to the application of existing Highway Administration, USA.
knowledge as to the creation of new knowledge. More of our Briaud, J-L., and Gibbens, R. 1999. Behavior of Five Spread Footings in
technical activities should involve practitioners and be aimed at Sand. Journal of Geotechnical and Geoenvironmental Engineering,
improving the state of practice within the industry. 125(9), 787 797.
Broms, B.B. 1964. Lateral resistance of piles in cohesive soils. ASCE J.
Soil Mechanics & Foundations 90(SM3), 27 63.
17 ACKNOWLEDGEMENTS Byrne, G., Chang, N., and Raju, V. 2019. A guide to practical
geotechnical engineering in Africa, 5th edition. Franki Africa,
The authors and organisers of the survey express thanks to all Johannesburg, South Africa.
respondents who gave their valuable time to participate in the Cao, Z-J., Gao, G-H., Shuku, T., Schweckendiek, T., and Redaelli, M.
survey. 2020. Probability-based analysis and design of soil nail wall.
Thanks is also due to the CAPG and YMPG for their guidance Probabilistic solutions for survey questions in “Are we
and support throughout the survey. Lucy Wu and Sukumar overdesigning? – a survey of international practice”, 75 84.
Pathmanandavel were key role-players. Thanks, too, to the Report by ISSMGE TC 304.
Ching, J., and Zhang, J. (editors) 2020. Probabilistic solutions for survey
members of TC 304 under the leadership of Jianye Ching for questions in “Are we overdesigning? – a survey of international
their very valuable analysis and report. The authors of the various practice”. Report by ISSMGE TC 304. DOI 10.53243/R0002.
sections of this report are acknowledged in the references. Flemming, W.G.K. 1992. A new method for single pile settlement
We also thank the organisers of the survey in the various prediction and analysis. Géotechnique, 42(3), 411 425.
regions and their team members. These include Jacobus Breyl, EN1990:2002. Eurocode: Basis of structural design. European
John Pavlakis, Bret Markides, Denis Kalumba, Linda Spies and Committee for Standardisation (CEN), Brussels.
EN1997-1:2004. Eurocode 7: Geotechnical design – Part 1: General
524
rules. European Committee for Standardisation (CEN), Brussels.
Jiang, S-H., Liu, X., and Ching, J. 2020a. Reliability-based design of
spatially varying cohesive slope. Probabilistic solutions for survey
questions in “Are we overdesigning? – a survey of international
practice”, 28 38. Report by ISSMGE TC 304.
Jiang, S-H., Liu, X., and Ching, J. 2020b. Reliability-based design of
spatially varying sandy slope. Probabilistic solutions for survey
questions in “Are we overdesigning? – a survey of international
practice”, 56 65. Report by ISSMGE TC 304.
King, P., Fernandez, A., and Pando, M.A. 2009. Post grouted drilled
shafts – a comprehensive case history from Texas. Proc. 2009
International Foundation Congress and Equipment Expo, 31 38,
ASCE.
Meyerhof, G.G., 1951. The ultimate bearing capacity of foundations.
Géotechnique, 2(4), 301 331.
Meyerhof, G.G. 1963. Some recent research on the bearing capacity of
foundations. Canadian Geotech. Journal, 1(1).
Meyerhof, G.G. 1965. Shallow foundations. ASCE Journal Soil Mech.
and Found. Div., 91(2), 21 31.
Morgenstern, N. R., and Price, V. E. 1965. The analysis of the stability of
general slip surfaces. Géotechnique, 15(1), 77 93.
Orr, T.L.L. 2005. Model solutions for Eurocode 7 – workshop examples.
Proc. international workshop on the evaluation of Eurocode 7,
Trinity College, Dublin, 31 Mar – 1 Apr 2005, 75 108.
Rollins, K.M., Evans, M.D., Diehl, N.B., and Daily, W.D. 1998. Shear
modulus and damping relationships for gravels. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 124(5)
396–405.
SAICE 1989. Lateral support in surface excavations, Code of practice
1989. South African Institution of Civil Engineering, Johannesburg.
Shuku, T., and Hachich, W. 2020a. Probabilistic analysis of axially
loaded pile in clay. Probabilistic solutions for survey questions in
“Are we overdesigning? – a survey of international practice”, 11
19. Report by ISSMGE TC 304.
Shuku, T., and Hachich, W. 2020b. Probabilistic analysis of a laterally
loaded pile in clay. Probabilistic solutions for survey questions in
“Are we overdesigning? – a survey of international practice”, 20
27. Report by ISSMGE TC 304.
Shuku, T., and Hachich, W. 2020c. Probabilistic analysis of axially
loaded pile in sand. Probabilistic solutions for survey questions in
“Are we overdesigning? – a survey of international practice”, 47
55. Report by ISSMGE TC 304.
Shuku, T., Cao, Z., Schweckendiek, T., and Redael, M. 2020.
Probabilistic analysis of propped embedded retaining wall.
Probabilistic solutions for survey questions in “Are we
overdesigning? – a survey of international practice”, 66 74.
Report by ISSMGE TC 304.
Stroud, M.A. 1989. The Standard Penetration Test Its Application and
Interpretation. Institution of Civil Engineers Conference on
Penetration Testing in the U.K., Birmingham. Thomas Telford
Limited.
Yang, Z., Xue, J., and Ching, J. 2020. Probabilistic analysis of vertically
loaded strip footing on clay. Probabilistic solutions for survey
questions in “Are we overdesigning? – a survey of international
practice”, 1 10. Report by ISSMGE TC 304.
Yang, Z., and Ching, J. 2020. Probabilistic analysis of vertically loaded
square footing on sand. Probabilistic solutions for survey questions
in “Are we overdesigning? – a survey of international practice”, 37
46. Report by ISSMGE TC 304.
525
ANNEXURE A: CLAY SITE PROPERTIES
A1 Laboratory tests
Figure A3. CLAY: undrained shear strength.
Figure A1. CLAY: moisture content and density.
A2 Field tests
Figure A4. CLAY: CPT point and friction resistances.
Figure A2. CLAY: grading and over-consolidation ratio.
526
ANNEXURE B: SAND SITE PROPERTIES
B1 Laboratory tests
Figure B1. SAND: grading analysis.
Figure A5. CLAY: Pressuremeter modulus and limit pressure.
Figure B2. SAND: SPT tests.
V G0
Figure A6. CLAY: SPT tests
Figure B3. cross-hole seismic tests.
527
KD ID
ED
Figure B4. SAND: pressuremeter tests.
Figure B5. SAND: dilatometer tests.
528
Figure B6. SAND: CPTu tests.
529
Invited Papers
Artificial Intelligence
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Deep Learning for Rainfall Prediction: Data Enhancement, Forecasting, and
Estimation
Deep Learning pour la prévision des précipitations : amélioration des données, prévision et
estimation
Lei CHEN, Jiachuan WANG & Jia LI
Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Hong
Kong, China, [email protected]
Charles Wang Wai NG
Department of Civil Engineering, The Hong Kong University of Science and Technology, Hong Kong, China
Yanyan SHEN
Department of Computer Science and Engineering, Shanhai Jiaotong University, China
ABSTRACT: High-quality rainfall prediction with large spatial coverage draws lots of attention these years. In this paper, we discuss
our solutions to it using deep learning, which is a framework that covers data enhancement, temporal prediction, and spatial
estimation. As the rainfall data suffer from noise, we propose a self-supervised model for data cleaning. To predict the future rainfall
data based on previous timestamps, we combine polynomial activation functions to approximate the non-linear physical relationships
in the rainfall system. To acquire accurate rainfall information at locations without rain gauges, we propose a new model to adaptively
capture the complex spatial relationships. Experimental results show the effectiveness of our methods.
RÉSUMÉ : La prévision des précipitations de haute qualité avec une large couverture spatiale attire beaucoup d'attention ces dernières
années. Dans cet article, nous discutons de nos solutions à l'aide de l'apprentissage en profondeur, qui est un cadre couvrant l'amélioration
des données, la prédiction temporelle et l'estimation spatiale. Comme les données pluviométriques souffrent de bruit, nous proposons un
modèle auto-supervisé pour le nettoyage des données. Pour prédire les données de précipitations futures sur la base des horodatages
précédents, nous combinons des fonctions d'activation polynomiales pour approximer les relations physiques non linéaires dans le
système pluviométrique. Pour acquérir des informations précises sur les précipitations à des endroits sans pluviomètres, nous proposons
un nouveau modèle pour capturer de manière adaptative les relations spatiales complexes. Les résultats expérimentaux montrent
l'efficacité de nos méthodes.
KEYWORDS: Deep Learning, Rainfall, Prediction
1 INTRODUCTION Rain gauge data provide direct measurements of liquid
precipitation and are usually considered as the most reliable data
Rainfall prediction is very important for civil life, including the source for ground rainfall (WMO 2008). However, such accurate
prevention of natural disasters, management of infrastructure rainfall information is sparse in space since it is only available at
systems, and the maintenance of the safe environment (Swain et the sampled locations with installing rain gauges. Therefore,
al. 2018). spatial estimation is required to obtain accurate rainfall
From the perspective of methods, as a famous chaotic system, information for other unsampled locations. Spatial interpolation
the theories of rainfall are still not explicitly resolved such as is a general way to obtain spatial estimates. The performance of
convection (Zhang et al. 2017). Traditional theory-driven models traditional spatial interpolation approaches is often limited by
are computational inefficient with complex and empirical unrealistic settings, e.g., deterministic formulations or statistical
assumptions (Fritsch & Chappell 1980, Grell & Dévényi 2002, assumptions (Li & Heap 2008, Ma et al. 2019, Rusu, C. & Rusu,
Grell & Freitas 2014), such as the widely used WRF (Weather V. 2006). Recently, researchers have exploited deep learning
Research and Forecasting) model (Skamarock 2018). In recent methods for spatial interpolation. But due to the lack of pertinent
years, deep learning models have also applied to rainfall analysis and improvements for practical interpolation tasks, they
prediction (Shi 2017, Wang2017, Wang 2019). Though their cannot handle the rainfall spatial interpolation well.
performances beat the traditional methods, they treat the problem In order to conduct fine-grained and high-accuracy rainfall
as a computer vision task. In contrast, rainfall is closely prediction, we propose methods from the following directions:
correlated with other weather data, while the underlying physical 1. Data enhancement. As the type and scale of noise in
relationships are highly non-linear, which are omitted by these radar images are not available, we propose methods to
deep learning methods. denoise images only based on noisy observations.
On the other hand, low-quality data sources also impair the 2. Model design for radar images. To fully make use of the
prediction performance. As a widely used indicator for rainfall, heterogeneous data and explore the underlying physical
many papers predict radar echo images alternatively (Shi 2017, relationships of rainfall system, we propose more
Wang 2017, Wang 2019). These radar images suffer from system powerful polynomial approximators for deep learning
noise and observation noise, where both the scale and type of model.
noise are unknown. 3. Model design for rainfall spatial interpolation. From the
perspective of discarding subjective assumptions, we
535
Figure 1. General overview of our research directions.
propose the new model and correction mechanisms to distribution of noise in the term, we replace it with a tractable
effectively solve the practical rainfall interpolation task distribution so that Monte Carlo integration can be used to
in a data-driven style. estimate a relaxed bound. Using arrangement inequality (Hardy
et al. 1952), we finally get an estimation of the upper bound of
In the following sections, we first given an overview of each only based on noisy images. We take turn to update the
direction and our methods in Section 2. Then each component is parameters of model and estimate , where the framework
detailed illustrated in Section 3. Experimental results are given helps to train convergent model with good performance.
in Section 4, and we conclude our paper in Section 5. In the second step, our target is to design more powerful
modules for deep learning to extract the non-linearity in physical
2 GENERAL OVERVIEW system. We apply polynomial activation functions, which are
more furious for approximation and vulnerable to data flow
We demonstrate a flow chart to overview our directions and explosion. To train a model stably, we select Chebyshev
methods in Figure 1. polynomials as basis functions. Its output are restricted if the
The data we used are from three sources. Precipitation from input is restricted. Prepended normalization layers can help to
rain-gauges is accurate, but observation stations are sparsely limit the range of input. We further propose Range Norm to
built. It has bad coverage, e.g., 123 monitoring stations in HK. strictly control the range of input. The polynomial activation
Another one consists of radar echo images, which are Constant functions and Range Norm are applied to rainfall prediction
Altitude Plan Position Indicator (CAPPI) images with a good networks and improve the state-of-art results.
correlation with rainfall (Shi et al. 2017). It has 480×480 In the third step, we aim to develop an effective model to
resolution and 6-minute time interval. Besides, we also use the solve the rainfall spatial interpolation task in a data-driven
predicted weather data of traditional numerical model WRF for manner. Since the key issue of spatial interpolation is on
rainfall prediction, including water vapor mixing ratio, accurately estimating spatial relationships, we put our focus on
temperature, zonal wind speed, and meridional wind speed. The the dynamic learning of graph structures. We propose a model
resolution is 162×213 and interval is one hour. named GSI to learn the complex spatial relationships and
One can see that, rain-gauge data are accurate but with bad implement direct interpolation. To further improve the accuracy
coverage, where radar images and WRF are of high resolution of rainfall spatial interpolation, we design two correction
but are inferior indicator of rainfall. These data are also mechanisms as the post-processing to reduce errors.
vulnerable to noise, missing value, abnormal value, etc. To
achieve accurate rainfall prediction with high-resolution, our
framework first enhances the input data (i.e., denoising), then 3 METHODOLOGY
radar image and WRF data are used for prediction of fine-grained
radar images. The last stage is rainfall estimation, which converts In this section, we discuss the details of our methods.
accurate sparse rainfall from rain-gauge into high-resolution
rainfall. 3.1 Noise2Info for image denoising
In the first stage, we design denoising method for radar
images. As the information of noise is known and the clean radar In recent years, deep learning methods have been used for image
images are not available, we need to use self-supervised models, denoising, where the hardness is continuously increasing.
which are trained only based on the noisy images. More Supervised deep learning models (He 2016, Ronneberger 2015)
information of noise such as its standard deviation can help are trained on datasets with both clean and noisy images and beat
to improve the performance. We propose a method named the traditional methods. When the clean ground truth is not
Noise2Same to estimate the upper bound of and design available, Noise2Noise is proposed to train a model with noisy
training algorithm based it. A theoretical bound for is first image pairs (Lehtinen et al 2018). Self-supervised models are
given, while one term of it is still intractable. For the unknown further designed, so that denoising models can be trained only
based on noisy observation (Batson & Royer 2019, Krull et al.
536
2019). Additional information of noise is still important for better
, (5)
denoising results. For example, with the distribution of noise, or
called noise model, many methods can improve their
performance (Moran et al. 2020, Guo et al. 2019); Noise2Same where are -size noise maps sampled from .
(Xie et al. 2020) can be trained better with a tighter upper bound
of the supervised learning loss if the scale of noise (i.e., standard 3.1.4 Relaxation of pixel correlations
deviation of noise, ) is available.
Now the distribution and removed noise are both tractable. The
However, for radar images, we only have noisy observations.
The clean images are not available, neither do we know the noise existing problem is that, among each pair of and , it is
model nor the scale of noise. To handle such a case, we propose unknown how the pairs of values are matched. We apply the
methods to estimate the standard deviation of noise . A case where model is well-trained, so that model is expected to
training framework is further designed to use the estimation for remove larger value when the noise is larger. The result is a tight
the training of denoising models. lower bound of Eq. (5). Theoretically, for each pair of and
, let be the indices of values of in
3.1.1 A theoretical upper bound increasing order, and be those for . We
have:
First, we propose a preliminary upper bound for the , which
is based on the previous statement in Noise2Same (Xie et al. (6)
2020):
(1) It can be proved using the arrangement inequality (Hardy
1952).
Till now, we can estimate the lower bound of only based
where is the number of dimensions of each image. is the on the noisy images. With a set of maps of removed noise, we
noisy image, and is the clean ground truth. and are first get the distribution of . Through Monte Carlo
two tractable terms from Noise2Same, which can be calculated integration, we sample noise maps from and calculate the
only based on the noisy images. refers to our denoising model lower bound estimation of according to Eq. (6). Finally, all
and is the denoising output. these results are fed into Eq. (1) to bound .
The inequation cannot be directly applied for upper bound
estimation, where is unknown and term 3.1.5 Training and updating algorithm
(2) Once we estimate the lower bound of standard deviation of noise,
is still intractable. To estimate the upper bound of , we need , we can feed it into the loss function of Noise2Same (Xie et
to estimate the lower bound of al. 2020):
3.1.2 Estimation of intractable term , (7)
to replace where the closer to is, the loss
As the real image is not available, we further transform
function is more effective for training. One the other hand, if a
into the relation of “noise” removed by the model (i.e.,
model is more powerful for denoising, the distribution of
) and the real noise (i.e., ):
removed noise is closer to the real noise, so that the error from
Eq. (5) is smaller. A more powerful model is also supposed to
denoise larger value when the pixel has larger noise, which
(3) reduces the gap in Eq. (6). In conclusion, a well-trained model
leads to tighter estimation, where a good estimation can help to
train a better model.
can be obtained merely based on the noisy observation,
Inspired by this fact, we take turn to train the model and update
where the real noise is unknown. Existing denoising works
the . With gradient descent, model is trained with a fixed ;
(Batson & Royer 2019, Krull et al. 2019, Xie et al. 2020) assume
after fixed number of iterations, the newest model is used to
that values of noise are independent and identically distributed.
estimate a new . As all the estimation are lower bound of ,
The transformation help use to deal with one underlying noise
distribution instead of an intercorrelated image distribution. we update if the new estimation is smaller. We name our
framework as Noise2Info.
3.1.3 Transfer to tractable distribution
3.2 Pacs for approximation of non-linearity
According to Eq. (3), if the removed noise is closer to the real
To model the rainfall system for prediction, previous deep
noise, the is smaller. To lower bound it, we use the maximum
learning methods still use the idea of image processing without
likelihood estimation (MLE) of removed noises to get a
considering the physical relationships (Shi 2017, Wang 2017,
smaller estimation of original term. Wang 2019). In contrast, many well-known governing equations
To be more specific, given noisy images for estimation, are polynomials, where polynomials are commonly used in
the model generates maps of removed noise, each numerical approximation (Powell 2015) but hardly used for deep
contains values. Denote the set of all the values as , we learning. Neural networks mainly rely on the activation function
have the maximum likelihood estimation of as , where for approximation of non-linearity, which are mostly
the distribution is: asymptotically linear, and their first-order derivatives are
bounded. Some works (Lee & Jeng 1998, Liying et al. 2005,
(4) Lopez-Rubio et al. 2019) studies the polynomial activation
functions, where the networks cannot go deep. According to our
Then we can instead estimate Eq. (3) with: experiments, polynomial activation functions are powerful but
unstable, which cannot support a deep model. To conquer this
problem, we propose Pacs (Polynomial activation functions)
537
based on Chebyshev polynomials, where a normalization layer Besides, two types of correction mechanisms are proposed as the
called Range Norm is further designed for stable training. post-processing to reduce estimation errors.
3.2.1 Basis functions: Chebyshev polynomials 3.3.1 GSI model details
Activation functions inject non-linearity for the model, while the The GSI model adopts a 2-layer GNN architecture and can be
scale of its output is vulnerable to be too large, which suffers defined as follows:
from data flow explosion. For example, if we use a naïve
polynomial function as activator, an input leads (9)
to an output . The property makes the model hard to
converge. To handle this, a widely used method is appending
normalization layers. Layer normalization (Ba et al. 2016) and (10)
Batch normalization enforce a group of inputs to be of zero-mean
and unit-variance layer- and batch-wise. However, Pacs are much Here, is the adjacency matrix, is the adjacency
more aggressive and cannot converge even with appended matrix with added self-loops, is the identity matrix. and
normalizations. are the degree matrix of and . , that is, the input.
and are the layer-specific trainable transition matrix.
To train it stably, we use Chebyshev polynomials as basis
is the learnable mask. is a non-linear activation
functions. An important property that we want to make use of is
function. In GSI, two improvements are proposed: constraining
that, -degree can be constructed from a trigonometric form: the flow of message passing and adaptive graph structure
when input and , we have learning.
. This guarantees that if input is restricted to Constraining the flow of message passing. To avoid trai
, the output is restricted. Follow this idea, we do not ning nodes taking the labels as input, the adjacency matrix
append normalization layers but prepend them to restrict the without self-loops is used in the first layer. In the sec
input. The experiments show that network can train stably under ond layer, the adjacency matrix with added self-loops is
such a setting. used so that the aggregated value from the 1-hop neighbor
On the other hand, Chebyshev polynomials has first-order s can be used again.
derivatives up to in the range [-1, 1], which is effective for Adaptive graph structure learning. In the first layer, we
feature extraction compared with traditional activation functions. use learnable masks to adjust the graph structure. To mode
l the complex spatial relationships, the multi-head mechanis
3.2.2 Range Norm
m (Veli kovi et al. 2017) is adopted. is -th mask
If we prepend previous normalization layers to Chebyshev of the first layer, is the corresponding transition m
polynomial activators (e.g., Layer Norm), the input is still atrix. denotes the concatenation operation.
not strictly within but close to it. This motivates u 3.3.2 Correction Mechanisms
s to further propose a stricter normalization layer called R
ange Norm: To further improve the accuracy, we design two types of
correction mechanisms. One is self-correction, the other one is
residual correction. In self-correction, we add back interpolated
(8) values of test nodes to the graph as the input and utilize the
interpolated values of test nodes for retraining to construct a
where is the variance of input . Range Norm res better graph. In residual correction, we firstly obtain the
estimated values of training nodes by GSI and then calculate the
trict the input within and thus ensures a restricted corresponding residuals, then implement interpolation on the
output. residuals, finally leverage the estimated residuals to correct the
estimated results of test nodes.
3.3 GSI for rainfall spatial interpolation
The focus of spatial interpolation is on how to accurately 4 PRELIMINARY EXPERIMENTS
estimate the strength of spatial relationships between known
points and unknown points. It inspires us naturally think of a In this section, we show the results of preliminary experiments
form of data: graph, of which each location can be regarded as a for our methods.
node, and the spatial relationships are the edge weights between
the nodes. With the development of deep learning, Graph Neural 4.1 Image denoising
Networks (GNNs) are getting more and more attention due to
As we do not have the ground truth of radar images, we need
their great expressive power. More recently, researchers have
other benchmarks to validate the effectiveness of our Noise2Info.
started to apply GNNs for spatial interpolation tasks (Appleby et
We choose 3 benchmarks including ImageNet ILSVRC 2012 Val
al. 2020, Wu et al. 2021). These existing GNNs-based methods
(Russakovsky et al. 2015), Hanzi (Batson & Royer 2019), and
just follow the general learning framework and misplace their
BSD68 (Martin et al. 2001), which are commonly used in self-
focus on learning the contractive mapping (i.e., transition matrix)
supervised denoising works, (Xie et al. 2020, Batson & Royer
related to node attributes, rather than the graph structure (i.e.,
2019, Krull et al. 2019).
spatial relationships). However, the practical rainfall spatial
We use the Peak Signal-to-Noise Ratio (PSNR) as the metrics,
interpolation usually lacks abundant auxiliary attributes and
owns complex spatial relationships. which is a logtransformation of the ratio between the square of
To effectively handle rainfall spatial interpolation, we propose maximum value of a clean image and the mean-squared error of
the GSI (Graph for Spatial Interpolation) model which focuses itself and its noisy version :
on the dynamic learning of the graph structure and can
implement direct interpolation without auxiliary attributes. (11)
538
Larger PSNR corresponding to better denoising result.
We show the results in Figure 2. Among self-supervised
models, our method (black rectangle with red edge) outperforms
the state-of-the art on all the three datasets.
4.2 Precipitation forecasting
We use the dataset for precipitation forecasting from HKO (Hong
Kong Observatory) (Shi et al. 2017). As heavy summer rainfall
has great impact to civil lives and triggers natural disasters such
as landslide, we focus on the summer rainfall with WRF data
from 2011 to 2015. Shi et al. propose new metrics called
Balanced Mean Square Error (B-MSE) and Balanced Mean
Absolute Error (B-MAE), which have larger weights for the error
from heavy rains. To be more specific, according to the rain rate
of each pixel, a weight is given as:
Figure 3. The B-MSE of the models based on our Pacs and Range Norm
(12) compared with original models and various of activation functions and
normalization. LN refers to Layer Norm and RN for Range Norm.
Six baselines are compared: the piecewise cubic polynomial
and the B-MSE for true radar image series against predicted function (PF) (Alfeld 1984), the triangular irregular network
series is B-MSE and (TIN), Inverse Distance Weighting (IDW), Ordinary Kriging
the B-MAR is . refers to (OK), KCN (Appleby 2020) and IGNNK (Wu et al. 2021). Also,
the pixel of out the image of pixels at a series of four GSI variants are evaluated: (1) GSI, the basic solution
radar images with timestamps. without correction; (2) GSI-SC, GSI with self-correction; (3)
GSI-RC-G, GSI with residual correction by GSI; (4) GSI-RC-K,
GSI with residual correction by Kriging.
The results are shown in Figure 4. From the results, we can see
that our GSI model and its variants achieve good interpolation
performance on the two datasets. One finding caught our
attention. On the Hong Kong dataset, our methods can beat all
baseline solutions by a large margin on the HK dataset. But on
the California dataset, even the simplest interpolation baseline
IDW can achieve satisfying performance, providing the lowest
RMSE and the largest NSE; the performance of our methods and
Ordinary Kriging is in the second tier. We conjecture this is
because the CA raingauge data generally shows smooth spatial
variation, thereby eliminating the need to estimate spatial
relationships dynamically. Here, we design a new metric, called
Figure 2. The PSNR of denoising results from our Noise2Info com MSV, to intuitively describe the spatial variation. For one point
pared with Noise2Void, Noise2Self, and Noise2Same. and its k-nearest neighbors , the local mean is:
We apply our modules, Pacs and Range Norm, on three newest (13)
rainfall prediction neural networks, ConvGRU (Shi et al. 2017),
PredRNN (Wang 2017), MIM (Wang 2019). We compare our Then MSV is the standard deviation of all local means:
Pacs of degree 2 and 3 with other activation functions such as
ReLU and Sigmoid, where normalization Range norm is (14)
compared with Layer Norm.
Figure 3 shows that our Pacs (slash) perform the best for all Table 1 shows the MSV values of two datasets. We can see that
the three frameworks. With Range Norm, the performance is HK raingauge data has a stronger spatial variation (larger MSVs)
more stable, especially for Pac3 with higher degree. than CA. Our GSI is designed to dynamically learn complex
spatial relationships. However, in the face of the California
4.3 Rainfall spatial interpolation dataset, such dynamic learning seems to be unnecessary, since
Experiments are conducted on two real-world hourly raingauge the spatial structure of the data is smooth and simple.
datasets from Hong Kong (HK) and California (CA). HK dataset
includes 123 raingauge and 3770 valid time steps from 2009 to Table 1. MSV values of rainfall on two datasets. =0 denotes the case
with no neighbor.
2015, and the data precision is 0.1-mm. CA dataset includes 215
raingauge and 1761 valid time steps from 2008 to 2010, and the 0 1 2 3 4 5 6 7 8
data precision is 0.01-inch. Three metrics are used to evaluate the
performance: Root Mean Squared Error (RMSE), Mean Absolute HK 2.71 2.56 2.45 2.35 2.27 2.19 2.12 2.06 2.00
Error (MAE), Nash–Sutcliffe’s Efficiency (NSE) coefficient CA 1.3 1.2 1.08 1.02 0.99 0.96 0.93 0.90 0.88
(Nash & Sutcliffe 1970). The range of NSE is from to 1;
the closer to 1, the better. The smaller the values of RMSE and
MAE, the better.
539
Grell, G. A. , & Freitas, S. R. . 2014. A scale and aerosol aware stochastic
convective parameterization for weather and air quality modeling.
Atmospheric Chemistry and Physics, 13(9), 5233-5250.
Guo, S., Yan, Z., Zhang, K., Zuo, W., & Zhang, L. 2019. Toward
convolutional blind denoising of real photographs. In Proceedings of
the IEEE/CVF conference on computer vision and pattern
recognition (pp. 1712-1722).
Hardy, G. H., Littlewood, J. E., Pólya, G., & Pólya, G. 1952. Inequalities.
Cambridge university press.
He, K., Zhang, X., Ren, S., & Sun, J. 2016. Deep residual learning for
image recognition. In Proceedings of the IEEE conference on
computer vision and pattern recognition (pp. 770-778).
Krull, A., Buchholz, T. O., & Jug, F. 2019. Noise2void-learning
denoising from single noisy images. In Proceedings of the
IEEE/CVF Conference on Computer Vision and Pattern Recognition
(pp. 2129-2137).
Lee, T. T., & Jeng, J. T. 1998. The Chebyshev-polynomials-based unified
model neural networks for function approximation. IEEE
Transactions on Systems, Man, and Cybernetics, Part B
(Cybernetics), 28(6), 925-935.
Lehtinen, J., Munkberg, J., Hasselgren, J., Laine, S., Karras, T., Aittala,
M., & Aila, T. 2018. Noise2Noise: Learning image restoration
without clean data. arXiv preprint arXiv:1803.04189.
Figure 4. The results of rainfall spatial interpolation on two datasets. Li, J., & Heap, A. D. 2008. A review of spatial interpolation methods for
environmental scientists. Geoscience Australia.
López-Rubio, E., Ortega-Zamorano, F., Domínguez, E., & Muñoz-Pérez,
5 FUTUREWORKS J. 2019. Piecewise polynomial activation functions for feedforward
neural networks. Neural Processing Letters, 50(1), 121-147.
Ma, J., Ding, Y., Cheng, J. C., Jiang, F., & Wan, Z. 2019. A temporal-
In this work, we explore the application of deep learning in
spatial interpolation and extrapolation method based on geographic
rainfall prediction. Are we there yet? The answer might be NO. Long Short-Term Memory neural network for PM2. 5. Journal of
The major issues that need to be resolved include: Cleaner Production, 237, 117729.
1. Abundant high-quality data is not ready. Rainfall is a Ma, L., & Khorasani, K. 2005. Constructive feedforward neural networks
complex system which is affected by various weather using Hermite polynomial activation functions. IEEE Transactions
data. Different data relies on different ways for collection on Neural Networks, 16(4), 821-833.
with varied granularity and accuracy. More high- Martin, D., Fowlkes, C., Tal, D., & Malik, J. 2001. A database of human
segmented natural images and its application to evaluating
resolution and clean data is still needed for good rainfall segmentation algorithms and measuring ecological statistics. In
prediction results. Proceedings Eighth IEEE International Conference on Computer
2. Proper model is still missed. Deep learning models are Vision. ICCV 2001 (Vol. 2, pp. 416-423). IEEE.
mostly designed for general tasks, where the physical Moran, N., Schmidt, D., Zhong, Y., & Coady, P. 2020. Noisier2noise:
system has its own properties. Rainfall-oriented model is Learning to denoise from unpaired noisy data. In Proceedings of the
a promising way for accurate prediction. How to correctly IEEE/CVF Conference on Computer Vision and Pattern Recognition
fuse heterogeneous data is an important issue to be (pp. 12064-12072).
Nash, J. E., & Sutcliffe, J. V. 1970. River flow forecasting through
considered in model design. conceptual models part I—A discussion of principles. Journal of
3. Measure is not well-established. Current benchmarks still hydrology, 10(3), 282-290.
do not combine heterogeneous weather data and give Powell, M. J. D. 1981. Approximation theory and methods. Cambridge
widely accepted metrics. A well-prepared benchmark is university press.
important for model design and performance comparison. Ronneberger, O., Fischer, P., & Brox, T. 2015. U-net: Convolutional
networks for biomedical image segmentation. In International
Conference on Medical image computing and computer-assisted
6 ACKNOWLEDGEMENTS intervention (pp. 234-241). Springer, Cham.
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., ... &
This work is supported by the Hong Kong RGC AOE Project Fei-Fei, L. 2015. Imagenet large scale visual recognition challenge.
International journal of computer vision, 115(3), 211-252.
AoE/E-603/18.
Rusu, C., & Rusu, V. 2006. Radial basis functions versus geostatistics in
spatial interpolations. In IFIP International Conference on Artificial
Intelligence in Theory and Practice. Springer, Boston, MA.
7 REFERENCES Shi, X., Gao, Z., Lausen, L., Wang, H., Yeung, D. Y., Wong, W. K., &
Woo, W. C. (2017). Deep learning for precipitation nowcasting: A
Alfeld, P. 1984. A trivariate clough—tocher scheme for tetrahedral data.
benchmark and a new model. Advances in neural information
Computer Aided Geometric Design, 1(2), 169-181.
processing systems, 30.
Appleby, G., Liu, L., & Liu, L. P. 2020. Kriging convolutional networks.
Skamarock, W. C. , Klemp, J. B. , Dudhia, J. , Gill, D. O. , &
In Proceedings of the AAAI Conference on Artificial Intelligence.
Powers, J. G. . 2018. A description of the advanced research wrf
Ba, J. L., Kiros, J. R., & Hinton, G. E. 2016. Layer normalization. arXiv
version 2. Ncar Technical, 113, 7--25.
preprint arXiv:1607.06450.
Swain, D. L., Langenbrunner, B., Neelin, J. D., & Hall, A. 2018.
Batson, J., & Royer, L. 2019. Noise2self: Blind denoising by self-
Increasing precipitation volatility in twenty-first-century California.
supervision. In International Conference on Machine Learning (pp.
Nature Climate Change, 8(5), 427-433.
524-533). PMLR.
Veli kovi , P., Cucurull, G., Casanova, A., Romero, A., Lio, P., &
Fritsch, J. , & Chappell, C. . 1980. Numerical prediction of convectively
Bengio, Y. 2017. Graph attention networks. Graph Attention
driven mesoscale pressure systems. part i: convective
Networks. In Proceedings of 6th International Conference on
parameterization. J.atmos, 37(3), págs. 198-202.
Learning Representations.
Grell, G. A., & Dévényi, D. 2002. A generalized approach to
Wang, Y., Long, M., Wang, J., Gao, Z., & Yu, P. S. 2017. Predrnn:
parameterizing convection combining ensemble and data
Recurrent neural networks for predictive learning using
assimilation techniques. Geophysical Research Letters, 29(14), 38-
spatiotemporal lstms. Advances in neural information processing
1.
systems, 30.
540
Wang, Y., Zhang, J., Zhu, H., Long, M., Wang, J., & Yu, P. S. 2019.
Memory in memory: A predictive neural network for learning
higher-order non-stationarity from spatiotemporal dynamics. In
Proceedings of the IEEE/CVF Conference on Computer Vision and
Pattern Recognition (pp. 9154-9162).
World Meteorological Organization. 2008. Guide to Instruments and
Methods of Observation Volume 1–Measurement of Meteorological
Variables. Geneva: WMO.
Wu, Y., Zhuang, D., Labbe, A., & Sun, L. 2021. Inductive graph neural
networks for spatiotemporal kriging. In Proceedings of the AAAI
Conference on Artificial Intelligence.
Xie, Y., Wang, Z., & Ji, S. 2020. Noise2same: Optimizing a self-
supervised bound for image denoising. Advances in Neural
Information Processing Systems, 33, 20320-20330.
Zhang, X., Zwiers, F. W., Li, G., Wan, H., & Cannon, A. J. 2017.
Complexity in estimating past and future extreme short-duration
rainfall. Nature Geoscience, 10(4), 255-259.
541
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Machine learning in geotechnical engineering: Opportunities and applications
Apprentissage automatique en génie géotechnique: opportunités et applications
& Suzanne Lacasse
ABSTRACT: Geotechnical engineering is benefitting from the rapid development of Machine Learning techniques. Machine
Learning is particularly relevant for geotechnical engineering, where data are often scarce, the variability in soil and rock can be high
and measurements are influenced by both testing and measurement techniques and by data transformation uncertainties. This paper
discusses recent advances in the application of Machine Learning in geotechnical engineering, outlining the opportunities the
approach offers, discussing challenges and providing several examples taken from real-life geo-infrastructure design. To achieve this
goal, the paper gives a very brief overview of Machine Learning, introduces emerging trends in the geotechnical profession, and
exemplifies how Machine Learning and Deep Learning were used for site characterization, underground excavations, tunnel design
and landslide mapping, including landslide susceptibility and temporal forecasting. A critical discussion of the approach for
geotechnical engineering is done before conclusions are drawn.
RÉSUMÉ : L'ingénierie géotechnique peut grandement bénéficier des techniques d'apprentissage automatique (" ").
L'apprentissage automatique est particulièrement pertinent pour la géotechnique parce que les données sont souvent rares, la
variabilité du sol et de la roche peut être très grande et les mesures sont influencées par les techniques d'essai et de mesure et par les
incertitudes de transformation des données. L'article traite des avancées récentes dans l'application de l'apprentissage automatique
en ingénierie géotechnique, décrivant les opportunités de l'approche, discutant ses défis et fournissant plusieurs exemples tirés de la
conception des géo-infrastructures. L'article fournit un bref aperçu de l'apprentissage automatique, présente les émergences dans la
profession géotechnique et illustre comment l'apprentissage automatique (" ") et l'apprentissage en profondeur
(" ") sont utilisés pour la caractérisation d'un site, excavations souterraines, conception de tunnels, et cartographie de
glissements de terrain (et prévision et susceptibilité). L'article se termine avec une discussion critique de l'approche.
KEYWORDS: Machine learning; Deep learning; Geotechnical engineering; Big data analysis.
(a systematic review of ML in geotechnical engineering was,
1 INTRODUCTION however, not done), introduce emerging opportunities within the
geotechnical community, and demonstrate, with examples of
In recent years, with the rapid development of Machine Learning ML/DL geotechnical applications in practice. A critical
(ML) and more recently, Deep Learning (DL) as a Data Science discussion of the approach and conclusions close the paper.
branch, and its implementation in many engineering fields,
researchers have started looking into disciplinary or thematic
applications of ML methods. For instance, in the geotechnical 2 MACHINE LEARNING
engineering community, the growing interest in ML is led to the
recent creation (2018) of a technical committee (TC) of the In contrast to traditional physics-driven methods, machine
International Society of Soil Mechanics and Geotechnical learning (ML), a sub-field of the artificial intelligence research,
Engineering ISSMGE (TC309) specifically on “Machine is a type of data-driven approach that ‘trains’ a regression or
Learning and Big Data”. classification model through a complex nonlinear mapping with
Machine Learning is particularly helpful and relevant for adjustable parameters based on the training data set. Deep
geotechnical engineering, where data are often sparse, the learning (DL) can be regarded as a subset of ML.
variability in the characteristics of soil and rock can often be The main distinction between DL and conventional ML
high, and the measurements – especially in situ – are influenced algorithms falls in the modality of learning from data, i.e. the way
by both measurement techniques and data transformation in which the learning or ‘training’ is done. Besides, DL is
uncertainties (Jaksa and Liu 2021). There exist a few State-of- primarily based on neural networks (NN), whereas conventional
Art reviews on machine learning/deep learning for specific ML methods include algorithms other than, and as well as, NN.
geoscience domains, including a short review on DL applications In general, DL algorithms typically require more data for training
in geotechnical engineering (Zhang et al. 2021), ML modelling than conventional ML algorithms, given their higher number of
of soil properties (Zhang et al. 2021), ANN applications hyperparameters. A hyperparameter is a parameter which value
(Moayedi et al. 2020), small to large scale porous media is used to control the learning process. The values of other
(Tahmasebi et al. 2020), underground soil-structure interaction parameters, on the other hand, are derived by ‘training’.
(Jong et al. 2021), underground excavations (Zhang et al. 2020), Empirical analysis has shown that the best accuracy estimates are
geohazard modelling (Dikshit et al. 2021), maintenance of attained if we allocate 20-30% of the entire dataset for testing,
railway track (Xie et al. 2020), landslide studies (Mergadi et al. and use the remaining 70-80% for training. Gholamy et al. (2018)
2020; Ma et al. 2020; Tehrani et al. 2022) and geological hazard provided a possible explanation why this ratio is the best division
analysis (Ma and Mei 2021). into the training and the testing datasets.
The purpose of this paper is to present the most recent Most of ML modelling falls under 3 categories: Supervised,
advances in the application of ML in geotechnical engineering, unsupervised (semi-supervised can be seen as a mix of the two),
and to discuss the challenges and opportunities provided by ML and reinforcement learning. Figure 1 illustrates the ‘containment
and DL for applications by geotechnical engineers. The relationship' from artificial intelligence to machine learning and
following sections present a very brief overview of ML methods
543
deep learning. and gives geotechnical engineering examples for When asked to identify the major obstacles ML/AI faces for
the three learning categories. implementation in geotechnical engineering practice (Figure 2),
1. Supervised Learning: The most common form of learning the respondents prioritized 'insufficient data' (over 55%), 'no
used in geotechnical applications. Supervised machine examples to follow' (42%), 'don't know how to use big data'
learning is all about giving data to learning algorithms in (38%), then 'need cheaper & better data acquisition hardware'
a way to provide context and feedback for learning. By (27%).
'training' with the given input variables and target Data-driven methods, such as ML algorithms, are not useful
variables, the machine learns how to map inputs to if the necessary data are not available. In geotechnical
corresponding outputs. The training process continues engineering, good quality data do not always exist freely, and this
until the model achieves a desired level of accuracy using condition surely limits the application of ML algorithms. It is
the training data. Typical geotechnical applications expected, however, that in the near future this limitation may be
include modelling the mechanical behaviour of soils overcome by the growing availability of geotechnical data
(Zhang et al. 2021), ground improvement (Ranasinghe et initiated by large research projects (e.g., the NGTS database:
al. 2017; Hosseini et al. 2020), geo-structure design (piles, https://www.geocalcs.com/datamap), and the freeing og
shallow foundations, soil nail walls) (Kordjazi et al. 2014; quantities of remote sensing (e.g., satellite) data.
Ray et al. 2021; Liu et al. 2021), and landslide The majority of the ML research publicly available did not
deformation process prediction (Yang et al. 2019; Liu et make the utilized computer scripts and original dataset available.
al. 2020). Therefore, it hampers the utility of these studies and it is a big
2. Unsupervised Learning: Unlike supervised algorithms, obstacle to finding the resources/benchmark examples to train
unsupervised learning data sets only include inputs, and and educate geotechnical engineers. A significant stumbling
the algorithm must learn from those inputs and find block to any machine learning is accessing data locked away in
patterns and commonalities between data points to cupboards and electronic folders (Pathmanandavel and
determine the next steps to take. Examples of this type of MacRobert, 2020).
unsupervised learning algorithms are hierarchical
clustering, K-means, Density-Based Spatial Clustering of
Applications with Noise (DBSCAN). Typical
geotechnical applications include clustering underground
stratification (Wang et al. 2019; Wang et al. 2021;
Molina-Gomez et al. 2021), rock mass clustering
(Erharter and Marcher 2021), and exploratory monitoring
data analysis (Zhou et al. 2019).
3. Reinforcement Learning (RL): Reinforcement learning
emphasizes learning agents, or programs acting within
environments. RL refers to the process of an agent
learning to achieve a specific objective through
interaction with its environment. RL has not been used for
many specific applications in geotechnical engineering.
One example of RL-based process optimization and
strategy development in conventional tunnelling was
presented by Erharter et al. (2021). Figure 2. ML survey: “What do you think are the major obstacles to using
ML/AI to geotechnical engineering? (Select all that apply)".
The survey highlighted that machine learning is expected to
have a significant impact on monitoring and the observational
method, site characterization, hazard and risk management and
material modelling in the coming decade. Site characterization is
one of the leading areas where machine learning can transform
practice over the next 10 years (Figure 3).
Figure 1. The containment relationship between artificial intelligence,
machine learning and deep learning and example applications of machine
learning in geotechnical engineering.
3 EMERGING OPPORTUNITIES WITHIN
GEOTECHNICAL COMMUNITY
3.1 Survey on ML in geotechnics
The TC309 (ISSMGE’s Technical Committee for Machine
Learning and Big Data) carried out an online survey in 2019,
receiving 114 responses. Over 85% of respondents thought
Figure 3. ML survey: “What are the machine learning applications that
ML/AI will likely or very likely impact large numbers of can transform value for our industry in the next 10 years? (Select all that
geoscientists in the next 10 years. apply)
544
4. Value and Future of Machine Learning in geotechnical
Machine learning was seen to have the least impact on future engineering: Many issues were raised regarding the value
practice of geotechnical design, which perhaps reflects the of machine learning for geotechnical applications. For
commonly held view that much design knowledge remains example, machine learning, if supervised, can act as a
empirical and is held by senior engineers (Pathmanandavel and communication tool between the inputs and outputs of the
MacRobert, 2020). Although the authors disagree with the first problem analyzed. However, in an unsupervised model,
part of this statement (design is usually not based mainly on machine learning has the ability to model cases that may
empiricism), it is true that geotechnical engineering is a "science not have been known before. Thus, the unsupervised
of experience": the more different types of geo-situations one has learning is perhaps one of the most valuable contributions
seen, the better equipped one is to solve a new problem. of machine learning. Concerning the future of machine
learning, it was concluded that the machine learning
3.2 Machine learning dialogue process becomes closer and closer to human learning over
time, machine learning provides smarter tools and will
ISSME TC309 (Machine Learning and Big Data) organised the gradually help the user understand geotechnical
first Machine Learning (ML) Dialogue in July 2019 before the phenomena more effectively.
second International Workshop on Machine Learning and Big The ISSMGE TC304/TC309/TC2101 Second ML Dialogue
Data in Geoscience at Tongji University, Shanghai, China. A for Geotechnics was held in December 2019 following the 7th
summary of the discussions containing views on data availability, International Symposium on Geotechnical Safety and Risk
machine learning implementation, application of machine (ISGSR 2019), at the National Taiwan University, Taipei, Taiwan.
learning in geotechnical engineering, value and future of The aim of the 2nd ML dialogue was to discuss the opportunities
machine learning was published in ISSMGE Bulletin 13(4), 2019, and challenges in developing and applying ML to geotechnical
8-14 (https://www.issmge.org/publications/issmge-bulletin/vol- engineering research and practice (Phoon et al. 2019a)
13-issue-4-august-2019 ). (https://www.issmge.org/publications/issmge-bulletin/vol-14-
The key conclusions of the discussions at this 1st ML issue-1-february-2020 ). The following recommendations were
Dialogue were: made during the 2nd Dialogue:
1. Data: The participants' views on data were focused on three 1. Industry is a necessary partner for machine learning-based
main areas: data scattering, data sharing and data quality. studies in geotechnical engineering. There are two
On data scattering, the participants believed that there was compelling reasons. One, machine learning must bring
typically an insufficient volume of available specific data transformative value to industry, not universities. This
that can be used. Concerning data sharing, the participants value is ultimately linked to the end users that the industry
remarked that there is currently an insufficient effort to serves. Two, real data are generated and owned by projects
share geotechnical data among the researchers and and real data (not simulated data) are absolutely necessary
practitioners. The volume of data produced in research is ingredients for machine learning development.
high, but the sharing is not organized in a coherent manner. 2. A “data sandbox” is needed to enable and accelerate data
Thus, it was believed that a context, perhaps TC309, should sharing between industry and research. Theoretical and
be provided to explain the benefits of data sharing to the pure methodology development of ML was not seen to be
researchers. The third concern was the data quality. It was an effective means of bringing forward the implementation
stated that the data to be made public should be reviewed of ML in geotechnical engineering.
in advance and checked for its relevance and quality, 3. Future discussions on this topic, including organization of
following appropriate standards of quality for geotechnical future workshops, should be tied to a specific flagship
data. The production of benchmark dataset should actually project with the industry. This “AlphaGeo” project (cf.
be targeted and documented in terms of quality, to ensure Google’s “AlphaGo” project 2 ) would ideally be game-
that the data can be used safely in research projects. changing, rather than improving some aspects of existing
2. Machine Learning Implementation: the capabilities of practice in an incremental way.
machine learning in the field of geotechnics were described. The ISSMGE TC309/TC304/TC222 3 organised the Third
Capabilities include: using machine learning to avoid Machine Learning in Geotechnics Dialogue (3MLIGD) in 2021,
subjective outputs when making objective decisions; with a focus on a specific flagship project, data-driven site
application of machine learning for the analysis of complex characterization, and to actively facilitate a discussion among
problems, including cases with heterogenic materials, and researchers and practitioners. The meeting was hosted online
analysis of nonlinear and multivariate problems. In this from Oslo, Norway, by the Norwegian Geotechnical Institute in
context, it was mentioned that each problem modelled December 2021. A total of 15 academics and 21 practitioners
requires its own algorithm which depends strongly on the took part in this online dialogue.
problem scale, quantity of data available and type of The objectives of the dialogue were for academia and
outputs. industry to jointly identify: (1) topics of potential and actual
3. Application of Machine Learning in geotechnical value to industry, (2) ideas to foster collaboration between
engineering: various machine learning geotechnical academia and industry, and (3) platforms to accelerate progress.
applications, including tunnelling and TBM applications, The common theme was to find novel ways to foster greater
geohazards and landslides and underground structures. The connectivity between researchers, between practitioners, and
views of participants on the use of machine learning were between joint researchers and practitioners to leverage on the
categorised in two phases: 'under construction' and 'during limited resources and talents at the disposal of a fledging “data-
operation'. The applications were broadly divided into three centric geotechnics” community.
categories: supervised, unsupervised, and semi-supervised
learning techniques.
1
The TCs are ISSMGE technical committees: TC304 is the technical including a version that competed under the name Master. After retiring
committee on 'Engineering Practice of Risk Assessment and from competitive play, AlphaGo Master was succeeded by an even more
Management' and TC210 is the technical committee on 'Embankment powerful version known as AlphaGo Zero, which was completely self-
Dams'. taught without learning from human games.
2 3
AlphaGo is a computer program that plays the board game Go. TC222 is the ISSMGE technical committee on 'Geotechnical BIM and
Subsequent versions of AlphaGo became increasingly powerful, Digital Twins'.
545
The following recommendations were gleaned from the varying soil and rock properties are indispensable for site
small group and general discussions. characterization and following design analyses. An improved
1. Notwithstanding the various limitations of geo-data, the definition of soil and rock characteristics, i.e., smaller
key limitation is data access. In the absence of a better uncertainty and more reliable mean values, can lead to more
platform, 304dB should be further enhanced by: (a) rational and economic designs. This is where ML algorithms can
encouraging more researchers/practitioners to share their help.
published databases that are already in public domain, (b) Three opportunities, as well as challenges, were readily
including more diverse data sources such as geophysical identified in data-driven site characterization: (1) stratification
and satellite data, and (c) listing links to other public and soil classification, (2) soil property profile, and (3) 3D
sources of geo-data. integrated ground models.
2. TC309/TC304/TC222 should make a contact list of people
in academia/industry and their current interests to 3.1.1 Stratification and soil classification
encourage people with common interests to dialogue and The piezocone penetration test (CPTU) allows cost-effective and
work together. rapid in situ measurements within soils . It is the most widely
3. TC309/TC304/TC222 should make a list of ML projects used in situ testing method for geotechnical site investigation,
(research) and practical applications to enable one to learn and is the best available today for identifying subsurface
from on-going projects and their outcomes, enhance stratification and soil type. One CPTU profiling provides nearly
awareness of the capabilities of machine learning in continuous data versus depth. Several profiles in one area
geotechnics and to create a synergy among the projects provides therefore a lot of data. With the large amount of CPTU
and/or funding. data, and existing soil classification charts, ML algorithms can
4. Novel collaboration modes should also be explored, for be used to classify soils into using such soil type classification
example: charts developed from earlier experience (e.g., Robertson, 1990).
a. A “shadow project” could further facilitate Figure 4 shows one such soil classification chart. It correlates the
industry/academia collaboration by allowing industry cone resistance (in tis instance the normalized corrected cone
and academia partners to work on the same project in resistance Qtn with the normalized friction ratio Fr, both
parallel. This mode could lower the cooperation barrier measured by the cone during essentially continuous penetration.
arising from contrasting goals between industry (finish a The Robertson (1990) chart in Figure 4, called the soil behaviour
project on time and on budget) and academia (discovery type (SBT) chart, is the most used today, worldwide.
of novel ideas and high impact publications). The “data
sandbox” could be viewed as a more restricted
implementation of a “shadow project”.
b. It is advantageous to focus on “ML supremacy” projects
rather than “high value” projects. ML supremacy
projects are important projects where physics-based
approaches are impractical or even impossible to apply,
but data-driven approaches can provide highly effective
solutions. One broad category can cover projects
involving large datasets (perhaps real time data) from
multiple sources that cannot be handled effectively and
efficiently using conventional methods. There is a strong
demand for data-driven methods in such projects.
Specific examples include site characterization maps or
susceptibility, vulnerability and/or risk maps covering
large spatial domains. These projects can demonstrate
“ML supremacy” more convincingly than conventional
projects where the value added from ML may be
incremental due to the lack of mature technologies
applicable to geotechnical engineering within 5G, IoT,
BIM/digital twinning modelling, robotics, AI etc.
c. Industry and academia can join hands to create “learn and Figure 4. Normalized CPT soil behavior chart (Robertson, 1990).
work” projects of more modest scales for educating the
next generation of young digital and ML talents. This Several machine learning algorithms have been applied for
could be implemented as internships that are already the interpretation of CPTU data in terms of soil type including
prevalent in most institutions of higher learning. Support Vector Machine, SVM (Rauter and Tschuchnigg 2021),
Artificial Neural Network, ANN (Reale et al. 2018; Rauter and
Tschuchnigg 2021; Sastre et al. 2021), Random Forest, RF
3 APPLICATIONS IN GEOTECHNICAL ENGINERING (Rauter and Tschuchnigg 2021; Liu et al. 2021), and Bayesian
Given that a large quantity of data have been collected and stored learning (Wang et al. 2019; Xiao et al. 2021).
over the past years, ML has been widely used to solve several Rauter and Tschuchnigg (2021) compared three different
rather complex geotechnical problems. Several applications of models to determine soil behavior type, using the SVM, ANN
ML algorithms are described below. and RF algorithms. They applied the algorithms to a database
consisting of 1339 cone penetration tests (CPT) and concluded
3.1 Site characterization that the ANN and RF models yielded very accurate results.
In geotechnical practice, the datasets nearly always show an
Because of the geological processes leading to the compounded imbalanced distribution among the nine soil type categories
deposition and subsequent changes of geomaterials, including (abbreviated here with ST) in Robertson’s 1990 chart. In such
layering, each stratum exhibits variation from point to point case, it is necessary to find a way of balancing the dataset
within a volume (Liu et al. 2015). Nevertheless, the identification (Erharter et al. 2021). Liu et al. (2021) compiled a database
of subsurface stratification and the characterization of spatially consisting of 357 CPT tests sampled at each 0.5 meter. There
546
were 12,921 CPT data points in total that could be used in the presented by Zhang et al. (2021). The implementation of ML
application of machine learning methods. The number of data techniques has shown an exponential growth since 2018.
points was counted for each category of soil type (ST). The However, important challenges still remain, such as how to
summary of the soil type categories in the soil dataset is shown assess an algorithm’s performance and how to select an optimal
in Figure 5. The soil types in categories ST 1, 2, 8 and 9 are algorithm.
highly underrepresented. The data in those categories sum up to Pham et al. (2018) compared the performance of four
only 5% of the entire dataset. On the other hand, ST 3 machine learning methods, Particle Swarm Optimization -
category was significantly overrepresented (Fig. 5). Adaptive Network based Fuzzy Inference System (PANFIS),
Genetic Algorithm - Adaptive Network based Fuzzy Inference
System (GANFIS), Support Vector Regression (SVR), and
Artificial Neural Networks (ANN) for predicting the undrained
shear strength of soft soils. Zhang et al. (2021) investigated the
application of ensemble ML methods, i.e., XGBoost-based and
RF-based models, for the prediction of undrained shear strength
of soft soils. The comparisons with baseline models (i.e., SVM,
Multilayer perceptron (MLP) and Multivariate adaptive
regression spline (MARS)) show the superiority of the XGBoost-
based and RF-based model in terms of predictive performance.
In Norway, the undrained shear strength interpretation from
CPTUs for soft clays has been developed based on laboratory
testing on soil samples of very high quality (large block samples),
index parameters and measured cone resistance qc and/or CPTU
pore pressure response u (Paniagua et al. 2019). These
interpretations work well when the required index parameters are
Figure 5. Chart of the nine Robertson (1990) soil type categories (ST) known. However, the correlations do not include an uncertainty
represented in a dataset of nearly 13,000 data points. (Modified from Liu estimate. An ANN-based model that takes into account
et al. 2021).
uncertainty estimate was developed in NGI (2020).
The database of CPTU and triaxial tests used contained 241
To balance the dataset, the over-sampling technique SMOTE
laboratory triaxial tests. All test specimens were of high quality,
algorithm (Synthetic Minority Over-sampling Technique after
including 180 undisturbed specimens taken with a 72-mm
Chawla et al. 2002) was used. The accuracy for different soil
diameter fixed piston sampler and 61 undisturbed specimens
types for the test set (80% of the entire dataset) before and after
taken with a 40-cm diameter block samples. Figure 6 present the
over-sampling is summarized in Table 1. The results indicate that
undrained shear strength in triaxial compression (su) derived
the prediction accuracy for the two minority classes (ST 1 and
from the corrected cone resistance (qt).
ST 2) are significantly improved after over-sampling. The
accuracy for the soils in class ST 8 and ST 9 remained
unchanged, however.
Taking into account this aspect of imbalance in the data can
be very important for some geotechnical designs where the
potential failure could happen in specific soil types or layers, e.g.
weaker layers.
Table 1. Accuracy for different soil types ST for the test set before and
after oversampling, database in Figure 5 (modified from Liu et al. 2021).
ST Prediction accuracy Prediction accuracy after
category before Over-sampling (%) Over-sampling (%)
ST 1 83.3 100
ST 2 89.7 94.3
ST 3 99 95.2
ST 4 94.7 98.7
ST 5 94.7 93.3
ST 6 97.2 97.9
ST 7 99 98 Figure 6. Undrained shear strength, su (triaxial compression), as a
function of corrected cone resistance (qt) for 180 specimens taken with
ST 8 72-mm diameter piston sampler (black) and 61 specimens taken with 40-
96.4 96.4
cm diameter block sampler (red).
ST 9 92.3 92.3
The ANN model requires the water pressure, and effective
3.1.2 Soil properties stresses in addition to the CPTU data. The estimate of the
A very recent overview of the application of ML algorithms to standard deviation was done based on the "drop out method"
the prediction of soil properties in the past ten years was
547
approach described by Gal (2016). The predicted and measured 3.1.3 Integrated 3D ground models
undrained shear strength in triaxial compression are compared in
Figure 7. An estimate of the standard deviation of the predicted For large geotechnical projects on land and geotechnical offshore
undrained shear strength is also shown. developments, it is current practice to collect both geophysical
and geotechnical data.
Sauvin et al. (2019) developed a generic workflow (Fig. 8) to
integrate the results of geotechnical and geophysical
investigations using ML (e.g., Genetic algorithm (GA) and
ANN) to obtain a quantitative ground model in three dimensions.
The workflow in Figure 8 includes using isolated
geotechnical data from boreholes (usually the results of in situ or
laboratory test on recovered samples of different quality) and
geophysical data (the combined set of data is often denoted G&G
data). One-D borehole data provide valuable information on soil
stratification (often called soil units) and soil characteristics.
Geophysical data, collected along 2D lines and/or 3D volumes,
provide the natural link to re-populate geotechnical
characteristics found in the 1D boreholes onto a larger area.
For the 1D geotechnical data, the integration is done by
matching (or correlating) the unit or horizon to establish a ground
model in 2D or 3D. Geostatistics, comparison of boreholes,
statistical analysis and parameterization are used to establish the
Figure 7. Predicted vs measured CPT undrained shear strength (in kPa)
model. The parameterization finds parametric equations,
using the training dataset. Vertical line shows estimated standard
deviation estimated. Black dashed line is 1:1 diagonal
preferably implicit, of the curve or surface.
For the geophysical data, the physical properties and volumes
are assessed using also geostatistics, borehole comparison and
calibration, rock physics models and empirical relationships.
Figure 8. Generic workflow for integrating geotechnical and geophysical data into quantitative ground models (Sauvin et al. 2019).
548
The synthetic ground model from the geotechnical data and Figure 9 shows how the three methods, i.e., geometrical-
the added information from the geophysical data are then fed to statistical (Forsberg et al. 2017), co-kriging and Machine
a ML algorithm to derive the in situ properties over a surface or Learning, compare at the one location HKZ3-CPT-5. In these
a volume. The integration of geophysical and geotechnical data plots: the black curves correspond to the measured qc data; the
thereby builds a consistent ground model based on all the green curves are the best-fit solutions when the CPTU data set is
information available that is as robust as possible, given the part of the training set, and the orange curve is the prediction
available information. It will also allow some degree of solution at the given location after removing the CPTU data from
extrapolation outside the site. The framework in Figure 8 the database. It is indicated that the ML approach provides the
provides a quantitative interpretation, based on G&G, for most accurate prediction of the geotechnical properties from the
establishing a holistic ground model to serve as input to design. geophysical data.
Figure 9. Example of the results from the geometric-geostatistic approach (left) and co-kriging (middle) and Machine Learning (right) on qc for one
location HKZ-CPT05 (Sauvin et al. 2019).
Klinkvort et al. (2020) demonstrated this quantitative model 10 shows the histogram of computed wall deflections. The
to obtain an estimate of geotechnical properties and their statistical distribution can be approximated by a lognormal
uncertainties using two seismic attributes, i.e. the seismic quality distribution, and most of the wall deflections were between 50
factor Qp and acoustic impedance AI. They use geostatistics and 300 mm. the mean ( ) and standard deviation ( ) were 138
(kriging) across an area of about 25 km2. The predicted values mm and 69 mm, respectively. The coefficient of variation ( )
were then used for geotechnical design of spudcan penetrations. was 50% (so quite large).
In this case, the geophysical data guided the spatial interpolation
of the soil properties in a more consistent way than if it had been
done without the integration.
Chen et al. (2021) also demonstrated a successful G&G
integration that uses geophysical data and the integration with
cone penetration test (CPT) data using ANN networks to create
synthetic CPT profiles.
3.2 Underground excavations
Zhang et al. (2020) presented an overview of soft computing
techniques and their applications for underground excavations.
Using a braced excavation case study (Zhang and Goh 2013),
they compared the predictive performances of soft computing
methods (SCMs) including eXtreme Gradient Boosting
(XGBoost), Multivariate Adaptive Regression Splines (MARS),
Artificial Neural Networks (ANN), and Support Vector Machine
(SVM) for estimating the maximum lateral wall deflection
induced by braced excavation.
The database included the results of 1120 plane strain finite
element (FE) analyses of diaphragm walls in a deep braced
excavation. The influences of different parameters such as Fig. 10. Histogram of lateral wall deflections predicted from 1120 plane
geometry, soil properties and wall stiffness on the computed wall strain finite element (FE) analyses of diaphragm walls in deep braced
deflections were investigated in Zhang and Goh (2013). Figure excavation (Zhang et al. 2020).
549
3.3 Tunnels
Figures 11 and 12 show the 'training' and 'testing' results of Digitalization in tunneling has attracted a lot of attention in the
the prediction of wall deflection by XGBoost, MARS, ANN and past years. In drill and blast tunnelling, deciding rock support and
SVM. The training data set comprised xxx% of the total dataset. blasting design based on in-situ rockmass characteristics is
For both the training and the testing data sets, the four ML crucial for safe and efficient tunnel driving. This is normally
methods achieved reasonably good results up to a predicted wall carried out by engineering geologists on the face, based on
deflection of about 300 mm. The model predicted data points rockmass characterization after blasting using a rock mass
below 300-mm deflection fit well on the 45°-line. The calculated classification system, e.g., the Q-system (Barton et al. 1974). For
Root Mean Square Error (RMSE) for each of the four ML conventional tunnelling projects in Norway, all drillholes
approaches are shown in Figures 11 and 12. Values close to or (percussion drilling) are required by contract to be logged with
equal to 0 indicate that the error in prediction is marginal. The the Measure While Drilling (MWD) method. MWD data
RMSE-values were slightly higher for the testing dataset than for collection consists of equipping drilling machines with sensors
the training dataset, but still quite small. The increase in RMSE recording various drilling parameters, with subsequent real-time
was about 10%, except for the XGBoost method, which was data processing (see Fig. 13) for access by the on-site personnel
slightly higher. and face-engineers to assist them in decision making. The
The prediction performance was improved by the 'ensemble growing volume of data (e.g., from sensors, scanning or images)
learning approach' XGBoost method, compared to the more generated by methods like MWD data collection, provides
conventional MARS, ANN and SVM methods. An 'ensemble opportunities for the application of advanced machine learning
method' is an ML technique that combines several base models (ML) models, capable of generalising information from big
to produce one optimal predictive model. datasets, thereby detecting and extrapolating patterns. Hansen et
It should also be mentioned that the developed ML models are al. (2022) attempted to use deep-learning-based computer vision
more apt to be updated to improve the models as more techniques to interpret MWD-data framed as images, thereby
information becomes available. The new data can be fed to the preserving the richness in data and opening for complex and
'training' datasets and thereby the predictive model can be nonlinear relationship between data and prediction target. The
adjusted and the RMSE reduced. main goal was to investigate the ability of ML to automatically
map MWD-data to condense rock mass characteristics in a
consistent and more reproducible way and which is useful for
making decisions at the rock face, concretely by inferring the Q-
class from one blasting-round of MWD-data. A random forest
classifier (RFC) and a convolutional neural network (CNN)
approach were utilized to characterize the rock mass, represented
by the Q-class, i.e. class A-E in the Q-system, where class A is
strong and competent, and E is a weak rock mass.
Figure 11. Training results of computed FEM wall deflection using four
ML algorithms (Zhang et al. 2020). Reference line is 1:1 diagonal.
Figure 13. Example of tunnel section where interpretation of MWD is
straightforward. Top: Diabase dyke crossing the tunnel. Bottom: The
Figure 12. Testing results of FEM wall deflection using four ML same dyke visualised with low penetration in purple colour before the
algorithms (Zhang et al. 2020). Reference line is 1:1 diagonal. round was blasted. Source: BaneNOR/ROCKMA.
550
3.3.1 Database down-sampling was preserved in both datasets. The up-sampling
was only carried out on the training set to avoid that the same
MWD-data and its corresponding rock mass classification labels
samples would be found in both the train and test tests, leading
are collected from the highway project E39 Kristiansand –
to wrong and too optimistic results. Table 2 summarises the
Mandal (https://www.nyeveier.no/prosjekter/e39-sorvest/e39-
kristiansand-vest-mandal-ost/, accessed Sep. 09, 2021). Fig. 14 different balancing techniques for each combination with the
describes the distribution of label values among the 2401 samples resulting number of samples in train- and test sets. After
in the dataset (5 m tunnel sections). The histogram shows the preprocessing, the total dataset had 2401 samples of 5 m sections
dataset as highly unbalanced, with most samples in classes B and for the total dataset, divided into 47 Class A, 1038 Class B, 987
C. A swarm plot in Fig. 15 describes the distribution of Q-values Class C, 220 Class D, and 109 Class E samples.
within each Q-class, indicating challenging discrimination
between classes for samples close to the somewhat arbitrary
boundaries for rock masses that do not inherently divide into
strictly exclusive categories. In all Q-classes, the samples within
one class are spread over the entire range of values.
Fig. 16. 5- m sections of penetration-images in random order. Each image
is linked to an adjusted label-value for that section. Values under each
image represent (Q-class, (Q-value, RQD)). RQD is Rock Quality
Designation, and is a measure of fracturing of the rock mass.
Table 2. Recombined datasets, describing balancing technique and
number of samples, total, and in training and testing sets. (modified after
Hansen et al. 2022).
Fig. 14. Frequency of rock quality Q-class labels in dataset. Total No. of samples/pr. class
Dataset
Balancing technique
(Class A-E) Total Train Test
ABCDE Not balanced 2400 1680/336 720/144
ABCDE Down-sampled to 1000 700/140 300/60
200 in class B, C, D.
Up-sampled to 200
(without
replacement) A and
E in trainset.
AB/DE Down-sampling 658 460/230 198/99
combined classes to
329 in each class
Fig. 15. Swarm plot of registered Q-values, categorized by Q-class A to 3.3.3 Results
E. Each point is patterned with the registered rock types.
A confusion matrix was used to evaluate the performance of the
The raw input for the ML algorithms is sensor measurements classification model. The matrix compares the actual target
collected every 2-3 cm along the 4-6 m long blasting drill holes values with those predicted by the machine learning model. Fig.
from the drilling machines. The normalised penetration was used 16 describes the validation confusion matrix and accuracy
for rock mass characterization. Input data is processed from raw measurements for each of the five algorithm-pipeline (1) full-
sensor-data to model ready images is described in Hansen et al. class unbalanced dataset denoted CNN ABCDE; (2) RFC
2022. Fig. 16 shows four randomly picked 5-m sections of image ABCDE dataset balanced; (3) CNN ABCDE dataset balanced;
and label from MWD-images and supplied to the ML-algorithms. (4) RFC AB/DE dataset balanced; and (5) CNN AB/DE dataset
Blue colour represents the lowest penetration (presumably strong
balanced, where CNN and RFC stand for Convolutional Neural
rock). High penetration (presumably weak rock) is represented
Network and RFC for Random Forest Classifier. In the figure,
by red and dark colours.
the shade of grey is an indicator of the goodness of the prediction,
3.3.2 Data balancing and splitting with the darkest being above 80% and the entirely while shades
less than 3%. In the figure, the "true" Q-class was the one set by
Figures 14 and 15 indicated that the data is highly unbalanced. humans, as normally done in practice.
Down-sampling to sample size in each class in the range 200-329 Looking at the five confusion matrices, the results are in
and up-sampling, if necessary, were done to balance all class- general unsatisfactory for all models trained including all five (5)
samples to the same size, both without replacement was utilized classes (Parts (1) to (3) in Fig. 17), with total accuracies for all
in this study (see details in Table 2). The down-sampled data was classes ranging from 0.35 – 0.52. The total accuracy gives an
split into train- and test datasets with randomly selected samples, overall evaluation of the number of correct predictions over the
carrying out random sampling such that the class balance after total number of predictions. The value ranges from 0 or no
551
correct predictions, to 1 where 100% of the predictions are
correct.
Based on Parts (2), (3), (4) and (5) of the figure, the CNN-
pipeline performs slightly better than the RFC pipeline. Further
discussion focuses on the CNN-results.
There is an increasing trend of improved performance, both
qualitatively in the confusion matrix structure and in the accuracy
values, when the classes were combined into groups including
more than one class. ABCDE in Fig. 17 (3) is the most
challenging case in predicting Q-class from MWD-images, due
to the necessary number of discrimination borders between
classes. ABCDE has a positive structure for further development,
but the classifier struggles to divide between classes, indicating
the need for more feature-information, e.g. including more
MWD-parameters in the analysis.
The resulting classifier of AB/DE in Fig. 17 (5), both in terms
of accuracy and confusion matrix, is closer to a useful classifier:
if seems to be able to discriminate between a strong and weak
rock mass, with the input of an objective MWD-image of
measured normalised penetration only. It can be argued, when
(3) CNN: ABCDE balanced
optimising classifiers, that it is preferable to divide between good
rock and bad rock, since this is crucial information for face-
engineers, especially for decisions involving the rock mass in
front of the tunnel face. The results for the combined dataset
AB/DE can in that respect be seen as the most promising result
for this dataset.
There are two inherent challenges in this balancing and
splitting task that will reduce or even prohibit the performance of
the models:
Label set by humans: the resulting Q-value from the rock
mass characterisation on the face is a product of many input
parameters, leading to a natural spread in outcome when
the same rock mass is evaluated by different face
engineers.
Imposing a classification on the inherently non-categorical
rock mass: the Q-values will then be spread continuously
along one distribution, meaning that there will be Q-values
at class-boundaries with similar feature information,
leading to difficulties in the discrimination of samples.
(1) CNN: ABCDE unbalanced
(4) RFC: AB/DE balanced
(2) RFC: ABCDE balanced
(5) CNN: AB/DE balanced
552
slides, debris flows and floods, snow avalanches. Rockfalls also
present a serious threat in certain locations.
The Norwegian landslide database indicates that past events
were often hydrometeorological in origin, with the timing of
release related to periods of rapid snowmelt or intense rainfall,
frequently in combination. The event database for the
investigation in this paper comprised landslide release zones and
runout areas triggered by a very intense rainfall on 9–10 June
2011. The inventory was compiled from field survey
observations in conjunction with high-resolution aerial
photographs taken in the weeks following the landslide events.
In total, 86 landslides were identified. This 2011 intensive
rainfall resulted in a comprehensive landslide inventory and was
selected for further investigation in this study. The locations of
the landslides recorded in the database are shown in Fig. 18.
Fig. 17. Confusion matrixes for validation data from recombined
datasets. The confusion matrix shows the number in "%-predicted-class"
for each of the classes, with true labels on the horizontal axes and
predicted label on the vertical axes (e.g., in the top figure (1), the trained
classifier correctly classified 67% of class B samples).
3.4 Landslides
Considering the increasing availability of observational data,
such as remote sensing satellite data, landslide studies adopting
ML algorithms and methods have emerged in the literature with
increasing frequency. Focus is set on landslide detection and
mapping (e.g., Stumpf and Kerle, 2011; Keyport et al. 2018), Fig. 18. Spatial extent of landslide release in study area during 2011
landslide susceptibility mapping (e.g., Pourghasemi and rainfall, Kvam, Norway.
Rahmati, 2018; Mergadi et al. 2020) and temporal forecasting of
landslides (e.g., Yoon et al. 2011; Huang and Xiang, 2018; Spatial landslide mapping is considered to be a binary
Stanley et al. 2020). The following two sections will present two classification in which the landslide index is separated into two
examples on landslide spatial prediction and landslide temporal classes, i.e., landslide and non-landslide. Landslide pixels are as-
forecasting. signed a value of "1", and non-landslide pixels are assigned a
value of "0".
3.4.1 Spatial prediction In this study, the area comprised 535 839 location points, of
which 3 399 were release areas. To avoid unbalanced data points
ML algorithms are frequently used to compute landslide
between landslide and non-landslide, 6 798 non-landslide points,
susceptibility, i.e. the “likelihood of a landslide occurring in a
i.e. twice the number of release areas, were randomly selected.
given area” (Brabb 1984), relying on two standard key
In total, 10 197 location points were considered in the dataset.
assumptions: i) future landslides are more likely to occur under
The 3 399 landslide points were then randomly split into two
conditions that led to slope instabilities in the past; ii) conditions
parts: 70% of the data were used for training the ML landslide
that are directly or indirectly linked to slope failures can be
susceptibility models and 30% were kept for model validation.
collected and used to build predictive models of landslide spatial
The 6 798 non-landslide points were also randomly split into the
occurrence. These two assumptions are common to both
same ratio of 70/30.
statistical and ML approaches developed for landslide
Eight static landslide conditioning factors and three time-
susceptibility analysis and zoning (e.g., Varnes 1984;
dependent triggering factors were selected for input into the
Reichenbach et al. 2018).
machine learning models. The landslide conditioning factors
A very recent overview of the most popular machine learning
were: slope angle, aspect, plan curvature, profile curvature, flow
techniques available for landslide susceptibility studies is
accumulation, flow direction, distance to rivers, and distance to
presented by Mergadi et al. (2020), who also statde that "only a
roads. The three time-dependent triggering factors were: total
handful of researchers use machine learning techniques in
water content, saturation, and rainfall These are illustrated in 11
landslide susceptibility mapping studies." They identified 10
vignettes on Fig. 19.
authors who were responsible for approximately 50% of the
The results of a Receiver Operating Characteristic (ROC)
published landslide susceptibility studies adopting neural
networks, 70% of studies adopting random forest (RF) analysis for three ML models and for a numerical analysis
algorithms, 85% of studies adopting decision tree (DT) (TRIGRS) are shown on Fig. 20. The three different landslide
algorithms, and 85% of studies adopting support vector machines models used the validation dataset with 2 040 non-landslide
(SVM) algorithms. points and 1 020 landslide points. An AUC indication (Area
Liu et al. (2021) introduced three machine learning (ML) Under the Curve) was calculated to measure the performance
algorithms, the 'ensemble' Random Forest (RF), the 'ensemble' measurement for the ML models.
Gradient Boosted Regression Tree (GBRT) and the MultiLayer The ROC curves show that the GBRT and RF models have
Perceptron neural network (MLP) and applied them to the spatial the highest AUC value as 0.99, followed by the MLP model
modelling of shallow landslides near Kvam in eastern Norway (0.97). The physically-based numerical model TRIGRS model
(Fig. 18). The major natural hazards in the study area are soil (0.87). TRIGRS is a coupled hydro-mechanical slope-stability
assessment model that combines modules for the infiltration and
553
subsurface flow of storm water with those for runoff routing and study area, as computed by TRIGRS is shown in the upper right
slope stability. The steady infiltration rate, saturated hydraulic panel. The three ML approaches resulted in accuracies above
conductivity, and slope angle determine the steady (initial) flow 90%. This accuracy was much better than the prediction accuracy
direction. The transient component assumes one-dimensional of the geomechanical model TRIGRS which accuracy was 87%.
(1D), vertical downward flow. TRIGRS is a strong tool for
shallow landslide estimation, particularly in large areas where
input data for simulation is not fully available, and has been
widely used in the past 15 years or so, e.g., Park et al. (2013).
Fig. 21. Prediction of landslide occurrence by RF, GBRT and MLP
models: black circles (True Negative) represent correct predictions of
non-landslide locations; red rectangles (False Negative) represent
incorrect predictions of actual landslide locations; blue rectangles (False
Positive) represent incorrect predictions of non-landslide locations; and
green rectangles (True Positive) represent correct predictions of actual
Fig. 19. Maps of landslide conditioning (8) and triggering factors (3) in landslide locations; the Factors of computed by TRIGRS are shown in
Kvam study area: (a) slope angle, (b) aspect, (c) plan curvature, (d) the upper right panel.
profile curvature, (e) flow accumulation, (f) flow direction, (g) distance
to rivers, (h) water content, (i) saturation, (j) rainfall from radar, and (k) 3.4.2 Temporal forecasting
distance to roads.
The prediction of the evolution of landslides in time, and more
generally forecasts of the time evolution of key factors affecting
the slope safety level, can be performed at both a global/regional
scale and at a local slope scale. The choice of the scale is usually
linked to the type of monitored parameters, which in turn is
related to the type of landslides. Typically, regional scale
predictions are based on rainfall monitoring, geomorphological,
and hydro-meteorological approaches, while local slope scale
predictions take advantage of a geotechnical engineering method
relating displacement or other monitoring data to the time of
failure (Intrieri et al. 2019).
There is a relationship between monitored parameters and
types of landslide; for example, for shallow landslides triggered
by extreme precipitation events or by a combination of hydro-
meteorological events, meteorological data dominates the
monitored parameters. For slow moving deep-seated landslides,
displacements can be a crucial input to assess slope behavior.
New data assembling methods and Internet of Things (IoT)
techniques have recently started to provide large datasets of
monitored data that can be advantageously be used by ML
techniques for temporal forecasting of landslides.
Landslide displacement forecasting is an essential component
for developing modern early warning systems. It can be used to
Fig. 20. The ROC curves and performance (AUC) of three ML and one set warning thresholds and to recognize when a landslide
numerical landslide models, with testing dataset only. undergoes a sudden acceleration and then failure. Time series of
Figure 21 illustrates the predictions of landslide occurrence real-time data collected from landslide monitoring systems, e.g.
with the RF, GBRT and MLP models. The factor of safety in the Geophones, Interferometric Synthetic Aperture Radar (InSAR),
554
and Global Navigation Satellite System (GNSS), along with MAPE (%) RMSE (mm)
triggering data, e.g. water level and precipitation, provide critical Landslide
inputs to ML modelling. However, the prediction of landslide LSTM GRU RF LSTM GRU RF
displacement that changes over time is very challenging and it is
Baishuihe 10.5 10.7 13.5 7.5 8.0 6.5
inevitably linked to complex deformation mechanisms in the
slope.
Bazimen 8.6 8.4 14.2 13.8 12.6 26.0
Liu et al. (2020) explored how several machine learning
algorithms (long short term memory (LSTM), random forest
Baijiabao 14.2 15.6 24.4 9.2 8.8 10.5
(RF), support-vector machine (SVM), and gated recurrent unit
(GRU)) might predict ground displacements for slopes with
distinct step-wise displacement characteristics in the Three
Gorges Dam reservoir (China). The predictions by the LSTM and GRU models were similar
Figure 22 compares the measured and predicted and had approximately the same margin of error. For the
displacements for three landslides with three of the ML Baishuihe and Baijiabao landslides, the results with the three
algorithms, using the testing dataset only. The periodic landslide machine learning algorithms were approximately equivalent.
displacement was initiated and aggravated by rainfall and by the The RMSE values are less than 10 mm and the MAPE values are
annually periodic lowering and raising of the water level in the between 10% and 16% for the LSTM and the GRU models. The
Three Gorges dam reservoir. Table 3 compares the accuracy LSTM neural network and the GRU models agreed especially
indicators MAPE (Mean Absolute Percentage Error) and RMSE well with the measurements during the period of the step-wise
(Root Mean Square Error) for each of the three ML models for increase in the displacement (May to July), as illustrated in Fig.
the three landslides. 22. The RF model gives higher values of MAPE, although the
RMSE values are similar to those from the other models. For the
Bazimen landslide, the LSTM and GRU models appear to predict
the displacement more reliably than the RF model. The reason
for this difference might be because the RF algorithm, as a
combination of binary decision trees, is not able to predict time
series observations as well as the other deep learning algorithms.
For the same input from the Baijiabao Landslide, the dynamic
model LSTM showed higher prediction accuracy than the static
model SVM. The LSTM model could establish connections
between the data in different time steps. The LSTM model learns
rules from historical information and then applies these rules to
the current step. The LSTM model can judge, filter and
remember the information from previous time steps. The useful
information is remembered; the useless information is forgotten.
For the SVM model, connections between different time steps
are not possible. The SVM model learn rules from a time step but
cannot use historical data, and therefore cannot model the
response of landslide deformation as well.
Although good performances have been achieved in the
deterministic prediction, no information was given on the
reliability of the prediction results because these studies did not
consider different types of uncertainties in the prediction.
However, many different uncertainties exist, due to both inherent
and external factors in the evolution process of step-like
landslides (Wu et al. 2013).
Ge et al. (2021) developed a novel approach for the interval
prediction (i.e. uncertainty) of landslide with step-like
displacement pattern in the Three Gorges Reservoir (TGR) area
using Density-Based Spatial Clustering of Applications with
Noise (DBSCAN), Synthetic Minority Oversampling Technique
and Edited Nearest Neighbor (SMOTEENN) based Random
Forest (RF) and bootstrap-Multilayer Perceptron (MLPs).
DBSCAN was employed to carry out clustering analysis for
different deformation states of the landslide with step-like
displacement pattern. The SMOTEENN based RF classifier was
trained to deal with imbalanced classification problems. A
dynamic switching prediction scheme to construct high quality
Prediction Intervals (PIs) using bootstrap-MLPs was established.
The final interval prediction results of the Bazimen landslide
are presented in Figure 23. It shows that the widths of the
constructed prediction intervals (i.e. the uncertainties) vary
significantly from one point in time to another. Since the
Fig. 22. Measured and predicted periodic displacements for the validation influence factors of landslide displacements vary with time, the
dataset only, using 3 ML models, the LSTM, GRU, and RF for (a) uncertainties keep changing accordingly. The results show that
Baishuihe landslide; (b) Bazimen landslide; and (c) Baijiabao landslide. the widths of the prediction intervals of the mutational points are
Table 3. Comparison of accuracy of predicted periodic displacement for larger than that of the stationary points, which indicates that the
three landslides. (Liu et al. 2020). risk level and the uncertainty of the predicted landslide
555
deformation are relatively high under the influence of the heavy
rainfall and rapid drawdown of the reservoir.
Training Testing
Width of the prediction interval (mm)
Accumulated displacement (mm)
Incorrect
classification
Point
Point not been
covered
Time (month-year)
Fig.23. Interval prediction results of at 96% confidence level of the Bazimen landslide.
calibration, but did not perform as well as the ML models. The
two approaches are, however complementary alternatives. Data-
4 DISCUSSION driven models are often called upon only when the use of
physically-based models is deemed unfeasible or prohibitive in
Machine Learning (ML) algorithms aim at making accurate costs. In landslide studies, ML algorithms are currently being
predictions, while the understanding of the "why" may be often adopted in lieu of statistical techniques. However, physical or
only come in second place (Therani et al. 2022). Applications of numerical models can be used in combination with ML models,
ML methods in landslide studies should therefore focus on for example, to 1) help explain the ML model, 2) complement
problems where there is a need for predictions. Features detected the number of data points to train the ML model, 3) produce
by ML, for instance related to relative significance of synthetic data for data-scarce problems.
conditioning or triggering factors, will however help understand ML and statistically-based approaches share many common
the undergoing processes. characteristics. The spatial landslide forecasting studies adopting
The paper presented positive ML applications on (1) site ML algorithms draw from the experience accumulated since the
characterization (stratification and soil characterization with the seminal publication by Varnes (1984) on bivariate and
piezocone, derivation of the undrained shear strength of clay, and multivariate statistical techniques and procedures for landslide
3D ground model integrating geotechnics and geophysics); (2) susceptibility assessment and zoning.
braced excavations (comparing four ML models with the results There is no consensus on an "optimal" ML (or deep learning
of finite element simulations); (3) tunneling (predicting the rock DL) algorithm for the different types of studies illustrated in this
mass quality at the tunnel face for rapid decision-making); and paper. There is a growing tendency in the literature to propose
(4) landslides assessment (spatial landslide susceptibility the systematic use of an ensemble of algorithms, rather than
prediction, temporal landslide forecasting, and displacement unique ML algorithms such as Random Forest (RF), and then
interval prediction). In each case, the ML models adopted choose the best-performing one. The general opinion is that more
performed well, and better than more traditional models, complex algorithms show higher accuracy, although simpler
although a number of data balancing techniques had to be trialed algorithms with proper feature can outperform them.
to obtain optimum results. The selection of an algorithm depends on the type and
With ML techniques, there will also be a need for sufficient quantity of available data. In general, DL algorithms are not
quantity of data. Data-driven methods, such as ML algorithms, expected to outperform conventional ML if the size of training
are not useful if the necessary data are not available. For data is not very large. Number of features, attributes and
landslide detection and landslide susceptibility mapping, where preference over feature engineering also affect this choice.
satellite images at various resolutions are available, data Tehrani et al. (2022) suggested that for structured data,
availability is less problematic. However, for many cases in conventional ML algorithms are to be preferred, whereas for
geotechnical engineering, good quality, comparable and relevant unstructured data (e.g. text, video, imagery, etc.), DL algorithms
data are often in limited quantity. This restricts the application of can be more suitable.
ML algorithms. The ML approach can be very useful to help forewarn of
Model generalization is an important aspect of ML modelling future scenarios, especially if these fall within the realm of the
that is often neglected. There is a lack of benchmark case studies available data body, because ML models have already learned
available for testing various ML methods. from a broad spectrum of data. For extrapolation, ML may not
The paper presented one comparison of ML models with a be the preferred approach.
geomechanical numerical analysis model. Compared to ML- Especially for landslide susceptibility mapping, and spatial
based models, the geomechanical model requires fewer data for and temporal forecasting, the ML approach offers promising
556
opportunities. Over the years, the complexity of ML algorithms 7 REFERENCES
used in landslide studies has been matching the rapid
development that is occurring in the AI/ML community. At the Barton,N., Lien, R. and Lunde, J. 1974. Engineering classification of
same time, ML still has a long path to follow in the landslide rock masses for the design of tunnel support. Rock Mech.
Felsmechanik Mécanique des Roches, 6(4), 189–236.
community. Brabb, E.E. 1984. Innovative approaches to landslide hazard mapping.
It is also expected that the application of ML algorithms to Proc. 4th Int. Symp. Landslides, 1, 307-323.
other areas of geotechnical research than those discussed in this Chawla, N.V., Bowyer, K., Hall, L. and Kegelmeyer, W. 2002. SMOTE:
paper will expand significantly in the coming years. synthetic minority over-sampling technique. Journal of artificial
intelligence research, 16, 321–357.
Chen, J., Vissinga, M., Shen, Y., Hu, S., Beal, E. and Newlin, J. 2021.
5 CONCLUSIONS AND OUTLOOKS Machine Learning–Based Digital Integration of Geotechnical and
Ultrahigh–Frequency Geophysical Data for Offshore Site
The paper gives an overview of some of the applications of Characterizations. Journal of Geotechnical and Geoenvironmental
Machine Learning (ML) in geotechnical engineering, mainly Engineering, 147(12), 04021160.
from the perspective of how using the ML techniques will help Dikshit, A., Pradhan, B. and Alamri, A.M. 2021. Pathways and
improve the practice of geotechnical engineering. The ML challenges of the application of artificial intelligence to geohazards
approach was briefly summarized to outline the emerging modelling. Gondwana Research, 100, 290–301.
Erharter, G. H., Hansen, T. F., Liu, Z. and Marcher, T. 2021.
opportunities and demonstrate their use. Reinforcement learning based process optimization and strategy
Machine Learning is particularly helpful and relevant for development in conventional tunneling. Automation in
geotechnical engineering, where data are often sparse, the Construction, 127, 103701.
variability in the characteristics of soil and rock can often be Erharter, G. H., Oberhollenzer, S., Fankhauser, A., Marte, R. and
high, and the measurements, especially in situ, are influenced by Marcher, T. 2021. Learning decision boundaries for cone penetration
both measurement techniques and data transformation
uncertainties. Time may be a key factor, for example, for the Engineering, 36(4), 489-503.
development of an Early Warning System, where it is crucial to Forsberg, C. F., Lunne, T., Vanneste, M., James, L., Tjelta, T. I., Barwise,
make a decision within in a limited time, based on real time A., and Duffy, C., 2017. Synthetic CPTs from Intelligent Ground
Models based on the integration of geology, geotechnics and
monitored data and using threshold values with uncertainties. geophysics as a tool for conceptual foundation design and soil
ML models can help reduce the uncertainties and obtain answers investigation planning. In Proceedings Offshore Site Investigation
to queried rapidly. Geotechnics 8th International Conference Proceeding, London, UK,
Machine Learning (ML) in geotechnical engineering is a 2017, Society for Underwater Technology, p. 1254-1259.
dynamic new frontier destined to have many more applications, Gal, Y. 2016. Uncertainty in Deep Learning. University of Cambridge.
including in the area of virtual reality testing (i.e. modelling Gholamy, A., Kreinovich, V. and Kosheleva, O. 2018. Why 70/30 or
virtually millions of laboratory tests) and the new developments 80/20 Relation Between Training and Testing Sets: A Pedagogical
within digital twinning. Explanation. University of Texas at El Paso, Departmental Technical
ML is an inspiration for the geotechnical community for Reports (CS). 1209.
Ge, Q., Sun, H., Liu, Z., Yang, B., Lacasse, S. and Nadim, F. 2021. A
developing and implementing innovative data-driven methods novel approach for displacement interval forecasting of landslides
for a variety of new applications, also beyond those given as with step-like displacement pattern, Georisk: Assessment and
examples herein. There is surely an initial challenge to start Management of Risk for Engineered Systems and Geohazards, DOI:
implementing ML algorithms appropriately, but the authors are 10.1080/17499518.2021.1892769.
confident that the efforts will advance the ability to predict geo- Hansen, T.F., Erharter, G.H., Marcher, T., Liu, Z. and Torresen, J. 2022.
behaviour and reduce risk, for example, with improved "Alert- Improving face decisions in tunnelling by machine learning based
Alarm-Action" levels in an early warning system. MWD-analysis. Geomechanics and Tunnelling, accepted for
It is also quite obvious that the younger generation has a keen publication.
Hosseini, S. A. A., Mojtahedi, S. F. F. and Sadeghi, H. 2020.
interest in ML techniques. The entire profession needs therefore
Optimisation of deep mixing technique by artificial neural network
to keep up with this new development. There is potential for based on laboratory and field experiments. Georisk: Assessment and
wider applications in practice in the future, but further research Management of Risk for Engineered Systems and Geohazards,
is also needed to meet this objective. 14(2), 142-157.
To move forward industry is a necessary partner for machine Huang L. and Xiang, L.Y. 2018. Method for meteorological early
learning-based studies in geotechnical engineering for two warning of precipitation-induced landslides based on deep neural
reasons. One, machine learning must bring transformative value network. Neural Process Lett 48, 1243–1260.
to industry, not universities. This value is ultimately linked to the Intrieri, E., Carlà, T., Gigli, G. 2019. Forecasting the time of failure of
end-users. Two, real data are generated and owned by projects landslides at slope-scale: A literature review. Earth-Science
Reviews. 193, 333–349.
and real data (not simulated data) are absolutely necessary
ingredients for machine learning development. Artificial Intelligence and Machine Learning in Geotechnical
Feedback arenas where industry can interact with the Engineering”. Geosciences, 11(10), 399.
researchers, discuss goals and results and match those with Jong, S.C., Ong, D.E.L. and Oh, E. 2021. State-of-the-art review of
industry needs. The aim is to create a 'revolving door' for industry geotechnical-driven artificial intelligence techniques in underground
and researchers in a learning partnership: industry to define needs soil-structure interaction. 113, 103946.
and give feedback, researchers to provide innovation to solve Keyport, R. N., Oommen, T., Martha, T. R., Sajinkumar, K. S. and
present needs and perspective to respond to future needs. Gierke, J. S. 2018. A comparative analysis of pixel-and object-based
detection of landslides from very high-resolution images.
International journal of applied earth observation and
6 ACKNOWLEDGEMENTS geoinformation, 64, 1-11.
Klinkvort, R. T., Sauvin, G., Forsberg C. F. and Vanneste, M. 2020.
For the preparation of this manuscript, the authors wish to thank Integrated Ground Modelling from a Geotechnical Perspective. In Z.
Westgate (Ed.), 4th International Symposium on Frontiers in
the following colleagues for their valuable input and insight:
Offshore Geotechnics, Austin, Texas. November 8-11, 2021.
Professor Wengang Zhang (Chong qing University), Mr Tom Hawthorne: Deep Foundations Institute.
Frode Hansen (NGI) and Dr Mingliang Zhou (Tongji Kordjazi, A., Pooya Nejad, F. and Jaksa, M. B. 2014. Prediction of
University). Ultimate Bearing Capacity of Axially Loaded Piles Using a Support
557
Vector Machine Based on CPT Data. Computers and Geotechnics, Reichenbach, P., Rossi, M., Malamud, B.D., Mihir, M. and Guzzetti,
55, (1), 91–102. F. 2018. A review of statistically-based landslide susceptibility
Liu, D., Lin, P., Zhao, C. and Qiu, J. 2021. Mapping horizontal models. Earth-Science Reviews, 180, 60-91.
displacement of soil nail walls using machine learning approaches. Robertson PK (1990) Soil classification using the cone penetration test.
Acta Geotechnica, 16, 4027–4044. Can Geotech J, 27(1), 151–158.
Liu, J., Sauvin, G. and Yang, S. 2021. Application of machine learning Sastre, C., Breul, P., Navarette, M.B. and Bacconnet, C. 2021. Automatic
on soil classification based on CPTU data. The Sixth International soil identification from penetrometric signal by using artificial
Conference on Geotechnical & Geophysical Site Characterization: intelligence techniques. Can. Geotech. J., 58, 1148–1158.
ISC'6, Budapest, Hungary, 26-29 September 2021. Sauvin, G., Vanneste, M., Vardy, M.E., Klinkvort, R.T. and Forsberg,
Liu, Z., Gilbert, G., Cepeda, J.M., Lysdahl, A.O.K., Piciullo, L., Hefre, C.F. 2019. Machine Learning and Quantitative Ground Models for
H. and Lacasse, S. 2021. Modelling of shallow landslides with Improving Offshore Wind Site Characterization. Paper to Offshore
Machine Learning algorithms. Geoscience Frontiers, 12(1), 385- Technology Conference (OTC 2019), 6-9 May 2019, Houston,
393. #OTC-29351-MS.
Liu, Z., Guo, D., Lacasse, S., Li, J., Yang, B., and Choi ,J.C. 2020. Stanley, T.A., Kirschbaum, D.B., Sobieszczyk, S., Jasinski, M.F., Borak,
Algorithms for intelligent prediction of landslide displacements. J.S. and Slaughter, S.L. 2020. Building a landslide hazard indicator
Journal of Zhejiang University Science A (Appl Phys & Eng), 21(6), with machine learning and land surface models. Environmental
412-429. Modelling and Software 129, 104692.
Liu, Z., Lacasse, S., Nadim, F., Vanneste, M. and Yetginer, G. 2015. Stumpf, A. and Kerle, N. 2011. Object-oriented mapping of landslides
Accounting for the spatial variability of soil properties in the using Random Forests. Remote sensing of environment, 115(10),
reliability-based design of offshore piles. Proceedings of the 3rd 2564-2577.
International Symposium on Frontiers in Offshore Geotechnics. Tehrani, F.S., Calvello, M., Liu, Z., Zhang, L. and Lacasse, S. 2022.
Oslo, Norway. pp. 1375-1380. Machine learning and landslide studies: recent advances and
Ma, Z. and Mei, G. 2021. Deep learning for geological hazards analysis: applications. Natural Hazards, under review.
Data, models, applications, and opportunities. Earth-Science Varnes, D.J. IAEG Commission on Landslides and other Mass-
Reviews, 223, 103858 Movements, 1984. Landslide Hazard Zonation: A Review of
Ma, Z., Mei, G. and Piccialli, F. 2021. Machine learning for landslides Principles and Practice. The UNESCO Press, Paris, 63 pp.
prevention: a survey. Neural Computing and Applications, 33, Wang, H., Wang, X, Wellmann, J. F. and Liang, R.Y. 2019. A Bayesian
10881–10907. unsupervised learning approach for identifying soil stratification
Mergadi, A., Yunus, A.P., Dou, J., Whiteley, J., ThaiPham, B., Bui, D.T., using cone penetration data. Canadian Geotechnical Journal, 56(8),
Avtar, R. and Abderrahmane, B. 2020. Machine learning methods 1184-1205.
for landslide susceptibility studies: A comparative overview of Wang, X, Wang, H. and Liang, R.Y. 2021. Unsupervised and
algorithm performance. Earth-Science Reviews, 207, 103225. simultaneous stratigraphic interpretation of CPT soundings at site
Moayedi, H., Mosallanezhad, M., Rashid, A.S.A., Jusoh, W.A.W. and scale. Int. J. Geomech., 21(8): 04021130.
Muazu, M.A. 2020. A systematic review and meta-analysis of Wu, Y. P., C. Cheng, G. F. He, Zhang, Q. 2013. Landslide Stability
artificial neural network application in geotechnical engineering: Analysis Based on Randomfuzzy Reliability: Taking Liangshuijing
theory and applications. Neural Comput & Applic, 32, 495–518. Landslide as a Case. Stochastic Environmental Research and Risk
Molina-Gómez, F., da Fonseca, A.V., Ferreira, C., Sousa, F. and Bulla- Assessment 28: 1723–1732.
Cruz, L.A. 2021. Defining the soil stratigraphy from seismic Xiao, T., Zou, H.F., Yin, K.S., Du, Y. and Zhang, L. 2021. Machine
piezocone data: A clustering approach. Engineering Geology, 287, learning-enhanced soil classification by integrating borehole and
106111. CPTU data with noise filtering. Bull Eng Geol Environ, 80, 9157–
NGI 2020. CPTU prediction model. Techncial Note 20190685-01-TN. 9171.
Paniagua, P., D'Ignazio, M., L'Heureux, J.S., Lunne, T. and Karlsrud, K. Xie, J., Huang, J., Zeng, C., Jiang, S.H. and Podlich, N. 2020. Systematic
2019. CPTU correlations for Norwegian clays: an update. AIMS Literature Review on Data-Driven Models for Predictive
Geosciences, 5(2), 82-103. Maintenance of Railway Track: Implications in Geotechnical
Park, D.W., Nikhil, N.V., and Lee, S.R. 2013. Landslide and debris flow Engineering. Geosciences, 10, 425.
susceptibility zonation using TRIGRS for the 2011 Seoul landslide Yoon, H., Jun, S., Hyun, Y., Bae, G. and Lee, K., 2011. A comparative
event. Nat. Hazards Earth Syst., 13(11), 2833–2849. study of artificial neural networks and support vector machines for
Pathmanandavel, S. and MacRobert, C.J. 2020. Digitisation, predicting groundwater levels in a coastal aquifer. J. Hydrol. 396(1),
sustainability, and disruption – promoting a more balanced debate 128–138.
on risk in the geotechnical community, Georisk: Assessment and Yang, B., Yin, K., Lacasse, S. and Liu, Z. 2019. Time series analysis and
Management of Risk for Engineered Systems and Geohazards, long short-term memory neural network to predict landslide
14(4), 246-259. displacement. Landslides, 16(4), 677-694.
Pejman, T., Serveh, K., Tao, B. and Muhammad S. 2020. Machine
learning in geo- and environmental sciences: From small to large of Soil Properties for Geotechnical Design: Review, Tool
scale. Advances in Water Resources, 142, 103619. Development and Comparison. Arch Computat Methods Eng 29,
Pham, B.T., Son, L.H., Hoang, T., Nguyen, D. and Bui, D.T. 2018. 1229–1245.
Prediction of shear strength of soft soil using machine learning Zhang, W. and Chee Goh, A. T. 2013. Multivariate adaptive regression
methods. CATENA, 166, 181-191. splines for analysis of geotechnical engineering systems. Computers
Pourghasemi, H.R. and Rahmati, O. 2018. Prediction of the landslide and Geotechnics, 48, 82-95.
susceptibility: Which algorithm, which precision? Catena, 162, 177- Zhang, W., Li, H., Li, Y., Liu, H., Chen, Y. and Ding, X. 2021.
192. Application of deep learning algorithms in geotechnical engineering:
Ranasinghe, R. A. T. M., Jaksa, M. B., Pooya Nejad, F. and Kuo, Y. L. a short critical review. Artif Intell Rev 54, 5633–5673.
2017. Predicting the Effectiveness of Rolling Dynamic Compaction Zhang, W., Wu, C., Zhong, H., Li, Y. and Wang, L. 2021. Prediction of
Using Genetic Programming and Cone Penetration Test Data. Proc. undrained shear strength using extreme gradient boosting and
of Institution of Civil Engineers – Ground Improvement, 170(4), random forest based on Bayesian optimization. Geoscience
193–207. Frontiers, 12(1), 469-477.
Rauter, S. and Tschuchnigg, F. 2021. CPT Data Interpretation Employing Zhang, W., Zhang, R., Wu, C., Chee Goh, A. T., Lacasse, S., Liu, Z. and
Different Machine Learning Techniques. Geosciences, 11, 265. Liu, H. 2020. State-of-the-art review of soft computing applications
Ray, R., Kumar, D., Samui, P., Roy, L. B., Goh, A. T. C. and Zhang, W. in underground excavations. Geoscience Frontiers, 11(4), 1095-
2021. Application of soft computing techniques for shallow 1106.
foundation reliability in geotechnical engineering. Geoscience Zhou, C., Kong, T., Zhou, Y., Zhang, H. and Ding. L. 2019.
Frontiers, 12(1), 375-383. Unsupervised spectral clustering for shield tunneling machine
Reale, C., Gavina, K., Libri b, L. and Juri -Ka uni b, D. 2018. monitoring data with complex network theory. Automation in
Automatic classification of fine-grained soils using CPT Construction, 107, 102924.
measurements and Artificial Neural Networks. Advanced
Engineering Informatics, 36, 207–215.
558
Planetary Geotechnics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
ABSTRACT: The success of the Chang' E 4 and Chang' E 5 missions marks a phased victory for China's lunar exploration project,
which has attracted worldwide attention. Chongqing University is responsible for the research and development of biological
equipment payload (BEP) in Chang' E 4 explorer. This paper describes the design and test of BEP, and introduces the germination
and growth of organisms in the tank. The successful germination of cotton seeds marks the first time that human beings have
cultivated crops on an alien planet. Chang' E 5 successfully sent rock and soil samples collected from the Moon back to the Earth.
Research on the geological and geotechnical properties of the Moon is ongoing. These lunar soil samples collected in the young
geological areas have unique physical and mechanical properties. This paper also introduces the basic situation of Chang' E 5 explorer
and the process of sample collection. On the basis of summary of the existing research, the authors put forward some outlooks in the
field of lunar geotechnical engineering in the future, including improving the biotechnology to achieve geotechnical fixation, lunar
geotechnical testing and numerical simulation, and engineering construction on the lunar surface with low gravity, high radiation and
vacuum environment.
RÉSUMÉ: Le succès des missions Chang’E-4 et Chang’E-5 a marqué la victoire progressive du projet chinois d'exploration
lunaire.L'Université de Chongqing est responsable de la recherche et du développement de la charge utile de l'équipement biologique de
la sonde Chang’E-4.Cet article présente la conception et l'essai de la charge utile de l'équipement biologique, ainsi que la germination et
la croissance des organismes dans le bassin.Le succès de la germination des graines de coton a marqué la première fois que l'homme a
cultivé des cultures extraterrestres.Chang’E - 5 a réussi à ramener sur terre des échantillons de roche et de sol prélevés sur la lune.L'étude
des caractéristiques géologiques et géotechniques de la lune est en cours.Ces échantillons de sol lunaire prélevés dans de jeunes régions
géologiques ont des propriétés physiques et mécaniques uniques.La situation de base et le processus d'échantillonnage de la sonde
Chang’E - 5 sont introduits.Sur la base d'un résumé des résultats et des lacunes de la recherche existante, certaines perspectives d'avenir
dans le domaine de l'Ingénierie géotechnique lunaire sont présentées, y compris l'amélioration de la technologie biologique pour réaliser
la fixation géotechnique, l'essai géotechnique lunaire et la simulation numérique.La surface de la lune est construite dans un
environnement de faible gravité, de rayonnement élevé et de vide.
KEYWORDS: Chang' E program; biological equipment payload; Chang' E explorer; lunar rock and soil; lunar environment.
1 INTRODUCTION. stimulate the enthusiasm and innovation potential of the public
in exploring the universe, the creative design of lunar exploration
The far side of the Moon has never been directly explored before payload was designed and organized to send biological
the Chang’E-4 probe landed on the back of the Moon. On equipment payload to the Moon. It is the first popular science
January 3, 2019, the Chang’E-4 probe arrived at the Von Kármá payload to land on the back of the Moon, and it is also the first
n crater, located in the South Pole-Aitken (SPA) basin, with the biological experiment on the Moon. The original idea of the
lunar rover Yutu-2 and the biological equipment payload (BEP) experiment was to build a small recycling ecosystem in the lunar
developed by Chongqing University. Chang’E-4 also became the lander, which imitated the Earth's ecosystem. The first step is to
first spacecraft soft-landing on the Moon's uncharted side never make sure that the plant seeds germinate or eggs hatch in the
visible from the Earth (Jones 2019; Wang and Liu 2016). lunar surface environment. The plan also includes showing the
The probe comprises of relay satellitia, a lander and a rover. whole process of seed germination, seedling growth and blossom
The relay satellite called “Queqiao” is located at the mission orbit of plants under the lunar surface environment, or egg hatching,
around the L2 translational point of the Earth & Moon. The relay larva growth and development, and even flying to complete a life
satellite connects the Chang’E-4 lander and the rover on the back cycle in the future.
of the Moon with the command center on Earth. The lander is In general, the biological experiment on lunar surface a
equipped with landing camera, laser three-dimensional imaging chieved the following four objectives:
sensor, terrain camera, low-frequency radio spectrometer, lunar 1. To explore the growth as well as the photosynthesis effect of
neutron and radiation dose detector and other equipment. The animals and plants on the lunar surface under conditions of
BEP is also fixed in the lander. The scientific detection low gravity, strong radiation and natural light;
instruments on the rover Yutu-2 include panoramic camera, lunar 2. Through the implementation of BEP, to promote the
radar, infrared imaging spectrometer and neutral atom detector. popularization of biological knowledge and deepen the
These scientific equipment ensures the smooth landing and understanding of biological principles;
operation of Chang’E-4, and can carry out physical, geological, 3. Publicize the achievements of the lunar exploration project in
imaging and biological tests on the Moon (Liu et al. 2019; Tang China, stimulate people's interest in cosmic exploration and
et al. 2020; Zhang et al. 2020).
scientific research, and enhance people's awareness of
In order to give full play to the driving role of the lunar
environmental protection;
exploration project in the field of science and technology, and
561
4. To achieve some hints or indications for future geotechnical combination of semiconductor cooler, electric heater, radiator
engineering on the Moon and heat insulation measures is finally determined to realize the
internal intelligent temperature control of BEP. Finally, the
device passed thermal vacuum test, low temperature storage test
2 BIOLOGICAL EXPERIMENT and heat balance test.
In order to realize the photosynthesis of organisms in the tank
2.1 Design of BEP (Xie et al. 2017), a light guide tube is left at the top of the device
to guide sunlight into the tank, and a water storage tank is placed
In order to complete the development of BEP project, the team,
at the bottom (as shown in Figure 1). These additional openings
after more than 50 times of design, optimization, analysis and
require high air tightness of the device. Therefore, we did some
other relevant work, finally determined the scheme, completed
sealing performance test, including helium mass spectrometry
nearly 150 reports and more than 100 technical reports, and
single point leak detection and completed integral leak detection
completed more than 100 engineering and biological verification
by bell jar pressure change method. Some light guide
tests. The team also completed 17 times of structural and thermal
performance test and fixation effect and dissolution test of water
optimization, 12 times of complete mechanical simulation
soluble cotton (shown in Figure 3) were also carried out in design
analysis, and 6 times of structural antecedent machining. The
processing.
final configuration of test tank is shown in Figure 1.
Figure 3. Photosynthesis performance test of BEP: (a) Fixation effect
of water soluble cotton and its dissolution in water (b) Light guide
performance test
2.2 Position and structure of BEP
BEP is located at the bottom of Chang’E-4 lander, as shown in
Figure 4. The payload is composed of structure module, thermal
control module, control module, light guide module, biological
module and water tank. In order to meet the requirements of
biological test, the load tank is equipped with air conditioning
system, light pipe and water tank. The external insulation layer
Figure 1. Structural design of BEP: (a) Plan view of the device in the tank of the tank can withstand the test of severe temperature difference
(b) Schematic diagram of the position of the device in the tank (c) Overall on the surface of the Moon; the internal air conditioning system
appearance drawing (d) Physical picture. with automatic temperature adjustment ensures that the
In order to ensure the safety of on-orbit work, the tank via the temperature sensor. In the
structure parts are all processed by high-performance aluminum tank, plants can absorb the natural light of the Moon through the
alloy, and the anti-corrosion treatment is carried out, which is light tube on the top cover plate of the load, and produce oxygen
verified by analysis and test. These tests include acceleration test, through photosynthesis. A water bag and electromagnetic pump
impact test, destructive pressure test, sinusoid and random with 18ML biological water were carried in the water tank. After
vibration tests. In addition, the structural safety and mechanical landing on the Moon and power on, the lander received the
properties of the tank in space environment are studied by ground instruction to render the electromagnetic pump work. The
numerical simulation. Some results of numerical simulation are water in the water bag was released to the surface of the
shown in Figure 2. biological cabin through the water pipe to provide water for
biological growth.
Figure 2. Numerical simulation of mechanical properties of tank: (a) The
contour plot of maximum displacement of tank body (b) The contour plot Figure 4. Position and components of BEP.
of stress of the seal cover.
The BEP carries six kinds of organisms including potato,
Aiming at the phenomenon of long-term weak air leakage at Arabidopsis, rape, cotton, Drosophila and yeast. The total weight
the low temperature of - of the load is 2.6kg. In addition, there are 18 ml of water, soil,
sealing material performance optimization and other methods. In air, thermal control and two cameras recording the growth status
order to maintain a stable temperature suitable for biological of organisms. The total weight is 2.608 kg, and the growth space
survival in the temperature range of - of organisms is about 1 liter. The location of the modules and the
than 20 times of thermal design and optimization, the
562
distribution of various organisms are shown in Figure 5. The six
organisms carried in the payload constitute a micro ecosystem
with producers, consumers and decomposers. Among them,
plants produce oxygen and food for all organisms to "consume";
as consumers, Drosophila and yeast decomposer produce carbon
dioxide by consuming oxygen for photosynthesis. In addition,
yeast can decompose plant and Drosophila waste to grow, and
yeast can be used as food for Drosophila.
Figure 7. Camera captures of the ground-controlled BEP and the BEP on
the Moon. A, B, C, D and E are the pictures captured 0 h, 22 h, 57 h, 82
h and 190 h after the WIT in the ground controlled BEP, respectively. F,
G, H, I and J are the pictures captured 0 h, 22 h, 57 h, 82 h and 190 h
after the WIT in the BEP on the Moon, respectively.
3 GEOTECHNICAL AND GEOLOGICAL
OBSERVATIONS
Figure 5. Layout of the modules and the distribution of various
organisms 3.1 Path design of uetu-2
2.3 Biological development process Figure 8 shows the route of Yuetu-2. The lunar soil in the impact
crater may become very loose and may make the wheels sink into
After the payload was powered on at 23:18 on the first day of it. In order to avoid these craters and get to the observation point
Chang’E-4 landing on the Moon (January 3), the micro designated by scientists, photos of the surrounding terrain at each
ecosystem in the payload began to enter the growth mode on the step are taken, and then put together to form a complete
lunar. From the start-up to 20:00 on January 12, the ground sent topographic map. According to the photos sent back by the lunar
the power-off command of the biological science popularization rover and the survey results of the Chang’E-3 probe, the
test load, and the load was shut down normally (Jones 2019). The headquarters optimized the driving path to avoid the crater and
photos sent back show that the seeds inside the payload guide the lunar rover to reach the target site with less energy
germinated and the first green leaf grew on the barren lunar consumption.
surface, which will provide research basis for human to establish
a lunar base in the future. The photo of cotton seed germination
is shown in Figure 6.
Figure 8. 3D image reconstruction and path optimization
3.2 Geological information of the far side of the Moon
For the Moon, there exist only two types of large-scale activity,
meteorite impact and volcanic activity. Based on the scientific
data obtained, the Chang’E-4 team revealed for the first time the
geological stratification within 40 meters of the landing area on
the back of the Moon, and expounded its role and evolution
mechanism. Li et al. (2020) used the local positioning radar (LPR)
Figure 6 The image restoration of cotton leaves in the BEP on the Moon
to draw the schematic representation of the subsurface geological
structure at the CE-4 landing site (as shown in Figure 9). The
To better understand how the plants, animals and subsurface can be divided into three units: Unit 1 (up to 12 m)
microorganisms grow in the lunar environment with 1/6 gravity, consists of lunar regolith, unit 2 (depth range, 12 to 24 m)
low magnetic force, bright sunlight and radiation on the Moon, consists of coarser materials with embedded rocks, and unit 3
the other two identical BEPs were set up on Earth and start the (depth range, 24 to 40 m) contains alternating layers of coarse
experiments simultaneously with BEP on Moon at same and fine materials.
temperature and intensity of illumination (as shown in Figure 7). As the longest working time probe on the Moon in history,
The results showed that one of the cotton seeds successfully Yuetu-2 lunar rover is still on the back of the Moon, exploring
sprouted and turned green in the BEP on Moon. The germination more lunar areas that have never been involved before. The
of the cotton seed on Moon is much faster than that on Earth. researchers hope it will be able to observe the size changes of
impact debris in the formation, thus revealing more details of the
563
Moon’s ancient impact history. It can greatly improve our 5 UNMANNED LUNAR SAMPLER
understanding of the history of lunar meteorite impact and
volcanic activity, and may provide new ideas for understanding Chang’E-5 is equipped with two "sampling" modes: drilling a 2m
the geological evolution of the back of the Moon. deep soil sample with drilling tools and grabbing topsoil with
robotic hand (as shown in Figure 10). The robotic hand could
mine soil and rock samples from the Moon. The total length of
the robotic hand is 3.7m, the structural weight is 3100g, and the
arm thickness is only 1.4mm (Li et al. 2018). Test results show
that the stiffness and dimensional stability of the robotic hand
exceed the design requirements, which ensures the highly
repeated positioning accuracy of the manipulator in the process
of digging lunar soil and transferring samples (Liu et al. 2021).
The collected lunar rock and soil were packaged like “sausage”
to avoid contamination. Then the collected samples were sent to
the ascender and returned to the Earth.
Lunar samples sent back to the Earth are mainly used for three
purposes: scientific research, exhibition in the museum and
international research cooperation. In the field of scientific
research, the retrieved lunar rocks and soil can be used to
investigate the origin of the Earth and the Moon, analyze the
Figure 9. Shallow underground structure of Chang’E-4 landing area (Li composition and the geotechnical properties of rock and soil, and
et al. 2020). prepare for the future development of lunar resources.
4 CHANG’E-5 EXPLORER 6 OUTLOOK OF GEOTECHNICAL ENGINEERING ON
THE MOON
The Chang’E-5 mission launched from Earth on 24 November
2020 and landed in the Storms Ocean of Moon on 1 December. For geotechnical engineering construction and production
This mission is China's first automated Moon surface sampling activities, the low gravity, inactive geological activities, lack of
probe. water and pollution-free environment on the Moon are favorable
The Chang’E-5 is comprised of four components: an orbiter, for some underground construction. But on the other hand, the
a lander, an ascender, and an Earth return capsule (as shown in high radiation, the large temperature difference and the lack of
Figure 10). After touching down, the lander placed lunar samples atmospheric protection make the conventional geotechnical
into a vessel in the ascender. Then the ascender took off from the development difficult. Based on the advantages and
lunar surface to dock with the orbiter and the returner, which was disadvantages of these lunar environments, we put forward
circling the Moon, and transferred the samples to the returner. several possible geotechnical related construction activities on
Having successfully transferred its Moon samples to the orbiter the Moon in the future.
for return to Earth, the ascent vehicle later crashed itself into the
lunar surface. The orbiter and returner then headed back to Earth, 1. Biological fixation technique in complicated mechanical
separating from each other when they were several thousand environment: In the future, scientists and design engineers
kilometers from Earth (Zhao et al. 2017). Finally, the returner will improve the ecological system and organisms of
reentered the Earth and landed in China at on 16 December. landing on the Moon. Using the ecosystem carried in BEP
Chinese scientists had hoped to collect up to 2 kg of pristine lunar and solar energy as inputs, the materials (such as water,
surface samples of pebbles and regolith to a depth of up to 2 silicon and nitrogen) on the Moon are fixed in the extreme
meters with a drill and scoops aboard the Chang’E-5 lunar lander. environment to form a positive cycle ecosystem. On the
basis of microbial immobilization of inorganic substances
They took out the container loaded with lunar samples from the
(Chen et al. 2021), we expect to cultivate green plants on
returner on 17 December. These samples collected could be the
the Moon, including grass and shrubs (Bordoloi and Ng
youngest samples retrieved from the Moon so far dating to 2020; Ng et al. 2021;Fu and Zhang 2019). The roots of these
roughly 1 billion years vs 3.1 to 4.4 billion years old for the plants can fix the soil on the surface of the Moon and be
Apollo samples(Qian et al. 2021; Qian et al. 2020). decomposed by microorganisms to produce organic matter,
which can improve the soil environment of the Moon and
provide a reliable geotechnical environment for large-scale
human migration to the Moon. In addition, the research on
interaction between geotechnical systems and the
environment as well as the bio-inspired geotechnics is
essential (Zhong et al. 2021).
2. Light guide technique of lunar surface under natural
conditions: The basis of life activities on the Moon is the
energy provided by sunlight. Since the optical guide tube
itself does not have the sealing performance, new
techniques would be developed to introduce sunlight into
living areas or improve the airtightness of light guide tubes
to reduce the internal gas leakage and ensure sufficient
lighting.
Figure 10. Composition of Chang’E-5, includes ascender, lander 3. Autonomous temperature control under small scale, high
(sampling), returner and orbiter. humidity and significant temperature difference: The BEP
thermal control module consists of heat insulation pad, heat
sink, multilayer, heat insulation ring and control module
shell. These temperature control devices have achieved
564
miniaturization, but there are still large deficiencies in the chief advisor Academician Zhong Zhihua for his support to
heating efficiency. In the future, it is hoped that the use of this project.
the helium-3 material in the lunar rock and soil as the raw
material for nuclear reaction to provide enough heat for
residential areas, which is important for biological and
growth geotechnical improvement (Xiao et al. 2021).
4. Soil mechanical and permeability properties under low
gravity and strong radiation environment: In order to study
the rock and soil under Moon’s environment, it is essential
to reproduce the dry, low gravity and high radiation
environment similar to the Moon in the Earth's laboratory.
In fact, it is very difficult to produce this kind of special
environment. The existing technical means can make the
environment with low gravity and high radiation, but the
cost is too high, and it is difficult for the test personnel and
equipment to work for long time. Therefore, numerical
simulation may be a better solution (Gustafson et al. 2008;
Xi et al. 2020).
5. Lunar rock and soil reproduction and simulation: The study
of physical and mechanical properties of rock and soil is the
basis of engineering construction. Scientists all over the
world will conduct laboratory analysis and testing by
collecting lunar rock and soil samples sent to Earth, and
generate numerical models according to the obtained
physical parameters (such as mechanical properties,
composition, water physical properties, etc.), and use Figure 11. The multi field coupling laws commonly used in the field of
numerical simulation to carry out large-scale research work geotechnical engineering may be different in the lunar environment.
(Wang et al. 2020; Zhou et al. 2020).
6. Thermal-hydraulic-microgravity-chemical-radiation multi
field and multi scale geotechnical problems: The 8 REFERENCES
distribution and coupling changes of stress field, chemical Bordoloi S, Ng CWW. 2020. The effects of vegetation traits and their
field, thermal field and migration field in the reconstructed stability functions in bio-engineered slopes: A perspective review.
lunar rock or lunar soil will be studied according to the low Engineering Geology 275
gravity and high radiation environment like it on the Moon. Chen Y, Gao Y, Ng CWW, Guo H. 2021. Bio-improved hydraulic
In fact, due to the complexity of the lunar environment, properties of sand treated by soybean urease induced carbonate
many of the multi field coupling laws verified on Earth will precipitation and its application Part 1: Water retention ability.
be different in the lunar environment (as shown in Figure Transportation Geotechnics 27
11). This requires scientists to carefully verify and propose Fu Y, Zhang W. 2019. Effects of soil moisture content and root depth on
anti-overturning performance of Ginkgo biloba seedings. Journal of
a suitable multi field coupling calculation method on the
Civil and Environmental Engineering 41 (5): 42-48.
premise of giving full play to their imagination. Gustafson RJ, White BC, Gustafson MA Development of a High Fidelity
7. Building with 3D printing of lunar materials: According to Lunar Soil Simulant. In: Space Technology & Applications
the studied properties of lunar rock and soil, combined with International Forum-staif, 2008.
the geological data of the landing site on the Moon, we can Jones A. 2019. China grew two leaves on the moon: The Chang'e-4
develop survival base on the Moon with 3D printing and spacecraft also carried potato seeds and fruit-fly eggs to the lunar far
other reinforcement technology (Lane et al. 2016). 3D side - [News]. IEEE Spectrum 56,9-10
printing technology is a kind of construction technology Lane B, Moylan S, Whitenton E, Ma L. 2016. Thermographic
especially suitable for the Moon environment. It can realize Measurements of the Commercial Laser Powder Bed Fusion Process
at NIST. Rapid Prototyp J 22,778-787
unmanned manufacturing, use the materials on the Moon to
Li C et al. 2020. The Moon's farside shallow subsurface structure
build structures, and improve the materials and structures to unveiled by Chang'E-4 Lunar Penetrating Radar. Sci Adv 6,eaay6898
create shelters that can absorb radiation(Ma et al. 2019; Ma Li Y, Lu W, Fang G, Shen S. 2018. The Imaging Method and Verification
et al. 2018). With the study of lunar geotechnical materials Experiment of Chang’E-5 Lunar Regolith Penetrating Array Radar.
and the development of 3D printing technique, 3D printing IEEE Geoscience and Remote Sensing Letters 15,1006-1010
products on the Moon are quite possible. Liu J et al. 2019. Descent trajectory reconstruction and landing site
8. Exploitation and utilization of lunar geological resources positioning of Chang'E-4 on the lunar farside. Nat Commun 10,4229
under microgravity conditions: Due to the lack of geological Liu T, Zhou J, Liang L, Zhao Y, Cao D. 2021. A novel torque analysis
activities and low gravity on the Moon, it may be easier to method for drilling deep lunar soil by DEM. Journal of
Terramechanics 94,23-37
achieve geotechnical stability on the Moon than on Earth.
Ma G, Li Z, Wang L, Wang F, Sanjayan J. 2019. Mechanical anisotropy
But at the same time, the high radiation environment and of aligned fiber reinforced composite for extrusion-based 3D
meteorite impact are the problems that must be considered printing. Construction and Building Materials 202,770-783
in the open excavation and underground roadway protection Ma G, Zhang J, Wang L, Li Z, Sun J. 2018. Mechanical characterization
on the Moon. of 3D printed anisotropic cementitious material by the
electromechanical transducer. Smart Materials and Structures 27
Ng CWW, Zhang Q, Ni J, Li Z. 2021. A new three-dimensional
7 ACKNOWLEDGEMENTS theoretical model for analysing the stability of vegetated slopes with
different root architectures and planting patterns. Computers and
The authors would like to thank the State Administration of Geotechnics 130
Science, Technology and Industry for National Defense, Qian Y, Xiao L, Head JW, van der Bogert CH, Hiesinger H, Wilson L.
Ministry of Education, Chinese Academy of Sciences, Chinese 2021. Young lunar mare basalts in the Chang'e-5 sample return
Association for Science and Technology, Central Committee of region, northern Oceanus Procellarum. Earth and Planetary Science
Letters 555
the Communist Youth League. The authors are also grateful to
565
Qian Y et al. 2020. The regolith properties of the Chang'e-5 landing
region and the ground drilling experiments using lunar regolith
simulants. Icarus 337
Tang Z et al. 2020. Physical and Mechanical Characteristics of Lunar Soil
at the Chang'E 4 Landing Site. Geophysical Research Letters 47
Wang Q, Liu J. 2016. A Chang’e-4 mission concept and vision of future
Chinese lunar exploration activities. Acta Astronautica 127,678-683
Wang Y, Wang L, Meng F, Chen K. 2020. Strength test of 3D printed
artificial rock mass with pre-existing fracture. Underground Space
Xi B, Jiang M, Cui L. 2020. 3D DEM analysis of soil excavation test on
lunar regolith simulant. Granular Matter 23
Xiao Y et al. 2021. Homogeneity and mechanical behaviors of sands
improved by a temperature-controlled one-phase MICP method.
Acta Geotechnica,1-11
Xie B et al. 2017. The water treatment and recycling in 105-day
bioregenerative life support experiment in the Lunar Palace 1. Acta
Astronautica 140,420-426
Zhang L, Li J, Zeng Z, Xu Y, Liu C, Chen S. 2020. Stratigraphy of the
Von Kármán Crater Based on Chang'E 4 Lunar Penetrating Radar
Data. Geophysical Research Letters 47
Zhao J, Xiao L, Qiao L, Glotch TD, Huang Q. 2017. The Mons Rümker
volcanic complex of the Moon: A candidate landing site for the
Chang'E-5 mission. Journal of Geophysical Research: Planets
122,1419-1442
Zhong W, Liu H, Wang Q, Zhang W, Li Y, Ding, Chen Longlong. 2021.
Investigation of the penetration characteristics of snake skin-inspired
pile using DEM. Acta Geotechnica. DOI: 10.1007/s11440-020-
01132-2
Zhou S, Lu C, Zhu X, Li F. 2020. Preparation and characterization of
high-strength geopolymer based on BH-1 lunar soil simulant with
low alkali content. Engineering
566
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
The mechanical properties of the Martian soil at the InSight landing site
Les propriétés mécaniques du sol martien sur le site de la mission InSight
P. Delage
Ecole des Ponts ParisTech, Laboratoire Navier/CERMES, CNRS, Marne la Vallée, France
E. Marteau, M. P. Golombek, K. Hurst, W. B. Banerdt, S. Piqueux, S. Kedar, M. Lemmon & N. R. Williams
Jet Propulsion Laboratory, NASA - California Institute of Technology, Pasadena, CA, USA
N. Brinkman, P. Edme, J. Robertsson, D. Sollberger, S. Stähler, C. Schmelbach
ETH Zürich, Institute of Geophysics, Department of Earth Sciences, Switzerland
P. Lognonné, K. Onodera
Universit! de Paris, Institut de Physique du Globe de Paris, CNRS, Paris, France
C. Vrettos
Technical University of Kaiserlautern, Department of Civil Engineering, Germany
V. Asan-Mangold
Nantes Universit , Laboratoire de Planétologie et Géodynamique, CNRS, Nantes, France
M. Grott, N. Mueller
DLR Institute of Planetary Research, Berlin, Germany
N. Murdoch
ISAE – SUPAERO, Toulouse, France
T. Spohn
DLR Institute of Planetary Research, Berlin, Germany
International Space Science Institute, Bern, Switzerland
N. Warner
SUNY Geneseo, Department of Geological Sciences, Geneseo, NY, USA
R. Widmer-Schnidrig
Stuttgart University, Black Forest Observatory, Institute of Geodesy, Wolfach, Germany
B. Caicedo-Hormaza.
Universidad de los Andes, Bogota, Colombia
J. P. Castillo-Betancourt
Ecole des Ponts ParisTech, Laboratoire Navier/CERMES, CNRS, Marne la Vallée, France
Universidad de los Andes, Bogota, Colombia
L. Lange
Sorbonne Universit!, Laboratoire de M!t!orologie Dynamique, CNRS, Paris, France
N. Verdier
Centre National d’Études Spatiales, Toulouse, France
ABSTRACT: The InSight mission is a NASA geophysical mission aimed at better understanding the structure of Mars and of the
other rocky planets of the solar system. To do so, various instruments are used, including a very sensitive seismometer (SEIS) and a
dynamic self-penetrating heat probe (HP3) that have been placed on the Mars surface by the Instrument Deployment Arm (IDA).
Besides geophysical data (which have definitely enriched and completed existing knowledge on the structure of Mars), the InSight
instruments, together with orbiter observations and tests carried out on the soil with the IDA, have significantly increased the
knowledge of the geological and geotechnical characteristics of the surface material at the InSight site, which is made up of a basaltic
sand. In-situ data were also successfully compared with terrestrial previous estimates from terrestrial lab tests, carried out on various
soil simulants. Small strain (elastic) parameters at small strains were derived from wave velocity measurements between the self-
penetrating probe and the seismometer. Strength data were derived from both IDA operations and penetration data. The soil includes
some pebbles within a somewhat cohesive sandy matrix, limiting the heat probe penetration to only 40 cm length. Thermal data were
567
also obtained, allowing for some thermo-elastic modelling of the effect of the Phobos (one of the “Moons” of Mars) eclipses. Elastic
data were also derived from the effects of wind on the ground, detected by SEIS.
RÉSUMÉ : La mission InSight est une mission géophysique de la NASA destinée à mieux comprendre la structure de Mars et des
planètes telluriques du système solaire à l’aide de divers instruments, incluant un sismomètre ultra-sensible et une sonde de pénétra-
tion dynamique destinée à des mesures thermiques, qui ont été placés sur le sol de Mars par un bras robotisé. A côté des données
géophysiques obtenues (qui ont largement enrichi et complété la connaissance de la structure de la planète), les données fournies par
les instruments, complétées par des observations satellitaires et des essais réalisés par le bras robotisé sur le sol, ont permis d’élargir
considérablement la connaissance des caractéristiques géologiques et géotechniques du sol martien du site d’InSight, constitué de
sable basaltique. Les données obtenues in-situ ont aussi été favorablement comparées à celle estimées préalablement sur Terre par
des essais de laboratoire sur des analogues de sol martiens. Les paramètres élastiques à petites déformations ont été obtenus à partir
de mesures de vitesses d’ondes sismiques entre la sonde et le sismomètre, lors des opérations de pénétration dynamique en surface.
Des caractéristiques à la rupture ont été obtenues par les sollicitations exercées sur le sol par le bras robotisé et par les données de
pénétration. Le sol s’est révélé être composés de cailloux contenus dans une matrice sableuse légèrement cohérente, ce qui n’a rendu
la pénétration de la sonde thermique possible que sur les 40 cm de sa longueur. Des données thermiques ont aussi été obtenues,
permettant la modélisation thermo-élastique des déformations induites par une éclipse de Phobos, une des ‘Lunes’ de Mars. Des
données élastiques ont été aussi déduites des effets du vent sur la surface, détecté par le sismomètre.
KEYWORDS: Mars, geophysics, regolith, seismometer, heat probe, seismic waves, elastic parameters, penetration, strength
1 INTRODUCTION >3.6 billion years, Pan et al. 2020) possibly containing
phyllosilicates.
The investigation of the mechanical properties of the surface
deposits on Mars has been considered in details since the landing
of Vikings 1 and 2 in 1976 (e.g. Moore and Jakoski 1989) and
completed based on the data obtained from landers and rovers of
other missions (see Christensen and Moore 1992, Golombek et
al. 2008, Herkenhoff et al. 2008).
The InSight mission on Mars (Interior Exploration using
Seismic Investigations, Geodesy and Heat Transport, Banerdt et
al. 2020) is a geophysical mission managed by NASA and aimed
at better understanding the structure of Mars and of the other
terrestrial planets of the Solar system. Just to recall, Mars is one
of the four terrestrial planets of the Solar system, with a Martian
year around as long as twice the Earth year, an average diameter
of 6 779 km (0.553 that of the Earth) and, as a result, a smaller
gravity of 3.711 m/s2. The atmosphere on Mars is mainly
composed of CO2 with an average atmospheric pressure of 610
Pa.
The main geophysical instruments deployed by the InSight
mission on Mars are i) a very-high-sensitivity seismometer
(SEIS, Lognonné et al. 2019, 2020), including a very-broad-band
seismometer (VBB) funded by the French Space Agency
(CNES) and a short-period (SP) one developed at Imperial
College and ii) the HP3 instrument (Heat flow and Physical
Properties Package, Spohn et al. 2018) a device designed that
includes a self-penetrating dynamic cylindric probe (40 cm long
and 2.7 cm diameter) nicknamed the mole and funded by the
German Space Agency (DLR). The HP3 instrument is designed
to perform thermal conductivity and thermal gradient Figure 1. Locations of the InSight lander (NSY) close to the equator
(4.502°N/135.623°E), not far from the dichotomy boundary that
measurements along the first 3 – 5 meters below the surface, separates, on Mars, the southern heavily cratered highlands from the
from which the heat flow from of the interior of the planet could northern lowlands (see color scale). The location of Hephaestus Fossae
be calculated. To do so, the mole was supposed to drag down a (see Figure 6) is also indicated. The map is a portion of the Mars Orbiter
captor tether (Spohn et al., 2018). Laser Altimeter (MOLA) shaded topographic map of Mars (after
The InSight lander also carries a weather station monitoring Golombek et al. 2018).
the changes in temperature, atmospheric pressure and wind. The
InSight lander the successfully landed on Mars in western The artist view of the InSight lander (Figure 2) shows both
Elysium Planitia on 26 November 2018 (NSY in Figure 1), a the SEIS (covered by a wind and thermal shield) and the HP 3
little bit north of the equator. This first day of the mission is instruments deployed. Both are linked to the lander by tethers
defined as sol 1: a sol is a Martian day, about 40 minutes longer that transmit energy and data. The Figure also shows the 2 m long
than an Earth day. Figure 1 also shows the landing sites of the Instrument Deployment Arm (IDA) and its grapple that deployed
Viking 2 Lander (VL2, 1976), in Utopia Planitia, of the Mars both SEIS and HP3 on the ground (Trebi-Ollennu et al., 2018).
Science Laboratory (MSL) Curiosity rover in the Gale crater The IDA also carries a medium-resolution Instrument
(2012) and the Mars Exploration Rover (MER) Spirit in Gusev Deployment (colour) Camera (IDC). The lander is also equipped
crater (2004). InSight landed on ~200 m thick lava flows from with a wide-angle Instrument Context Camera (ICC), a
Early Amazonian-Hesperian period (1.8-3.6 billion years), meteorological station including temperature, windspeed and
underlain by sedimentary rocks of likely Noachian age (around pressure sensors, two X-band RISE antennas (Rotation and
568
Interior Structure Experiment) providing a precise location of the as discussed in details in Spohn et al. (2021a), it has not been
lander to accurately monitor the movements of the planet from possible to drive the mole as deep as planned enough due to
the Earth (Folkner et al. 2018, Golombek et al., 2020c), together unexpected lack of friction with the regolith, to counteract the
with an Ultra High Frequency antenna for data transmission. rebound occurring in the self-penetrating system during dynamic
Energy is provided by two twin 1.8-m diameter solar panels with penetration. The mole only barely buried its entire length (40
a power of 700 W each, on clear days. cm), which still allowed wave velocity measurements (Lognonné
et al. 2020, Brinkman et al. 2022) and thermal conductivity
measurements near the surface (Grott et al. 2021). In an attempt
to help the mole penetration, the HP3 surface structure observed
in Figure 2 and Figure 3 has been withdrawn by the IDA,
revealing a pit that developed around the mole during the many
hammering sessions carried out. The IDA was afterwards used to
fill the pit with regolith so as to improve the regolith/mole
friction, and to apply some pressure aimed at facilitating
penetration, as seen below in Figure 8. All these attempts were
unfortunately unsuccessful.
Figure 4 shows two selfies reconstituted from images taken
by the IDC just after landing (the SEIS and HP3 instruments are
still on the lander) and after around 100 days, with evidence of
dust coating on the lander and the solar panels. Dust coating of
Figure 2. InSight lander artist’s concept showing its scientific payload.
https://mars.nasa.gov/insight/spacecraft/instruments/summary/ the solar panels is quite an issue since it results in a regular
decrease of the available power. As further commented below, it
SEIS is a high-sensitivity inertial accelerometer, that has also confirms the presence of very fine particles transported by
yielded impressive geophysical results in the detection of seismic Mars winds.
signals and impacts on the planet since landing (Giardini et al.
2020). As an inertial accelerometer, SEIS is also sensitive to tilts
that change the projection of the gravity vector onto the
measurement direction of the accelerometer. In acceleration, its
ASD (amplitude spectral density) noise floor in the band 0.01 - 2
Hz is slightly below 1 (nm/s2 h as
a very sensitive tilt meter and a seismometer, identifying the
consequences of various mechanical actions like those due to
changes in atmospheric pressure due to wind and dust devils
(small local storms), temperature changes (including those from
the eclipse caused by Phobos, one of the two Mars moons) or
interactions between the IDA or the mole and the regolith. Some
of these will be considered in this paper. Further actions are
planned, like for example the monitoring by SEIS of a vertical
force applied by the IDA to the ground, with the scoop used as
an indenter.
Figure 4. Comparative selfies of the InSight lander reconstituted from
images taken by the IDC: a) at the beginning of the mission (Dec. 6, 2018,
sol 10), with the instruments (SEIS and HP3) still on the lander; b) after
around 100 days (March 15, 2019, sol 106 and April 11, 2019, sol 133),
with evidence of dust coating on the lander and solar panels (Image credit
JPL/NASA).
2 CHARACTERISATION OF THE SURFACE REGOLITH
The InSight landing site was selected from orbiter observations
including high-resolution images from the HiRISE camera
Figure 3. Photo taken from the Instrument Context Camera (ICC) show-
(Golombek et al. 2017). It was chosen for its smooth, flat
ing the location of SEIS and of the HP3 support structure, 1.18 meter apart
from each other (image credit JPL/NASA). relatively rock free surface mainly composed of sand-like
regolith. The regolith was formed by the long-term and
Figure 3 shows both the HP3 surface structure containing the successive impacts of meteors into the lava flows of Elysium
mole and the SEIS covered by the thermal and wind shield, both Planitia, that progressively broke the basalt parent rock into
having been successfully deployed by the IDA close to the smaller and smaller rocks and particles, to end up with a sand-
lander. The distance between them is close to 1.1 m. Note that like surface regolith.
the waves emitted by the HP3 driving system have been detected This formation process can be illustrated by the HiRISE
by SEIS, providing interesting information about wave velocities image (Figure 5, see location in Figure 1) of a steep scarp of
and elastic properties at the surface (see below). Note also that, Hephaestus Fossae exposure, a fracture ~900 km northwest of
569
the InSight landing site that cuts through a potentially analogous estimated angle of repose of loose granular materials). The
terrain (Golombek et al. 2018). The photo shows 4-5 m thick, estimation of the grain distribution of the talus material is not
relatively fine-grained regolith overlying blocky ejecta that straightforward, but it is probable that it is composed from part
grades into strong, jointed bedrock. Given that the bedrock is of fine regolith that flowed down from the top, together with
composed of basalt lava, the regolith is mainly basaltic as well. other rocky debris with size comprised between 1 and 10 m
coming from the upper layer of blocky ejecta. The entire
stratigraphic sequence is consistent with an impact-comminuted
lava plain with a gardened regolith that is up to 10 meters thick.
The regolith fines upwards, a result of the higher frequency of
small (order of 1 to 10 m) impacts.
Figure 5. A portion of the exposed steep scarp of Hephaestus Fossae in
southern Utopia Planitia (Golombek et al. 2018).
Figure 7. HiRISE image showing fresh crater A (112 m diameter), with
rocky ejecta around and B (75 m) with no rocky ejecta around. The depth
at which ejecta is sourced in fresh craters is 0.084 times the diameter
(Golombek et al. 2017). Note that Golombek et al. (2020b) has a more
specific treatment of the regolith depth.
a) In selecting the best landing site (Golombek et al. 2017), one
challenge was to make sure that the thickness of the surficial
regolith layer was larger than the 5 m depth at which the mole
was planned to be driven. This was done by considering the
various fresh craters located in the area of interest, and to check
whether or not they had rocky ejecta around. In Figure 7, the
small black dots around crater A (112 m diameter) are rocky
ejecta, indicating that the meteorite hit the bedrock. This is not
the case of crater B (75m diameter). As recalled by Golombek et
al. (2017), observations of fresh craters show that the ratio
between their depth and diameter is around 0.084 (75 x 0.084 =
6 m). Thus, the rocks around crater A have been extracted from
a depth of around 9 m. So, the average regolith thickness in the
area is between 6 and 9 m, which is satisfactory for driving the
mole.
A significant feature of the regolith on Mars, compared to
those of other terrestrial planets or bodies of the solar system
b) where there is no atmosphere, is that surface particles are
Figure 6. a) HiRISE image (ESP_052638_2020) of the Hephaestus Fos- entrained in the wind, with an average atmospheric pressure of
sae exposure, a fracture ~900 km northwest of the InSight landing site;
610 Pa (Spiga et al., 2018; Banfield et al., 2020). Wind effects
b) Slopes determined along profile 1 (Warner et al. 2022a, b).
result in saltation, in which sand particles are moved in parabolic
As shown in Figure 6, (Warner et al. 2022a, b), the profile has hops by the wind. Particles have been shaped, transported and
been further analyzed, showing, on the top, a 4 – 5 m thick fines- accumulated either in aeolian bedforms or in depressions like the
dominated layer (with a maximum slope of 20°, smaller than the so-called Homestead Hollow, in which the lander is located. As
angle of repose estimated at 30°), underlain by a 13 – 29 m thick a consequence, soil particles on Mars are rounded to subrounded,
boulder-dominated cliff-forming unit, underlain by another cliff- as will be discussed below in more details (Golombek et al.,
forming unit (112 – 133 m thick) that may be the primary 2020d).
bedrock, with a talus at its base inclined at 32° (close to the
570
Based on the formation mode described above and on Assuming the soil to be cohesionless and non-cemented, the
geological observations from various previous Martian missions, above thermal conductivity values indicate that 85%–95% of all
the density of the sandy regolith of the InSight landing site is particles are smaller than 104–173 m (Grott et al. 2021). It is
estimated to be low, with values between 1.1 and 1.4 Mg/m3, as worth noting, however, that if cementation plays a significant
shown in Table 1, that describes the nature and properties of role, this estimate needs to be viewed as an upper limit (Piqueux
various surficial soil type deposits on Mars (Golombek et al. and Christensen, 2009). In that case, the volume of bonding
2008, Ch. 8). The mechanical properties have been derived from between the grains has been modelled to be substantially less
scraping, trenching and dumping experiments carried out with than 1% in volume (Piqueux et al. 2021).
scoops as well as interactions with rover wheels. The Table also As will be seen below, various observations show that,
provides the properties of drift, or dust, a very fine material with besides the average particle size estimated from thermal inertia
particle size between 1 and 10 µm, that will be further discussed measurements, a significant proportion of finer particles
below. (including dust) exist in the surface layers of Mars, similar to
those observed on the solar panels (see Figure 4b) and
Table 1. Characteristics of some Martian surface materials (after transported by winds.
Golombek et al. 2008)
Material Bulk Grain Cohesion Friction Ther-
2.2 Density Estimate from thermal conductivity
Density Size (kPa) Angle mal Previous geological observations suggested that the regolith
Mg/m3 µm ° Inertia1
Drift 1–1.3 1–10 0–3 15–21 40–125
density on Mars is low, with density estimated between 1.1 and
Sand 1.1–1.3 60–200 0–1 30 60–200 1.3 Mg/m3 for sand-type regolith (see Table 1). While there are
Crusty to 1–1.6 5–500 0–4 30–40 200–326 no direct measurements of soil density at the InSight landing site,
cloddy sand some constraints may be derived based on the thermal
Blocky, in- 1–1.2 50–3000 31–11 40–60 369–410 measurements performed by the HP3 TEM-A (Thermal
durated soil Excitation and Measurement-Active mode, Grott et al. 2021).
See definition in main text; units are J m 2 K 1 s 1/2
While these measurements are aimed at deriving thermal
conductivity using a modified line heat source approach, the full
2.1 Thermal inertia measurements inversion of the data also contains information on the soil thermal
The average size of the regolith particles on Mars was initially diffusivity = k/ p and thus the soil density. Values of soil
estimated from thermal inertia measurements, which have been density compatible with TEM-A measurements are
conducted from orbit - including those of the THEMIS (Thermal Mg/m3, in accordance with estimates of Table 1.
Emission Imaging System) instrument on board of the Mars Od- This corresponds to a significant soil porosity of the order of
yssey spacecraft launched in 2001 (e.g., Kieffer et al., 1977; Pal- % (Grott et al. 2021). This value is in excess of the value
of 42% appropriate for a random loose packing, indicating that
luconi & Kieffer, 1981; Mellon et al., 2000; Putzig & Mellon,
soil cohesion has played a role during soil deposition, enhanced
2007; Golombek et al., 2008, Edwards et al. 2018) - and from the
by low gravity.
ground (Fergason et al., 2006; Golombek et al., 2020a; Hamilton
et al., 2014).
The principle that allows derivation of average particle size 3 DIRECT OBSERVATIONS ON MARS
from thermal inertia is simple: for the same mass, the changes in
temperature of a particle (the average temperature of the atmos- Figure 8 presents an IDC photo of the scoop of the Instrument
phere on Mars changes as much as 100°C between days and Deployment Arm (IDA) compressing the regolith above the hole
nights) are slower for larger particles (small specific surface) and that developed around the self-penetrating mole during
faster for small particles (high specific surface). This trend is hammering, with the tether appearing on the left side of the
clearly observed in the data of Table 1. Note however that inter- scoop.
grains cementation in slightly cohesive granular materials may
increase the thermal inertia.
Thermal inertia (unit Jm 2K 1s 1/2) is given by:
(1)
where k is the thermal conductivity, the density, and cp the spe-
cific heat capacity. While density and specific heat capacity vary
little for different soils, thermal conductivity can vary by orders
of magnitude, depending on bulk porosity, composition, grain
size and the state of cementation or induration.
At the InSight landing site, orbital thermal inertia
measurements indicated values between 160 and 230 J m K
s (Golombek et al. 2020a), which were later confirmed by
local measurements using the InSight HP3 Radiometer (Spohn et
al. 2018, Golombek et al. 2020a, Müller et al. 2020, Piqueux et
al. 2021). Such thermal inertias correspond to thermal Figure 8. IDC photo showing the scoop of the Instrument Deployment
conductivities of 0.041 ± 0.013 W m K , consistent with direct Arm compressing the regolith, once the pit filled. The tether appears on
thermal conductivity measurements using the HP3 mole as a the left side. The scoop width is 7.6 cm (image credit JPL/NASA).
modified line heat source (Grott et al. 2019, 2021), which yielded
thermal conductivities of 0.039 ± 0.002 W m K for the upper The photo confirms the sand-like aspect of the surface
regolith, with no rock in the area and some centimeter-sized
3-37 cm of the soil. Due to low atmospheric pressure and gravity
typical of Mars, these are very low values compared to those of pebbles around. Note also the perfectly smooth imprint of the
current terrestrial soils, where k is close to 1. scoop observed between the right side of the scoop and the
elongated pile on the right. Such smooth surfaces have also been
observed on rover tracks in other missions and indicate that the
571
fine particles contained in the sandy layer provide some degree
of cohesion (Golombek et al. 2020a) that will be further
discussed later on. As in other Martian missions since those
performed from the Viking lander (Moore et al. 1982, 1987,
Moore and Jakosky 1989) and as shown below, the scoop has
been used to perform scraping, trenching and piling operations,
from which some soil mechanics parameters have been derived.
It is well-known that the surface of Mars is regularly covered
by small particles of dust with size of a few microns (2-5 m)
similar to those observed on the solar panels (see Figure 4b).
Dust is transported over thousands of kilometers during regional
to global dust storms (e.g., Cantor, 2007). During the InSight
mission, a regional dust storm occurred (e.g., Viudez-Moreiras et
al. 2020). Note that the deposition of this red dust on the solar Figure 10. IDC images taken beneath the lander on sol 18. (a) The pits
panels of rovers and landers has serious consequences on their and depressions were excavated by the lander’s retrorockets. P1 and P2
energy supply, including for the InSight lander (Figure 4). are the largest and deepest pits. (b) P1 is ~50 cm in diameter and reveals
Actually, orbiter observations indicated that a thin layer (<1 cm ~12 cm of the shallow stratigraphy. The pit walls are vertically striated
thick) of dust was removed by the rockets during landing to an and steep (up to 65 ) and expose an up to ~10-cm-thick cohesive duricrust
average distance of 20 m around the InSight lander (Golombek comprised of weakly cohesive fine sand. The floor of each pit contains
et al. 2020a, b). So, the average 170 µm diameter particles abundant reddish, pebble-size clasts or clods of material that broke off
estimated from thermal inertia is only a global indication about the steep pit walls. Dark-gray pebbles of likely basaltic composition are
visible within the matrix of fine sand (Warner et al. 2022).
grain sizes, and smaller particles also exist in the regolith.
Further evidence on the nature of the regolith particles on Further observation of the near-surface has been made
Mars has been provided by the microscopic imagers of the Spirit, possible by the pit that developed around the mole, observable
Opportunity and Curiosity rovers, the Phoenix lander and by the once the support structure was removed by the IDA.
atomic force microscope of the Phoenix mission (Pike et al.
2011). They revealed sorted dark grey basaltic sub-rounded to
rounded fine sand particles due to wind saltation, with diameter
between 80 and 200 µm, as seen in Figure 9 from the Phoenix
mission (Goetz et al. 2010). This range is compatible with the
average diameter (170 µm) derived from thermal inertia
measured at the InSight landing site. Bright red dust particles (2
– 5 µm) of slightly chemically altered basalt are also observed
(Arvidson et al. 2004a, b, Golombek et al. 2006a, b and Goetz et
al. 2010).
Figure 11. Image of the hole created by the HP3 mole during hammering
(inclined 15°), with the scoop appearing on top right (image credit
JPL/NASA).
The photo of Figure 11 shows the pit created around the
inclined mole (inclined 15°), with an almost vertical wall made
up of resistant layers containing some pebbles. The steep edge
and some overhangs indicate the existence of cohesion. Pebbles
are embedded in a fine-grained matrix, indicating cementation.
Consistent with the formation process of the regolith (see
Figure 9. Microscopic observation of sub-rounded to rounded particles Figure 6) and from observing the photos taken around the InSight
of Mars regolith at the Phoenix site. Picture width is 500 µm (Goetz et landing site by the two cameras, Golombek et al. (2021)
al. 2010). proposed the sketch of Figure 12 to describe the structure of the
regolith in the landing area. A depression of ~27 m in diameter
The colour of the particles in Figure 9 varies from dark gray (in which the lander sits) interpreted as an ancient degraded and
(black) to slightly reddish. This probably depends on the degree infilled impact crater in the regolith is underlain by blocky ejecta
of contamination of their surface by red fines, or on surface stains and fractured basaltic bedrock. The crater has been filled
(Goetz et al. 2010). dominantly by sand moved by the wind, producing an upper layer
Figure 10 shows the walls formed during the excavation of of fine material three meters thick, consisting mostly of sand.
the surface by the rocket exhaust beneath the lander. It provides
more evidence of the nature of the regolith, confirmed afterwards
by observing the pit created around the mole during hammering. 4 SMALL STRAIN PROPERTIES OF THE SURFACE
At meter scale, the nature of the surficial regolith is spatially REGOLITH
heterogeneous within the Homestead Hollow. ~10 cm-deep pits
show ~ centimeter-size pebbles embedded in a finer grain matrix The small strain investigation of the regolith mechanical proper-
which maintains a steep slope, suggesting some degree of ties presented below has been carried out either by wave velocity
cohesion. The slopes of regolith material piled at the bottom of measurements or by small strain plate loading, with quite reversi-
the excavation particles moved by the rocket exhaust are also ble responses. As a consequence, one considers that they reason-
observed. ably provide some elastic parameters of the regolith, that will be
572
considered here isotropic and homogeneous, for sake of simplic-
ity.
Figure 12. Interpretative cross-section of the shallow surface beneath the a)
InSight lander (Golombek et al. 2020a, Creative Commons CC BY li-
cense). 1: Fractured basalt flow; 2: Blocky ejecta; 3: Fine-grained impact
generated regolith; 4, 5: Overlapping craters; 6: Rockier area; 7: Rocks
embedded in regolith; 8: Pits opened by retro rockets during landing; 9:
Surface divots; 10: Lens of ejecta from other craters (Golombek et al.
2020a).
4.1 Previous estimation from lab tests
Some first estimations of possible elastic values of the Mars reg-
olith were derived from wave velocity measurements carried out
on Earth for Mars regolith simulants provided by both NASA
(Mojave sand) and DLR (MSS-D and Eifelsand). In a standard
fashion, wave velocities were measured by using piezo-ceramic
bender elements on triaxial samples (100 mm in diameter and b)
200 mm in height), yielding the change in velocities with confin- Figure 13. Bender elements measurements of velocity waves: a) changes
ing pressure shown in Figure 13a (Delage et al. 2017) for com- in Vp with confining stress; b) linear relation between Vp and Vs for the
pression waves (Vp) in the Mojave simulant. three simulants investigated. A common value = 0.22 is obtained for
A power law fit corresponding to the following expression of the three simulants tested (Delage et al. 2017).
the changes in wave velocity with stress (see Santamarina et al.
2001) is drawn in Figure 13a:
(2)
From this fit, a value b = 0.3 can be deduced. Note that, given
the angular nature of the simulants investigated, this value might
be somewhat overestimated. For rounded/subrounded grains like
those observed on Mars, a value of 0.28 could be more
appropriate. Figure 13b shows a linear relation between Vp and Figure 14. Shape of the SEIS foot (Delage et al. 2022). The conical spike
Vs for all three simulants, according to the following relation, in has a maximum diameter of 10 mm.
which n is the Poisson’s ratio:
As seen in Figure 15, the container is placed on the plate of a
(3) triaxial press that is slowly moved upwards (0.08 mm/min) so as
to cause the penetration of the foot (spike + disk) that is fixed to
A common value n = 0.22 is derived for the three simulants the top of press through a sensitive force transducer that provides
from these data, at densities around 1.5 Mg/m3. the change in penetration force while the plate is moved upwards.
Further lab investigations, carried out to investigate the The maximum applicable load is defined by the weight of the
SEIS/regolith elastic interaction, also provide some information suspended foot, that is equal to 10 N, i.e., the force supported by
on possible elastic properties at the surface (Delage et al. 2022). one on the three SEIS foot under Mars gravity. Special care was
This interaction was experimentally examined by using the taken during the setting up of this experiment to ensure sufficient
device presented in Figure 15, in which a model of the SEIS foot stiffness of the device, as well as satisfactory thermal insulation
(that consists of a 60 mm diameter disk with a 20 mm long spike (not shown in the Figure) to get satisfactory precision in
in its center, see Figure 14), is slowly penetrated into a loose sand displacement measurements. Displacements were measured by
mass (unit mass around 1.4 Mg/m3) contained in a 240 mm using high precision LVDT sensors.
diameter and 120 mm depth container. These tests were carried Figure 16 shows 10 force/displacement curves obtained on
out on samples made up of a new rounded/subrounded simulant, various samples of around 1.4 Mg/m3 unit mass. Note that a
the Fontainebleau sand. This sand is well-sorted, with an average comparison between plate loading tests (with disk only) and foot
diameter of 220 µm, slightly larger than the average estimated loading tests (with disk and spike) didn’t show any significant
size of 170 µm on Mars derived from thermal inertia effect of the spike. A first series of curves exhibit a maximum
measurements. penetration of around 450 µm under the maximum force of 10 N,
whereas a second series displays a maximum penetration of 600
µm. These variations are linked to possible changes in the density
573
at surface of the sample. When compared to the average grain
size (220 µm), the penetration under the maximum force
corresponds to the size of two to three grains only, showing quite
a small strain level.
1 kg
mass
SEIS
foot
Figure 17. Force cycle performed around the maximum force corre-
sponding to the force supported by a SEIS foot. The cycle is fairly re-
versible illustrating satisfactory elastic response under the imposed dis-
placement (around 2 µm). Its slope provides the elastic spring constant
kv used in the analysis of the dynamic interaction between SEIS and the
ground (Delage et al. 2022)
Sand
container By considering an average response of the various tests
performed, an average value of 20 MPa was deduced for the
Young modulus. This value is typical of loose sand on the Earth.
Triaxial
Based on relation (5) with n = 0.22 and r = 1.4 Mg/m3, a
press
value Vp = 128 m/s is obtained, not far from those measured at
Figure 15. Experimental investigation of the interaction between the the InSight site (Lognonné et al. 2020, Brinkman et al. 2022, see
SEIS foot and a loose sample of Fontainebleau sand. The container is below).
slowly moved upwards to cause penetration of the foot (Delage et al.
2022).
(5)
It is clear however that terrestrial lab experiments can only
provide a magnitude estimate of the elastic properties of the
regolith, given the possible differences in density, homogeneity,
grain size distribution and shape (see Figure 10 and Figure 11)
and gravity with the actual state of the regolith at the InSight site,
which is discussed in the next session.
4.2 Direct estimation from Mars data
4.2.1 SEIS-HP3 interaction
As seen in Figure 18, the SEIS was placed at a distance of around
1.18 m from HP3. Recording the HP3 mole hammering noise with
Figure 16. Force/displacement curves of the foot penetration tests
(Delage et al. 2022).
SEIS provided a unique opportunity to estimate the seismic
velocities of the shallowest regolith in situ (Kedar et al. 2017;
The curves also exhibit some sudden repeated drops in forces Sollberger et al. 2021; Brinkman et al. 2022). Brinkman et al.
followed by force recovery paths that return the curves to the (2022) determined effective P- and S-wave velocities of
same levels of force, that are related to stick-slips phenomena Vp = m/s and Vs = m/s respectively, from around
between the foot and the loose sand. 2,000 HP3 hammer stroke recordings. These seismic velocities
The interpretation of these data was made by using Sneddon’s likely represent bulk values for the uppermost several 10's of cm
(1946) elastic solution of plate loading on a homogeneous semi- of regolith. The incidence angles of the P-wave first-arrivals
infinite half mass. This provides the following simple relation provided an independent Vp/Vs ratio estimate of which
between the Young modulus E, the Poisson ratio n, the spring is in good agreement with the travel-time based estimate of
constant kv and the plate radius R:
Assuming a density of 1.2 Mg/m3 (Spohn et al. 2021a), the Vp
(4)
and Vs estimates can be converted into a bulk, shear, and Young’s
modulus of 9.8 6.8 MPa, 4.3 1.0 MPa, 11.3 2.9 MPa,
respectively, and a Poisson’s ratio of 0.31 0.15 (actually larger
The spring constant kv = 1.47 x 106 N/m was deduced from than that derived for lab measurements equal to 0.22). When
the slope of force cycles carried out around the maximum force, interpreting these numbers, one has to keep in mind that they
as shown in Figure 17. Note that the cycle mobilizes a small represent values for a homogeneous half-space and were derived
displacement of 1.5 µm, to be compared to the average grain size from seismic waves with a dominant frequency of around 60 Hz.
of 220 µm, ensuring a fairly reversible response, as observed in
Figure 7. kv values were used to investigate the dynamic
interaction between the SEIS and the ground (Fayon et al. 2018).
574
speculated that some electric cross talk as in the case of the
magnetometer might be responsible for the VBB signal.
However, a subsequently performed experiment with an equally
well shielded seismometer at the Black Forest Observatory in
Germany, that simulated the radiation time history of the Phobos
transit, showed that also the terrestrial seismometer responded
with a purely horizontal acceleration, but with a 12s delay
relative to the radiation onset.
c SEIS
e
d c 2020-03-05 eclipse
Ne ar RAD spot Far RA D s pot
2 92
Figure 18. a) Image taken on June 28, 2019 (sol 209) showing how the
Ground Temperature (K)
HP3 support structure was lifted and revealing the partially buried mole
2 90
2 88
T=1.3o
(see Figure 11), located around 1.18 m away from InSight’s seismometer 2 86
SEIS (image credit JPL/NASA); b) Conceptual illustration of the HP3 2 84
mole generating seismic waves (blue and red colors correspond to nega- 2 82
tive and positive amplitudes) propagating through the shallow subsurface 2 80
at the InSight landing site (Brinkman et al. 2022). 2 78
0 1 00 2 00 3 00 4 00 5 00 6 00 7 00 8 00 9 00
Seconds since 2019-03-05T10:45:00
4.2.2 Thermal effects
The observation of the Phobos transits by the different sensors
on InSight led to an unexpected soil mechanics experiment e
(Stähler et al. 2020). The largest observed Phobos transit on d Leg 1
05/03/2019 lasted 27 seconds and lead to a peak drop in solar
Z
array current of 10%. While no visual observation of the Sun was
made during the transit, we can take the drop in solar array Leg 3 Leg 2
current as a proxy for the fraction of the solar disc covered at the N 3.5e-8 =4.1e-8 m/s2
peak of the transit, and that is consistent with simulations of the -9.2e-9 =1.2e-8 rad
Modeled Tilt 12 nrad 178o
transit as depicted in Figure 19a, that shows how the solar disk is Obs Tilt 12 nrad 180o
El
hidden by Phobos on sol 96 (3 March 2019). A complementary
o
Az o 71
E Sun 243
=0 096
observation comes from the radiometer on board the lander. It 03-
05_0
provides a direct observation of the cooling of the Martian
surface during the transit by 1.3°K. The skin depth of such a short
negative heat pulse is only 0.5-1.0 mm. Thermal modeling of the Figure 19. a) Simulated view of the transit of Phobos at maximum on 3rd
temperature response indicates that the equivalent thermal inertia March 2019, showing how the solar disk is partially hidden by Phobos;
in this layer is only 103 J m-2K-1s-1/2, significantly lower than that b) Projected shadows of the lander and the SEIS Wind and Thermal
of the diurnal skin depth (Mueller et al. 2021). The atmospheric Shield. Blue dots are the locations of the Lander feet and the SEIS feet.
sensors measuring wind and air pressure did not record a c) Ground temperature as measured by the HP3 RAD radiometer. d) Ac-
celeration time series measured by SEIS along vertical (Z), north (N) and
detectable response to the transit while the magnetometer located east (E). e) FEM model output of vertical motions at the seismometer.
under the deck of the lander recorded a drop in the magnetic field
components completely in phase with the drop in the solar array With subsequent thermoelastic finite element modeling we
currents. These magnetometer signals can be fully explained by established that thermoelastic strains caused by cooling of the top
the array currents and do not indicate any variations of the 0.5-1.0 mm of regolith everywhere except the area shadowed by
ambient magnetic field. the WTS dome can explain the observed tilting of the
From the above information, one might expect that the VBB seismometer. The very thin (skin depth: 0.5-1mm) top layer of
seismometer, which is installed directly on the Martian surface the regolith is elastically coupled to the material below and as
and is covered by multiple thermal shields, should not respond to soon as the top layer contracts, the medium below must follow
a Phobos transit, particularly because the thermal time constant elastically. Stähler et al. (2020) invoked local inhomogeneities in
between the ambient air temperature and the VBB core sensor the regolith to convert the nearly uniform strain field into tilts.
temperature is in excess of 8 hours. However, this is not what Since the Stähler et al. (2020) paper, we found that the tilts can
happened. In fact, what we observe with high signal-to-noise also be modeled by a regolith with a more rigid top layer of 15-
ratio is that the VBB detects a purely horizontal acceleration 25 cm thickness. This is consistent with other observations of a
toward the north. Interpreted as a tilt, we observe a downward more cohesive duricrust layer seen in the pits under the lander
tilting to the south by 12 nrad, that are correctly predicted by descent engines and around the HP3 mole. (Figure 10 and Figure
FEM calculations, as seen in Figure 19. The onset of this 11, Spohn et al. 2021, Golombek et al. 2020a).
response is delayed relative to the solar array currents by ~5 This three-layer solution is, at the moment, non-unique and
seconds. If it weren’t for this time delay, one could have there is therefore large uncertainty in the duricrust thickness and
575
the rigidity contrast between it and the material below. The range where E is the Young's modulus and is the Poisson's ratio) of
of parameters that can fit the data by trading off against each the Martian ground in the region around SEIS is 239 +/- 140 MPa
other is given in Table 2. Glitches in the seismometer data (see from seismic and pressure data filtered in the (0.02 – 0.3 Hz)
Kim et al 2021) likely caused by thermoelastic stick-slip events frequency band. These average elastic properties are similar to
within the instrument complicate the extraction of the tilts. We those found in studies focusing on compliance analysis
hope to be able to constrain these parameters better with mentioned previously (Garcia et al. 2020; Kenda et al. 2020;
additional data from future Phobos transits. It is worthy of note Lognonné et al. 2020). These values, significantly larger than
that the Lander shadow does not contribute to the tilt at the those determined at the first centimeters by in-situ wave velocity
seismometer. Tilts are confined to within the shadow in which measurements or preliminary lab tests on Earth, correspond to a
they are generated and measured. It is fortuitous that the larger depth and integrate the effects of the blocky ejecta
seismometer is deliberately offset from the center of the WTS illustrated in Figure 12 below 12 m.
dome, and that the shadow is slightly asymmetric due to the sun By studying the direction of the ground tilt at the closest
not being directly overhead. If the geometry were fully approach (when radial tilt amplitude is largest), it would appear
symmetric and the duricrust homogeneous, there would be no that the majority of seismically detected dust devils have a
tilts to observe at the seismometer. closest approach to the east of SEIS and that there is a distinct
lack of detections to the west (Golombek et al. 2020c). However,
Table 2. The parameters of the thermoelastic model. the vortices that lift dust also leave behind tracks on the surface
Top skin thickness 0.5 mm of Mars that are visible from satellites in orbit. Perrin et al. (2020)
Skin temperature change Depends on which transit use post-landing high-resolution satellite images to monitor dust
Duricrust thickness 15 to 25 cm devil activity during the first eight months of the mission. They
GDURICRUST/GHALFSPACE 1.5 to 2.2 find that the dust devil tracks observed around the InSight
Poisson Ratio -0.1 to 0.25 landing site from orbit indicate a preferential direction, aligned
Thermal Expansion 6.9e-6 to 1.0e-5 per K closely with the most common wind direction (~N145°E +/-
30°), but are not preferentially located to the east of SEIS.
4.2.3 Convective vortices Murdoch et al. (2021) demonstrate that a plausible explanation
Convective vortices (named dust devils when the vortex for the observed close-approach azimuth bias in the seismic data
transports dust particles) are detected as a sharp dip in local is that the ground is softer to the east thus increasing the number
pressure in the time series (e.g., Murphy et al. 2016). The day of seismic detections of vortices on that side. Such an
time turbulence at the InSight landing site is very active, leading interpretation of ‘softer’ ground to the east is consistent with
to a vast population of 6000 recorded convective vortex geomorphological surface interpretations (Golombek et al.
encounters over the first 400 sols of operations (Banfield et al. 2020).
2020; Spiga et al. 2020). During its passage, the relatively low
atmospheric pressure at the center of a convective vortex pulls
the elastic ground up causing the surface to tilt away from the 5 FAILURE PROPERTIES
vortex. This leads to a tilt signature on the horizontal Soil mechanics experiments aiming at characterizing the
component of a seismometer in contact with the ground. The mechanical failure properties (i.e., the internal friction angle and
isolated seismic signature of a convective vortex was first the cohesion) have been performed using the IDA. Specifically,
detected on Earth in 2015 (Lorenz et al. 2015) and has since been the IDA and its scoop (see Figure 8) were used around sol 673
detected on Mars with SEIS (Lognonné et al. 2020; Banerdt et (i.e., the 673rd day of the mission) to apply pressure on the ground
al. 2020), including at high frequencies (Lorenz et al., 2021). near HP3, as well as scrape, scoop, and dump the loose surface
In the context of the InSight mission, vortex modelling of the material (Marteau et al. 2021). They also have been more
seismic and pressure data has contributed to providing the first recently used to bury the SEIS tether (from sol 802), in an attempt
constraints on the average properties of the shallow elastic to get an improved thermal insulation of the tether from the
structure of Mars (Lognonné et al. 2020). These initial analyses atmospheric temperature changes (~100°C from day to night),
used the plane wave approach of Sorrells' theory (Kenda et al. respectively), so as to reduce the glitches and noise affecting the
2020), that is based on elastic calculations of the effects of SEIS seismological data. Observation of scrapes, trenches and
change in gas pressure at the surface of a semi-infinite layered dumped piles provided interesting information on the mechanical
half-space. The point-source modelling approach has also been properties of the regolith at the InSight site.
used to interpret joint observations of a dust devil vortex made
by orbital imaging, in-situ imaging, SEIS and the InSight 5.1 IDA pressing over the mole pit
pressure and wind sensors (Banerdt et al. 2020). Through
waveform fitting, the simple vortex model allowed the dust devil Shear strength parameters for the regolith at the mole pit have
pressure drop event to be identified among the multiple pressure been estimated by a back-analysis of the observed geometry at
drops present in the data. The Young’s modulus of the upper few the pit wall during soil loading by the lander’s robotic arm scoop
meters of the Martian subsurface in a specific known location at (Marteau et al. 2021). Depending on the orientation of the scoop,
a distance of ~20 m from the lander was then determined to be two different pushes were applied on the regolith: a flat push (see
~270 MPa by combining the information from the multiple Figure 8) and a tip push. The width and depth of the hole were
instruments (Banerdt et al. 2020). More recent analyses set equal to 0.045 m and 0.07 m, respectively, with a wall incli-
(Onodera, 2022) have been made on more than 500 pressure nation equal to 85° (see the shape of the pit in Figure 11).
drops and for different wind levels. This allows a better For the analysis, a 3D finite element model has been em-
resolution with depth in the ground rigidity determination, as ployed, assuming an elasto-plastic Mohr-Coulomb failure crite-
wind and period are the two key parameters enabling the control rion for the material. Soil unit weight soil in the Mars environ-
of the depth sensitivity of the compliance method. This confirms ment was set equal to 4.526 kN/m³ (corresponding to a unit mass
the gradient with depth of previous studies and shows that the of 1.216 Mg/m3 under Mars gravity of 3,721 m/s2). A friction
first upper meter below SEIS is made with material of about 20 angle of = 30 ° was fixed for all calculations and the required
MPa, while deeper, the Young’s modulus rises from 300-500 cohesion was determined iteratively so that the stability criterion
MPa at a few meters to 500-800 MPa deeper than 50 m. is just fulfilled, with a
In another analyses studying almost 500 vortices, Murdoch The applied force of 29 N derived from the motor currents
et al. (2021) estimate that the mean value of ( = E 2], activating the IDA for the flat push was distributed over a square
576
footprint area (side length = 0.07 m) corresponding to the scoop model (see Figure 21), one observes that the slopes of the
width (Figure 8), with a trapezoidal intensity starting with zero bulldozed mounds of regolith can be as high as 38-39° on the
along the front line, attaining a maximum of 7.89 kPa at the upstream side (S1), where the particles have been pushed by the
centre and remaining constant afterwards. For the tip push, the scoop. Once the scoop is no longer in contact with the regolith,
measured force of 45 N has been converted to a rectangular load particles likely cascade down the slope. On the downstream side,
of intensity 91.84 kPa distributed over a footprint area of 0.07 m the slopes of the piles are between 49 to 53°. The walls formed
by 0.007 m. The inferred minimum cohesion values are by the sides (S2) of the scoop have steep slopes with values of
c’ = 0.38 kPa for the flat push and 5.8 kPa for the tip push. Figure 78° and 70°.
20 displays the model with displacement contours for the flat
push.
e
a)
-
-
-
r
-
- Figure 21. (a) Digital elevation model of the HP3 pit based on the stereo
pair taken on Sol 673 after scraping; (b) Elevation profile S1 of the
scraped pile (c) Elevation profile S2 of the scraped wall (Marteau et al.
2022, Spohn et al. 2021a).
Starting on sol 803, the arm and scoop were used to bury the
SEIS tether by creating a large number of scrapes, scooping and
dumping the scraped regolith. The arm scraping activities created
several scrapes from which the regolith was extracted (like in
b) Figure 21). The regolith extracted in the scoop was then dumped
Figure 20. a) Photo of the scoop pressing above the pit; b) Finite element in piles above the tether close to SEIS along line S2 in Figure 22,
model with incremental displacement contours for the flat push (Vrettos with some proportion blown away by the wind, as commented
2022, pers. comm.). previously.
Elevation profiles were obtained from digital elevation mod-
5.2 Scraping and piling operations els acquired between Sols 803 and 822 to measure the slopes of
the scraped piles and walls. The average slope of the regolith
During scraping operations, it is important to remind that the
mounds is between 42° ± 2.7° (with a range from 40° to 45.8°),
Martian soil is defined by three successive and different unit
comparable to what is observed along line S1 in Figure 21. The
masses: i) the intact and densest one, prior to scraping; ii) the unit
side walls scraped by the vertical sides of the scoop have a slope
mass once extracted in the scoop, related to the previous one by
value of 54.7° ± 6.6° (with a range from 46° to 63°) smaller than
an expansion ratio, estimated for sands on the Earth around 10%
those determined in Figure 21.
and iii) once dumped into a pile. In this case, it was observed, at
Subsequently, the IDA scooped and dumped the scraped
least during the tether burial operations described below, that
material from a height of 40 cm on top of the SEIS tether. The
some proportions of finer particles were blown by the wind away
digital elevation model and elevation profiles of the dumped pile
from the pile. In other words, the material of the pile is somewhat
obtained on Sol 877 are presented in Figure 22. At its highest
coarser than that contained in the scoop. Investigations are pres-
point, the dumped pile is ~3 cm high and the material rests at
ently carried out (Verdier et al. 2022) to estimate the proportion
slope value of 24.1° ± 6.1° (with a range from 16.6° to 32°).
of fines submitted to the combined horizontal action of the wind
The difference in slope angle between the dumped and
(Stokes law, based on the wind speed and direction monitored by
scraped piles is first due to the difference between the materials,
the weather station of the lander) and gravity, based on the photo
with less fine particles in the dumped piles. In the intact state of
identification of the surface along which these particles have
the regolith, these fine particles, mixed to sand, are known to
been blown away.
provide some degree of cohesion also increased by low gravity
On sol 673, two overlapping 12-cm long scrapes were
(Walton et al. 2007). This would explain that the imprints of the
commanded to bring regolith from the far side of the HP3 pit
scoop in the unconsolidated sand layer, like rover tracks on other
towards the lander (Marteau et al. 2022). The scrapes created two
missions, are perfectly smooth (Golombek et al. 2020a). This
piles near the HP3 mole by bulldozing mounds of particles. From
small cohesion explains that the wall along slope S2 in Figure 21
the elevation profiles extracted from the sol 673 digital elevation
have such a high angle compared to the angle of repose. Another
577
possible reason explaining the differences in slopes is most likely have dragged down (Spohn et al. 2018). Upon penetration, a ther-
linked to the methods used to create the piles. Piles produced by mal conductivity and thermal diffusivity profile would have been
pouring a material (from which finer particles have been blown measured using sensors embedded in the mole (e.g., Grott et al.
out by the wind) from a given height yield different bulk densities 2021). From the product of the temperature gradient at a depth of
and geometries than piles generated by scraping. Larger densities at least 3 m and the thermal conductivity, the surface heat flow
are obtained with larger heights. The fall height from the scoop of Mars would have been calculated. It would have been required
was here around 45 cm. to reach that minimum depth to avoid the perturbation of the tem-
perature gradient caused by seasonal variations of the surface
temperature. The thermal conductivity and diffusivity measure-
ments would have allowed an estimate of the density, and the
mechanical strength of the soil would have been derived from the
penetration rate that would have been recorded. Unfortunately,
the mole did not penetrate deeper than a total of 40 cm. There-
fore, the surface heat flow could not be determined, and the phys-
ical properties could only be derived down to that depth.
The mole was designed to penetrate cohesionless soil like
quartz sand, which was expected to provide a good analogue
material for Martian sand. The sand would provide friction to the
buried mole hull to balance the remaining recoil of the mole
hammer mechanism that drives the mole forward. Although a
suppressor mass and spring in the hammer mechanism absorbed
much of the recoil, the available mass did not allow a system that
would have eliminated the recoil altogether (Spohn et al. 2021b).
The root cause of the failure - as was determined through an
extensive, almost two years long campaign (Spohn et al. 2021a,
b) - was a lack of friction in an unexpectedly thick cohesive and
brittle duricrust. In addition, it was found that the Martian soil
provided unexpected levels of penetration resistance. By
precessing about a point midway along its hull, the mole carved
a 7 cm deep and 5-6 cm wide pit during initial hammering and
reached a depth of initially 31 cm aided by friction springs in the
mole support structure.
During the campaign - described in detail in Spohn et al
(2021a) - the mole penetrated further to a final depth of about 40
cm aided by friction applied using the scoop at the end of the
IDA and by direct pressure by the latter. That brought the mole
body 1 - 2 cm below the surface at which depth thermal
conductivity measurements could be performed (Grott et al.
Figure 22. (a) Digital Elevation Model of the surface between HP3 and 2021). The mole reversed its downward motion twice during
SEIS based on the stereo pair taken on Sol 877 after scraping, scooping
and dumping; (b) Elevation profile S1 of the dumped pile; (c) Elevation
attempts with the IDA to provide friction through pressure on the
profile S2 of the dumped pile (Marteau et al. 2020). regolith instead of directly with the scoop to the hull.
The penetration record of the mole, the thermal conductivity
The average slope value of 42° obtained from the scraped and diffusivity measurements, and seismic data recorded during
piles is rather high with respect to the typical friction angle of a the hammerings were used to derive a model of the properties of
loose sub-rounded/rounded granular regolith (estimated around the first 40 cm of the Martian soil at the landing site (Spohn et al.
30° at undisturbed state, Golombek et al. 2008, Morgan et al. 2021a). Accordingly, a duricrust of about 20 cm thickness is
2018). This high slope value could be related to i) a larger friction found underneath a one-centimeter-thick unconsolidated sand
angle of a soil mass that has been moved into a denser state as a and dust layer. Beneath the duricrust a sand layer of about ten-
result of pushing and compacting during scraping and ii) the centimeter thickness is found followed by a layer of a sand mixed
effects of cohesive forces acting in conjunction with the presence with gravel, possibly consisting of debris from a small impact
of fines within the sand grains under low gravity conditions crater.
(Walton et al. 2007) and low atmospheric pressure (see The penetration resistance of the sand/gravel layer is best
Bromwell’s (1966) work in lunar conditions). In addition, the constrained and was found to be 5.3 MPa while the duricrust has
slopes of the side walls created by the scrapes are larger than a 5 to 10 times smaller penetration resistance. Applying cone
those of the piles, whereas slope failure was not seen on these penetration theory, the resistance of the duricrust was used to
walls. This confirms the presence of some inter-grains bounding estimate its cohesion to be 4 - 25 kPa, depending on the assumed
forces, mainly due to the fine fraction, that provide some internal friction angle of the duricrust. Pushing the scoop with its
cohesion to the undisturbed soil. Finally, the average slope value blade into the surface and chopping off a piece of duricrust
of 24.1° obtained for the dumped piles can be interpreted as a provided another estimate of the cohesion of about 6 kPa.
lower bound estimate of the angle of repose of the soil (with no The density of the topmost sand layer is estimated to be
fines) in a loose remolded state. 1.3 Mg/m3 while that of the duricrust is 1.1 Mg/m3, followed by
1.3 Mg/m3 for the sand layer underneath and by 1.6 Mg/m3 for
the sand/gravel layer. The thermal conductivity increases from
6 PENETRATION ISSUES 14 mWm-1K-1 to 34 mWm-1K-1 through the one-centimeter
sand/dust layer, keeps the latter value in the duricrust and the
The HP3 was designed to measure the temperature profile, a ther-
sand layer underneath and then increases to 64 mWm-1K-1 in the
mal conductivity profile and related physical properties down to
sand/gravel layer below (Grott et al. 2021).
a depth of 3 – 5 m. The temperatures would have been measured
using sensors on a tether that a small penetrator (the mole) would
578
7 CONCLUDING REMARKS Arvidson, R. E., et al. (2009), Results from the Mars Phoenix Lander
Robotic Arm experiment, Journal of Geophysical Research: Planets,
Increased interest in the physical and mechanical properties of 114, E00E02.
the Martian soil started during the early Viking 1 and 2 missions Banerdt, W.B., Smrekar, S.E., Banfield, D. et al. (2020). Initial results
(1976), which provided significant contributions from photos from the InSight mission on Mars. Nature Geosciences 13, 183–189.
and trenching/scraping activities with their robotic arms. Many Banfield, D., Spiga, A., Newman, C. et al. (2020). The atmosphere of
Mars as observed by InSight. Nature Geosciences 13, 190–198.
years later, the same approach has been adopted by the InSight Brinkman, N., Schmelzbach, C., Sollberger, D. et al (2022) In-situ
mission and its IDA robotic arm. These data have been supported regolith seismic velocity measurement at the InSight landing site on
by better quality photos and, more importantly, completed by pi- Mars. submitted to Journal of Geophysical Research: Planets
oneering measurements of wave velocities at the surface of the Fergason, R., Christensen, P., Bell, J., Golombek, M., Herkenhoff, K., &
planet during the hammering session of the HP3 and the monitor- Kieffer, H. (2006). Physical properties of the Mars Exploration
ing of their arrival by SEIS, providing valuable elastic parame- Rover landing sites as inferred from Mini-TES-derived thermal
ters of the regolith at small strains. These parameters have been inertia. Journal of Geophysical Research: Planets 111(E2), E02S21.
further constrained by compliance calculations allowed by the Bromwell. L. (1966). The friction of quartz in high vacuum. Materials
monitoring of the ground deformations induced by convective Science Report, MIT, Cambridge Mass.
Cantor, B.A., (2007). MOC observations of the 2001 Mars planet
vortexes. Interestingly, previous estimates from through labora- encircling dust storm. Icarus 186, 60–96.
tory testing on Earth on regolith simulants provided good predic- Christensen P.R., H.J. Moore (1992). The martian surface layer, in
tions of the parameters, including with respect to strength param- MARS, ed. by H.H. Kieffer, B.M. Jakosky, C.W. Snyder, M.S.
eters. Of course, photo observations also demonstrated that the Matthews (University of Arizona Press, Tucson, 1992), pp. 686–727
regolith on Mars is different from the simpler regolith simulants Delage P., Karakostas F., Dhemaied A., Belmokhtar M., Lognonné P.,
used in laboratory testing. Instead of being a well-sorted sand, it Golombek M.et al. 2017. An investigation of the mechanical
appears that it also contains some pebbles embedded in a sandy properties of some Martian regolith simulants with respect to the
matrix with some degree of cohesion that might be partly related surface properties at the InSight mission landing site. Space Science
to the fine particles contained in the regolith, and also to the low Reviews, 211, 191-213
Delage P. , Castillo-Betancourt J.P., Caicedo-Hormaza B., Karakostas F.,
gravity and atmospheric pressure conditions. Photo observations De Laure E., Lognonné P., Antonangeli D., Banerdt B. (2022). The
could also be completed by Digital Image modelling that pro- interaction between the SEIS seismometer of the InSight Martian
vided accurate morphologies of the trenches and dumps made mission and a regolith simulant. Géotechnique, accepted.
with the IDA scoop, both to support the attempts of mole pene- Edwards, C. S., Piqueux, S., Hamilton, V. E., Fergason, R. L.,
tration and the SEIS tether burying operations. The differences Herkenhoff, K. E., Vasavada, A. R., et al. (2018). The
in slopes observed between trenches and piles were related to the thermophysical properties of the Bagnold dunes, Mars: Ground-
way they were set up, evidencing some degree of cohesion due, truthing orbital data. Journal of Geophysical Research: Planets, 123,
among other things, to the fines contained in the sand/fine mix- 1307– 1326.
ture. Interestingly, it was also observed that the fine fraction of Fayon L., Knapmeyer-Endrun B., Lognonné P., Bierwirth P., Kramer A.,
Delage P., Karakostas F. et al. (2018). A numerical model of the
the regolith was dispersed by the wind. This feature is presently SEIS leveling system transfer matrix and resonances: application to
analyzed to better constrain the proportion of fines contained in SEIS rotational seismology and dynamic ground interaction. Space
the surface regolith, which will provide interesting further infor- Science Reviews, 214: 119.
mation with respect to the previous estimation get by thermal in- Folkner, W. M., Dehant V., Le Maistre S., Yseboodt M., Rivoldini A. et
ertia measurements. al. (2018). The rotation and interior structure experiment on the
Complementary information has also been gained from InSight mission to Mars. Space Science Reviews 214, 100.
thermal measurements, that locally confirmed the preliminary Garcia, R. F., Kenda, B., Kawamura, T., Spiga, A., Murdoch, N.,
orbiter thermal inertia measurements. In spite of the difficulties Lognonné, P. H., et al.
instrument, improvement of seismic records, and characterization of
met by the mole, that could not be driven as planned, the (very
long period atmospheric waves from ground displacements. Journal
low) thermal conductivity at surface could be directly determined of Geophysical Research: Planets, 125, e2019JE006278.
by the HP3 probe. Thermal changes and their thermo-elastic Giardini D., Lognonné P., Baerdt W.P., Pike W.T., Christensen et al.
effects on the surface were also detected during Phobos eclipses, (2020). The seismicity of Mars. Nature Geosciences 13, pages 205–
together with the effects of change in atmospheric pressure, 212.
thanks to the very high sensitivity of SEIS. Goetz W. et al., Microscopy analysis of soils at the Phoenix landing site,
Further soil mechanics tests are now planned for the Mars: classification of soil particles and description of their optical
remaining of the mission, in which the robotic arm will be used and magnetic properties. Journal of Geophysical Research: Planets
to apply controlled forces on the ground, while measuring the 115, E00E22 (2010).
resulting tilts by SEIS. In other words, this successful Golombek, M., Haldemann, A. F. C., Simpson, R. A., Fergason, R. L., et
al. (2008). Martian surface properties from joint analysis of orbital,
geophysical mission also provided – and will provide – Earth-based, and surface observations. In I. Bell Jim (Ed.), The
significant geological and geotechnical results. Martian surface composition, mineralogy, and physical properties
(Ch. 8).
Golombek, M., Kipp D., Warner I.J., Daubar I.J., Fergason R.L., Kirk
8 ACKNOWLEDGEMENTS R.L. et al. (2017). Selection of the InSight landing site. Space
Science Review 211, 5–95.
We acknowledge NASA, CNES, their partner agencies and Golombek M., · M. Grott,· G. Kargl, · J. Andrade, · J. Marshall, · N.
Institutions (UKSA, SSO, DLR, JPL, IPGP-CNRS, ETHZ, IC, Warner, · N.A. Teanby, · V. Ansan et al. (2018). Geology and
MPS-MPG) together with Ecole des Ponts ParisTech and Physical Properties Investigations by the InSight Lander. Space
Universidad de los Andes (Colombia) for supporting this work. Science Review 214:84.
This paper is InSight Contribution Number ICN 253. Golombek, M., Warner, N. H., Grant, J. A., Hauber, E., et al., (2020)a,
Geology of the InSight landing site on Mars: Nature
Communications, 11, 1014.
9 REFERENCES Golombek, M., Kass, D., Williams, N., Warner, N., Daubar, I., Piqueux,
S., Charalambous, C., and Pike, W. T., (2020)b, Assessment of
Arvidson, R. E., et al. (2008), Spirit Mars Rover Mission to the Columbia InSight landing site predictions: Journal of Geophysical Research:
Hills, Gusev crater: Mission overview and selected results from the Planets, v. 125, e2020JE006502.
Cumberland Ridge to Home Plate, Journal of Geophysical Research: Golombek, M., Williams, N., Warner, N. H., et al, (2020)c, Location and
Planets, 113, E12S33. setting of the Mars InSight lander, instruments and landing site,
Earth and Space Science, v. 7, e2020EA001248.
579
Golombek, M., Charalambous, C., Pike, W. T., and Sullivan, R., (2020)d, Moore, H. J., Jakosky, B. M. (1989). Viking landing sites, remote sensing
The origin of sand and dust on Mars: Evidence from the InSight observations, and physical pproperties of Martian surface materials.
landing site: 51st Lunar and Planetary Science Conference, Abstract Icarus 81, 164 – 184.
#2744, Lunar and Planetary Institute, Houston. Morgan P., Grott M., Knapmeyer-Endrun B., Golombek M., Delage P.,
Grott, M., Spohn, T., Knollenberg, J., Krause, C., Scharringhausen, M., Lognonné P. et al. (2018). A pre-landing assessment of regolith
Wippermann, T., et al. (2019). Calibration of the heat flow and properties at the InSight landing site, Space Science Reviews
physical properties package (HP3) for the InSight Mars mission. 214:104.
Earth and Space Science, 6(12), 2556–2574. Mueller, N. T., Knollenberg, J., Grott, M., Kopp, E., Walter, I., Krause,
Grott, M., T. Spohn, J. Knollenberg et al. (2021) Thermal conductivity C., et al. (2020). Calibration of the HP3 Radiometer on InSight. Earth
of the Martian soil at the InSight landing site from HP3 active and Space Science, 7(5), e01086.
heating experiments. Journal of Geophysical Research: Planets, Mueller, N., S. Piqueux, M. Lemmon, J. Maki, R. D. Lorenz, M. Grott,
doi:10.1029/2021JE006861 et al. (2021). Near Surface Properties of Martian Regolith Derived
Hamilton, V. E., Vasavada, A. R., Sebastián, E., Torre Juárez, M., From InSight HP³-RAD Temperature Observations During Phobos
Ramos, M., Armiens, C., et al. (2014). Observations and preliminary Transits. Geophysical Research Letters 48, e93542.
science results from the first 100 sols of MSL Rover Environmental
Monitoring Station ground temperature sensor measurements at Schnidrig, R., et al. (2021). Constraining Martian regolith and vortex
Gale Crater. Journal of Geophysical Research: Planets, 119(4), parameters from combined seismic and meteorological
745–770. measurements. Journal of Geophysical Research: Planets, 126,
Herkenhoff, K. E., Golombek, M. P., Guiness, E. A., Johnson, J. B., e2020JE006410.
Kusack, A., Richter, L., Sullivan, R. J., and Gorevan, S. (2008). In Murphy, J. et al., Field Measurements of Terrestrial and Martian Dust
situ observations of the physical properties of the martian surface: Devils. (2016) Space Science Reviews, vol. 203 (n° 1). pp. 39-87.
Chapter 20 in The Martian Surface: Composition, Mineralogy and Onodera, K. (2022). Subsurface structure of the Moon and Mars deduced
Physical Properties, J. F. Bell III editor, Cambridge University from 3D seismic wave propagation simulation and analysis of Apollo
Press, p. 451-467. and InSight seismic data, PhD Thesis, Université de Paris.
Kedar, S., Andrade, J., Banerdt, B., Delage, P., Golombek, M., Grott, M., Palluconi, F. D., & Kieffer, H. H. (1981). Thermal inertia mapping of
et al. (2017). Analysis of regolith properties using seismic signals Mars from 60°S to 60°N. Icarus, 45(2), 415–426.
generated by InSights HP3 penetrator. Space Science Reviews, Pan, L., Quantin, C., Tauzin, B., Michaut, C., Golombek, M., Lognonné,
211(1-4), 315–337. P.et al. 2020, Crust stratigraphy and heterogeneities of the first
Kenda, B., Drilleau, M., Garcia, R. F., Kawamura, T., Murdoch, N., kilometers at the dichotomy boundary in western Elysium Planitia
Compaire, N., et al. (2020). Subsurface structure at the InSight and implications for InSight lander: Icarus, v. 338, 113511.
landing site from compliance measurements by seismic and Perrin, C., Rodriguez, S., Jacob, A., Lucas, A., Spiga, A., Murdoch, N.,
meteorological experiments. Journal of Geophysical Research: et al. (2020). Monitoring of dust devil tracks around the InSight
Planets, 125, e2020JE006387. landing site, Mars, and comparison with in situ atmospheric data.
Kieffer, H. H., Martin, T. Z., Peterfreund, A. R., Jakosky, B. M., Miner, Geophysical Research Letters, 47, e2020GL087234.
E. D., & Palluconi, F. D. (1977). Thermal and albedo mapping of Piqueux, S., & Christensen, P. R. (2009). A model of thermal
Mars during the Viking primary mission. Journal of Geophysical conductivity for planetary soils: 2. Theory for cemented soils.
Research, 82(B28), 4249–4291. Journal of Geophysical Research: Planets, 114(E9), E09006.
Kim, D., Davis, P., Lekic, V., Maguire, R., Compaire, N., Schimmel, M., Piqueux, S., Mueller, N., Grott, M., Siegler, M., Millour, E., Forget, F.,
Stutzmann, E., Irving, J., Lognonné, P., Scholz, J-R., et al. (2021). et al. (2021). Regolith Properties near the InSight Lander Derived
Potential pitfalls in the analysis and structural interpretation of from 50 Sols of Radiometer Measurements. Journal of Geophysical
seismic data from the Mars InSight mission. Bulletin of the Research: Planets, submitted.
Seismological Society of America, 111 (6), 2982-3002. Pou, L., Nimmo, F., Lognonné, P., Mimoun, D., Garcia, R. F., Pinot,
Lognonné P., W.B. Banerdt, D. Giardini, W.T. Pike, U. Christensen, P. B., A. Rivoldini D. Banfield W. B. Banerdt (2021). Forward
Laudet, S. de Raucourt et al. (2019). SEIS: InSight’s Seismic Modeling of the Phobos Tides and applications to the first Martian
Experiment for Internal Structure of Mars. Space Science Review, year of the InSight mission. Earth and Space Science, 8,
https://doi.org/10.1007/s11214-018-0574-6. e2021EA001669
Lognonné, P., W.B. Banerdt, W.T. Pike, et al. (2020). Constraints on Putzig, N. E., & Mellon, M. T. (2007). Apparent thermal inertia and the
the shallow elastic and anelastic structure of Mars from InSight surface heterogeneity of Mars. Icarus, 191(1), 68–94.
seismic data, Nature Geoscience, http://doi.org/10.1038/s41561- Pike, W. T., U. Staufer, M. H. Hecht, W. Goetz, D. Parrat, H. Sykulska-
020-0536-y. Lawrence, S. Vijendran, and M. B. Madsen (2011), Quantification
Lorenz, R. et al., (2015). Seismometer Detection of Dust Devil Vortices of the dry history of the Martian soil inferred from in situ
by Ground Tilt, Bulletin of the Seismological Society of America, microscopy, Geophysical Research Letters, 38, L24201.
105(6): 3015. Santamarina J.C., K.A. Klein, M.A. Fam, Soils and Waves (Wiley, New
Lorenz, R.D. Spiga, A. Lognonné, P., Plasman, M., Newman, C.E, York, 2001)
Charalambous, C. (2021). The whirlwinds of Elysium: A catalog Sollberger, D., Schmelzbach, C., Andersson, F., Robertsson, J. O. A.,
and meteorological characteristics of “dust devil” vortices observed Brinkman, N., Kedar, S., et al. (2021). A reconstruction algorithm
by InSight on Mars, Icarus, 355, doi : 10.1016/j.icarus.2020.114119. for temporally aliased seismic signals recorded by the InSight Mars
Marteau, E., Golombek, M., Vrettos, C., Garvin J. and Williams, N. lander. Earth and Space Science, e2020EA001234.
(2021). Soil Mechanical Properties at the InSight Landing Site on Spiga, A., Murdoch, N., Lorenz, R., Forget, F., Newman, C., Rodriguez,
Mars. 52nd Lunar and Planetary Science Conference (LPSC), S., et al. (2020). A study of daytime convective vortices and
Abstract #2067, Lunar and Planetary Institute, Houston
Marteau E., M. Golombek, C. Vrettos, P. Delage, N.R. Williams, and V . -Eddy
Ansan (2022). Soil strength properties derived from scraping and Simulations. Journal of Geophysical Research: Planets 125,
dumping activities at the InSight landing site on Mars. 53nd Lunar e2020JE006511.
and Planetary Science Conference, Abstract 1253, Houston, USA. Spohn, T., Grott, M., Smrekar, S. E., Knollenberg, J., Hudson, T. L.,
Mellon, M. T., Jakosky, B. M., Kieffer, H. H., & Christensen, P. R. Krause, C., et al. (2018). The heat flow and physical properties
(2000). High-resolution thermal inertia mapping from the Mars package (HP3) for the InSight mission. Space Science Reviews,
global surveyor thermal emission spectrometer. Icarus, 148(2), 437– 214(5), 96.
455. Spohn T., Hudson T., Marteau E. et al (2021a) The HP3 penetrator (Mole)
Moore, H. J., G. D. Clow, and R. E. Hutton (1982). A summary of Viking on Mars: Soil properties derived from the penetration attempts and
sample-trench analyses for angles of internal friction and cohesions, related activities. Submitted to Space Science Reviews,
Journal of Geophysical Research: Planets 87, 10043–50. http://arxiv.org/abs/2112.04438
Moore H.J., R.E. Hutton, G.D. Clow, and C.R. Spitzer (1987). Physical Spohn, T., Hudson, T., Witte, L. et al (2021b) The InSight HP 3 mole on
properties of the surface materials at the Viking landing sites on Mars: Lessons Learned from attempts to penetrate to depth in the
Mars. U.S. geological survey professional paper 1389. Martian soil. Advanced Space Research, in press,
http://arxiv.org/abs/2112.03234
580
Stähler R.C., R. Widmer-Schnidrig, J.-R. Scholz, M. van Driel, A.
Mittelholz, K. Hurst, et al. (2020). Geophysical Observations of
Phobos Transits by InSight. Geophysical Research Letters, 47,
e2020GL089099.
Trebi-Ollennu A, W Kim, K Ali et al (2018). InSight Mars Lander
Robotics Instrument Deployment System. Space Science Reviews,
doi: 10.1007/s11214-018-0520-7
Verdier, N., Asan-Mangold V., Delage P., et al. (2022). Using wind
dispersion effects during the InSight tether burial activities to better
constrain the regolith grain size distribution. In prep.
Viudez-Moreiras D, Newman, CE, Forget, F, Lemmon, M, Banfield, D
et al. (2020). Effects of a Large Dust Storm in the Near-Surface
Atmosphere as Measured by InSight in Elysium Planitia, Mars.
Comparison With Contemporaneous Measurements by Mars
Science Laboratory, Journal of Geophysical Research: Planets,
(125) 9, e2020JE006493.
Walton O.R., De Moor P. C., Gill K.S. (2007). Effects of gravity on
cohesive behavior of fine powders: implications for processing
Lunar regolith. Granular Matter 9:353–363.
Warner N. H., M.P, Golombek, V. Ansan, E. Marteau, N. Williams et
al. (2022a). In-situ and orbital stratigraphic observations from the
insight landing site, Mars. 53rd Lunar and Planetary Science
Conference, Houston, USA.
Warner, N.H., Golombek, M.P, Ansan, V., Marteau, E., Williams, N. et
al. (2022b). In Situ and Orbital Stratigraphic Characterization of the
InSight Landing Site – The Type Example of a Regolith-Covered
Lava Plain on Mars. Submitted to Journal of Geophysical Research:
Planets.
581
Engineering Geology
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Xuanmei Fan
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology,
Chengdu, Sichuan, People's Republic of China, [email protected]
ABSTRACT: As the Third Pole of the Earth, Tibet Plateau (TP) is sensitive to climate change which induces complex geo-
environmental changes and hydrological patterns. In recent years, with increasing environmental temperatures, there have been
growing numbers of geological hazards that produced significant threats and damage to the settlements, hydropower stations, and
transportation infrastructures in this tectonically active region. However, the mechanism causing such chains of geological hazards
is not well understood. Therefore, studies on glacial-related geohazards induced by climate change in the Tibetan Plateau region
have become an emerging hot issue. In this paper, we introduce five common types of cascading hazards induced by climate change
and their possible mechanisms. Some representative cases are discussed. We provide a brief review of studies related to the impact
of climate change on chains of geohazards on the TP and its margins, and highlight future research directions.
RÉSUMÉ : En tant que troisième pôle de la Terre, le plateau du Tibet (PT) est sensible au changement climatique qui induit des
changements géo-environnementaux et formes hydrologiques complexes. Au cours des dernières années, avec l’augmentation des
températures environnementales, il y a eu un nombre croissant de risques géologiques qui ont produit des menaces et des dommages
importants pour la population, les centrales hydroélectriques et les infrastructures de transport dans cette région tectoniquement active.
Cependant, le mécanisme à l’origine de telles chaînes de risques géologiques n’est pas bien compris. Par conséquent, les études sur les
géorisques liés aux glaciers et donc induits par le changement climatique dans la région du plateau tibétain sont devenues très
importantes. Dans cet article, nous présentons cinq types de dangers en cascade induits par le changement climatique et leurs mécanismes
possibles. Certains cas représentatifs sont discutés. Nous fournissons un bref examen des études liées à l’impact du changement
climatique sur les chaînes de géorisques sur le PT et le long de ses frontières, et mettons en évidence les orientations futures de la
recherche.
KEYWORDS: Climate change, Tibetan Plateau, Geohazard chains, Rock-ice avalanche
satellite images, it nearly slumbered between the 1960s and 2012.
1 INTRODUCTION However, it has been activated since 2012, especially after the
2017 Ms 6.9 Milin earthquake. Many studies in the Sedongpu
Tibetan Plateau (TP), as the Third Pole of earth and the Water region proposed that frequent geohazard events have a close
Tower of Asia, acted as the driver and amplifier of climate connection to climate change (Chen et al. 2020; Jia et al. 2019;
change (Yang et al. 2014). The average warming rate of the TP Li et al. 2021).
-1) is approximately 1.5 times of the global In addition, TP is a region with intensive tectonic deformation
-1) (Zhang et al. 2020). It and high-level earthquake activity (Huang and Fan 2013; Wang
means the changing of temperature in the TP is more sensitive to et al. 2011). Strong earthquakes, caused by the continental
climate change. The TP has been experiencing significant collision, weakened and fractured the steep mountain slopes,
environmental changes in the past few decades. The major increasing the probability of geological disasters (Fan et al.
changes on TP are driven by glacier retreat, glacial lake outburst 2019a; Jiang et al. 2021). The influence of the Wenchuan
floods (GLOFs), landslides and debris flows caused by warming- earthquake on geological hazards has been well documented,
induced changes, etc., with severe repercussions on the socio- especially on the spatio-temporal evolution of landslides and
economic condition in the region. debris flows (Fan et al. 2018; Fan et al. 2019b), the formation of
Accompanied by climate change, earthquakes, and landslide dams and dammed lakes (Fan et al. 2021b; Fan et al.
rainstorms, there are increasing numbers of geological hazards 2020), the enhancing sedimentary transportation (Dai et al. 2021;
on the Tibetan Plateau and its margins. The Yigong landslide in Fan et al. 2019a; Fan et al. 2021a), the vegetation recovering
2000 was one of the most catastrophic landslides worldwide, (Yunus et al. 2020), etc. In the TP region, the freeze-thaw process
resulting in enormous casualties and property losses in China and caused by climate change can further cause strength degradation
India. 300 Mm3 of debris blocked the Yigong River for about 62 of the rock mass in high mountain regions and enhance the
days, resulting in reservoir volume increasing to about 2.015 possibility of slope instability (Krautblatter et al. 2013).
Gm3. The corresponding catastrophic breaching caused a Furthermore, post-seismic debris flows induced by extreme
massive outburst of flood in the Yarlung Zangpo (in Tibet) and rainfall events are of significant concern in the seismically
the Dihang (in India) rivers (Delaney and Evans 2015; Kang et affected zones (Jiang et al. 2021). Under the effects of climate
al. 2017; Zhou et al. 2016). Adjacent to the Yigong landslide, change, along with strong seismic activities, geological hazards
Tianmo Valley did not suffer from debris flows for could be prone to present a considerable threat to society.
approximately 100 years before 2007. However, approximately TP experienced a significant increase in air temperature
1.34 Mm3 sediment was transported from the channel in 2007, especially in the past 50 years, the increasing rate of air
causing eight persons missing. Subsequently, two debris flow temperature has been twice the global average, reaching 0.3-0.4
events occurred in 2010, generating a dam lake with a volume of every ten years (Yao et al. 2019). With such a rapid rate of
90 Mm3, inundating the highway G318 (Deng et al. 2017; Wei et climate change in the past decades, there has been an accelerated
1) since the
al. 2018). Sedongpu ice-rock avalanche-debris flows, located on rate of glacier mass loss
the left bank of Yarlung Zangbo River, have temporarily blocked mid-1990s (Bolch et al. 2012; Brun et al. 2017). The retreat of
the river many times since 2012. Interestingly, according to glaciers and excessive rains gradually expanded glacial lakes at
the low-lying lands and formed a potential source of flooding
585
(Clague and Evans 2000; Mool et al. 2001; Song et al. 2016;
Wang et al. 2020). Climate change has increased the natural
disasters in the region at an alarming rate. These rapid changes
on TP and their effects on the downstream area, although studied
extensively, lack a clear understanding of the cascading effects.
This paper introduces five types of geohazard chains induced
by climate change and the mechanisms associated with them on
the Tibetan Plateau and its margins. We briefly reviewed recent
studies related, presented a few representative cases of chains of
geological hazards, and highlighted future research directions.
2 TYPES OF CLIMATE CHANGE INDUCED CHAINS OF
GEOLOGICAL HAZARS ON TIBETAN PLATEAU
REGIONS Figure 1. The distribution of strong earthquakes in the Tibetan Plateau
and its margins. Earthquakes in this region are according to the
Earthquake Hazards program dataset
2.1 Climatically induced cascading hazards after strong
(https://earthquake.usgs.gov/earthquakes). Size of filled circles
earthquakes corresponds to earthquake magnitude. The background Terrain is from
Strong earthquakes are typical of TP region, resulted by the Stamen Design open-source maps (http://maps.stamen.com/).
continent-continent collision. International Seismological Center
(ISC) database reports over 7200 earthquakes between 1975 and 2.2 Climate change-induced slope failures in permafrost
2015 on the Tibetan Plateau margin. Well known examples regions on TP
include 2008 Wenchuan (Mw 7.9), and 2015 Nepal (Mw 7.5) Permafrost related slopes failures are typical examples of
cases, both of them caused numerous landslides in the mountain climate-change-induced geohazards. The 0.5 additional
slopes, loss of life and infrastructure damages, see Figure 1. warming is a severe threat to cold permafrost regions like the
Remobilizations of the coseismic landslide deposits in the form Qinghai-Tibet plateau. The frequency of extreme weather-related
of debris flows during subsequent rainy periods are the common geohazards is increasing, and freeze-thaw induced slope failures
geological hazard in areas affected by earthquake-induced in permafrost regions are not exempt. Over 1,500,000 km2 area
landslides (Fan et al. 2019a). of Qinghai-Tibet plateau consists of permafrost (Zhou et al.
The 2008 Wenchuan earthquake triggered the largest number 2018; Niu et al. 2005). During thawing, shear strength reduction
of coseismic landslides on record: 200,000 (Xu et al. 2014). occurs in ice-rich sediments within the active layer and triggers
After this event, extensive debris remobilizations were occurred slope failures. Driven by the increasing demand for infrastructure
continuously for several years, even multiple times in the same development, transport networks encounter thaw induced-
deposit (Fan et al. 2021a). They often resulted in human losses permafrost slope failures along the new engineering corridors
and damage to property and infrastructure. The prediction of (Niu et al. 2014). Studies in the recent decades' efforts on
post-seismic landslides is almost based on analysing the understanding the types of these slope failures as retrogressive
relationship between debris flows initiation and cumulative flows, thaw slumps, and gelifluction (Wei et al. 2006). Of these
rainfall. We noticed that the triggering rainfall intensities of post- types, retrogressive failures and active layer detachment thaw
seismic debris flows evolve with time. Based on 172 triggering slumps are prominent. Terrian factors, i.e., slope, aspect, deposits
rainfalls and 2396 non-triggering rainfalls of debris flows from characteristics, temperature, and ground-ice content, control
2008 to 2013 after the Wenchuan earthquake, we analysed the these slope failures' spatial distribution (Niu et al. 2005). External
evolution of probabilistic rainfall thresholds for post-seismic triggering factors of these slope failures are increased
debris flows using a Bayesian technique and found that rainfall temperature (thawing), heavy summer precipitation, and
thresholds significantly decrease compared with pre-earthquake occasionally an earthquake. Studies have identified a 253%
events initially and later tend to increase annually (Jiang et al. increase in thaw slumps in the last decade (Luo et al. 2019). With
2021). Meanwhile, the triggering rainfall characteristics tend to the rise of air temperature during the thawing season and
gradually change from a short-duration high-intensity pattern to abundant precipitation related to extreme-weather events under
a long-duration and low-intensity pattern. climate change, it is vital to expect the trend to prolong.
Elevated frequency of debris flows has been reported also
after the 2015 Gorkha earthquake in Nepal (Dahlquist & West 2.3 Climate change-induced ice/rock mass strength
2019) and 2017 Juizhaigou earthquake in China (Fan et al. degradation and cascading hazards
2021a). Nevertheless, studies show that the rates of debris
remobilisations and post-seismic landslides eventually return to The temperature of ice is one of the most important factors
a pre-earthquake level in a few years (Marc et al. 2015; Fan et al. affecting the strength of ice (Petrovic 2003). The influence of
2019a). The decay timescale of major post-seismic landslide temperature rise on ice strength mainly includes three aspects:
disturbance appears to be a few years to several decades (Fan et directly reducing the strength of ice, the temperature stress that
al. 2019a). For example, after Mw 7.6 Chi-Chi earthquake of causes cracks in the ice, and the melting of ice reduces the
1999, the affected areas progressively returned to the pre- confining pressure of the ice. The existing research results show
earthquake condition within seven to ten years; whereas that for that when the ice temperature is above -30°C, its strength
Wenchuan is more than 10 years. It is also worth to note that there decreases significantly with the increase of the ice temperature
were catastrophic reactivations during major rainstorm events in (Chang et al. 2021). Uneven temperature changes will produce
Wenchuan (e.g., 12th August 2010 rainstorm, 4th July 2011 temperature stress, causing cracks in the ice, thereby reducing the
storm and 9th July 2013 storm). Hence, extreme rainfall strength of the ice layer. In addition, the mechanical properties
variability needs to be taken into account in projections of of ice, like rocks, are also significantly affected by the stress
landslide decay assessments. environment in which it is placed. When the upper ice layer
melts, the confining pressure of the lower ice layer decreases,
which will also cause the loss of its strength.
On the other hand, the increase in temperature will also lead
to changes in the structure of the ice layer. It can be clearly found
in the field survey that the snow line has raised as the temperature
586
rises. As for the ice layer on the slope surface, the lower part
becomes thinner due to the rising temperature, and the upper part
becomes thicker due to further snowfall, which causes the slope
of the ice layer to increase, and the ice layer is prone to instability
(Kääb et al. 2018). In addition, the melting of the lower ice layer
also produces an effect similar to "excavating the foot of a slope",
which easily induces the collapse of the ice layer.
Due to above reasons, ice avalanches have continued to occur
on TP in recent years. For example, in 2016, two large-scale ice
avalanches occurred in the Ngari area of the Tibet Autonomous
Region, causing serious casualties and property losses (Kääb et
al. 2018). Ice avalanches can also cause cascading disasters such
as glacial lake outbursts. The ice body falling into the glacial lake
can stir up huge surges, swelling through the dam and then
causing the dam to break (Cook et al. 2016).
The increase in temperature not only leads to a decrease in the
strength of the ice, but also causes damage to the mechanical
properties of the rock mass. Freeze-thaw is generally considered
to be one of the leading mechanisms for the deterioration of rock
mechanical properties in cold regions (Huang et al. 2020).
Climate warming will gradually increase the severity and area of
freeze-thaw damage on TP. The infiltration of ice and snow melt
is another major factor that causes rock mass instability in cold
regions. Water infiltration not only lubricates the joints and
cracks of the rock mass, but also generates hydrodynamic
pressure and hydrostatic pressure, and even leads to liquefaction
of the sliding surface, which further promotes the generation and
development of landslides or collapses. Figure 2. (a) The catastrophic Yigong Landslide occurred in 2000 in TP,
On the TP and its margin regions, active tectonic movements and (b) the Chamoli ice-rock avalanches occurred in 2021 in India
coupled with long-term geological evolution processes such as
freeze-thaw cycles, dry-wet cycles, loading and unloading, have 2.4 Climate change-induced debris flows and flash floods
formed potentially unstable rock masses in many areas. In As a common type of natural hazard in mountainous areas, debris
addition to the infiltration of ice and snow melt water, climate flow is usually a mixed flow of rocks, mud, water and air (Iverson
change can also lead to extreme weather such as heavy rainfall, 1997; Hürlimann et al. 2019). The velocity and impact force of a
rapid changes in hydrological conditions often become a direct debris flow could be tremendous, imposing serious threats to
factor inducing the instability of dangerous rock masses. The people, properties, and infrastructure in the affected areas. Due
aforementioned Yigong landslide that occurred in 2000 was to favourable natural environments for geohazards, the southeast
caused by the coupling of such long-term and short-term of the Tibetan Plateau suffers serious debris flow hazards, which
inducing factors (Li et al. 2020; Zhou et al. 2016), see Figure 2a.
are initiated by intensive rainstorms, maritime monsoon glacier
The Chamoli ice and rock avalanche that occurred in 2021 in
melting, glacial avalanches and glacial lake outburst floods
India also has a similar initiation mechanism, Figure 2a (Shugar
(GLOFs) (Wei et al. 2018). Especially, with the effect of climatic
et al. 2021). Existing research results have shown that in the years
before the event completely collapsed, obvious seasonal warming, the occurrence of debris flows becomes more frequent.
deformation characteristics have been revealed: the ice rock mass Previous studies have demonstrated that three key factors may
deforms in summer, but remains stable in other seasons (Pandey contribute to the triggering of debris flows, including steep slope,
et al. 2021). In the winter when the avalanche occurred, availability of sediment and input water flow (Mcguire et al.
significant abnormal temperature changes also appeared. This 2017). On one hand, climatic warming will induce glacier retreat
further confirms that the hydrological driving factors caused by and permafrost degradation, which will expose more moraine
climate change have an important influence in the evolution of and rock slopes. The new-exposed moraine will be the
such events. supplement sediment for debris flows, which will promote the
activities of debris flows (Walter et al. 2020). On the other hand,
the high temperature will accelerate the melting of the glaciers,
the meltwater will be transformed into the runoff in the channel.
The effect of meltwater for the initiation of debris flow is similar
to that of antecedent rainfall. The melted water will saturate the
moraine, which makes the moraine into a water-filled state.
Under the action of heavy rainfall in the later period, the moraine
deposits in the channel will be easily washed away to transfer a
debris flow (Kumar et al. 2018). At the same time, in the
background of climate warming, the climatic characteristics in
the TP increasingly show a trend that the rainfall and high
temperature will occur in the same period, that is, the strong
melting of glaciers occurs in summer, and this is also the season
of concentrated precipitation. The superposition of concentrated
rainfall and high temperature will enhance the volume of water
supply in the channel and promote the occurrence of debris flows
under the coupling effect of glacial-melting and intensive rainfall
(Deng et al. 2017). In addition, climatic warming will also
increase the numbers and volumes of moraine-dammed lakes
587
with the possibility of hazardous debris flows induced by Glacial Table 1. Details of rock-ice avalanche in the Tibetan Plateau
Locatio Latitu Longitu Date Type Referen
Lake Outburst Floods (GLOFs) in the Tibetan Plateau. ID
n de de ces
Due to climate change, glaciers in the TP have shown an Yigong 30.23 95.00 2000.04 Rock-ice Li et al.
accelerating retreat (Dussaillant et al. 2019; Fujita and Ageta 1
.09 avalanche 2020
2000; Kääb et al. 2018; Ke et al. 2017; Yao et al. 2012). By 2015, 2
Aru 34.03 82.25 2016.07 Glacier/ice Kääb et
the area of glaciers has decreased by about 178 km2 since 1999 .17 avalanche al. 2018
34.01 82.26 2016.09 Glacier/ice
(Bhattacharya et al. 2021). The increased melting volume of .02 avalanche
glaciers has largely increased the frequency of mass movements, Yarlon 29.81 94.92 Before Liqiang
especially debris flows (Aggarwal et al. 2017; Muneeb et al. g 2014 et al.
2021; Ashraf et al. 2012; Cook et al.2018; Veh et al. 2019). In Tsangp 2018
3
u
the past sixty years, many large debris flows took place in the (Sedon
Tibetan Plateau, especially along the Yarlung River (Wang et al. gpu)
2021). These debris flows severely impacted the Sichuan-Tibet 2014
highway (State Highway G318) along the right bank of Parlung 2017.10
River, and the highway was frequently closed for long periods .22
2017.11
(Zou et al. 2019). On 29th Sep 1953, a mega glacier debris flow
with a leading head of 40m occurred in the Guxiang Gully, the 2017.12
sediment transport volume of the debris flow was about 17.1×106 2018.01
m3 (Zou et al. 2020; Dang et al. 2009). A huge debris flow 2018.07
accumulation fan with an area of 4.2 km2 also formed at the .06
2018.10 Rock-ice
mouth of the gully and blocked the Parlung River, a tributary of .17 avalanche
the Yarlung River. The debris flow event caused more than 140 2018.10 Rock-ice
deaths and a large amount of arable land was buried (You 2001). .29 avalanche
On September 4 of 2007, July 25–31 and September 5–8 of 2010, Amney 34.82 99.43 2004.02 Glacier/ice Paul
4 Mache .10 avalanche 2019
three large-scale debris flows took place in the Tianmo gully, a n
tributary of the Parlung River (Figure 3) (Wei et al. 2018; Deng 2007.10
et al. 2017; Wang et al. 2018). All the three debris flow events .08
blocked Parlung river and produced dammed lakes. The outburst 2016.10
.06
flow intensively scoured the foot of the high terrace at opposite Yulong 27.09 100.19 2004.03 Rock-ice Zhang
bank and made it collapse, resulting in the ruin of State Highway 5 Mount .12 avalanche et al.
G318 base and the interruption of traffic (Zou et al. 2020; Ge et ain 2007
al. 2014). The distribution of large-scale geohazards in the south- Peilong 1968- Glacier Li et al.
gully 1977? landslide 2021;
eastern of the TP is shown in Figure 4 and Table 1. 6
Wang et
al. 1999
Jiubie 38.71 75.27 2015.05 Glacier Li et al.
Peak in surge and 2016
7 Kongur ice
Mount avalanche
ains
Zelong 29.62 94.99 1950.08 Glacier
8 nong .15 surge/avala
glacier nche
1968.09
.02
1984.04
.13
Kunlun 35.75 93.53 2001.11 Ice van der
Mount .14 avalanche Woerd
9
ain et al.,
2004
North 34.33 77.16 2000.04 Rock-ice Bhutyan
Terong .05 avalanche i and
10
glacier Mahto
Figure 3. The Tianmo debris flow destroyed the G318 Highway in TP 2018
2.5 Climate change-induced glacier retreat and chains of
hazards
Due to climate change, glacier retreating, as well as increasing
number of glacial lakes nurturing by glacier retreating process,
are hot issues emerging during recent years, especially regional
inventories and spatial-temporal evolution analysis (Dehecq et
al. 2018; Chen et al. 2021; Dou et al. 2021a; Wang et al. 2020).
The TP climate is under the combined effects of the East Asian
and South Asian monsoons and of the westerlies, resulting in
inhomogeneous distribution and variation of glacier and glacial
lakes (Chen et al. 2021; Yang et al. 2014).
The satellite images with the least cloud cover and the least
mountain shadows were selected by the relevant algorithms, and
Figure 4. Distribution of large-scale chains of geohazards in the southeastern based on Google Earth Engine using image fusion techniques
Tibetan Plateau (Gorelick et al. 2017; Kumar and Mutanga 2018), we fused the
42833 Landsat satellite images into three images of different
periods (1990-1999, 2000-2012 and 2013-209) over the past
588
three decades, followed by subsequent processing using © glacial lake on TP, and limited accessible cases seemed crucial
ArcGIS Pro and © ENVI software, including manual cross- to acknowledge the characteristics of glacial lakes on TP.
checking and correction by image interpreters (Figure 5). Based
on this we produced an updated inventory of TP glacial lakes
with complete data for three periods (Dou et al. 2021b). In the 3 MECHANISM OF CLIMATE CHANGE INDUCED
last thress decades, our inventory showed that the number and CASCADING HAZARDS
area of glacial lakes in TP region had increased by 3285 and
258.82 km2, respectively (Figure 6). There area, however, 3.1 Post-seismic landslides and debris flows driven by climatic
decreased in the western Pamir and the eastern Hindu Kush due forces
to reduced rainfall rates (Dou et al. 2021a). The coseismic deposits are typically constituted by loose
materials with significant amounts of fines, hence are susceptible
to sudden collapse and liquefaction upon loss of suction or pore
water pressure increase (Hu et al. 2017, 2018). Hence the major
factor for the post-seismic landslides and debris remobilization
in earthquake affected areas are found to be the volumes of
coseismic deposits and rainfall intensities (Dadson et al. 2004;
Hovius et al. 1997; Marc et al. 2016). Other than the above,
multiple mechanisms may also be at play (Kincey et al. 2021);
essential controls seem to be the sizes, topography, landcover,
and hydrological conditions, the healing or removal of
seismically damaged layers, climatic factors, and the legacy of
2021b; Tian et al. 2020). The decay of post-seismic landslide
Figure 5. Flowchart of the glacial lake automatic extraction and mapping
activities is controlled by various processes, including the
workflow
progressive exhaustion of the hillslope and channel material,
precipitation regime, and hillslope healing due to revegetation,
grain coarsening and consolidation (Domènech et al. 2019; Fan
et al. 2018, 2019a). An important one among these is the recovery
of vegetation growth (Lin et al. 2006; Saba et al. 2010; Yang et
al. 2017, 2018; Fan et al. 2018). Yunus et al. (2020) by analyzing
decadal evolution of vegetation recovery in the affected region
showed that post-earthquake landslide activity returned to the
pre-earthquake level within 18 years. Tanyas et al. (2021a)
reported that quick recovery of post-seismic landslides, also
correlated with the seasonality of precipitation. According to
them, a prolonged precipitation makes the recovery process
quicken.
3.2 Initiation and runout mechanism of rock-ice avalanches-
debris flows
Rock-ice avalanches and resulting debris flows origination in
Figure 6. Area change rate of glacial lakes from the first period (1990- mountain permafrost regions are more significant threats of
1999) to the last period (2013-2019) on 1°×1° grids. The circle size climate change regarding the scale and nature of impacts of these
represents the total glacial lake area in the first period of each grid geohazards. In comparison, the magnitude of the long-runout
rock-ice avalanches were always enormous; their frequency in
The accompanied GLOFs aggravated with glacier retreat and the recent decade has increased, possibly due to climate change
glacial lakes expansion. Benefit from remote sensing (Huggel et al. 2010; Schneider et al. 2011; Schneider et al. 2013;
development, archived glacial lakes and GLOF database are Cloutier et al. 2017). Unlike the progress in understanding the
established to explore their spatial-temporal evolution (Zheng et initiation and runout mechanisms of warmer region rock-
al. 2021a; Veh et al. 2019; Veh et al. 2020). Since 1900, more avalanches (Hungr 2006; Korup 2011; Shugar et al. 2013;
than 100 GLOFs have originated in the Himalayas, causing Dufresne et al. 2016), there are critical gaps in our knowledge on
damage to bridges, road, houses and other infrastructure (Zhang rock-ice avalanches (Pudasaini & Krautblatter 2014; Yu et al.
et al. 2022; Nie et al. 2018; Zheng et al. 2021b). More than 50% 2020; Sansone et al. 2021). Similar to rock avalanches triggered
of catastrophic moraine dam failures ascribed to rock/ice by an earthquake, studies explain a few instances of earthquake-
avalanche in the Himalaya (Allen et al. 2016). Similarly, some triggered rock-ice avalanches on TP (van der Woerd et al. 2004).
research revealed that 26 out of 33 glacial lakes outburst are However, the initiation mechanism is still speculative for many
related to rock/ice avalanche in TB region (Liu et al. 2019). climate-weather driven rock-ice avalanches (Martha et al. 2021;
Therefore, exploring the process and mechanisms of those chains Shugar et al. 2021). Debris flows origination from rock-ice
of hazards is essential for risk assessment. avalanches are a more significant threat than rock-ice avalanches
With the development of urbanization and hydropower (Yu et al. 2020). Cold region permafrost debris flows multiply
projects booming in TB, downstream impact of GLOFs is urgent the impact of rock-ice avalanches into glaciers. Permafrost debris
for policy-makers and research scientists (Dubey and Goyal flows are more common in Tibet (Ge et al. 2014; Deng et al.
2020). Some large-scale analyses have been made to assess and 2017; Wei et al. 2018). Yu et al. (2020) separate the initiation
predict dangerous glacial lake (Allen Simon et al. 2019; Zheng mechanism of debris flow into geomechanically related and
et al. 2021a). Apart from regional analysis, individual case also hydraulically related groups. Geomechanically triggered debris
made detailed assessment by numerical simulation or other flows originate through rock-ice avalanches and liquefaction of
methods (Westoby et al. 2014; Mergili et al. 2017). However, moraine deposits (Noetzli et al. 2006; Jiskoot 2011; Carey et al.
accurate assessment was impeded by nearly inaccessible of 2012), and hydraulically triggered debris flow originate through
589
heavy rainfall and/or rapid ice/snowmelt runoff. evolution responses to climate change and tectonic
movements needs more studies in the future.
3.3 Mechanism of moraine dam breach Early recognition and warning of cascading hazards on TP
Moraine dam breach and subsequent flooding are typical based on multi-source remote sensing data and monitoring
examples of climate change induced-cascading hazards on TP system are crucial to control their potential risk. Seismic
(Montgomery et al. 2004; Korup & Montgomery 2008). Recent monitoring system might be a useful method to monitoring
events of rock-avalanche and rock-ice avalanche landslide potential hazards in the inaccessible regions.
damming events highlight the climate emergency on TP and also The initiation and runout mechanisms of large rock-ice
signifies the importance and threat of these transboundary avalanches and debris flows, especially the thermo-hydro-
geohazards. The impacts of moraine dam breach are cascading in mechanical coupling mechanism need to be further studied
nature due of sequential triggering of secondary hazards i.e., by field monitoring and laboratory tests.
flash floods, and debris flows (Cui et al. 2010; Cui et al. 2013;
Xiangang et al. 2017; Bazai et al. 2021). Recent reviews on the Risk prediction of future hazards by considering different
formation and stability and breach mechanisms of landslide dams magnitudes of earthquakes, climate change and their
highlight the present state of knowledge and gaps in coupling effect is in high need of risk control of cascading
understanding of dam breach mechanisms and modelling (Fan et hazards on TP. To this end, numerical models based on
al. 2020; Fan et al. 2021b; Zhong et al. 2021). However, detailed mechanism research will be very helpful.
studies on moraine dam failure are few (Neupane et al. 2019).
Moraine dams are mechanistically weaker compared to landslide
dams due to their mixed grain composition, morphology and 5 ACKNOWLEDGEMENTS
sloping positions (Neupane et al. 2019). The width to height ratio
This research is financially supported by the National Science
of moraine dams is found to be smaller than landslide dams
Fund for Outstanding Young Scholars of China (Grant No.
(Evans 1986; Clague & Evans 1994; Evans et al. 1996). Neupane
42125702). The author would like to thank ICSMGE for the
et al. (2019) differentiate types of moraine dams using their
invitation to present the work.
sediment height and composition. Triggering or initiation of a
moraine dams are referred to thawing of ice-rich sediments,
sudden impact of snow/ice/rock/rock-ice avalanches into glacial 6 REFERENCES
lakes, tectonic events i.e., earthquake triggering these mass
movements, and extreme-hydro-meteorological events like Aggarwal, S., Rai, S.C., Thakur, P.K. and Emmer, A. 2017. Inventory
abrupt snowmelt, and heavy precipitation. While the onset of and recently increasing GLOF susceptibility of glacial lakes in
failures is known for the reasons above, the instability of dam Sikkim, Eastern Himalaya. Geomorphology 295, 39-54.
breach is driven by wave overtopping, erosion (both internal and Allen Simon, Keith, Guoqing Zhang, Weicai Wang, Tandong Yao, and
Tobias Bolch. 2019. 'Potentially dangerous glacial lakes across the
external), seepage and slope failures. These are similar
Tibetan Plateau revealed using a large-scale automated assessment
mechanisms of landslide dam breach. However, moraine-dam approach', Science Bulletin 64: 435-45.
breach is characterised by thermo-hydro-mechanical changes Ashraf, A., Naz, R. and Roohi, R. 2012. Glacial lake outburst flood
that dynamically take place during failure (Hewitt 1999; Korup hazards in Hindukush, Karakoram and Himalayan Ranges of
& Tweed 2007; Neupane et al. 2019). Pakistan: implications and risk analysis. Geomatics Natural Hazards
& Risk 3(2), 113-132.
Bazai, N. A., Cui, P., Carling, P. A., Wang, H., Hassan, J., Liu, D. 2021.
4 CONCLUSIONS Increasing glacial lake outburst flood hazard in response to surge
glaciers in the Karakoram. Earth-Science Reviews 212, 103432.
In the Tibetan Plateau and it's margins, the impact of climate https://www.sciencedirect.com/science/article/pii/S0012825220304
change is evident from the concurrent, cascading, and chains of 785
geohazards in recent decades (Lu et. al. 2019). Both tectonic Bhattacharya, A., Bolch, T., Mukherjee, K., King, O., Menounos, B.,
Kapitsa, V., Neckel, N., Yang, W. and Yao, T. 2021. High Mountain
driven and climate driven geological processes have converted to Asian glacier response to climate revealed by multi-temporal
geohazards due to enhanced activity i.e., glacier retreat, glacial satellite observations since the 1960s. Nat. Commun 12(1), 4133.
lake outburst floods (GLOFs), cloud outburst events, flooding Bhutiyani, M. R. & Mahto, R. 2018. Remote-sensing-based study of
and landslides caused by warming-induced changes. With the impact of a rock avalanche on North Terong Glacier in Karakorum
rise of air temperature during the thawing season and abundant Himalaya. Int. J. Remote Sens. 39(22), 8076–8091.
precipitation related to extreme-weather events under climate Bolch, T., Kulkarni, A., Kääb, A., Huggel, C., Paul, F., Cogley, J. G., &
change, it is vital to expect the trend to prolong. This brief review Stoffel, M. 2012. The state and fate of Himalayan glaciers. Science
336(6079), 310-314.
emphasizes the following directions for future research outlook: Brun, F., Berthier, E., Wagnon, P., Kääb, A., & Treichler, D. 2017. A
The Tibetan Plateau is one of the most sensitive regions in spatially resolved estimate of High Mountain Asia glacier mass
the world to climate change, and also one of the most balances from 2000 to 2016. Nature geoscience 10(9), 668-673.
tectonically active regions with frequent strong earthquakes. Carey, M., Huggel, C., Bury, J., Portocarrero, C., & Haeberli, W. 2012.
An integrated socio-environmental framework for glacier hazard
The climatic and tectonic conditions and their coupling management and climate change adaptation: lessons from Lake 513,
effects on causing geological hazards on TP are not well Cordillera Blanca, Peru. Clim Change 112(3), 733-767.
understood. https://doi.org/10.1007/s10584-011-0249-8
Efforts have been made in generating glacier and glacier Chang X, Liu W, Zuo G, et al. 2021. Research on ultrasonic-based
investigation of mechanical properties of ice. Acta Oceanol. Sin
lake inventories, but there is still a data gap in all different
40(10), 97-105.
types of geological hazards for the whole TP. Most of Chen, C., L. M. Zhang, T. Xiao, and J. He, 2020. Barrier lake bursting
database cover only specific areas. Both spatial and and flood routing in the Yarlung Tsangpo Grand Canyon in October
temporal resolution of database need to be improved. 2018. Journal of Hydrology 583. doi:
Based on the abovementioned multi-temporal inventory of 10.1016/j.jhydrol.2020.124603
Chen, F., M. M. Zhang, H. D. Guo, S. Allen, J. S. Kargel, U. K.
multi-hazards, the spatio-temporal evolution of geological Haritashya, and C. S. Watson. 2021. 'Annual 30m dataset for glacial
hazards for the whole TP can be analysed. How their lakes in High Mountain Asia from 2008 to 2017', Earth System
Science Data, 13: 741-66.
590
Clague, J. J., & Evans, S. G. 1994. Formation and failure of natural dams Evans, S. G., Hu, X. Q., & Enegren, E. G. 1996. The 1973 Attachie slide,
in the Canadian Cordillera: Geological Survey of Canada Bulletin. Peace River Valley, near Fort St. John, British Columbia, Canada: a
Clague, J. J., & Evans, S. G. 2000. A review of catastrophic drainage of landslide with a high-velocity flowslide component in Pleistocene
moraine-dammed lakes in British Columbia. Quaternary Science sediments. Paper presented at the Proceedings, 7th International
Reviews 19(17-18), 1763-1783. Symposium on Landslides.
Cloutier, C., Locat, J., Geertsema, M., Jakob, M., & Schnorbus, M. 2017. Fan, X. M., et al., 2021b. Recent technological and methodological
Potential impacts of climate change on landslides occurrence in advances for the investigation of landslide dams. Earth-Science
Canada. In Slope Safety Preparedness for Impact of Climate Change Reviews 218. doi: 10.1016/j.earscirev.2021.103646
(pp. 71-104): CRC Press. Fan, X. M., G. Scaringi, G. Domenech, F. Yang, X. J. Guo, L. X. Dai, C.
Cook S J, Kougkoulos I, Edwards L A, et al. 2016. Glacier change and Y. He, Q. Xu, and R. Q. Huang, 2019b. Two multi-temporal datasets
glacial lake outburst flood risk in the Bolivian Andes. Cryosphere that track the enhanced landsliding after the 2008 Wenchuan
10(5), 2399-2413. earthquake. Earth System Science Data 11(1), 35-55. doi:
Cook K.L., Andermann, C., Gimbert, F., Adhikari, B.R. and Hovius, N. 10.5194/essd-11-35-2019
2018. Glacial lake outburst floods as drivers of fluvial erosion in the Fan, X., Domènech, G., Scaringi, G., Huang, R., Xu, Q., Hales, T.C., Dai,
Himalaya. Science 362(6410), 53-57. L., Yang, Q., Francis, O., 2018. Spatio-temporal evolution of mass
Cui, P., Dang, C., Cheng, Z., & Scott, K. M. 2010. Debris flows resulting wasting after the 2008 Mw 7.9 Wenchuan earthquake revealed by a
from glacial-lake outburst floods in Tibet, China. Physical detailed multi-temporal inventory. Landslides 15, 2325–2341.
Geography 31(6), 508-527. https://doi.org/10.1007/s10346-018-1054-5
Cui, P., Zhou, G. G. D., Zhu, X. H., & Zhang, J. Q. 2013. Scale Fan, X., Dufresne, A., Subramanian, S. S., Strom, A., Hermanns, R.,
amplification of natural debris flows caused by cascading landslide Stefanelli, C. T., et al. 2020. The formation and impact of landslide
dam failures. Geomorphology 182, 173-189. dams – State of the art. Earth-Science Reviews 103116.
Dadson, S.J., Hovius, N., Chen, H., Dade, W.B., Lin, J.-C., Hsu, M.-L., http://www.sciencedirect.com/science/article/pii/S00128252193047
Lin, C.-W., Horng, M.-J., Chen, T.-C., Milliman, J., Stark, C.P., 63
2004. Earthquake-triggered increase in sediment delivery from an Fan, X., Scaringi, G., Korup, O., West, A.J., van Westen, C.J., Tanyas,
active mountain belt. Geology 32, 733–736. H., Hovius, N., Hales, T.C., Jibson, R.W., Allstadt, K.E., Zhang, L.,
https://doi.org/10.1130/G20639.1Dahlquist, M.P., West, A.J., 2019. Evans, S.G., Xu, C., Li, G., Pei, X., Xu, Q., Huang, R., 2019a.
Initiation and Runout of Post-Seismic Debris Flows: Insights From Earthquake-Induced Chains of Geologic Hazards: Patterns,
the 2015 Gorkha Earthquake. Geophys. Res. Lett. 46, 9658–9668. Mechanisms, and Impacts. Rev. Geophys 57, 421–503.
https://doi.org/10.1029/2019GL083548 https://doi.org/10.1029/2018RG000626
Dai, L. X., G. Scaringi, X. M. Fan, A. P. Yunus, J. Liu-Zeng, Q. Xu, and Fan, X., Yunus, A.P., Scaringi, G., Catani, F., Siva Subramanian, S., Xu,
R. Q. Huang, 2021. Coseismic Debris Remains in the Orogen Q., Huang, R., 2021a. Rapidly Evolving Controls of Landslides
Despite a Decade of Enhanced Landsliding. Geophysical Research After a Strong Earthquake and Implications for Hazard Assessments.
Letters 48(19). doi: 10.1029/2021gl095850 Geophys. Res. Lett 48, 1–12.
Dang, C., Cui, P. and Cheng, Z.-l. 2009. The formation and failure of https://doi.org/10.1029/2020GL090509
debris flow-dams, background, key factors and model tests: case Fujita, K. and Ageta, Y. 2000. Effect of summer accumulation on glacier
studies from China. Environmental geology 57(8), 1901-1910. mass balance on the Tibetan Plateau revealed by mass-balance
Dehecq, Amaury, Noel Gourmelen, Alex S. Gardner, Fanny Brun, Daniel model. Journal of Glaciology 46(153), 244-252.
Goldberg, Peter W. Nienow, Etienne Berthier, Christian Vincent, Ge, Y.-g., Cui, P., Su, F.-h., Zhang, J.-q. and Chen, X.-z. 2014. Case
Patrick Wagnon, and Emmanuel Trouvé. 2018. 'Twenty-first century history of the disastrous debris flows of Tianmo Watershed in Bomi
glacier slowdown driven by mass loss in High Mountain Asia', County, Tibet, China: Some mitigation suggestions. Journal of
Nature Geoscience 12: 22-27. Mountain Science 11(5), 1253-1265.
Delaney, K. B., and S. G. Evans, 2015. The 2000 Yigong landslide Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D. and
(Tibetan Plateau), rockslide-dammed lake and outburst flood: Moore, R. 2017. Google Earth Engine: Planetary-scale geospatial
Review, remote sensing analysis, and process modelling. analysis for everyone. Remote Sens. Environ 202, 18-27.
Geomorphology, 246, 377-393. doi: Hewitt, K. 1999. Quaternary moraines vs catastrophic rock avalanches in
10.1016/j.geomorph.2015.06.020 the Karakoram Himalaya, northern Pakistan. Quaternary Research
Deng, M. F., N. S. Chen, and M. Liu, 2017. Meteorological factors 51(3), 220-237.
driving glacial till variation and the associated periglacial debris Hovius, N., Stark, C.P., Allen, P.A., 1997. Sediment flux from a
flows in Tianmo Valley, south-eastern Tibetan Plateau. Natural mountain belt derived by landslide mapping. Geology 25, 231–234.
Hazards and Earth System Sciences, 17(3), 345-356. doi: https://doi.org/10.1130/0091-
10.5194/nhess-17-345-2017 7613(1997)025<0231:SFFAMB>2.3.CO;2
Domènech, G., Fan, X., Scaringi, G., van Asch, T.W.J., Xu, Q., Huang, Hu, W., Scaringi, G., Xu, Q., Huang, R., 2018. Internal Erosion Controls
R., Hales, T.C., 2019. Modelling the role of material depletion, grain Failure and Runout of Loose Granular Deposits: Evidence From
coarsening and revegetation in debris flow occurrences after the Flume Tests and Implications for Postseismic Slope Healing.
2008 Wenchuan earthquake. Eng. Geol 250, 34–44. Geophys. Res. Lett 45, 5518–5527.
https://doi.org/10.1016/j.enggeo.2019.01.010 https://doi.org/10.1029/2018GL078030
Dou, X., Fan, X., Yunus, A.P., Xiong, J., Tang, R., Wang, X. and Xu, Q. Hu, W., Scaringi, G., Xu, Q., Pei, Z., Van Asch, T.W.J., Hicher, P.Y.,
2021a. Spatio-temporal evolution of glacial lakes in the Tibetan 2017. Sensitivity of the initiation and runout of flowslides in loose
Plateau over the past 30 years. Earth Syst. Sci. Data Discuss 2021:1- granular deposits to the content of small particles: An insight from
31. flume tests. Eng. Geol 231, 34–44.
Dou, X., Fan, X., Yunus, A.P., Xiong, J., Tang, R., Wang, X. and Xu, Q. https://doi.org/10.1016/j.enggeo.2017.10.001
2021b. Spatio-temporal glacial lakes dataset in the Tibetan Plateau: Huang S, Lu Z, Ye Z, et al. 2020. An elastoplastic model of frost
1990 – 2019 (TPGL), Zenodo. deformation for the porous rock under freeze-thaw. Engineering
Dubey, S., and M. K. Goyal. 2020. 'Glacial Lake Outburst Flood Hazard, Geology 278, 105820.
Downstream Impact, and Risk Over the Indian Himalayas', Water Huang, R. Q., and X. M. Fan, 2013. The landslide story. Nature
Resources Research 56. Geoscience 6(5), 325-326. doi: 10.1038/ngeo1806
Dufresne, A., Bösmeier, A., & Prager, C. 2016. Sedimentology of rock Huggel, C., Salzmann, N., Allen, S., Caplan-Auerbach, J., Fischer, L.,
avalanche deposits–case study and review. Earth-Science Reviews Haeberli, W., et al. 2010. Recent and future warm extreme events
163, 234-259. and high-mountain slope stability. Philos Trans A Math Phys Eng
Dussaillant, I., Berthier, E., Brun, F., Masiokas, M., Hugonnet, R., Sci 368(1919), 2435-2459.
Favier, V., Rabatel, A., Pitte, P. and Ruiz, L. 2019. Two decades of http://www.ncbi.nlm.nih.gov/pubmed/20403836
glacier mass loss along the Andes. Nature Geoscience 12(10), 802- Hungr, O. 2006. Rock avalanche occurrence, process and modelling. In
808. Landslides from massive rock slope failure (pp. 243-266): Springer.
Evans, S. G. 1986. The maximum discharge of outburst floods caused by Hürlimann, M., Coviello, V., Bel, C., Guo, X.J., Berti, M., Graf, C., Hübl,
the breaching of man-made and natural dams. Canadian J., Miyata, S., Smith, J.B., Yin, H.Y. 2019. Debris-flow monitoring
Geotechnical Journal 23(3), 385-387. and warning: Review and examples. Earth-Science Reviews 199,
102981
591
Iverson, R.M. 1997. The physics of debris flows. Review of Geophysics Liqiang, T., Chang, Z., Jienan, T. U., Lixin, P. E. I., Zhaocheng, G. U.
5(3):245–296 O., Xiongwei, Z., Hong, C., Jinghui, F. A. N., Chunling, L. I. U.,
Jia, H. C., F. Chen, and D. H. Pan, 2019. Disaster Chain Analysis of Shanshan, W., & Peng, H. E. 2018. preliminary discussion of the
Avalanche and Landslide and the River Blocking Dam of the frequently debris flow events in sedongpu basin at gyalaperi peak
Yarlung Zangbo River in Milin County of Tibet on 17 and 29 yarlung.
October 2018. International Journal of Environmental Research and Liu, Jiankang, Jiajia Zhang, Bo Gao, Yuanling Li, Mengyu Li, Duoji
Public Health 16(23). doi: 10.3390/ijerph16234707 Wujin, and Luxu Zhou. 2019. 'An overview of glacial lake outburst
Jiang, Z., Fan, X., Siva Subramanian, S., Yang, F., Tang, R., Xu, Q., flood in Tibet,China', Journal of Glaciology and Geocryology 41:
Huang, R., 2021. Probabilistic rainfall thresholds for debris flows 1335-47.
occurred after the Wenchuan earthquake using a Bayesian technique. llen, Simon K, Andreas Linsbauer, Christian Huggel, SS Randhawa,
Eng. Geol. 280, 105965. Yvonne Schaub, and Markus Stoffel. 2016. 'Current and future
https://doi.org/10.1016/j.enggeo.2020.105965 glacial lake outburst flood hazard: application of GIS-based
Jiskoot, H. 2011. Long-runout rockslide on glacier at Tsar Mountain, modeling in Himachal Pradesh, India.' in, Climate change, glacier
Canadian Rocky Mountains: potential triggers, seismic and response, and vegetation dynamics in the Himalaya (Springer).
glaciological implications. Earth Surface Processes and Landforms Luo, J., Niu, F., Lin, Z., Liu, M., & Yin, G. 2019. Recent acceleration of
36(2), 203-216. https://doi.org/10.1002/esp.2037 thaw slumping in permafrost terrain of Qinghai-Tibet Plateau: An
Kaab A, Leinss S, Gilbert A, et al. 2018. Massive collapse of two glaciers example from the Beiluhe Region. Geomorphology 341, 79-85.
in western Tibet in 2016 after surge-like instability. Nature Marc, O., Hovius, N., Meunier, P., 2016. The mass balance of
Geoscience 11, 114-120. earthquakes and earthquake sequences. Geophys. Res. Lett 43, 3708–
Kääb, A., Leinss, S., Gilbert, A., Bühler, Y., Gascoin, S., Evans, S.G., 3716. https://doi.org/10.1002/2016GL068333
Bartelt, P., Berthier, E., Brun, F., Chao, W.-A., Farinotti, D., Marc, O., Hovius, N., Meunier, P., Uchida, T., Hayashi, S., 2015.
Gimbert, F., Guo, W., Huggel, C., Kargel, J.S., Leonard, G.J., Tian, Transient changes of landslide rates after earthquakes. Geology 43,
L., Treichler, D. and Yao, T. 2018. Massive collapse of two glaciers 883–886. https://doi.org/10.1130/G36961.1
in western Tibet in 2016 after surge-like instability. Nature Marc, O., Sens-Schönfelder, C., Illien, L., Meunier, P., Hobiger, M.,
Geoscience 11(2), 114-120. Sawazaki, K., Rault, C., Hovius, N., 2021. Toward using seismic
Kang, C., D. Chan, F. H. Su, and P. Cui, 2017. Runout and entrainment interferometry to quantify landscape mechanical variations after
analysis of an extremely large rock avalanche-a case study of earthquakes. Bull. Seismol. Soc. Am 111, 1631–1649.
Yigong, Tibet, China. Landslides 14(1), 123-139. doi: https://doi.org/10.1785/0120200264
10.1007/s10346-016-0677-7 Martha, T. R., Roy, P., Jain, N., Vinod Kumar, K., Reddy, P. S., Nalini,
Ke, L., Ding, X., Li, W. and Qiu, B. 2017. Remote Sensing of Glacier J., et al. 2021. Rock avalanche induced flash flood on 07 February
Change in the Central Qinghai-Tibet Plateau and the Relationship 2021 in Uttarakhand, India—a photogeological reconstruction of the
with Changing Climate. Remote Sensing 9(2). event. Landslides https://doi.org/10.1007/s10346-021-01691-9
Kincey, M.E., Rosser, N.J., Robinson, T.R., Densmore, A.L., Shrestha, Mcguire, L. A., Rengers, F. K., Kean, J. W. and Staley, D. M., 2017.
R., Pujara, D.S., Oven, K.J., Williams, J.G., Swirad, Z.M., 2021. Debris flow
Evolution of Coseismic and Post-seismic Landsliding After the 2015
Mw 7.8 Gorkha Earthquake, Nepal. J. Geophys. Res. Earth Surf 126. Geophysical Research Letters 44: 7310–7319.
https://doi.org/10.1029/2020JF005803 Mergili, Martin, Jan-Thomas Fischer, Julia Krenn, and Shiva P
Korup, O. 2011. Rockslide and rock avalanche dams in the Southern Pudasaini. 2017. 'r. avaflow v1, an advanced open-source
Alps, New Zealand. In Natural and Artificial Rockslide Dams (pp. computational framework for the propagation and interaction of two-
123-145): Springer. phase mass flows', Geoscientific Model Development 10: 553-69.
Korup, O., & Montgomery, D. R. 2008. Tibetan plateau river incision Montgomery, D. R., Hallet, B., Yuping, L., Finnegan, N., Anders, A.,
inhibited by glacial stabilization of the Tsangpo gorge. Nature Gillespie, A., & Greenberg, H. M. 2004. Evidence for Holocene
455(7214), 786. megafloods down the Tsangpo River gorge, southeastern Tibet.
Korup, O., & Tweed, F. 2007. Ice, moraine, and landslide dams in Quaternary Research 62(2), 201-207.
mountainous terrain. Quaternary Science Reviews 26(25-28), 3406- Mool, P. K., Wangda, D., & Bajracharya, S. R. 2001. Inventory of
3422. glaciers, glacial lakes and glacial lake outburst floods: monitoring
Krautblatter, M., D. Funk, and F. K. Gunzel, 2013. Why permafrost rocks and early warning systems in the Hindu Kush-Himalayan region-
become unstable: a rock-ice-mechanical model in time and space. Bhutan. International Centre for Integrated Mountain Development
Earth Surface Processes and Landforms 38(8), 876-887. doi: (ICIMOD).
10.1002/esp.3374 Muneeb, F., Baig, S.U., Khan, J.A. and Khokhar, M.F. 2021. Inventory
Kumar, A., Bhambri, R., Tiwari, S.K., Verma, A., Gupta, A.K., and GLOF Susceptibility of Glacial Lakes in Hunza River Basin,
Kawishwar, P. 2019. Evolution of debris flow and moraine failure in Western Karakorum. Remote Sensing 13(9), 1794.
the Gangotri Glacier region, Garhwal Himalaya: Hydro- Neupane, R., Chen, H., & Cao, C. 2019. Review of moraine dam failure
geomorphological aspects. Geomorphology 333:152-166. mechanism. Geomatics, Natural Hazards and Risk 10(1), 1948-
Kumar, L. and Mutanga, O. 2018. Google Earth Engine Applications 1966.
Since Inception: Usage, Trends, and Potential. Remote Sensing Nie, Y., Q. Liu, J. D. Wang, Y. L. Zhang, Y. W. Sheng, and S. Y. Liu.
10(10), 1509. 2018. 'An inventory of historical glacial lake outburst floods in the
Li J, Chen N, Zhao Y, et al. 2020. A catastrophic landslide triggered Himalayas based on remote sensing observations and
debris flow in China's Yigong: factors, dynamic processes, and geomorphological analysis', Geomorphology 308: 91-106.
tendency. Earth Sciences Research Journal 24(1), 71-82. Niu, F., Cheng, G., Ni, W., & Jin, D. 2005. Engineering-related slope
Li Junli, Chen Xi, Bao Anming, Shen Zhanfeng, Ji Shunping. 2016. failure in permafrost regions of the Qinghai-Tibet Plateau. Cold
Glacier hazard emergency monitoring of the Jiubie Peak in Kongur Regions Science and Technology 42(3), 215-225.
Mountains using unmanned aerial vehicle photogrammetry. Arid Niu, F., Luo, J., Lin, Z., Fang, J., & Liu, M. 2016. Thaw-induced slope
Land Geography 39(2), 378-386 (In Chinese) failures and stability analyses in permafrost regions of the Qinghai-
Li, W. L., B. Zhao, Q. Xu, G. Scaringi, H. Y. Lu, and R. Q. Huang, More Tibet Plateau, China. Landslides 13(1), 55-65.
frequent glacier-rock avalanches in Sedongpu gully are blocking the Niu, F., Luo, J., Lin, Z., Liu, M., & Yin, G. 2014. Thaw-induced slope
Yarlung Zangbo River in eastern Tibet. Landslides doi: failures and susceptibility mapping in permafrost regions of the
10.1007/s10346-021-01798-z Qinghai–Tibet Engineering Corridor, China. Natural Hazards 74(3),
Li, Z., Zhou, F., Han, X., Chen, J., Li, Y., Zhai, S., Han, M., & Bao, Y. 1667-1682.
2021. Numerical simulation and analysis of a geological disaster Noetzli, J., Huggel, C., Hoelzle, M., & Haeberli, W. 2006. GIS-based
chain in the Peilong valley, SE Tibetan Plateau. Bulletin of modelling of rock-ice avalanches from Alpine permafrost areas.
Engineering Geology and the Environment, 80(4), 3405–3422. Computational Geosciences, 10(2), 161-178.
Lin, C.W., Liu, S.H., Lee, S.Y., Liu, C.C., 2006. Impacts of the Chi-Chi P.Stark, S.J.D.N.H.H.C.W.B.D.J.-C.L.M.-L.H.C.-W.L.M.-J.H.T.-
earthquake on subsequent rainfall-induced landslides in central C.C.J.M.C., 2004. Earthquake-triggered increase in sediment
Taiwan. Eng. Geol 86, 87–101. delivery from an active mountain belt. Geology.
https://doi.org/10.1016/j.enggeo.2006.02.010 Pandey P, Chauhan P, Bhatt C M, et al. 2021. Cause and process
mechanism of rockslide triggered flood event in Rishiganga and
592
Dhauliganga river valleys, Chamoli, Uttarakhand, India using faults in the 2008 Wenchuan earthquake and uplift of the Longmen
satellite remote sensing and in situ observations. Journal of the Shan. Nature Geoscience 4(9), 634-640. doi: 10.1038/ngeo1210
Indian Society of Remote Sensing, 49(5), 1011-1024. Wang, S., Yang, Y., Gong, W., Che, Y., Ma, X. and Xie, J. 2021. Reason
Paul, F. 2019. Repeat Glacier Collapses and Surges in the Amney Analysis of the Jiwenco Glacial Lake Outburst Flood (GLOF) and
Machen Mountain Range, Tibet, Possibly Triggered by a Potential Hazard on the Qinghai-Tibetan Plateau. Remote Sensing
Developing Rock-Slope Instability. Remote Sensing 11(6), 708-. 13(16), 3114.
Petrovic J J. 2003. Review mechanical properties of ice and snow. Wang, X., Guo, X., Yang, C., Liu, Q., Wei, J., Zhang, Y& Tang, Z. 2020.
Journal of Materials Science 38(1), 1-6. Glacial lake inventory of high-mountain Asia in 1990 and 2018
Pudasaini, S. P., & Krautblatter, M. 2014. A two phase mechanical derived from Landsat images. Earth System Science Data 12(3),
model for rock ice avalanches. Journal of Geophysical Research: 2169-2182.
Earth Surface, 119(10), 2272-2290. Wei, M., Fujun, N., Satoshi, A., & Dewu, J. 2006. Slope instability
Sansone, S., Zugliani, D., & Rosatti, G. 2021. A mathematical framework phenomena in permafrost regions of Qinghai-Tibet Plateau, China.
for modelling rock–ice avalanches. Journal of Fluid Mechanics, 919. Landslides, 3(3), 260-264.
Schneider, D., Huggel, C., Haeberli, W., & Kaitna, R. 2011. Unraveling Wei, R., Zeng, Q., Davies, T., Yuan, G., Wang, K., Xue, X. and Yin, Q.
driving factors for large rock–ice avalanche mobility. Earth Surface 2018. Geohazard cascade and mechanism of large debris flows in
Processes and Landforms, 36(14), 1948-1966. Tianmo gully, SE Tibetan Plateau and implications to hazard
Schneider, J. F., Gruber, F. E., & Mergili, M. 2013. Recent cases and monitoring. Engineering Geology 233, 172-182.
geomorphic evidence of landslide-dammed lakes and related hazards Wei, R.Q., Zeng, Q.L., Davies, T., Yuan, G.X., Wang, K.Y., Xue, X.Y.,
in the mountains of Central Asia. In Landslide science and practice Yin, Q.F. 2018. Geohazard cascade and mechanism of large debris
(pp. 57-64): Springer. flows in Tianmo gully SE Tibetan Plateau and implications to
Shugar, D. H., Clague, J. J., & Giardino, M. 2013. A quantitative hazard monitoring. Engineering Geology, 233:172-182.
assessment of the sedimentology and geomorphology of rock Westoby, M. J., N. F. Glasser, J. Brasington, M. J. Hambrey, D. J.
avalanche deposits. In Landslide Science and Practice (pp. 321-326): Quincey, and J. M. Reynolds. 2014. 'Modelling outburst floods from
Springer. moraine-dammed glacial lakes', Earth-Science Reviews, 134: 137-
Shugar, D. H., Jacquemart, M., Shean, D., Bhushan, S., Upadhyay, K., 59.
Sattar, A., et al. 2021. A massive rock and ice avalanche caused the Xiangang, J., Peng, C., Huayong, C., & Yayong, G. 2017. Formation
2021 disaster at Chamoli, Indian Himalaya. Science, eabh4455. conditions of outburst debris flow triggered by overtopped natural
http://science.sciencemag.org/content/early/2021/06/09/science.abh dam failure. Landslides, 14(3), 821-831.
4455.abstract Xu, C., Xu, X., Yao, X., Dai, F., 2014. Three (nearly) complete
Song, C. Q., Sheng, Y. W., Ke, L. H., Nie, Y., and Wang, J. D. 2016. inventories of landslides triggered by the May 12, 2008 Wenchuan
Glacial lake evolution in the southeastern Tibetan Plateau and the Mw 7.9 earthquake of China and their spatial distribution statistical
cause of rapid expansion of proglacial lakes linked to glacial- analysis. Landslides 11, 441–461. https://doi.org/10.1007/s10346-
hydrogeomorphic processes, J. Hydrol 540, 504-514, 013-0404-6
https://doi.org/10.1016/j.jhydrol.2016.06.054. Yang, Kun, Hui Wu, Jun Qin, Changgui Lin, Wenjun Tang, and
Yingying Chen. 2014. 'Recent climate changes over the Tibetan
2021. New Insight into Post-seismic Landslide Evolution Processes Plateau and their impacts on energy and water cycle: A review',
in the Tropics. Front. Earth Sci. 9. Global and Planetary Change, 112: 79-91.
https://doi.org/10.3389/feart.2021.700546 Yang, W., Qi, W., Wang, M., Zhang, J., Zhang, Y., 2017. Spatial and
temporal analyses of post-seismic landslide changes near the
Lombardo, L., 2021. A closer look at factors governing landslide epicentre of the Wenchuan earthquake. Geomorphology 276, 8–15.
recovery time in post-seismic periods. Geomorphology 391. https://doi.org/10.1016/j.geomorph.2016.10.010
https://doi.org/10.1016/j.geomorph.2021.107912 Yang, W., Qi, W., Zhou, J., 2018. Decreased post-seismic landslides
Tian, Y., Owen, L.A., Xu, C., Ma, S., Li, K., Xu, X., Figueiredo, P.M., linked to vegetation recovery after the 2008 Wenchuan earthquake.
Kang, W., Guo, P., Wang, S., Liang, X., Maharjan, S.B., 2020. Ecol. Indic. 89, 438–444.
Landslide development within 3 years after the 2015 Mw 7.8 Gorkha https://doi.org/10.1016/j.ecolind.2017.12.006
earthquake, Nepal. Landslides 17, 1251–1267. Yao, T., Xue, Y., Chen, D., Chen, F., Thompson, L., Cui, P & Li, Q.
https://doi.org/10.1007/s10346-020-01366-x
van der Woerd, J., Owen, L. A., Tapponnier, P., Xiwei, X., Kervyn, F., melt and water cycle intensification and interactions between
Finkel, R. C., & Barnard, P. L. 2004. Giant, M8 earthquake- monsoon and environment: Multidisciplinary approach with
triggered ice avalanches in the eastern Kunlun Shan, northern Tibet: observations, modeling, and analysis. Bulletin of the American
Characteristics, nature and dynamics. Bulletin of the Geological Meteorological Society 100(3), 423-444.
Society of America, 116(3–4), 394–406. Yao, T.D., Thompson, L., Yang, W., Yu, W.S., Gao, Y., Guo, X.J., Yang,
https://doi.org/10.1130/B25317.1. X.X., Duan, K.Q., Zhao, H.B., Xu, B.Q., Pu, J.C., Lu, A.X., Xiang,
Veh, G., Korup, O., von Specht, S., Roessner, S. and Walz, A. 2019. Y., Kattel, D.B. and Joswiak, D. 2012. Different glacier status with
Unchanged frequency of moraine-dammed glacial lake outburst atmospheric circulations in Tibetan Plateau and surroundings.
floods in the Himalaya. Nature Climate Change 9(5), 379-383. Nature Climate Change 2(9), 663-667.
Veh, G., O. Korup, and A. Walz. 2020. 'Hazard from Himalayan glacier You, Y. 2001. Transporting characteristics of debris flow at Guxiang
lake outburst floods', Proceedings of the National Academy of accumulation fan in Tibet Autonomous Region. Bulletin of Soil and
Sciences of the United States of America, 117: 907-12. Water Conservation 21(2):28-30. (In Chinese)
Veh, G., O. Korup, S. von Specht, S. Roessner, and A. Walz. 2019. Yu, G.-A., Yao, W., Huang, H. Q., & Liu, Z. 2020. Debris flows
'Unchanged frequency of moraine-dammed glacial lake outburst originating in the mountain cryosphere under a changing climate: A
floods in the Himalaya', Nature Climate Change, 9: 379-+. review. Progress in Physical Geography: Earth and Environment
Walter, F., Amann, F., Kos, A., Kenner, R., Phillips, M., Preux, A., Huss, 0309133320961705. https://doi.org/10.1177/0309133320961705
M., Tognacca, C., Clinton, J., Diehl, T., Bonanomi, Y. 2020. Direct Yunus, A.P., Fan, X., Tang, X., Jie, D., Xu, Q., Huang, R., 2020.
observations of a three million cubic meter rock-slope collapse with
almost immediate initiation of ensuing debris flows. reveals the spatio-temporal evolution of post-seismic landsliding
Geomorphology, 351:106933.
Wang YC, Wu JS, Zhao HL. 1999. Chapter 5 Glacier-rainfall-type debris Environment (2020) 236, (S003442571930495X),
flow in the Peilong Valley of Sichuan-Tibet Highway. Research on (10.1016/j.rse.2019.111476)). Remote Sens. Environ 237.
typical mountain disasters of Sichuan-Tibet Highway. Chengdu: https://doi.org/10.1016/j.rse.2019.111541
University of Science and Technology Press 103-157. Zhang, G.Q., Yao, T.D., Xie, H.J., Yang, K., Zhu, L.P., Shum, C.K.,
Wang, J., Cui, Y., Choi, C.E. and Ng, C.W. 2018. The effect of climate Bolch, T., Yi, S., Allen, S., Jiang, L.G., Chen, W.F. and Ke, C.Q.
change on alpine mountain hazards chain: a case study in Tianmo 2020. Response of Tibetan Plateau lakes to climate change: Trends,
Ravine, Tibet, China, pp. 461-470, Springer. patterns, and mechanisms. Earth-Science Reviews 208, 1-22.
Wang, Q., X. J. Qiao, Q. Q. Lan, J. Freymueller, S. M. Yang, C. J. Xu, Zhang, N., He, Y., HE, X., Pang, H., and Zhao, J. 2007. The analysis of
Y. L. Yang, X. Z. You, K. Tan, and G. Chen. 2011. Rupture of deep ice fall at Mountain Yulong. Journal of Mountain Science 25(4):412-
418. (In Chinese)
593
Zhang, Taigang, Weicai Wang, Tanguang Gao, Baosheng An, and
Tandong Yao. 2022. 'An integrative method for identifying
potentially dangerous glacial lakes in the Himalayas', Science of the
Total Environment, 806.
Zheng, G. X., S. K. Allen, A. Bao, J. A. Ballesteros-Canovas, M. Huss,
G. Q. Zhang, J. L. Li, Y. Yuan, L. L. Jiang, T. Yu, W. F. Chen, and
M. Stoffel. 2021a. 'Increasing risk of glacial lake outburst floods
from future Third Pole deglaciation', Nature Climate Change 11:
411-+.
Zheng, Guoxiong, Anming Bao, Simon Allen, Juan Antonio Ballesteros-
Canovas, Ye Yuan, Guli Jiapaer, and Markus Stoffel. 2021b.
'Numerous unreported glacial lake outburst floods in the Third Pole
revealed by high-resolution satellite data and geomorphological
evidence', Science Bulletin 66: 1270-73.
Zhong, Q., Wang, L., Chen, S., Chen, Z., Shan, Y., Zhang, Q., et al. 2021.
Breaches of embankment and landslide dams - State of the art
review. Earth-Science Reviews 216, 103597.
https://www.sciencedirect.com/science/article/pii/S0012825221000
970
Zhou J W, Cui P, Hao M H. 2016. Comprehensive analyses of the
initiation and entrainment processes of the 2000 Yigong catastrophic
landslide in Tibet, China. Landslides 13(1), 39-54.
Zhou, T.-J., Sun, N., Zhang, W.-X., Chen, X.-L., Peng, D.-D., Li, D.-H.,
et al. 2018. When and how will the Millennium Silk Road witness
1.5 °C and 2 °C warmer worlds? Atmospheric and Oceanic Science
Letters 11(2), 180-188.
Zou, Q., Cui, P., Jiang, H., Wang, J., Li, C. and Zhou, B. 2020. Analysis
of regional river blocking by debris flows in response to climate
change. Sci. Total Environ 741, 140262.
Zou, Q., Cui, P., Zhang, G. and Wang, D. 2019. Multi-scale hazard
assessment of debris flows in eastern Qinghai-Tibet Plateau area,
Colorado School of Mines. Arthur Lakes Library.
594
Rock Mechanics
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Application of satellite technology (GPS and SAR) for monitoring ground surface
displacements in rock and geotechnical engineering
Application de la technologie satellitaire (GPS et SAR) pour la surveillance des déplacements de
la surface du sol en génie des roches et géotechnique
ABSTRACT: Monitoring is important for assessing the stability of the ground, confirming the validity of the design, predicting risks,
managing safe operations, and reducing project costs. The ideal monitoring system in Rock and Geotechnical Engineering should be
able to monitor the rock and soil behavior of small to extensive areas continuously and automatically with high accuracy and at a
low cost. Satellite technology (GPS and SAR) has the potential to realize such monitoring by combining it with conventional
geotechnical instruments. This paper firstly provides outlines of GPS and SAR as tools for displacement monitoring and then
describes the concept of “spatio-temporal continuous displacement monitoring”. Practical applications of GPS and SAR for
monitoring an unstable steep slope, two cases of landslides, land subsidence due to the extraction of groundwater, another subsidence
due to underground salt mining, and a rockfill dam during an earthquake are introduced. Some of the applications were conducted
as international collaborative research with Asian and Balkan countries. Finally, case studies of the assessment of the stability of
slopes based on the monitoring results are demonstrated.
RÉSUMÉ : La surveillance est importante pour évaluer la stabilité du sol, confirmer la validité de la conception, prévoir les risques,
gérer la sécurité des opérations et réduire les coûts du projet. Le système de surveillance idéal en génie des roches et géotechnique devrait
être capable de surveiller le comportement de la roche et du sol de zones petites à étendues en continu et automatiquement avec une
grande précision et à faible coût. La technologie satellitaire (GPS et SAR) a le potentiel de réaliser une telle surveillance en la combinant
avec des instruments géotechniques conventionnels. Cet article présente d'abord les grandes lignes du GPS et du SAR en tant qu'outils
de surveillance des déplacements, puis décrit le concept de “surveillance spatio-temporelle continue des déplacements”. Des applications
pratiques du GPS et du SAR pour surveiller une pente raide instable, deux cas de glissements de terrain, un affaissement de terrain dû à
l'extraction d'eaux souterraines, un autre affaissement dû à l'extraction souterraine de sel et un barrage en enrochement lors d'un
tremblement de terre sont présentés. Certaines des applications ont été menées dans le cadre d'une recherche collaborative internationale
avec des pays d'Asie et des Balkans. Enfin, des études de cas de l'évaluation de la stabilité des pentes sur la base des résultats de la
surveillance sont démontrées.
KEYWORDS: displacement monitoring, GPS, SAR, landslide, subsidence.
1 INTRODUCTION Practical applications of GPS and SAR for monitoring an
unstable steep slope, landslides, land subsidence, and a rockfill
Monitoring is important for assessing the stability of the ground dam during an earthquake are introduced. Some of the
and for confirming the validity of the design during the applications were conducted as international collaborative
construction and operation of structures. Monitoring is also research with Asian and Balkan countries. Finally, case studies
useful for predicting risks, for managing safe operations, and for of the assessment of the stability of slopes based on the
reducing project costs. monitoring results are demonstrated.
The ideal monitoring system for projects in Rock and
Geotechnical Engineering should be able to monitor the behavior
of small to extensive areas continuously and automatically with 2 SATELLITE TECHNOLOGY FOR DISPLACEMENT
high accuracy. In addition, the costs should be low, and the MONITORING
system should be reliable and easy to handle.
There are various types of instruments for taking
displacement measurements in Rock and Geotechnical 2.1 Displacement monitoring by GPS/GNSS
Engineering, such as extensometers, inclinometers, etc.
Although they are useful, these instruments may not be adequate GPS is a satellite-based positioning system that was developed
for monitoring large slopes or extensive areas because they can in the USA. It was established as a method for navigation and
only be applied to limited areas within dozens of meters or at long baseline surveys (Hoffmann-Wellenhof et al. 2001, Misra
most around 100 m. and Enge 2006). The advantage of GPS is that it can easily
On the other hand, satellite technology, GPS (Global provide three-dimensional displacements with mm accuracy over
Positioning System) and SAR (Synthetic Aperture Radar), is extensive areas. Recently, other satellite navigation systems,
capable of overcoming the above problems since it can be GLONASS, Galileo, BeiDou, QZSS, etc., have been developed;
applied to monitor the displacements of the ground and structures they are generally termed GNSS (Global Navigation Satellite
over extensive areas. System).
This paper firstly outlines GPS and SAR as tools for A GPS displacement monitoring system using the L1 signal
displacement monitoring. Then, the concept of “spatio-temporal was developed by the author and his colleagues (Iwasaki et al.
continuous displacement monitoring”, which uses satellite 2003, Masunari et al. 2003, Shimizu and Matsuda 2002). It was
technology and geotechnical instruments together, is described. then upgraded to a wireless data transfer system, as illustrated in
Figure 1. In this system, sensors are set on monitoring points and
597
a reference point, respectively. The data from the satellites are
received at each sensor and then transferred to a control box into
which a computer, data memory, and network device are
installed. The server computer, located in an office away from
the monitoring area, automatically controls the entire system to
acquire and analyze the data. Then, three-dimensional ALOS-2(JAXA) Sentinel-1(ESA)
displacements are obtained for all the monitoring points. The
Ascending
monitoring results are provided to users through the Internet in
real time. Users only need to access the home page to see the N
monitoring results. SAR SAR
The most important issue in the practical use of GPS is how
to improve the measurement accuracy. The author and his E
colleagues have proposed methods for removing errors and for
estimating the real values of the measurements. Those methods Descending
have succeeded in providing measurement results that are a few
times higher (i.e., mm) in accuracy than the standard GPS Figure 2. SAR satellites and their orbits.
(Shimizu et al. 2011, Shimizu and Nakashima 2017). The
procedure has been approved as “the ISRM suggested method for 2.3 Comparison of GPS and DInSAR
monitoring rock displacements using the Global Positioning
System” (Shimizu et al. 2014). The features of GPS and DInSAR are compared in Table 1 and
Figure 3. GPS can continuously monitor three-dimensional
displacements at certain points 24 hours a day with mm accuracy,
whereas DInSAR can take one-dimensional displacement
measurements of much greater areas usually once every few
Sensor days/weeks with cm accuracy in spatial resolution areas of 3-30
m. DInSAR does not require any sensors on the ground, while
GPS requires a sensor at each monitoring point. Therefore, GPS
and DInSAR are complementary to each other.
Table 1. Features of GPS and DInSAR.
GPS DInSAR
Required devices Receivers Not necessary
for user
Control box Entire area
Observable Point(s) (3-30 m spatial
displacements
resolution)
Server
Every hour or
Continuous shorter periods Periodic: a week ~
monitoring (available 24 hours a several months
day)
Figure 1. GPS fully automatic continuous displacement monitoring Dimension of 3-dimensional 1-dimensional
system. measurements
Accuracy mm level cm level
Boldface: advantages
2.2 Displacement monitoring by DInSAR
SAR is a radar device mounted on an aircraft or artificial satellite. (a) (b)
It is one type of active remote sensing that uses microwave GPS/GNSS
GPS/GNSS
SAR
SAR
radiation, enabling observations 24 hours a day under all weather
©JAXA
@JAXA
conditions (Hanssen 2002). Nowadays, there are several SAR
satellites orbiting the Earth, i.e., COSMO-SkyMed (ASI: Italian
Space Agency), TerraSAR-X (DLR: German Aerospace Center), Slope monitoring
RADARSAT-2 (CSA: Canadian Space Agency), Sentinel-1
(ESA: European Space Agency), ALOS-2 (JAXA: Japan
Aerospace Exploration Agency), etc. (Figure 2).
Interferometric SAR (InSAR) is a method for taking the
signal phase difference (interference) from two SAR data images Google Earth
which are observed in the same area at different periods by the
same SAR satellite in the same orbit. GPS satellite GPS satellite
SAR satellite
Differential Interferometric SAR (DInSAR) is the term for the Displacement
method of producing interferograms from which the influence of (Three-dimensional)
topography and the baseline length of the satellite positions has
been removed (Ferretti et al. 2007). The advantage of DInSAR is
that it can provide the displacement distribution of the ground
surface over a vast area without the use of any sensors on the
ground (Ferretti 2014). The observed displacement is one- Displacement
dimensional along the Line of Sight (LOS), which is the direction Surface
Sensor
Surface (One-dimensional)
from the satellite to the observed Earth’s surface (Figure 2). The Sensor (Monitoring point)
(Reference point)
accuracy is around cm level. When temporal continuous
displacement monitoring is conducted by DInSAR, SBAS (Small Figure 3. GPS and DInSAR, and their monitored displacements: (a)
Baseline Subset) DInSAR (Berardino et al. 2002) and PS InSAR GPS and (b) InSAR (photo: Google Earth).
(Ferretti et al. 2001) are often used.
598
3 SPATIO-TEMPORAL CONTINUOUS DISPLACEMENT 4 CASE STUDIES OF DISPLACEMENT MONITORING
MONITORING
This section introduces case studies of displacement monitoring
Figure 4(a) shows a schematic diagram of the relationship using satellite technology (GPS and SAR).
between the measurement accuracy and the size of a monitoring
area expressed by the representative length. The applicable 4.1 Unstable steep slope along traffic road
ranges of the geotechnical instruments and conventional
surveying methods are illustrated in this figure. It can be seen The monitoring site is a steep slope, with a height of about 100
that there is a gap in accuracy between the two methods. GPS can m, along a national road in Japan (see Figure 5). The slope is
cover this gap. In addition, DInSAR is able to expand the validity composed primarily of rhyolite and granite which formed in the
of displacement monitoring to huge areas. Cretaceous period of the Mesozoic era, and its surface is partially
Looking at the relative accuracies of these methods, they are covered with a colluvial deposit. Vegetation extends over the
almost the same, in the order of about 10 -6, i.e., 0.1 mm slope and creates an obstruction above the antennas of the GPS
(accuracy)/100 m (baseline length) for geotechnical instruments, sensors.
1 mm/1 km for GPS, and 10 mm/10 km for DInSAR. This means In July 1972, the slope collapsed in large areas and deposits
that displacements could be measured with almost the same flowed, causing damage to a railway and a national road. The
accuracy as 10-6 over small to large areas by applying both railway track was later replaced by a new one with a tunnel
geotechnical instruments and satellite technology (GPS and running through the mountain. A check dam was then
DInSAR). Spatially continuous (spatio-continuous) monitoring constructed to collect deposits from the upper part of the slope,
can be realized. and a rock shed was built above the road to protect passing
On the other hand, Figure 4(b) shows a diagram of the vehicles and people. In 2010, a failure with a volume of 1600 m3
relationship between the measurement interval (period) and the was recorded due to heavy rainfall. Other small surface collapses
size of a monitoring area. Geotechnical instruments can measure and runoffs occurred in 2011 after heavy rainfall.
displacements at any interval. In the case of GPS, the interval of GPS sensors were set at reference point K1 and monitoring
the measurements is usually an hour when the baseline length is points G1 and G2 in 2012. Since the displacements at G1
within 1 km. Recently, however, the measurement interval can increased significantly, due to heavy rainfall events, and surface
be a second or less if accuracy of a few cm is acceptable. collapses occurred just beneath G1, the sensor at G1 was
In the case of DInSAR, the interval of the measurements removed in August 2014. In the same period, two new
depends on the satellite regression cycle. It is usually several monitoring points, G3 and G4, were set on the right-hand side
days or a few weeks or more. Although there may be a limitation (southwest side) of the slope (Figure 5). All the sensors were
in the measurement interval for monitoring extensive areas at firmly fixed at the monitoring points on the ground surface by a
present, there are plans to launch several SAR satellites which tripod with anchors (Figure 6). The height differences between
will improve this issue. Therefore, temporally continuous monitoring points G1, G2, G3, and G4 and reference point K1
monitoring can be done. Consequently, spatio-temporal were 103 m, 112 m, 103 m, and 94 m, respectively. The upper
continuous displacement monitoring can be realized by row of Figure 6 shows sky views above the sensors. It was found
combining the two types of satellite technology with that the vegetation and the slope surface behind the sensors
geotechnical instruments. partially cover the sky area from the northeast to southwest. The
system automatically provides the three-dimensional
(a) displacement every hour at each point.
Accuracy
10cm
DInSAR
1cm Surveying Methods
GPS/GNSS mm/1km=10-6
1cm/10km=10-6
1mm
Geotechnical 0.1mm/100m=10-6
instruments
0.1mm
10m 100m 1km 10km 100km
Baseline Length (size of monitoring area)
(b)
Interval
Surveying methods DInSAR
106sec
(11.5day)
104sec GPS/GNSS
(3hr)
Geotechnical GPS/GNSS
102sec
(1.7min)
instruments
10m 100m 1km 10km 100km
Baseline length (size of monitoring area)
Figure 4. Spatio-temporal continuous monitoring using geotechnical
instruments and satellite technology: (a) measurement accuracy and
baseline length, and (b) measurement interval and baseline length.
Figure 5. Monitoring site and locations of GPS sensors (dates below
GPS points indicate monitoring periods).
599
(a) (b) (c) (d) (e) August 2014 to December 2018. During this period, the
N N N
displacement at G3 headed to the northeast, which is the direction
N N
W
90o 60o 30o 0o
E W
90o 60o 30o 0o
E W
90o 60o 30o 0o
E W
90o 60o 30o 0o
E W
90o 60o 30o 0o
E of the collapsed portion up to 9 May 2016, after which it turned
to the northwest (downward movement). The displacement at G4
constantly increased in the inclination direction of the slope,
S S S S S
while the displacement at G2 subsided. Therefore, this portion,
including monitoring points G2, G3, and G4, seems to show a
rotational slide.
(a)
(a) (b) (c) (d) (e)
Figure 6. GPS sensors with sky view above sensors: (a) reference
point K1, (b) monitoring point G1, (c) G2, (d) G3, and (e) G4.
GPS measurement results generally include random errors
and bias errors. Random errors are brought about by random
fluctuations in the measurements. These errors are caused by (b)
receiver noise, the operating limit of the receiver (resolution),
and other accidental errors. Typical bias errors, on the other hand,
are tropospheric delays and other signal disturbances caused by
obstructions above the antennas. Since both random and bias
errors influence the monitoring quality, they should be corrected
by appropriate methods which can reduce or eliminate them.
Measurement value y is assumed to be composed of exact (c)
value u0, bias errors P and T, and random error R, as shown in
equation (1).
y = u0 + p + T + R (1)
where P and T are signal disturbances caused by obstructions
above the antennas and tropospheric delays affecting the signal Figure 7. Time-transition of three-dimensional displacement: (a)
traveling through the troposphere from the satellite, respectively. latitude, (b) longitude, and (c) height (Sato et al. 2021).
In the ISRM suggested method, the author proposed that
signal disturbances due to obstructions above the antennas can Table 2. Annual average displacement at monitoring points
be reduced by the mask procedure, and that errors due to (mm/year).
tropospheric delays can be corrected by a tropospheric model Year 2014 2015 2016 2017 2018 2019 2020 2021
using meteorological data (Shimizu et al. 2014). After removing G2 19 7 6 5 3 5 4 6
the bias errors from the measurement value, the “trend model” is
G3 40 18 18 18 24 12 6 5
applied to estimate the exact displacement (Shimizu et al. 2014).
The monitoring results from March 2013 to August 2014 G4 24 9 6 7 3 2 2 6
were already shown and discussed in previous papers (Furuyama
et al. 2014, Kien et al. 2017). They are summarized as follows.
The displacement at G-1 increased at a rate of 2-3 mm/month in
the beginning, and then gradually increased every time it rained.
Finally, it reached 355 mm in the north direction and 234 mm in
the west direction, and the settlement reached 137 mm. The
displacement velocity exceeded the safety criteria (Level 2: 20
mm/day) two times during the rainy season, and the road was
closed at those times for the sake of traffic safety. It was proven
that the GPS displacement monitoring system was useful and
effective for assessing slope stability and managing road safety.
Figure 7 shows the monitoring displacements at G2, G3, and
G4 from 12 August 2014 to 1 May 2021 together with the
amounts of rainfall (mm/day). The standard deviations in the
measured displacements in the directions of latitude, longitude,
and height are 1.6 mm, 1.4 mm, and 3.6 mm at G2, 1.5 mm, 1.6
mm, and 3.7 mm at G3, and 1.8 mm, 1.1 mm, and 3.2 mm at G4,
respectively. Continuous three-dimensional displacement
monitoring was performed for almost seven years without
missing any data, except during the maintenance period (Sato et
al. 2021).
Table 2 shows the annual average displacement at all the
monitoring points, G2, G3, and G4. The value is seen to
gradually decrease within several mm per year, although there
have been heavy rainfall events with more than 100 mm/day
several times since 2014. The decrease might have been caused
by the effect of horizontal drainage holes, drilled in 2014 in the
upper area around G4, on the stability of the slope.
Figure 8 shows the transition of the displacement vectors in
the plan view and in the vertical sections of the slope from
600
(a) (a)
(b) (b)
(c) (c)
Figure 8. Transition of displacement vectors August 2014 to Figure 9. Transition of displacement vectors January 2019 to May
December 2018: (a) plan view, (b) vertical section (B-B’), and (c) 2021: (a) plan view, (b) vertical section (B-B’), and (c) vertical
vertical section (C-C’) (Sato et al. 2021). section (C-C’) (Sato et al. 2021).
On the other hand, the slope behavior has changed since SBAS-DInSAR using ALOS-2 data was applied to this slope
January 2019, as shown in Figure 9. The displacements at G3 and (Nagasaki et al. 2021). All the SAR data processing in this paper
G4 have been heading in the same direction as that at G3 since 9 was conducted using Envi-SARScape 5.5.3.
May 2016. In addition, G2 has also turned and is moving in the Figures 10(a) and (b) show the spatial distributions of the
same direction as that at G3 and G4 in the plan view and as that LOS displacements obtained using the ascending and descending
at G3 in the vertical section. This portion seems to show ALOS-2 data for about 5 years from 2014/15 to 2021 (16
translational movement along the boundary between the rhyolite ascending and 26 descending images) in the target area,
(surface rock) and the granite (bedrock). respectively. The horizontal displacement vectors obtained by
As seen above, GPS displacement monitoring system is GPS during the same period are also drawn in Figure 10. The
useful for detecting the detailed three-dimensional displacements resolution of the ground surface (unit of pixels) is 10 m x 10 m.
of a slope. However, the slope behavior cannot be found except The LOS displacements are represented by the range in color
at monitoring points set with a GPS sensor. SAR can find the from red
surface displacement distribution, and the combination of SAR The red and blue mean negative (extension) and positive
and GPS could be one of desirable methods for monitoring the (compression) LOS displacements (Figure 3(b)), respectively.
behavior of an entire slope. In the area with low coherence, below the threshold value
(0.3 is employed in this study), the results of the LOS
displacement have been removed from the map. The coherence
601
is an index, taken as 0 to 1, to represent whether a good (a) (b)
interferogram has been produced. A higher value indicates a
better condition for obtaining reliable results. Results are not
shown in some of the middle portion of the slope due to low LOS LOS
coherence in the descending results (Figure 10(b)). (Ascending) Up Up (Descending)
Descending
(a) (b) azimuth
Ascending
azimuth North East
North
East
(West) (West) (South)
(South)
Figure 11. Geometry of three-dimensional displacements projected
into LOS displacements in ascending and descending right-looking
directions: (a) ascending case and (b) descending case.
(a)
Figure 10. LOS displacement distribution by DInSAR and (b)
displacement vectors by GPS: (a) ascending results (Nov. 6, 2014 to
Mar. 4, 2021) and (b) descending results (Feb. 9, 2015 to Mar. 1,
2021).
In order to investigate the validity of the results of DInSAR,
the LOS displacements are compared with ones measured by the (c)
GPS displacement monitoring system. Since the GPS monitoring
system provides three-dimensional displacements, they must be
converted to the LOS direction by the following equations:
(2) Figure 12. Time-transition of LOS displacements by SBAS-DInSAR
(ascending) and GPS: (a) G2, (b) G3, and (c) G4.
(a)
(3)
where and are the ascending and descending
LOS displacements, respectively, and are the incidence (b)
angles of the microwaves transmitted from the satellite on the
ascending and descending orbits, respectively, and and
are the angles of the ascending and descending satellite flight
directions, respectively, as shown in Figure 11. , ,
and are the displacement components of the longitude,
latitude, and height direction, respectively. (c)
Figures 12 and 13 present comparisons between the
displacements obtained by DInSAR and GPS at G-2, G-3, and
G-4 in the ascending and descending LOS directions,
respectively. The discrepancies between the two sets of
displacements are within 10 mm and 20 mm for the ascending
and descending results, respectively. Figure 14 shows the
Figure 13. Time-transition of LOS displacements by SBAS-DInSAR
correlation between DInSAR and GPS at all the GPS monitoring (descending) and GPS: (a) G2., (b) G3, and (c) G4.
points, namely, G-2, G-3, and G-4. Although a few results are
largely scattered, it is found that the DInSAR results agree well
with the GPS results. This means that DInSAR can be applied for
monitoring displacements with cm accuracy even at steep slopes
in areas with dense vegetation.
602
0
SBAS-DInSAR(mm)
-40
-80
Ascending
Descending
-120
-120 -80 -40 0
GPS(mm)
Figure 14. Correlation of LOS displacements obtained by DInSAR
and GPS.
4.2 Landslides in Vipava River Valley, Slovenia
The Vipava River Valley, located in the southwest part of
Slovenia, is well known as a landslide-prone area (see Figure 15).
There are four remarkable landslides in this area (Bizjak and
Zupan i 2009, Jemec Aufli et al. 2017, Verbovšek et al. 2018),
namely, Rebrnice, Stogovce, Slano blato, and Selo. The area
extends 40 km in length and several km in width. The landslides
are characterized by the various volumes and velocities of their
movement. The occurrence of debris flows and landslides
sometimes causes damage to residences, infrastructures,
farmland, etc.
It is vital that the current situation of these landslides be
known and that better mitigation plans be designed by
monitoring the landslide behavior and conducting geological and
geotechnical studies. An effective monitoring method that is
capable of covering extensive areas is needed. SBAS-DInSAR
has been applied to this area (Yastika et al. 2019b).
Complex landslides, which have been occurring from the
Pleistocene to recent times, are related to the geological structure
- Mesozoic carbonates have been thrust over folded and
tectonically fractured Tertiary siliciclastic flysch (Jemec Aufli Figure 15. Vipava River Valley and surrounding area: (a) study area
et al. 2017). Such overthrusting has caused steep slopes and the and (b) geomorphological map of study area (Yastika et al. 2019b).
fracturing of rocks, producing intensely weathered carbonates The “Slano blato landslide area” is taken to see the detailed
and large amounts of carbonate scree deposits. transition of the LOS displacements from Figure 16. Figure 17
The differences in elevation here are significant and range shows LOS displacement distribution maps and the temporal
from 100 m at the valley bottom to over 1200 m on the high transition at several points in the Slano blato landslide area. The
karstic plateau. The combination of unfavorable geological spatial distribution of the surface displacements is clearly visible.
conditions and periods of intense short or prolonged rainfall has Several deep wells were constructed at the top of the slope as a
led to the formation of different types of complex landslides. The countermeasure to the landslide.
blue polygon in Figure 15(a) shows the study area.
This study used 134 and 139 Sentinel-1 SAR images in the In Figures 17(a) and (b), the temporal transitions of the LOS
ascending and descending passes, respectively, observed from displacements at W6, N20, and UD2 were taken from the upper,
September 2016 to January 2019 by Sentinel-1 satellites. Figures middle, and lower portions of the slope, respectively. Figure 18
16(a) and (b) show the LOS displacement distributions in the illustrates the geometrical relationship between the actual three-
ascending and descending passes, respectively. The magenta dimensional displacement vector and the LOS (ascending and
color on the maps represents areas of geometrical distortion in descending) displacements.
the SAR data, such as radar layover or shadowing. Some areas
had no results (no color) due to low coherence.
603
The results in Figure 17 indicate that the displacement of the
Slano blato landslide area seems to still be moving downward at
(a) the top and the toe, and the middle portion is moving slightly
toward the front of the slope at a rate of at least 50 mm/year.
(a)
Ascending SAR Descending SAR
Descending LOS
Actual
displacement
displacement
Actual
Descending LOS displacement
Ascending LOS
displacement displacement
Ascending LOS
displacement
(b)
Actual
displacement
Descending
(b) displacement
Ascending
displacement
Figure 18 Schematic diagram of geometrical relationship between
actual displacement vector and LOS (ascending and descending)
displacements: (a) W6 and UD2, and (b) N20.
4.3 Landslides in northern Black Sea coast of Bulgaria
The Black Sea coast of Bulgaria is well known as an active
landslide area across which various landslides extend (Bruchev
et al. 2007). Monitoring the real behavior of landslides has not
often been conducted due to technical difficulties and cost issues,
although it is essential to reveal the mechanisms of these
landslides and to evaluate the risks they cause. SBAS-DInSAR
was applied to monitor the landslide displacements along the
northern Black Sea coast (Yamaguchi et al. 2021). The results
Figure 16. LOS displacement distribution maps in Vipava River are compared with the actual ground behavior described in the
Valley from 2016 to 2019: (a) results in ascending pass and (b) results official annual reports of landslides in Bulgaria
in descending pass (Yastika et al. 2019b).
2017, 2018,
(a) (b) 2019).
Ascending Figure 19(a) shows the distribution of landslide locations at
W6 W6Descending Balchik and Kavarna municipality along the Black Sea coast in
northern Bulgaria (
). In this region, there are four landslide
hotspots, namely, Topola village, Alley Echo (EXO), Sviloza,
N20 N20 and Fish-Fish, which are represented by the squared regions in
Figure 19(a).
Figure 19(b) is a geological map with the landslides in this
area (Evstatiev et al. 2013). The regions denoted by “V” and “IV”
UD2 UD2
along the coast represent the target areas including Topola village
and Fish-Fish, respectively. These regions are composed of
limestone (kvN1s), aragonite clay with limestone interaction
(toN1s), and diatomaceous clay (evN1kg-s). The northern part
50 50
from the crown line of the main scarp comprises limestone of the
N20 Kavarna Formation (kvN1s), and the more northern area
UD2
0
UD2 0 comprises loess (yellow soil) complex (eoIQp). North of the
W6
LOS Disp. (mm)
LOS Disp. (mm)
crown line is a limestone plateau called the Lower Romanian
-50
W6 -50
Level (LRL).
Ascending Descending
-100 -100
The Sentinel-1 data are applied to create the interferogram for
W6
UD2 N20 W6
UD2 the SBAS-DInSAR analysis. Two datasets, namely, 273 scenes
N20 N20
-150
06/2016 06/2017 06/2018
ASCENDING
06/2019
-150
06/2016 06/2017 06/2018
DESCENDING
06/2019
on the descending pass observed from 8 October 2014 to 9 May
Date Date 2020, and 196 scenes on the ascending pass observed from 9
Figure 17. LOS displacement distribution maps and temporal October 2014 to 7 September 2020, are used.
transition at selected points in Slano blato landslide area: (a) results Figure 20 shows the LOS displacement distributions of the
in ascending pass and (b) results in descending pass (Yastika et al.
2019b).
target area of the Balchik and Kavarna (Figure 19(a)) from 2014
to 2020 for the descending and ascending passes. The
604
displacements are represented by the range in color from red to sediments and a sloping formation are visible. The scarp may
The large have been affected by sea erosion.
LOS displacements, shown in red and blue, are found at the The border of the landslide area, shown in Figure 22(a), is
eastern and western ends of Figure 20. drawn on the displacement map in Figure 21. The area with the
large displacement observed by DInSAR coincides with the
(a) actual landslide area (Figure 21).
Balchik Kavarna (a)
Bulgaria Municipality Municipality
Golf
course
EXO Topola village Massive
Meeting point buildings
& Sviloza
Fish-Fish &
Western Karamanlii Black Sea
Legend:
Stabilized landslide
Dobrich Province Potential landslide
Active landslide
Varna Province
(b)
(b)
eoIQp kvN1s
VI
V
toN1s IV
kvN1s
III
evN1kg-s aQh Black Sea
eoIQp Loess evN1kg-s Diatomaceous
II complex clays
kvN1s Limestones Delapsium
Aragonite
s clays with Oldest
toN1 limestone landslide
I interaction scarp
odN1s Limestones
with sand
interactions
frN1s Sands Figure
- : (a) descending pass (8 Oct 2014 to 5 May
Figure 19. Study area along coastal area in northern Black Sea, 2020) and
Bulgaria: (a) m
and (b) geological map
(Evstatiev et al. 2013). (a)
28 16’30’’E
To evaluate the results of SBAS-DInSAR and to discuss its
applicability for monitoring the landslide behavior, the DInSAR P_I7
43 25’0’’N
P_I6
results are compared with the actual behavior observed by field P_I5 P29
P_II7
investigations P_II4 P_II5
P_II6
P30
. In this paper, only P31
Topola village (the eastern end of the area in Figure 20) is taken P27
P28
P33
for comparison as an example.
Figure 21 is an enlarged view showing the LOS displacement
43 24’30’’N
distribution in the descending and ascending cases at the eastern
end of Topola village. DInSAR detected large displacements in
the eastern portion; they are denoted in red, orange, and blue.
According to the annual reports (b)
28 16’30’’E
2017, 2018,
2019), a landslide developed in the coastal area near a golf course
around the eastern end of Topola village. Large deformation has P_I7
43 25’0’’N
P_I5
P_I6
P29
been seen among the massive buildings located inside of the P_II6
P_II7
landslide area (Figure 22(a)). For example, one building is P_II4 P_II5 P30
inclined backward and the front of it is visibly raised, while P27
P27
P31
others have moved forward and are twisted. P28
P33
This landslide was registered in 2014 as No. DOB 17.05009-
01-03 in Bulgaria (
43 24’30’’N
). The landslide has cut off the road to
the golf course, and transverse ridges have appeared, as shown in
Figure 22(b). There are terraces and cliffs at the landslide toe up
to a height of about 3.0 to 4.0 m along the coastline. On the Figure maps around eastern end of
and
western side, near the coastal cliff, a scarp with preserved
605
(a) (a)
200
LOS Disp. (mm)
0
Golf course
P29_Asc
P29_Dsc
-200
2014-06 2015-06 2016-06 2017-06 2018-06 2019-06 2020-06
Date
(b)
200
Small landslide indentation
Boundary of landslide DOB 17.05009-01-03 - tentatively
LOS Disp. (mm)
(b)
0
P30_Asc
P30_Dsc
-200
2014-06 2015-06 2016-06 2017-06 2018-06 2019-06 2020-06
Date
(c)
200
Figure 22. Landslide around eastern end of Topola village: (a)
boundary of landslide, and (b) damaged buildings and transverse
ridges (red line)
LOS Disp. (mm)
. 0
Figure 23 shows the time-transition of the LOS displacements
P31_Asc
in both ascending and descending cases at four points (pixels) of
P31_Dsc
interest, P29, P30, P31, and P33, as shown in Figure 21. -200
Point P29, on the plateau above the slope, is stable because 2014-06 2015-06 2016-06 2017-06 2018-06 2019-06 2020-06
Date
almost no displacement appears, as shown in Figure 23(a). (d)
At P30, located in the area of the massive buildings (Figure 200
22), both descending and ascending LOS displacements are seen
to significantly increase toward the negative side (extension)
over time (Figure 23(b)). Such behavior indicates that the actual
LOS Disp. (mm)
displacement is dominantly subsidence. The eastern component 0
of the displacement is estimated to be small.
At P31, in the middle of the slope, the descending LOS
displacement is almost zero or slightly positive, while the P33_Asc
ascending one is negative (Figure 23(c)). This indicates that the P33_Dsc
-200
actual displacement dominantly moves to the east and 2014-06 2015-06 2016-06 2017-06 2018-06 2019-06 2020-06
downward. Date
At P33, near the coastal area, both descending and ascending Figure 23 Time transition of LOS displacements by ascending and
LOS displacements are compression, and the descending one is
larger than the ascending one (Figure 23(d)). Therefore, the
actual displacement seems to move upward and slightly to the
east at the landslide toe. P29
From the above results, it is found that this area shows typical
landslide behavior along the vertical section of P29-P30-P31- Stable
P33, as shown in Figure 24. The SBAS-DInSAR results P30
correspond to the actual behavior described in the official reports P31 Uplift
, and they verify its Sliding P33
applicability for monitoring landslides in extensive areas along Subsidence
the northern Black Sea coast.
Figure 24. Schematic illustration of estimated displacement
directions at selected points.
606
4.4 Subsidence induced by underground salt mining in Tuzla,
Bosnia and Herzegovina
The subsidence in Tuzla has been creating a large hazard for a
long period, mainly since the 1950s. It was reported that the main
factor of this subsidence is the underground salt mining activities
(Mancini et al. 2009a).
The subsidence was measured by traditional topographic
surveys from 1956 to 2003. The GPS surveys produced
subsidence information in three periods, namely, 2004 to 2005,
2005 to 2006, and 2006 to 2007 (Mancini et al. 2009b). DInSAR
was applied to know the present situation and to update the
subsidence data by the author and his colleagues (Parwata et al.
2020a, 2020b).
The Tuzla salt deposit is located beneath Tuzla City in an area
of approximately 2 km2 (Figure 25). It consists of five separate
salt series embedded in syncline with one of the wings close to
the surface of the city's center. The maximum thickness of the
salt formation is 600 meters (Figure 25(b)). The lithological
composition is halite, gypsum, anhydrite, laminated and thin-
layered marls, tuff, poriferous limestone, etc.
The total amount of subsidence during the period of 1956 to
Figure 26. Subsidence in Tuzla measured by traditional topographic
2003 was up to 12 m, while it was more than 1 m in the area of
surveys: (a) contour lines of subsidence in 1-meter intervals (1956 to
about 2 km2 of the city (Figure 26) (Mancini et al. 2009a). The 2003), and (b) annual average subsidence and amount of extracted
official date of termination of the salt deposit exploitation was in saltwater (1956 to 1992) (Mancini et al. 2009a).
May 2007. Up to then, salt deposits were being developed in two
separate mining fields, i.e., underground salt mining (excavation) For the SBAS-DInSAR analysis, 145 SAR data images
in Tušanj (north-west portion of the shaded area in Figure 25(a)) observed by Sentinel-1 from October 2014 to July 2018 in the
and borehole salt mining (extraction of salt water after injecting descending orbit were used.
water into the underground through boreholes) in Hukalo– Figures 27(a)-(d) show the distribution of subsidence
Trnovac (south-east portion of the shaded area in Figure 25(a)). obtained by DInSAR. The maps show that the incremental
However, subsidence still continued at a rate of 10-20 cm/year subsidence started on October 9, 2014. The subsidence area is
from 2004 to 2007 according to GPS surveys using the static indicated in yellow and red (negative values). The location of the
method (Mancini et al. 2009b). large subsidence is almost the same as that obtained by GPS (see
Figures 27(c) and (d)).
Comparisons of the LOS displacements by GNSS and
DInSAR at Tušanj and Pannonica stations are shown in
Figure28. GNSS monitoring was conducted by the local
municipality. Since the GNSS results were greatly scattered, very
few results could be utilized. The two sets of displacements agree
well.
Figure 29 illustrates the relationship between the DInSAR
results and the GNSS results at Tušanj and Pannonica stations. It
is seen that the DInSAR results coincide with the GNSS results
with a discrepancy of less than 10 mm. The RMSE (root mean
square error) is 4.3 mm for Pannonica station, 1.6 mm for Tušanj
station, and 3.9 mm totally for both stations. These results are
similar to, or better than, the ones seen in Figure 14.
Figure 28 also presents the LOS displacement by DInSAR at
Tuzla station, which is located far from the subsidence area. The
displacement at this station did not increase much over the
monitoring period and the accumulated value is less than a few
mm. This means that no remarkable displacement occurred at
this point and that Tuzla station was suitable for the reference
point of the GNSS monitoring.
Figure 25 Geological condition of Tuzla: (a) plan view (modified
from Mancini et al. (2009a)) and (b) vertical X-X’ cross section
(Parwata et al. 2020a).
607
(a)
(b)
Figure 29. Scatter plot of LOS displacements obtained by SBAS-
DInSAR and GNSS at (a) Pannonica and (b) Tušanj (Parwata et al.
2020a).
From the above results, it is found that the DInSAR results are
acceptable for monitoring the subsidence in this area.
The temporal transitions of the subsidence at several selected
July 2, 2018 (c)
points, P-5, P-7, …, P-38 (Figure 30(a)) are shown in Figure
29(b). The subsidence rate at P-38 is the smallest among these
points, -1 mm/y, while the rate at P-16 is the largest among them,
-36 mm/y (Parwata et al. 2020b).
Figure 31 shows the subsidence trends from 2004 to 2019 at
P-10, P-16, P-25, and P-38. The results from 2004 to 2007 were
obtained by GPS (Mancini et al. 2009b) and those from 2014 to
2019 were obtained by DInSAR. From these figures, the changes
in the rate over the last 15 years can be found.
(d)
Figure 27. Transition of subsidence distribution obtained by SBAS-
DInSAR: (a) Oct. 21, 2014, (b) Oct. 22, 2016, (c) July 2, 2018, and
(d) contour of subsidence obtained by GPS (2006 to 2007) (Parwata
et al. 2020a).
Figure 30. Transition of subsidence: (a) distribution map and locations
of selected points, and (b) temporal transition of subsidence at
selected points (Oct. 9, 2014 to May 28, 2019) (Parwata et al. 2020b).
Figure 28. Comparison between DInSAR and GPS results (Parwata
et al. 2020a). The subsidence rates at P-25 and P-38 were -20 mm/y and -
47 mm/y, respectively, from 2004 to 2007, and recently the rates
decreased to -4 mm/y and -1 mm/y, respectively. The subsidence
has almost come to an end at P-38. The rate at P-10 from 2004 to
2007 was -73 mm/y, and it has decreased considerably to -10
mm/y. It is anticipated that the subsidence around this point will
be stable in the near future. On the other hand, although the rate
608
of subsidence at P-16 has decreased from -133 mm/y, the present
rate is -36 mm/y, which is still not small. Therefore, subsidence (a)
monitoring should be carefully continued for the whole area of
the city.
(GPS: Mancini et al. 2009b)
Semarang
9km
(b)
Figure 32. Monitoring site: Semarang: (a) study area (photo: Google
Earth) and (b) damage in city area due to subsidence (photos by the
author and Yastika).
To monitor the subsidence, 71 SAR images were employed.
The data were classified into three groups acquired by three
different satellites, namely, Envisat-ASAR (ESA), ALOS-1
PALSAR-1(JAXA), and Sentinel-1A. The first (oldest) data
provided by Envisat were taken in March 2003, while the last
were taken in October 2007. ALOS-1 provided data from
January 2007 to February 2011. Then, Sentinel-1A provided data
from the end of 2014 to the beginning of 2015. No SAR data are
available for the period of 2011 to 2014.
The total subsidence distribution maps, generated by each
satellite, were superimposed onto a geological map (Figure
Figure 31. Changes in subsidence transition from 2004 to 2019 at P- 33(a)), as shown in Figures 33(b), (c), and (d). Generally, the
16, P-10, P-25, and P-38 (Parwata et al. 2020b). spatial distribution pattern of the subsidence in Semarang does
not change much. These maps reveal that large subsidence has
occurred in the area of alluvium. The area where the subsidence
4.5 Long-term land subsidence in Semarang, Indonesia transits from large to small clearly matches the boundary
between the Damar Formation (sedimentary rock) and the
Land subsidence is a critical issue to be addressed in large cities alluvium.
around the world. Semarang, Indonesia, is one of those cities due Alluvium has the potential to subside due to consolidation. On
to the extraction of groundwater. The subsidence in Semarang the other hand, the Damar Formation is firm. This means that the
has led to serious and costly damage to urban infrastructures, geological conditions influence the different effects on the
such as houses, buildings, roads, railway tracks, etc. (Abidin et subsidence process in Semarang.
al. 2013, Marfai et al. 2008).
The monitoring of land subsidence is a vital task in predicting
and managing disasters that might occur. An economic
monitoring method, that can be applied quickly and
continuously, and produces accurate measurements over large
areas, is necessary. SBAS-DInSAR has the potential to be a
powerful technique for the above task, and it has been applied for
long-term continuous subsidence monitoring (Yastika et al.
2019a).
Semarang City (Figure 32) is located in the north of Java
Island, Indonesia. Geologically, Semarang is underlain by three
lithological units: volcanic rocks of the Damar Formation in the
southwest, sedimentary marine rocks in the north, and alluvial
deposits, also in the north. The volcanic rocks consist of volcanic
breccias, lava flow material, tuff, sandstone, and claystone
(Figure 33(a)).
609
(a) (b)
(a) (b)
DInSAR (cm)
DInSAR (cm)
Figure 35. Comparison of DInSAR and GPS results at GPS
measurement points: (a) 2008 to 2009, and (b) 2009 to 2010 (Yastika
(c) (d) et al. 2019a).
Figure 36(a) shows the long-term transition of subsidence at
a point (RMPA) located in the alluvium at the seaside as an
example. The subsidence was determined using the hyperbolic
function (Tan et al. 1991) to connect the results obtained by three
different SAR data in different periods (Yastika et al. 2019a). It
shows a good agreement with the results of GPS represented by
bullets (Figure 36(a)).
Figure 36(b) shows the number of registered deep-water wells
and the total extraction in the Semarang-Demak groundwater
basin (Putranto and Rüde 2016, Semarang Municipality
Government 2015). Since 1999, the yearly extraction of
Figure 33. Subsidence distribution map superimposed on map of
geological conditions: (a) geological conditions, (b) subsidence from groundwater has been constant, except for in 2004, and it
2003 to 2007, (c) subsidence from 2007 to 2011, and (d) subsidence increased a little in 2010. From 2011 to 2013, the amount of
from 2015 to 2017 (Yastika et al. 2019a). groundwater extraction increased rapidly. The recent behavior of
the acceleration in subsidence seems to be particularly related to
the trend in groundwater extraction activities, as shown in Figure
The subsidence distribution along cross-section line A-A’ in
36(b). Although no data on groundwater extraction has been
Figure 33(a) is shown in Figure 34 for the three periods. The
found since 2016, the extraction seems to be continuing because
elevation of the ground surface is also illustrated. The area in the
the subsidence is still accelerating (Figure 36(b)).
north, composed of alluvium, exhibits large subsidence. The
subsidence around the boundary between the alluvium and the
(a)
Damar Formation becomes smaller than the subsidence in the 20
I
north. Almost no subsidence is seen in the areas of the Damar 0 II III
and Kaligetas Formations. The subsidence in the area of alluvium
Subsidence (cm)
-20
becomes larger year by year. -40 RMPA
In order to verify the validity of the DInSAR results, they -60 Envisat
HM_Envisat
-80
were compared with GPS results at 33 measurement points ALOS
HM_ALOS
-100
obtained from 2008 to 2009 and from 2009 to 2010 (Abidin et al. -120
Sentinel
HM_Sentinel
GPS by Abidin et al. (2013)
2013). Figures 35(a) and (b) show the correlation between the -140
DInSAR results and the GPS results at all the GPS measurement 2002 2004 2006 2008 2010 2012 2014 2016 2018
Date
points. The correlation values for the two above periods are 0.67
and 0.88, respectively. It was found that the two sets of results (b)
agree well, namely, the discrepancy between them was within a 20 120
Groundwater extraction (x106 M3/y)
4000
few cm, and the validity of DInSAR was confirmed. 0
100 3500
Subsidence (cm)
-20
Number of wells
3000
80
-40
2500
-60 60 2000
-80 1500
40
-100 1000
20
-120 500
-140 0 0
01/1993 01/1998 01/2003 01/2008 01/2013 01/2018
RMPA Date
Envisat ALOS Sentinel
HM_Envisat HM_ALOS HM_Sentinel
GPS by Abidin et al. (2013)
Groundwater extraction (Putranto and Rüde, 2016) Groundwater extraction (Semarang Govt., 2015)
Number of wells (Putranto and Rüde, 2016) Number of wells (Semarang Govt. 2015)
Figure 36. Relationship between temporal transition of subsidence
and total extraction of ground water: (a) temporal transition of
subsidence at RMPA, and (b) number of registered deep wells and
total extraction in Semarang-Demak groundwater basin (Yastika et al.
Figure 34. Relationship among elevation, geological conditions, and 2019a).
subsidence in cross section (A-A’) (Yastika et al. 2019a).
4.6 Rockfill dam behavior due to 2011 Great East Japan
Earthquake
Exterior deformation is one of the most important items to
monitor when managing the safety of a rockfill dam. It is usually
measured by a conventional surveying method (geodetic total
station) using targets set on the crest and the slope of the dam.
610
It is not easy, however, to measure the deformation by such a represent the time of the main shock (14:46 p.m. on March 11,
method immediately after an earthquake or other emergent event. 2011) and the time of the largest aftershock (11:32 p.m. on April
In order to overcome this issue, the GPS displacement 7, 2011), respectively.
monitoring system has been employed for continuous It is found that the main shock caused displacements of about
displacement monitoring at more than 30 dams in Japan. It was 10 mm in the downstream direction and about 3 mm in the right
able to detect the displacements of the dam embankment due to bank direction, as well as settlement of about 10 mm at G-9.
the Great East Japan Earthquake (9.1 MW) (Kobori et al. 2014). The largest aftershock caused displacements of about 2 mm in
The GPS monitoring system was installed at Ishibuchi Dam, the downstream direction and settlement of about 2 mm.
with a height of 53 m, which is an old concrete-face rockfill dam The displacement vectors due to the main shock and the
constructed in 1953. It is located 216 km away from the epicenter largest aftershock are shown on the plan view in Figures 39(a)
of the Great East Japan Earthquake. Figure 37 shows the plan and (b), respectively. The direction of the vectors due to the main
view and a vertical section of the dam with the monitoring points. shock were different from those due to the largest aftershock.
Sensors were set at 17 monitoring points, denoted by G-1 to G- However, the displacement vectors caused by both shocks were
17, and at one point as the reference point denoted by K-1 directed to the downstream side, which was the direction toward
(Figures 37(b) and (c)). The sensors on the dam were fixed to the the epicenter of the earthquake. This means that the dam
surface of the concrete facing with anchor bolts. embankment moved in the downstream direction relative to the
The measurements were started on June 26, 2008, which was reference point due to the earthquake.
12 days after the Iwate-Miyagi Nairiku Earthquake (6.9 MW) and
three years before the Great East Japan Earthquake. (a)
The monitoring results at G-4, G-9, and G-14 on section line
No. 20, which is the section of the dam with the maximum area 40 Reservoir water level
Great East Japan Earthquake
(main shock)
Largest aftershock
320
(Figure 37), are shown along with the water level of the reservoir
Reservoir water level EL. [m]
30
Disp. in up-downstream
310
20
in Figure 38. The three-dimensional displacements in the up-
direction [mm]
G-4 (Upstream face) 300
10
G-9 (Crest)
0 290
downstream, dam axis, and vertical direction were continuously -10
-20
G-14 (Downstream face)
280
measured every hour. -30
-40
270
260
7/1 1/1 7/1 1/1 7/1 1/1 3/11 7/1 1/1
2009 2010 2011 2012
(a) (b)
60 320
Reservoir water level EL. [m]
50
310
40
Disp. in dam axis
300
direction [mm]
30
20 290
10
280
0
270
-10
-20 260
7/1 1/1 7/1 1/1 7/1 1/1 3/11 7/1 1/1
2009 2010 2011 2012
(c)
30 320
Reservoir water level EL. [m]
20
310
Vertical disp. [mm]
10
300
0
-10 290
(b)
-20
280
-30
270
-40
-50 260
7/1 1/1 7/1 1/1 7/1 1/1 3/11 7/1 1/1
2009 2010 2011 2012
Figure 38 Displacement monitoring results by GPS at G-4, G-9, and
G-14 located on cross section (line No. 20) and water level of
reservoir: (a) up-downstream direction, (b) dam axis direction, and (c)
vertical direction (Kobori et al. 2014).
(a)
(c)
(b)
Figure 37. Ishibuchi dam and layout of GPS sensors: (a) dam and
sensor set on the dam embankment, (b) plan view, and (c) cross
section (line No. 20) (Kobori et al. 2014).
Figure 39. Displacement vectors at monitoring points on dam
Upon the occurrence of the Great East Japan Earthquake on embankment: (a) displacement due to main shock March 11, 2014
March 11, 2011 (denoted by “the main shock”), a seismograph and (b) displacement due to aftershock (April 7, 2014) (Kobori et al.
on a right bank terrace recorded a maximum acceleration value 2014).
of 184 gal. When the largest aftershock occurred on April 7, the
acceleration value of 140 gal was recorded by a seismograph in
the same place. The broken and dashed-dotted lines in Figure 38
611
Figure 40 focuses on the monitoring results at G-9 during the 5.1 Assessment of stability of large landslide colluvial deposit
periods before and after the main shock. The displacements in all during disaster recovery works
directions were almost zero until 14:00 on March 11 before the
earthquake. The earthquake occurred at 14:46. After 15:00 on the A large-scale landslide occurred in a mountainous area of
same day, the GPS monitoring system clearly detected western Japan due to heavy rainfall brought about by a typhoon.
displacements due to the earthquake (bullets in Figure 40), i.e., Cumulative rainfall reached 1200 mm within a few days after it
both the displacement in the direction of the downstream side and started to rain. The huge volume of colluvial deposit blocked a
the settlement were about 10 mm, and the displacement in the river channel and created landslide dams (Figure 42).
direction of the right bank side was about 3 mm. A debris flow occurred due to the erosion of part of the
For the other GPS monitoring points, the main shock and the deposit. In order to prevent the collapse of the landslide dam and
largest aftershock caused displacements similar to those shown the occurrence of further debris flows, disaster recovery works
in Figure 38. No remarkable displacements occurred after the were carried out. The GPS monitoring system was applied to
largest aftershock, and the displacements gradually decreased monitor the behavior of the unstable colluvial deposit during
and became stable. these disaster recovery works (Satoh et al. 2014).
Figure 41 shows the relationship between the water level of The landslide occurred over an area that was 650 m in length,
the reservoir and the horizontal displacement of the dam 410 m in width, 60 m in depth, and 300 m in height (Figure 42).
embankment before and after the earthquake (Nakashima et al. The volume of the colluvial deposit was estimated to be 4.1
2021). The slope of the relationship between the water level and million m3. The slope was formed by alternating beds of
the dam displacement represents a kind of stiffness of the dam sandstone and mudstone.
embankment. It is found from Figure 41 that the slope does not
change before or after the earthquake. This means that the dam
embankment maintained the same level of stiffness and that it Length: 650 m
was not damaged by the earthquake. Width: 410 m
It is seen that GPS monitoring is effective for detecting Depth: 60 m
displacements and for assessing the damage due to an earthquake Unstable slope
immediately after it has occurred. Unstable colluvial deposit
River channel blockage
(a) (b)
10 10
GPS
up-downstream
before earthquake before earthquake 5
5
after earthquake
Disp.ininup/downstream
River
[mm]
0 Debris flow
Disp. in dam axis
0
direction [mm]
[mm]
-5
direction
-5
direction
-10 -10
-15 -15 Upper dam
Disp.
-20 -20
3/13 3/10 3/11 14:46 3/12 3/13 3/13 3/10 3/11 3/12 3/13
Lower dam
(c) Disaster recovery works area
10 Figure 42. Large landslide and unstable colluvial deposit (Satoh et al.
5
2014).
Vertical disp. [mm]
-5
-10
-15
GPS sensors were set at 10 monitoring points on the slope of
3/13
-20
3/10 3/11 3/12 3/13
the colluvial deposit denoted by G-1, G-2, G-3, G-3’, …, and G-
Figure 40. Detection of three-dimensional displacements at G-9 due 8, and at a reference point outside of the unstable area (Figure
to the Great East Japan Earthquake: (a) up-downstream direction, (b) 43). The distances between the monitoring points and the
dam axis direction, and (c) vertical direction (Kobori et al. 2014). reference point were about 100 m to 700 m. Electric power was
supplied to the monitoring system by solar panels, and the
measurement data were transferred via a cellular phone from the
Main: Main shock
Max.: Maximum aftershock monitoring site to a server computer at the control office.
G4 G9 G14
Extensometers, denoted by S-1 to S-4, were also set in the upper
320 320 320
Water level of reservoir [EL. m]
Water level of reservoir [EL. m]
Water level of reservoir [EL. m]
+: Downstream
315 315 315
area of the unstable portion (Figure 43).
310 310 310
Unstable slope
305 305 305
300 Main Max. 300 Main Max. 300 Main Max.
-5 0 5 10 15 20 25 -5 0 5 10 15 20 25 -5 0 5 10 15 20 25
Horizontal displacement [mm] Horizontal displacement [mm] Horizontal displacement [mm]
(up-downstream direction) (up-downstream direction) (up-downstream direction)
Figure 41. Relationship between water level of reservoir and
horizontal displacement of dam embankment at G4, G9, and G14 Unstable colluvial deposit
before and after Great East Japan Earthquake (Nakashima et al. 2021).
5 CASE STUDIES OFASSESSING STABILITY OF
SLOPES BASED ON MONITORED DISPLACEMENTS
This section demonstrates two case studies of assessing the
stability of slopes based on displacement monitoring.
Disaster recovery works area
Figure 43. Plan view of monitoring site and locations of GPS sensors
(Satoh et al. 2014).
612
The temporal displacement monitoring results for G-5 over Warning Criteria line Criteria line
to “Caution”: to “Warning”:
three years are shown in Figure 44(a), and the displacement 10mm/day 25mm/day
vectors are given on the plan view in Figure 44(b). The
displacement vectors show that the total displacements in the
lower area of the slope (G-1 to G-8, except for G-5) were not
large and ranged from a few mm to 34 mm.
On the other hand, the displacements in the upper area of the
unstable portion (G-5, and S-1 to S-4) significantly increased up
to 540 mm and to 3354 mm. When the rate of displacement Figure 45. Warning process based on monitored displacement by GPS
exceeded the criteria for early warning (Table 3), the monitoring at G-5 (Sato et al. 2014).
center declared “Warning” or “Caution” for the evacuation of
workers and the suspension of the recovery works, as shown in
Figure 45. 5.2 Assessment of stability of slope along expressway by back
The use of GPS monitoring was validated for the safe analysis of monitored displacements
management of disaster recovery works.
The second case study was a cut slope at a ramp section of an
expressway (Shoji et al. 2021). The height and width of the slope
(a)
are 25.4 m and 45.7 m, respectively. The slope is composed of
1500 500
sedimentary rocks significantly affected by weathering, as shown
Rainfall (mm/day)
Latitude (mm)
0 400
-1500 300 in Figure 46 (Shitano et al. 2014). After detecting several cracks
-3000 200 on the surface of the slope, continuous displacement monitoring
-4500
-6000 0
100
using GPS was conducted.
05/10/11 05/2/12 05/6/12 05/10/12 05/2/13 05/6/13 05/10/13 05/2/14 01/6/14 Figure 47 shows the monitoring results of the three-
D/M/Y
dimensional displacement at G-1. Although the displacement
1500 500
increased up to about 20 mm, due to rainfall in early July 2009,
Rainfall (mm/day)
Longitude (mm)
0 400
-1500 300 the slope was stable. Eventually, displacement significantly
-3000 200 increased and exceeded 200 mm in July 2010 due to continuous
-4500 100
heavy rainfall over a few days. Then, the slope collapsed.
-6000 0
05/10/11 05/2/12 05/6/12 05/10/12 05/2/13 05/6/13 05/10/13 05/2/14 01/6/14 Fortunately, as the road was closed to traffic when the
displacement began to increase quickly, no accidents occurred.
1500 500
In order to assess the stability of the slope during the
Rainfall (mm/day)
0 400
Height (mm)
-1500 300 monitoring period, a back analysis method (Sakurai 2017) was
-3000 200 applied. Since the detailed process of this back analysis method
-4500
-6000
100
0
is described in a reference (Shoji et al. 2021), only the results are
05/10/11 05/2/12 05/6/12 05/10/12 05/2/13 05/6/13 05/10/13 05/2/14 01/6/14 introduced here.
(b) A back analysis was applied to the monitored displacements
at G-1 and G-3 three times when the displacement rate changed.
The cases are denoted as CASE-B1, CASE-B2, and CASE-B3,
as shown in Figure 47.
To: Shimonoseki JCT
G-1
A
B ramp
Ramp
Toll gate
Tollgate
To: Ube JCT Borehole
Borehole
Figure 44 Monitoring results: (a) displacements at G-5 and daily
amount of rainfall, and (b) displacement vectors in plan view (Sato et
al. 2014).
Table 3. Criteria of early warning system.
Rate of displacement
Criteria Precipitation Extensometer Figure 46. Plan view of the slope at ramp of a highway and the cross
(GPS)
section “A-A” (Shitano et al. 2014).
Caution 5 mm/hour >10 mm/day -
Warning 10 mm/hour >25 mm/day 4 mm/hour
613
CASE-B1 (01/2009) CASE-B2 (12/2009) CASE-B3 (7/2010)
(mm) July 2009 July 2010 G-1
Back analysis
Displacement
North Measurement
in latitude
Slope collapsed
South
G-3
Displacement
in longitude
East
West Assumed slip zone
in height direction
20mm
Displacement
Up
Figure 48. Comparison between calculated and measured
Down displacements for CASE-B3.
Rainfall (mm)
2008 2009 2010
Figure 47. Monitored displacement at G-1 by GPS (Shitano et al. 2014,
Shoji et al. 2021).
The displacements were 1 mm, 20 mm, and more than 200
mm in CASE-B1, CASE-B2, and CASE-B3, respectively. After
Figure 49. Transition of factor of safety.
applying the back analysis, the Young’s modulus of the slope
ground and anisotropic parameter (Sakurai 2017) of the slip
zone were obtained as shown in Table 4. The unit volume weight 6 CONCLUDING REMARKS
and Poisson’s ratio were assumed to be 16 kN/m3 and 0.3,
respectively Satellite technology, namely, GPS and SAR, have become
To confirm the validity of the back analysis, the important engineering tools for monitoring ground
displacements calculated using the back-analyzed Young’s displacements in Rock and Geotechnical Engineering. The
modulus and anisotropic parameter were compared with the concluding remarks are as follows:
measured displacements. Figure 48 presents a comparison
between them for CASE-B3. The displacements at G-1 and G-3 Displacement monitoring using GPS/GNSS can provide the
show a good agreement. three-dimensional displacements of the ground automatically
Shear modulus is estimated by , and critical and continuously with mm accuracy.
shear strain is obtained by the empirical relationship The International Society for Rock Mechanics and Rock
(Sakurai 2017). Finally, shear strength and cohesion Engineering approved “The ISRM suggested method for
are derived from equations and monitoring rock displacements using the Global Positioning
( :internal friction angle). Those values are System”. GPS/GNSS has become a standard tool for
shown in Table 4. displacement monitoring.
Finally, the factor of safety is calculated using the DInSAR analysis using SAR data can provide the
conventional limit equilibrium method. Figure 49 shows the displacement distribution of the ground surface in extensive
transition of the factor of safety during the monitoring period. It areas without the necessity for any devices on the ground. The
decreases from 1.2 to 1.0, and finally becomes less than 1.0. At time transition of the displacement at points of interest can
that moment, the slope collapsed. The results of the back analysis also be found.
coincide with the real slope behavior. Case studies of applications of GPS and DInSAR to
Monitoring the displacements and applying a back analysis monitoring an unstable steep slope, landslides, land
method are confirmed as useful for safe road management. subsidence, and a rockfill dam during an earthquake were
introduced in this paper. All cases indicated that both
Table 4. Results of back analysis. methods, GPS and DInSAR, are very effective for
displacement monitoring over extensive areas.
CASE-B1 CASE-B2 CASE-B3
Young’s modulus: (kPa) 1,000 1,000 1,000 Satellite technology can make invisible ground behavior,
Anisotropic parameter:
which could not be detected up to now, visible. Its contribution
0.260 0.080 0.005 to finding the mechanism of the ground behavior, assessing the
(Slip zone)
stability, predicting risks, and preventing and mitigating disasters
Shear modulus: (kPa) 260.0 70.0 3.7
is highly anticipated.
Critical shear strain: (%) 5.9 8.3 18.5
Shear strength: (kPa) 15.3 6.5 0.9
7 ACKNOWLEDGEMENTS
Internal friction angle (assumed):
20 20 20
(°) This research was partially supported by JSPS KAKENHI
Cohesion: (kPa) 10.7 4.6 0.6 (Grant-in-Aid for Scientific Research by the Japan Society for
0.8 the Promotion of Science) Grant Number 16H03153.
Factor of safety 1.2 1.0
(collapse) The author wishes to express his sincere appreciation to his
colleagues, students, and partners in international collaborative
research for their support and cooperation in achieving these
614
studies. He also expresses his gratitude to the European Space Misra P. and Enge P. 2006. Global Positioning System - signals,
Agency (ESA) and the Japan Aerospace Exploration Agency measurements, and performance. 2nd Ed. Ganga-Jamuna Press.
(JAXA) for providing the Sentinel-1A/B and ALOS-2 SAR data
used in these studies, respectively. (Ministry of Regional Development and Public Works). 2017. 2018.
2019. . -02-29-
The author also thanks Ms. H. Griswold for proofreading the -02-29- -02-29-
manuscript of this paper.
8 REFERENCES egister). http://gz-varna.mrrb.
government.bg/map/. Accessed on 31 December 2021 (in
Abidin H.Z., Andreas H., Gumilar I., Sidiq T.P. and Fukuda Y. 2013. Bulgarian).
Land subsidence in coastal city of Semarang (Indonesia): Nakashima S., Shimizu N., Kobori T. and Yamaguchi Y. 2021.
Characteristics, impacts and causes. Geomatics, Nat. Hazards Risk Utilization of displacement monitored by GPS for Safety
4, 226–240. https://doi.org/10.1080/19475705.2012.692336 Management of rockfill dams. Journal of Japan Society of Dam
Berardino P., Fornaro G., Lanari R. and Sansosti E. 2002. A new Engineers 31(1), 53-63. https://doi.org/10.11315/jsde.31.53 (in
algorithm for surface deformation monitoring based on small Japanese)
baseline differential SAR interferograms. IEEE Trans Geosci Nagasaki K., Kameyama M., Yoshimoto M., Nakashima S. and Shimizu
Remote Sens 40, 2375–2383. N. 2021. Applicability of SBAS-DInSAR for Monitoring
Bruchev I, Dobrev N, Frangov G, Ivanov P, Varbanov R, Berov B, Displacements of a Local Steep Slope. Proceedings of 5th
Nankin R and Krastanov M 2007. The landslides in Bulgaria – International Workshop on Rock Mechanics and Engineering
factors and distribution. Geologica Balcanica 36(3-4), 3–12. Geology in Volcanic Fields (RMEGV2021), ISRM Specialized
Bizjak K. F. and Zupan i A. 2009. Site and laboratory investigation of Conference, OS6-4.
the Slano blato landslide. Eng. Geology 105, 171-185. Parwata I N. S., Shimizu N., Gruji B., Zekan S., elikovi R. and
Evstatiev D. and Evlogiev Y. 2013. Landslides along the northern Black Vrkljan I. 2020a. Monitoring the Subsidence Induced by Salt Mining
Sea coast between Varna city and Kavarna town (Bulgaria). Geo- in Tuzla, Bosnia and Herzegovina by SBAS-DInSAR Method. Rock
Eco-Marina 19, 39–57. Mechanics and Rock Engineering 53(11), 5155–5175.
Ferretti A., Prati C., Rocca F. 2001. Permanent scatterers in SAR Parwata I N. S., Shimizu N., Gruji B., Zekan S., elicovi R., Imamovi
interferometry. IEEE Trans. Geosci. Remote Sens. 39, 8–20. E. and Vrkljan I. 2020b. Monitoring of subsidence induced by salt
Ferretti A. 2014. Satellite InSAR data: reservoir monitoring from space. mining in Tuzla, Bosnia and Herzegovina by means of SBAS-
EAGE Publication, The Netherlands. DInSAR. Proceedings of ISRM International Symposium: Eurock
Ferretti A., Monti-Guarnieri A., Prati C. and Rocca F. 2007. InSAR 2020, Paper No. ISRM-EUROCK-2020-130.
Principles: Guidelines for SAR interferometry processing and Putranto T. and Rüde T. 2016. Hydrogeological Model of an Urban City
interpretation. ESA Publications, The Netherlands. in a Coastal Area, Case study: Semarang, Indonesia. Indones. J.
Furuyama Y., Nakashima S. and Shimizu N. 2014. Displacement Geosci. 3(1), 17-27. https://doi.org/10.17014/ijog.3.1.17-27
monitoring using GPS for assessing stability of unstable steep slope Sakurai S. 2017. Back Analysis in Rock Engineering. ISRM Book Series
by means of ISRM suggested method. Proceedings of the ISRM vol 4, CRC Press.
International Symposium - 8th Asian Rock Mechanics Symposium Sato, T., Kien, N. T., Nakashima, S. and Shimizu, N. 2021. Monitoring
(ARMS8), Sapporo, 1897-1904. the three-dimensional displacement by GPS - a case study in a large
Hanssen R. F. 2002. Radar Interferometry. New York: Kluwer Academic steep slope, Proceedings of 5th International Workshop on Rock
Publisher. Mechanics and Engineering Geology in Volcanic Fields
Hoffmann-Wellenhof B., Lichtengger H. and Collins J. 2001. GPS – (RMEGV2021), ISRM Specialized Conference, Paper No. OS6-5.
Theory and Practice. 5th revised edition, Springer. Satoh W., Iwasaki T., Sakurai W., Fujii A. and Shimizu N. 2014.
Iwasaki T., Takechi K., Takeishi A., Masunari T., Takechi Y. and Monitoring the stability of a large-scale colluvium deposited by
Shimizu N. 2003. Web-based displacement monitoring system using slope failures due to heavy rainfall using a GPS automatic
GPS for the maintenance of roadside slopes. Proceedings of 6th monitoring system. Proceedings of the ISRM International
International Symposium on Field Measurements in Geomechanics, Symposium - 8th Asian Rock Mechanics Symposium (ARMS8),
Oslo, 137-143. Sapporo, 1778-1783.
Jemec Aufli M. ., Popit T., Košir A., Ma ek M., Logar J., Shimizu N., Masunari T. and Iwasaki T. 2011. GPS displacement
Petkovšek A., Mikoš M., Calligaris C., Boccali C., Zini L., Reitner monitoring system for the precise measuring of rock movements.
J. and Verbovšek T. 2017. The variety of landslide forms in Slovenia Proceedings of 12th International Congress on Rock Mechanics,
and its immediate NW surroundings. Landslides 14(4), 1537-1546. Beijing, 1117-1120.
Kien, N. T., Hayashi Y., Nakashima S. and Shimizu N. 2017. Long-term Shimizu N. and Matsuda H. 2002. Practical applications of the Global
displacement monitoring using GPS for assessing the stability of a Positioning System for the assessment of slope stability based on the
steep slope. Proceedings of ISRM Young Scholars’ Symposium on Displacement Monitoring Approach. Proceedings of the 3rd Korea-
Rock Mechanics & International Conference on New Development Japan Joint Symposium on Rock Engineering, ISRM Regional
in Rock Mechanics and Geotechnical Engineering, Jeju, 165-168. Symposium, Seoul, 57-70.
Kobori T., Yamaguchi Y., Nakashima S. and Shimizu N. 2014. Shimizu N. and Nakashima S. 2017. Review of GPS displacement
Displacement monitoring of a rockfill dam before, during and after monitoring in rock engineering. Rock Mechanics and Engineering.
the Great East Japan Earthquake using GPS. Proceedings of the Volume 4, CRC Press, Chapter 19, 593-626.
ISRM International Symposium - 8th Asian Rock Mechanics Shimizu N. Nakashima S. and Masunari T. 2014. ISRM suggested
Symposium (ARMS8), Sapporo, 930-938. method for monitoring rock displacements using the Global
Mancini F., Stecchi F., and Gabbianelli G. 2009a. GIS-based assessment Positioning System. Rock Mech and Rock Engineering 47, 313-328.
of risk due to salt mining activities at Tuzla (Bosnia and Shitano M., Murakami T. and Shimizu N. 2014. Application of the GPS
Herzegovina). Engineering Geology 109, 170–182. Displacement Monitoring System for the Preventive Maintenance of
Mancini F., Stecchi F., Zanni M. and Gabbianelli G. 2009b. Monitoring a Cut Slope along a National Expressway - a Case Study.
ground subsidence induced by salt mining in the city of Tuzla Proceedings of the ISRM International Symposium - 8th Asian Rock
(Bosnia and Herzegovina). Environmental Geology 58, 381–389. Mechanics Symposium (ARMS8), Sapporo, 1877-1886.
Marfai M. A., King L., Sartohadi J., Sudrajat S., Budiani S. R. and Shoji Y., Takabatake M., Suzuki H. and Shimizu N. 2021. Monitoring
Yulianto F. 2008. The impact of tidal flooding on a coastal slope stability based on the factor of safety estimated by back
community in Semarang. Indonesia. Environmentalist. 28, 237–248. analysis of measured displacements. Proceedings of EUROCK-
https://doi.org/10.1007/s10669-007-9134-4 Torino 2021. IOP Conf. Series: Earth and Environmental Science
Masunari T., Tanaka K., Okubo N., Oikawa H., Takechi K., Iwasaki T., 833, Mechanics and Rock Engineering, from Theory to Practice,
and Shimizu N. 2003. GPS-based continuous displacement Turin, Paper No. 012144. doi:10.1088/1755-1315/833/1/012144
monitoring system. Proceedings of 6th International Symposium on Semarang Municipality Government 2015. Laporan Keterangan
Field Measurements in Geomechanics, Oslo, 537-543. Pertanggungjawaban (LKPJ) Walikota Semarang Akhir Tahun
Anggaran 2014. Indonesia. (in Indonesian)
615
Tan T. S., Inoue T. and Lee S. L. 1991. Hyperbolic method for
consolidation analysis. J. Geotech. Eng. ASCE 117, 1723–1737.
Verbovšek T. and Popit T. 2018. GIS-assisted classification of litho-
geomorphological units using Maximum Likelihood Classification,
Vipava Valley, SW Slovenia. Landslides 15(7), 1415-1424.
Yamaguchi M., Yastika P. E. Shimizu N., Milev N., and Vrkljan I. 2021.
Application of SBAS-DInSAR to monitoring landslides along the
northern Black Sea coast in Bulgaria, Proceedings of EUROCK-
Torino 2021, IOP Conf. Series: Earth and Environmental Science
833, Mechanics and Rock Engineering, from Theory to Practice,
Turin, Paper No. 012151. doi:10.1088/1755-1315/833/1/012151
Yastika P. E., Shimizu N. and Abidin H. Z. 2019a. Monitoring of long-
term land subsidence from 2003 to 2017 in coastal area of Semarang,
Indonesia by SBAS DInSAR analyses using Envisat-ASAR, ALOS-
PALSAR, and Sentinel-1A SAR data. Advances in Space Research
63, 1719–1736.
Yastika P. E., Shimizu N. and Verbovšek T. 2019b A case study on
landslide displacement monitoring by SBAS-DInSAR in the Vipava
River Valley, Slovenia. Proceedings of ISRM Specialized
Conference on 5th ISRM Young Scholars' Symposium on Rock
Mechanics and International Symposium on Rock Engineering for
Innovative Future, 406-411.
616
Bright Spark Papers
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Prediction of earthquake-induced ground motions using simulation methods and their
validation against observations
Brendon A. Bradley
ABSTRACT: Earthquake-induced ground motion simulation is rapidly being adopted in research, and increasingly considered in
practice, for seismic hazard analysis. Despite the potential of these simulation methods, many challenges remain for the consistent
and uniform application as well as advancing their prediction accuracy and precision. This paper provides a brief view on efforts to
validate such methods in a New Zealand context, as well as on-going challenges with modelling more of the salient physics that is
qualitative understood, but has been elusive to quantitively model with sufficient accuracy in a practical manner.
RÉSUMÉ : La simulation des mouvements du sol induits par les tremblements de terre est rapidement adoptée dans la recherche, et
de plus en plus considérée dans la pratique, pour l'analyse des risques sismiques. Malgré le potentiel de ces méthodes de simulation,
de nombreux défis subsistent pour une application cohérente et uniforme ainsi que pour faire progresser leur exactitude et leur
précision de prédiction. Cet article donne un bref aperçu des efforts pour valider de telles méthodes dans un contexte néo-zélandais,
ainsi que des défis en cours avec la modélisation d'une plus grande partie de la physique saillante qui est qualitativement comprise,
mais qui a été difficile à modéliser quantitativement avec une précision suffisante dans un contexte pratique. manière.
KEYWORDS: ground-motion simulation, validation, New Zealand earthquakes.
1 INTRODUCTION provide a parametric prediction of the Fourier spectra of the
ground motion, from which a time series is obtained through
Earthquake-induced ground motion prediction is presently convolution with random-phase white noise (Baker et al., 2021).
under-going a paradigm shift from the empirical prediction of Both comprehensive and simplified physics approaches have
ground motion intensity measures (IMs, e.g., PGA, SA), based merits, with the former generally preferred for the prediction of
on regression analysis of observed IMs from past earthquakes, low-frequency ground motion (typically f<1Hz) and the latter for
toward the use of physics-based simulation methods that directly high-frequency ground motion. As a result, the most common
predict the ground motion time series (i.e., multi-component predictions presently used in engineering applications are a
acceleration as a function of time). This progression in modelling ‘hybrid’ of these two methods in order to leverage their relative
complexity and potential is akin to the ‘quiet revolution’ in benefits.
weather prediction that occurred in the 1950’s (Bauer et al., 2015;
Bradley, 2019).
This paradigm shift is presently occurring as a result of: (i)
diminishing advances in conventional empirical predictions
(refs); (ii) recent well-recorded earthquakes illustrating that, even
now, physics-based simulation methods provide predictions that
are comparable to, or even superior than, those from empirically-
based predictions (refs); and (iii) a substantially greater volume
of data from seismological observations can be synthesised,
enabling the incorporation of region and site-specific features
and the potential to reduce prediction uncertainties. (refs).
This paper is intended to provide a brief summary of the
common ground motion simulation methods currently being
employed in large-scale research, and increasingly in practice;
the critical role of validation to demonstrate the predictive
capability of these methods and their constituent model inputs,
and highlight some of the on-going challenges. Due to space
limitations, a comprehensive overview of these topics is omitted,
but the cited references provide a broader coverage for the
interested reader. 000
090
2 GROUND-MOTION SIMULATION BASICS
10 s
10 cm/s
ver
A distinguishing element of ground motion simulation methods, Figure 1. Example wavefield of ground velocity at the earth’s surface
from a M6.8 event. The hypocentre is denoted by the star, and the
from conventional empirical ground-motion models, is that they
resulting time series at Site 1 is shown at the base of the figure (after
directly provide a ground motion time series, rather than simply Baker et al., 2021).
intensity measures (Figure 1). While there are numerous different
simulation techniques, they can broadly be broken down into
‘comprehensive physics’ and ‘simplified physics’ approaches.
While the use of these two qualifiers is relative, comprehensive 3 VALIDATION OF SIMULATION METHODS
methods directly solve the 3D equation of motion that governs
seismic wave propagation, while simplified methods typically Ground motion simulation methods have been employed in
619
seismological research for many decades, but it is only recent increase in the adoption of simulated motions, hence on-going
advances in theory, model resolution, and computing resources simulation validation is essential.
that have enabled such predictions to provide a sufficient
representation of reality. Validation of the predictive capability
of ground-motion simulations is an essential element for its
adoption in practical seismic hazard analysis applications.
In addition to individual validation studies (Frankel, 2017;
Lee et al., 2020; Paolucci et al., 2015; Ramírez-Guzmán et al.,
2012; Taborda and Bielak, 2014), more recently larger
coordinated efforts to verify and validate a multitude of
‘broadband’ ground motion simulation methods have occurred
under the auspices of the Southern California Earthquake Center
(SCEC) through the Broadband Platform (BBP) validation
exercise (Dreger and Jordan, 2015).
3.1
Simulation validation needs to evaluate the predictive
capabilities of the overall simulation methodology (i.e., physical
assumptions), as well as the input models and their parameters.
For ground motion simulation, in particular, the adequacy of
input models describing the source rupture, 3D crustal structure,
Figure 2. Ground motion validation matrix and relation to the intended
and surficial site conditions are themselves complex and usage of ground motion simulations. The vertical axis indicates the
regionally varying. As a result, the predictive capability of increasing spatial resolution from generic to region and site-speci c
ground motion simulations is region-, and even site-specific. validation. The horizontal axis indicates the increasing complexity of
Furthermore, ground-motion time series are complex transient IM metrics used in quantifying simulation validation, which is a
signals, and the ability of simulations to adequately reproduce the function of the speci c engineered system considered (Bradley,
salient features of these signals varies depending on which aspect Pettinga, et al., 2017).
is of particular interest. The engineering representation of ground
motion severity generally refers to a collection of different 3.2
ground motion IMs. Some IMs are ubiquitous, such as elastic
Figure 3 illustrates the spatial distribution of simulated peak
response spectra; some are seeing increasing awareness and
ground velocity for the 2010 Darfield and 2011 Christchurch
utilisation (e.g., parameters representing the duration and
earthquakes (as part of the Canterbury earthquake sequence) and
cumulative nature of the motion); and others are problem-
the 2016 Kaikoura earthquake, which were the most significant
specific (e.g., induced displacement response of a specific
events to occur over the past decade in New Zealand. The figure
building typology).
highlights, in particular, the importance of source directivity and
Figure 2 provides a graphical validation matrix of the spatial-
sedimentary basin effects on amplifying surface ground motions.
and IM-dependence of ground motion simulation validation. The
The role of source directivity is least pronounced in the 2011
vertical axis represents the transition from generic through to
Christchurch earthquake because of its moderate magnitude and
site-specific locations where simulated ground motions are
the general misalignment of the direction of rupture propagation
desired, while the horizontal axis represents the complexity of
and slip vectors. Directivity in the 2010 Dar- field earthquake
IM metrics used to quantify predictive capability in validation.
was most pronounced in central and northern Christchurch as a
Both of these axes are continuous in nature, however, they are
result of the west-to-east rupture of the Greendale fault. Finally,
discretised here for practical application.
while the 2016 Kaikoura earthquake was the result of an
Figure 2 identifies three principal domains of the validation
exceptionally complex multi-segment rupture, strong directivity
matrix in the context of intended utilisation. The first being that
effects were present in the north eastern part of the South Island
if only qualitative validation is performed by comparing the
and the Wellington region as a result of the general south-to-north
nature of simulated and observed waveforms then those
rupture propagation direction.
simulations are not appropriate for utilisation in practice, i.e.
quantitative validation is essential. In the context of seismic
hazard analysis, in which an accurate and precise prediction of
the distribution of IM is needed, the specifics of the particular
region and site of interest are essential components. Therefore,
simulation methodologies that have been validated using only
data in generic regions would not be considered appropriate for
use in determining the seismic hazard at another region/location
at which no specific validation has been performed. Ground
motion simulations undertaken in generic regions that have been
validated would, however, still provide simulated time series that
could be utilised for response history analyses once scaled to the
target design ground motion intensity (this is similar to the
current conventional use of as-recorded ground motions from
past worldwide earthquakes for response history analysis).
As would be expected, a framework for validation, and
accompanying documentation [9], is critical for a bi-directional
understanding between earthquake engineers on the
appropriateness of a suite of ground motion simulations for Figure 3. Simulated peak ground velocities for the 2010 Dar eld, 2011
utilisation in a site-specific context, as well as ground motion Christchurch, and 2016 Kaikoura events. Strong motion station
simulators to understand the context in which their results will be locations, which recorded the consequent ground motions, are shown in
utilised. The increased realisation of the benefits of utilising white triangles (Bradley, 2019).
simulated ground motions is likely to provide a demand-side
620
3.3
Figure 4 illustrates the ground motion spectral amplitudes for
two vibration periods (T = 0.0 and 3.0 s) as a function of source-
to-site distance for the 2016 Kaikoura event. The observed and As noted with reference to Figure 2, the ability to explicitly
simulated ground motion amplitudes for the 162 stations within model region- and site-specific phenomena is a key potential
the simulation domain are shown. In addition to the simulation strength of physics-based simulation methods. Use of the word
predictions, for reference, the NZ-specific empirical ground ‘potential’ is important because simulations should be validated
motion model (GMM) of Bradley (2013) is also shown based on to ensure that such effects can in fact be predicted by the adopted
a single representative Vs30 = 250 m/s value. model(s). The consideration of region- and site-specific effects
naturally requires the evaluation of simulation predictions over a
smaller geographical region than possible to evaluate predictive
performance for generic regions, which entails a reduction in
validation data for all other things equal, making statistically-
significant statements sometimes challenging.
The consideration of ground motions from smaller magnitude
events naturally enables a significant increase in the number of
data which can be utilised for simulation validation. The
appropriateness of using ground motion records from small
magnitude earthquakes for ‘testing’ the applicability of empirical
ground motion models for larger magnitude events (which are of
primary concern for seismic design) has long been a topic of
debate. However, the physics-based nature of ground motion
simulation methods provide a rational framework in which to
evaluate the predictive capability of various model ‘components’.
Of course, the use of small magnitude earthquakes for validation
offers limited (if any) ability to examine finite-fault kinematic
source rupture or nonlinear near-surface site response modelling,
but the observations can obviously be used for validation of the
considered crustal (velocity) model, such as demonstrated by Lee
et al. [28].
Despite the obvious need, extensive region- and site-specific
simulation validation has not yet become commonplace. Figure
5 provides one example from Lee et al. (2020) in which nearly
2000 ground motions recorded from events in the Canterbury,
NZ, region were used to validate the commonly-adopted Graves
and Pitarka (Graves and Pitarka, 2010) simulation method. This
dataset enabled systematic identification of several features in
the simulation method that can be refined in order to provide
improved predictions in this region, as well as likely biases in the
adopted crustal model. Such improvements are also likely to
result in improved validation outcomes for the larger magnitude
event simulations.
Figure 4. Observed, physics-based and empirically-predicted geometric
mean ground motion intensities as a function of source-to-site distance
( ) for the 2016 Kaikoura Earthquake for peak ground acceleration
(top) and T=3s spectral acceleration (bottom) (Bradley, Razafindrakoto,
et al., 2017)
Figure 4 illustrates that the simulations provide a good
comparison with the observed amplitudes. In particular, the
observed distance attenuation of short period amplitudes (i.e., T
= 0.0 s) is consistently predicted by the simulations, while the
empirical model predicts a slower attenuation; conversely at long
periods, the empirical model predicts a faster attenuation than
exhibited by both the observed and simulated amplitudes. Figure 5. 148 Earthquake sources and 53 strong motion stations
Figure 4 also separately annotates the simulated and observed considered within the ground motion simulation validation of Lee et al.
amplitudes based on their location in either the North or South (2020). Schematic ray paths of observed ground motions are also shown
Island, for which both observation and simulation consistently as black lines. A total of 1896 ground motions were used for simulation
illustrate higher- than-average amplitudes of North Island ground validation.
motions relative to those in the South Island for the same source-
to-site distance - because of the aforementioned effect of rupture
directivity in this event.
621
4 ON-GOING CHALLENGES FOR PREDICTION AND 4.3
UTILIZATION
Physics-based ground motion simulations commonly have
4.1 neglected surficial site effects (including soil nonlinearity) as a
result of using viscoelastic soil models, coarse spatial grids, and
Quantitative understanding of the earthquake sources is the most minimum shear wave velocities corresponding to stiff soils. Four
difficult piece of the ground-motion prediction problem, and is different methods have been used or proposed to incorporate soil
likely to remain so (with advances in crustal and site response nonlinearity into simulations: (i) fully-coupled 3D simulation
modelling occurring much more readily). Two notable problems models that directly allow soil nonlinearity in surficial soils; (ii)
hindering ground motion simulation are: (a) source generation the domain reduction method for decomposing the physical
for simulation at high frequencies (f < 1 Hz); and (b) source domain into multiple subdomains for separate simulation; (iii)
generation for multi-segment earthquake ruptures (a third conventional 1D wave propagation site response analysis
problem of uncertainties in seismic source modelling is discussed uncoupled from the simulations; and (iv) the use of simple
separately in a later section). Both of these problems gain empirically-based site amplification factors dependent on 30-m-
significant insight from the use of dynamic rupture simulations, averaged shear wave velocity, Vs30 (de la Torre et al., 2020).
however, present computational demands require that regional While the explicit consideration of site response using
ground motion simulations generally utilise kinematic re- approaches (i)-(iii) above would be considered preferable over
presentations of earthquake rupture (referred to as pseudo- the use of empirically-based site amplification factors, there has
dynamic models if they are parametrised based on dynamic previously been limited systematic examination of the benefits
considerations). of explicitly modelling site response based on validations against
In relation to high frequency source generation, for example, historical earthquakes. de la Torre et al. (2020) recently used the
Graves and Pitarka (2016) present a further iteration of a prior 10 most significant earthquake events in the 2010–2011
kinematic rupture generator that emulates the effect of geometric Canterbury earthquake sequence to examine site response
fault complexity (i.e. non-planar faults), fault damage zone (a modelling effects in ground motion simulations. The empirical
lower velocity region in the vicinity of the fault), and also results of de la Torre et al. illustrate that the explicit consideration
random perturbations to the underlying crustal velocity model. of site response in ground motion simulation is important (as
The principal need for these factors relates to the observed expected), however, the use of 1D site response analysis methods
homogenisation of the ground motion radiation pattern (i.e., still leave substantial room for improvement, with
general orientation-independence of ground motion amplitudes) comprehensive statistics from site response analyses at downhole
at frequencies f > 1 Hz, which was not present in prior simulated array sites indicating that appreciable biases with 1D site
ground motions. More recently, observations from the 2019 response analysis do exist.
Ridgecrest earthquake resulted in future proposed changes to the In parallel, and increasing number of studies are beginning to
slip-time function in the shallow crust (Pitarka et al., 2022). In comprehensively consider the practical prediction of seismic site
the context of validation, such changes in source modelling result response using 2D and 3D methods. Early examples that directly
in either direct validation improvements against data and/or the examined performance against observations include Thompson
need to adjust other heuristic parameters in order to avoid et al. (2009), who considered the effects of 3D stochastic
subtractive cancellation (i.e., there are many adjustable perturbations, and more recently de la Torre et al. (2021)
parameters that are set with limited constraints, and therefore generalized the exploratory study of Thompson et al. to propose
continual model improvements and validation result in iterative an approach that can be used based on available 1D geotechnical
refinement and constraint of such parameters). information. Additional challenges with the consideration of 2D
or 3D methods of site response analysis include: (i) the need for
4.2 robust 3D constitutive models, particularly for considering
effective stresses – for which the majority of models adopt a
3D crustal velocity models provide the 3D domain over which plane strain formulation; and (ii) the additional site
the wave equation is solved. Models of crustal velocity exist in characterization data that is required to be collected in order to
the majority of the world's seismic regions for the purpose of reasonable describe the spatial variation in soil properties over
earthquake location determination, and are often of insufficient the depths of importance for site response.
spatial resolution for their direct use in ground motion simulation.
This is because the latter involve both higher frequencies of
interest (thus requiring a finer spatial resolution), and also full 5 CONCLUSIONS
ground motion waveforms rather than just arrival times, for
which crustal impedance contrasts, particularly in the presence Earthquake-induced ground-motion simulation provides the
of sedimentary basins, have more significance. potential to significantly increase the accuracy and precision of
The explicit modelling of sedimentary basins in crustal predictions – with significant benefits to downstream
models is generally achieved via the use of geological surfaces engineering assessment and interventions toward resilience of
which are ‘embedded’ into regional 3D crustal, with the the built and natural environments. The increasing adoption of
sediments between each layer having their constitutive properties these simulation methods will be accelerated by a greater
assigned in a number of approaches (e.g., Thomson et al., 2020). emphasis on robust validation to illustrate their applicability on
However, arguably the most promising approach involves the a regional- and even site-specific basis, as validation with a large
direct use of recorded seismological data (either earthquake- number of events and observed ground motions ensures that
induced ground motions, ‘passive’ ambient vibrations, or ‘active’ iterative advancements in simulation methods and their
vibrations from geophysical experiments) to iteratively improve constituent models is more likely to converge toward the global
the crustal model using the full seismological waveform, so- maxima of prediction performance, rather than a local maxima.
called ‘full waveform tomography’ (Tape et al., 2010). The This paper has briefly highlighted on-going work, from the
principal benefit of full waveform tomography is that full ground author’s perspective, focused on validation, as well as the major
motion waveforms are used to invert the crustal model, which is challenges that remain over the coming years in order to make
the same features of the ground motion that are of interest in the appreciable advances in prediction performance. Much like the
forward problem. In contrast, the construction of crustal models ‘quiet revolution’ in weather forecasting – no single
through the use of geological surfaces and parameterised improvement is likely to be game-changing, but the cumulative
constitutive relations does not directly validate the resulting effect of many small advances combined with a robust and
model against seismic waveforms. comprehensive validation enterprise will ultimately yield results.
622
6 ACKNOWLEDGEMENTS Lee, R. L., Bradley, B. A., Stafford, P. J., Graves, R. W., and
Rodriguez-Marek, A., 2020. Hybrid broadband ground motion
This research is supported by QuakeCoRE: The New Zealand simulation validation of small magnitude earthquakes in
Centre for Earthquake Resilience. This is QuakeCoRE Canterbury, New Zealand. Earthquake Spectra, 36(2), 673–699.
Publication number 0742. SAGE Publications Ltd STM. DOI: 10.1177/8755293019891718
Paolucci, R., Mazzieri, I., and Smerzini, C., 2015. Anatomy of strong
ground motion: near-source records and three-dimensional physics-
7 REFERENCES based numerical simulations of the Mw 6.0 2012 May 29 Po Plain
earthquake, Italy. Geophysical Journal International, 203(3), 2001–
Baker, J., Bradley, B., and Stafford, P., 2021. Seismic Hazard and Risk 2020. DOI: 10.1093/gji/ggv405
Analysis. Cambridge: Cambridge University Press. Pitarka, A., Graves, R., Irikura, K., Miyakoshi, K., Wu, C., Kawase, H.,
Bauer, P., Thorpe, A., and Brunet, G., 2015. The quiet revolution of Rodgers, A., et al., 2022. Refinements to the Graves–Pitarka
numerical weather prediction. Nature, 525(7567), 47–55. DOI: Kinematic Rupture Generator, Including a Dynamically Consistent
10.1038/nature14956 7.1 Ridgecrest
Bradley, B. A., 2013. A New Zealand-Specific Pseudospectral Earthquake. Bulletin of the Seismological Society of America,
Acceleration Ground-Motion Prediction Equation for Active 112(1), 287–306. DOI: 10.1785/0120210138
Shallow Crustal Earthquakes Based on Foreign Models. Bulletin of Ramírez-Guzmán, L., Boyd, O. S., Hartzell, S., and Williams, R. A.,
the Seismological Society of America, 103(3), 1801–1822. DOI: 2012. Seismic Velocity Model of the Central United States
10.1785/0120120021 (Version 1): Description and Simulation of the 18 April 2008 Mt.
Bradley, B. A., 2019. On-going challenges in physics-based ground Carmel, Illinois, Earthquake. Bulletin of the Seismological Society
motion prediction and insights from the 2010–2011 Canterbury and of America, 102(6), 2622–2645. DOI: 10.1785/0120110303
2016 Kaikoura, New Zealand earthquakes. Soil Dynamics and
Earthquake Engineering, 124, 354–364. DOI: Validation of the 2008 Chino Hills, California, Earthquake Using
10.1016/j.soildyn.2018.04.042 Different Velocity Models. Bulletin of the Seismological Society of
Bradley, B. A., Pettinga, D., Baker, J. W., and Fraser, J., 2017. America, 104(4), 1876–1898. DOI: 10.1785/0120130266
Guidance on the Utilization of Earthquake-Induced Ground Motion Tape, C., Liu, Q., Maggi, A., and Tromp, J., 2010. Seismic tomography
Simulations in Engineering Practice. Earthquake Spectra, 33(3), of the Southern California crust based on spectral-element and
809–835. SAGE Publications Ltd STM. DOI: adjoint methods. Geophysical Journal International, 180, 433–462.
10.1193/120216eqs219ep Thompson, E. M., Baise, L. G., Kayen, R. E., and Guzina, B. B., 2009.
Bradley, B. A., Razafindrakoto, H. N. T., and Polak, V., 201 Impediments to Predicting Site Response: Seismic Property
Motion Observations from the 14 November 2016 Mw7.8 Estimation and Modeling Simplifications. Bulletin of the
Kaikoura, New Zealand, Earthquake and Insights from Broadband Seismological Society of America, 99(5), 2927–2949. DOI:
Simulations. Seismological Research Letters, 88(3), 740–756. DOI: 10.1785/0120080224
10.1785/0220160225 Thomson, E. M., Bradley, B. A., and Lee, R. L., 2020. Methodology
Dreger, D. S., and Jordan, T. H., 2015. Introduction to the Focus and computational implementation of a New Zealand Velocity
Section on Validation of the SCEC Broadband Platform V14.3 Model (NZVM2.0) for broadband ground motion simulation. New
Simulation Methods. Seismological Research Letters, 86(1), 15–16. Zealand Journal of Geology and Geophysics, 63(1), 110–127.
DOI: 10.1785/0220140233 Taylor & Francis. DOI: 10.1080/00288306.2019.1636830
de la Torre, C. A., Bradley, B. A., and Lee, R. L., 2020. Modeling
Mw 8.8 Maule, Chile, Ea nonlinear site effects in physics-based ground motion simulations
Subevents and Background Slip. Bulletin of the Seismological of the 2010–2011 Canterbury earthquake sequence. Earthquake
Society of America, 107(1), 372–386. DOI: 10.1785/0120160127 Spectra, 36(2), 856–879. SAGE Publications Ltd STM. DOI:
10.1177/8755293019891729
Simulations on Rough Faults Including Effects of 3D Stochastic de la Torre, C. A., Bradley, B. A., and McGann, C. R., 2021. 2D
Velocity Perturbations. Bulletin of the Seismological Society of Geotechnical site-response analysis including soil heterogeneity
America, 106(5), 2136–2153. DOI: 10.1785/0120160088 and wave scattering. Earthquake Spectra, 87552930211056670.
Graves, R. W., and Pitarka, A., 2010. Broadband Ground-Motion SAGE Publications Ltd STM. DOI: 10.1177/87552930211056667
Simulation Using a Hybrid Approach. Bulletin of the
Seismological Society of America, 100(5A), 2095–2123. DOI:
10.1785/0120100057
623
Proceedings of the 20th ICSMGE-State of the Art and Invited Lectures– Rahman and Jaksa (Eds)
© 2022 Australian Geomechanics Society, Sydney, Australia, ISBN 978-0-9946261-6-5
Soft peaty clay stabilization in Sri Lanka: sustainability through utilizing secondary
raw materials
Stabilisation de l'argile tourbée molle au Sri Lanka: durabilité grâce à l'utilisation de matières
premières secondaires
Ashani Ranathunga & Athula Kulathilaka
ABSTRACT: In situ deep mixing of soft peaty clay with a binder such as cement can improve strength and compressibility
characteristics significantly in a short period of time. However, due to the higher cost and environmental impact of cement as the
binder, reducing the cement content using secondary raw materials such as Calcium carbide residue (CCR) and Fly ash (FA) is
needed and is focused on this study. Improvements in strength for various proportions of CCR, FA and cement were assessed by
conducting unconsolidated undrained triaxial tests for short-term (28 days) and long-term (90 days) curing. Further, behaviour of
compressibility was investigated by conducting loading, unloading, and preloading one-dimensional consolidation test on oedometer.
Adding only cement has the best improvement while 10% CCR, 15% FA with 5% cement mixture showed the best results among
alternative mixtures. The CCR fixation point was 30% and was independent from time. Further, having similar proportions for FA
with CCR & cement mixtures provided better results. However, permitting pozzolanic reaction under an overburden stress can
noticeably improve the strength & stiffness of peaty clay.
RÉSUMÉ : Le mélange profond in situ d'argile de tourbe molle avec un liant tel que le ciment peut améliorer considérablement les
caractéristiques de résistance et de compressibilité en peu de temps. Cependant, en raison du coût plus élevé et de l'impact
environnemental du ciment en tant que liant, la réduction de la teneur en ciment à l'aide de matières premières secondaires telles que les
résidus de carbure de calcium (CCR) et les cendres volantes (FA) est nécessaire et se concentre sur cette étude. Les améliorations de la
résistance pour diverses proportions de CCR, FA et ciment ont été évaluées en effectuant des essais triaxiaux non consolidés non drainés
pour une cure à court terme (28 jours) et à long terme (90 jours). De plus, le comportement de la compressibilité a été étudié en effectuant
des tests de consolidation unidimensionnels de chargement, de déchargement et de préchargement sur un œdomètre. L'ajout uniquement
de ciment a la meilleure amélioration tandis que 10 % de CCR, 15 % de FA avec un mélange de ciment à 5 % ont montré les meilleurs
résultats parmi les mélanges alternatifs. Le point de fixation du CCR était de 30 % et était indépendant du temps. De plus, le fait d'avoir
des proportions similaires pour l'AF avec des mélanges de CCR et de ciment a donné de meilleurs résultats. Cependant, permettre une
réaction pouzzolanique sous une contrainte de mort-terrain peut sensiblement améliorer la résistance et la rigidité de l'argile tourbeuse.
KEYWORDS: calcium carbide residues, compressibility, fly ash, shear strength, soft peaty clay
1 INTRODUCTION compaction piles and crushed stone piles; and excavation and
removal have been used to enhance stiffness and shear strength
Rapid development in the construction industry has compelled of peaty clays (Kulathilaka 2015). These techniques involve
the geotechnical engineers to utilize lands underlain by weak consolidation of the peaty clay or transfer of loads to a stronger
soils forcing to improve strength and compressibility intrusion (granular columns). These additional steps in ground
characteristics of the soil. Lands underlain by thick layers of preparation leads to an extension of the duration of projects
peaty clay were often encountered in many new highway projects specially when it involves high embankments.
spanning across flood plains in the country forcing a viable Since the speed of construction is a major concern of current
solution for soil improvement. development projects, geotechnical engineers must look for rapid
Peaty clays encountered in flood plains of Sri Lanka have methods which are sustainable. Deep mixing with cement, lime
undergone complete degradation (classified as amorphous or with a mix of other industrial binders in-situ is a technique that
granular or H10 (Hobbs 1986)) are mixed with alluvial deposits. can be adopted to improve the engineering properties of soft
As such, the organic contents are in the order of 30-40%. They peaty clay in a short period of time.
are characterised by a very high-water contents (over 300%) and
therefore of a very high compressibility and very low shear 1.1 Deep mixing with cement
strength (as low as 2 kN/m2). Peaty clays possess very high
secondary consolidation settlements with the highest secondary The deep mixing process is beneficial over other techniques
compression index to compression index ratio (C /Cc) for all because when peaty clay is deep mixed in-situ with a binder like
geo-materials (Mesri 1997). cement, a hydration process (refer to Eq. 1) will occur followed
Therefore, to prevent possible shear failures, high road by a subsequent pozzolanic reaction which increases strength and
embankments constructed on sites underlain by peaty clays consolidation characteristics in a short period of time (around 28
should be done in stages (or sufficiently slowly) allowing for days).
strength gain with consolidation. The embankments are to be Cement + water CSH + CAH + CASH+ Ca(OH)2 (1)
preloaded during the construction to ensure that the settlements
in service are within acceptable limits. In the cement hydration process, hydrated calcium silicates
In Southern Expressway and Colombo Katunayake (CSH), hydrated calcium aluminates (CAH), hydrated calcium
Expressway projects, different techniques such as pre- aluminium silicates (CASH), hydrated lime [Ca(OH)2] are the
consolidation with preloading; pre-consolidation with vacuum main products (formation of primary cementitious materials).
consolidation and heavy tamping followed by preloading; sand There is an increase in pH value of the medium due to the
625
dissociation of Ca(OH)2 which let the silica and alumina from noticed that the plasticity index reduction is marginal at a certain
the soil to be dissolved (as an acid dissolves in a strong base). CCR content (7% for Thai silty clay) and denoted this point as
The hydrous silica and hydrous alumina then gradually react with the CCR fixation point which implies the maximum adsorption
Ca2+ ions to form secondary cementitious materials. Formation capacity of Ca+2 ions by the soil.
of cementitious materials which get hardened with time and The strength improvement in peaty clay with CCR can be
dissipation of pore water for the hydration process make the classified into three zones: active, inert, and deterioration
peaty clay to be stabilized rapidly (Horpibulsuk et al. 2012). (Horpibulsuk et al. 2012). In the active zone, strength increases
Dry cement powder (dry mixing) is used for improvement of with the CCR input and ends with CCR fixation point. Then the
organic clays of high initial water content in the field mixing inert zone starts with a slower strength development due to the
process. The deep mixing is conducted in the form of columns lack of natural pozzolanic materials in the soil to react with
(20 to 30 m depth) in a designed grid arrangement where these excess amount of CCR input. Adding another pozzolanic
columns may be overlapped to create walls or blocks. material rich source like FA, in this zone can accelerate the
Alternatively, the complete mass of peat also can be stabilized process again. In the deterioration zone strength decreases due to
over a shallow depth according to the availability of machinery the loss of pozzolanic materials and unsoundness caused by
(Porbha, Tanaka and Kobayaska 1998). excess free lime (Kampala & Horpibulsuk, 2013). Hence, using
Simple machineries were developed in Sweden by modifying CCR and FA together should provide better strength
a crawler mounted crane and mass stabilization of thick peat improvement in peaty clay. Vitharana (2019) studied the use of
layers were carried out successfully for the construction of CCR with FA for stabilization of peaty clay and the results are
highways and railways (Jelisic 2004). Machinery of this type can shown in the Figure 1. However, according to the figure, the
be developed locally, and the process of deep mixing would be improvement of strength of the stabilized soil samples were not
practically viable. As such, mixing with cement or lime could be sufficient for the field applications. Hence, Vitharana (2019)
a plausible method for improvement of soft peaty clay layers proposed to use a smaller amount of cement with CCR and FA
encountered in Sri Lankan Road projects. to improve soft peaty clay as a partial replacement rather than the
However, laboratory studies done at University of Moratuwa, full replacement of cement.
Sri Lanka have shown that a cement content of the order of 20-
25% is required to achieve a desired level of improvement in
organic clay with properties ranging between 260 – 600% of CCR20+FA20
moisture content, 20 – 40% organic content, 3 – 4 pH value and CCR20+FA10
1.5 to 2.1 specific gravity (Vitharana 2019). This is comparable CCR15+FA20
with the cement weights of 200-250 kg per cubic meter of CCR15+FA10 after 90 days
organic clay reported in literature (Kulathilaka 2015). Hence,
CCR10+FA20 after 28 days
using cement is not economically and environmentally feasible
CCR10+FA10
for Sri Lankan organic clay stabilization. Hence, the use of
secondary raw materials for full/partial replacement for cement CCR20
to reduce the environmental impact and higher costs is essential CCR15
CCR10
1.2 Deep mixing with Fly ash (FA) and Calcium Carbide Cement20
Residues (CCR) Soil only
Fly ash (FA) is one such conventional secondary raw material 0 10 20 30 40
produced at coal power plants and is successfully used as a
Shear strength, Cu (kN/m2)
cementitious material in many countries. FA mainly includes
SiO2, Al2O3, and Fe2O3 which makes FA a pozzolanic material. Figure 1. Strength of peaty clay samples after 28 days of curing
However, Sri Lanka has class “F” FA which is low in calcium (Vitharana 2019)
(~6%) (Vitharana 2019). Therefore, Sri Lankan FA needs to be
mixed with a high calcium additive such as Calcium Carbide Thus, the main objective of this study was to investigate the
Residues (CCR) to generate better pozzolanic reactions. applicability of CCR and FA as a partial replacement of cement
CCR is a by-product of acetylene gas production factories, during dry mixing to improve the shear strength of soft peaty
and it is a waste product in the industry derived as shown in Eq. clay.
2.
2 METHODOLOGY
CaC2 + 2H2O C2H2 + Ca(OH)2 (2)
2.1 Materials
Ca(OH)2 is disposed in slurry form and after being sun dried
Peaty clay samples were extracted from the Outer Circular
it becomes dry form which is known as CCR. 64 g of calcium
Expressway at Kerawalapitiya, Sri Lanka. Debris were removed
carbide (CaC2) provides 26 g of Acetylene gas (C2H2) and 74 g
and a homogeneous sample was prepared. The basic properties
of CCR, in terms of Ca(OH)2 that is an abundance as a raw
of peaty clay are presented in Table 1. Soil samples were mixed
material (Horpibulsuk et al. 2012). There is an increasing
with CCR (from Ovin Gas (pvt.) Ltd, Homagama) and FA (from
demand to find a way to dispose CCR since it is challenging to
Lakvijay Power Plant, Norochcholai) on a wet weight basis as in
dispose to the nature being a highly basic material. Therefore,
Table 2. The chemical composition of FA and CCR used are in
utilization of CCR as a secondary raw material for stabilization
Table 3 and the particle size distribution curves are shown in
of peaty clay will be a sustainable solution.
Figure 2. The moisture content of CCR was 42.14% (CCR was
Horpibulsuk et al. (2012), studied the change in index
air-dried before use) and the specific gravity was 1.875. The
properties of the CCR stabilized silty clay from Thailand and
moisture content of FA was 0.15% while the specific gravity was
observed a considerable increase in the plastic limit and a smaller
2.270. The total additive percentage was kept at 30%. Five
reduction in liquid limit when the CCR content is increased in
percent cement content was selected as a fixed amount and
the clay samples. Hence, increase CCR content caused a
remaining 25% was replaced with CCR and FA. Further, three
reduction of the plasticity index which caused by the adsorption
samples were prepared with 15%, 30% and 45% of CCR to
of Ca+2 ions from the cation exchange process indicating the
investigate the CCR fixation point of the soil.
flocculation of clay particles. However, Horpibulsuk et al. (2012)
626
Table 1. Basic Properties of Peaty clay a surcharge of 10 kN/m2 was applied to simulate the hardening
Basic Property Value of the soil mix under its own weight in the field (Figure 4). The
Moisture Content (%) 614.12 samples were cured for 28 days and 90 days separately for short-
Bulk Density (kN/m3) term and long-term stabilization respectively.
10.75
After curing, undisturbed samples were obtained and
Specific Gravity (Gs) 1.514 Unconsolidated Undrained Triaxial test was conducted
pH Value 6.0 according to ASTM D2850 standards under 50 kPa, 100 kPa and
Organic Content (%) 150 kPa cell pressures to investigate shear strength parameters of
39.64 stabilized peaty clay samples.
Fibre Content (%) 27 One-Dimensional Consolidation test was carried out to
Ash Content (%) 59.5 evaluate the improvements achieved in the compressibility
characteristics. The Oedometer apparatus was used to conduct
Initial Void Ratio 5.44 the experiment according to the ASTM – D 2435 standard test
method. The test was done with both loading, unloading and
Table 2. Mix Proportions reloading stages on the undisturbed samples extruded from the
Additives % by weight buckets. In the loading stage, started from 5 kN/m2 stress and
Combination
CCR FA Cement increasing to 10, 20, 40, 80 kN/m2 stresses respectively. Then
Peat only - - - unloading was done for stress levels of 40, 20 and up to 10
Peat + Cement kN/m2. After that reloading was started from 20 to 160 kN/m2.
- - 30
The loading duration was one day for all stages. Stress levels of
Peat + CCR + FA + Cement 20 5 5 these magnitudes were selected because these soft peaty clay
15 10 5 layers are encountered near the surface levels and are of
maximum thicknesses in the range of 10-12 m.
10 15 5
5 20 5
Peat + CCR + FA 15 15
Peat + CCR 15, 30, 45
Table 3. Chemical Composition of FA and CCR
Material Composition in FA (%) Composition in CCR (%)
SiO2 45 6.49
Al2O3 31.8 2.55
Fe2O3 4.4 3.25 Figure 3. Mechanical Figure 4. Loading Mechanism
mixer
CaO 9 70.78
MgO 1.1 0.69 3 RESULTS AND DISCUSSIONS
K 2O - 7.93
SO3 0.66
3.1 Optimum CCR content
0.5
LOI 3.7 1.35 Similar to the observations by Horpibulsuk et al. (2012) for Thai
silty clay, it was be interesting to check the behaviour of CCR for
Sri Lankan peaty clays. However, O’Kelly (2015) concluded that
100 the Atterberg limits of peat are not reliable indicators of its
consistency and natural water content, organic content, fibre
80
content and degree of humification of peat are more suitable for
assessing its likely engineering behaviour. Hence, for this study,
% Passing
60
the water content and the organic content of the peaty clay
CCR
40 samples were used for investigating the optimum CCR content.
FA
Here, water and organic content of both short-term (28 days) and
20 long-term (90 days) curing were used for comparison and are
shown in Figure 5.
0
0.001 0.01 0.1 1 10 Moisture content @ 28 days Moisture content @ 90 days
Particle Diameter (mm) Organic content @ 28 days Organic content @ 90 days
Figure 2. Particle Size Distribution of FA and CCR 500 50
400 40
2.2 Laboratory Simulation of the deep mixing process
Moisture content (%)
Organic content (%)
300 30
The process of complete liquefaction of the in-situ soil and
mixing with stabilizers was simulated in the laboratory by mixing 200 20
the soil with dry stabilizers with the help of a mixing tool.
Peaty clay was mixed with cement, CCR and FA according to 100 10
CCR fixation point
different mix proportions by weight (Table 2). The mixing
0 0
percentages are the percentage weight of each additive out of the 0 10 20 30 40 50
weight of the peaty clay used for the sample. Mixing was done CCR content (%)
with an aid of a mechanical mixer (Figure 3) to simulate the Figure 5. Moisture content and organic content of the treated peaty clay
process of complete liquefaction of the in-situ soil and mixing samples
with dry stabilizers. Prepared samples were kept in buckets and
627
According to the Figure 5, the optimum CCR content where from 18% to 40% which is much higher than with only CCR.
the maximum adsorption of Ca+2 ions to the peaty clay occurs This explains the potential to improve the strength of the peaty
when CCR content is 30%. Because, both moisture content and clay with time when FA and/or cement is available which
the organic content of the treated peaty clay samples has minimal provides additional pozzolanic material for the excess lime
changes after 30% of CCR content. Interestingly, this CCR generated by CCR to react.
fixation point does not change with time as both 28 days curing, This can be further explained by the shear strength increment
and 90 days curing indicates similar behaviour. Hence, it with the FA content of the sample with 15% CCR and 15% FA.
concludes that the CCR fixation point is independent from the Addition of 15% FA has increased the shear strength by around
curing time. However, this observation should be confirmed with 6% and 14% in 28 days and in 90 days consequently compared
more experiments. to only 15% CCR sample. This observation further confirms the
Strength behaviour of the short-term and long-term stabilized importance of having rich aluminosilicates as secondary raw
samples is represented in Figure 6. The graphs were extended materials in the binder for continuation of pozzolanic reactions
using best fit (R2 > 0.99) polynomial trend lines to clearly for longer-durations.
observe the inert and the deterioration zones. Similar to the Further, the strength has been increased with the increasing
findings in literature, shear strength of peaty clay has improved FA content up to 15% and then has reduced for 20% of FA when
rapidly at the beginning. Because initially, pozzolanic reaction 5% cement is added to the samples. For 20% FA, there is only
occurs rapidly and makes the cementitious material, which 5% cement and 5% CCR and the amount was not sufficient to
densify the peaty clay by filling pores in it. Next, an inert region generate enough pozzolanic reactions to strengthen the soil
is displayed, where increment slows down as the pozzolanic matrix. The maximum strength additive combination has an
material in peaty clay gets depleted. However, further increase in equal amount of FA (15%) and CCR (10%) with cement (5%).
CCR%, caused the shear strength to decline due to the excess Further, only this combination was able to surpass the 30% CCR
lime created. strength value during short-term curing (refer to Figure 7).
Compared to short-term, long-term curing exhibits around However, addition of 5 to 15% of FA to the mixture could exceed
3.5% increment in shear strength of peaty clay samples. In the 30% CCR strength value after 90 days curing. This
addition, the active region of both short-term and long-term observation suggests using an equal amount of FA with CCR and
curing falls up to 30% CCR. Conversely, long-term curing cement mixtures for better results in short-term.
displays a longer inert region (30% - 60%) while that for short-
term curing is comparatively shorter (30% - 45%). These results
30
indicate that long-term curing allows more room for the Short-term Long-term
Shear strength Cu, (kN/m2)
28.2
flocculation of peaty clay particles to densify the soil matrix. 25
20.2
30% CCR strength reference line
Therefore, in field application, we can expect higher strength
16.0
20
15.0
improvement in peaty clay with time. However, a similar
14.3
13.7
13.5
12.7
12.2
12.2
12.1
optimum CCR content of 30% was observed for both short-term 15
10.1
10.0
and long-term curing similar to the results of CCR fixation point
8.0
10
in Figure 5, indicating that the maximum adsorption capacity of
Ca2+ ions by soft peaty clay will not change with time for curing. 5
0
short term long term
16
y = -0.0019x2 + 0.1763x + 9.945
Additive Combination
14 R² = 0.994
Figure 7. Strength variation of soil samples with CCR, FA and
Shear strength Cu, (kN/m2)
12 cement
10 y = -0.0037x2 + 0.287x + 8.005
R² = 1
In addition, Horpibulsuk et al. (2012) observed that addition
8 of FA for silty clay stabilization was more effective when FA is
added excess to the CCR fixation point. However, this
6
Active* Inert* Deterioration
observation was not applicable for the present study as the CCR
4 * fixation point was 30% and the total additive mixture was
maintained at 30% considering the field application which
2
Optimum CCR content warrants future studies.
0
0 20 30% 40 60 80 100 3.3 Compressibility of peaty clay with secondary raw materials
CCR content %
The preloading process was simulated by conducting Oedometer
Figure 6. Shear Strength variation with CCR (*shaded areas show the
tests with loading, unloading and reloading increments. The
different stages of strength variation for long-term curing)
specimen in the reloading increments resembles a preloaded soil
(at a stress level of 80 kN/m2). The coefficient of volume
3.2 Strength behaviour of peaty clay with CCR, FA and cement compressibility (m) and the coefficient of secondary
Second objective was to investigate the best mixture of CCR and consolidation (C ) were evaluated in all the loading and
FA as a partial replacement of cement which gives the maximum reloading increments. The values of compression index (Cc) and
improvement in shear strength of soft peaty clay samples and the recompression index (Cr) were also evaluated using the e vs log
shear strength behaviour of the soil samples are shown in Figure plot.
7.
According to Figure 7, maximum improvement is achieved 3.3.1 Compressibility of peaty clay with CCR
when 30% cement is added to the peaty clay. Among the other Figure 8 shows the compressibility behaviour of peaty clay after
additive combinations, CCR 10%, FA 15% and 5% cement gave addition of CCR and 30% cement. According to the figure,
the highest shear strength. However, it is almost half of the shear addition of CCR has improved both primary and secondary
strength obtained for 30% cement. In addition, when the curing consolidation of peaty clay for all the CCR contents. However,
period extends, all the soil samples showed an increment ranging 30% cement shows the best improvement similar to the
observations in shear strength probably due to the availability of
628
higher aluminosilicates in the media to develop more Table 4. Compression and recompression index
cementitious products. Interestingly, still CCR could provide a Cc/(1+eo) Cr/(1+eo)
better improvement than the preloaded natural peaty clay (refer Sample
28 days 90 days 28 days 90 days
to the gradients of natural peat in Figure 8(a), (b), (c) and (d)).
Peaty clay 0.627 0.328 0.091 0.065
CCR 15% 0.048 0.035 0.012 0.009
0.0030 0.0015
CCR 30% 0.041 0.032 0.007 0.006
CCR 45% 0.039 0.021 0.005 0.004
0.0020 0.0010
CCR 20% + FA 5% + cement 5% 0.197 0.196 0.005 0.004
mv
mv
CCR 15% + FA 10% + cement 5% 0.192 0.094 0.005 0.004
0.0010 0.0005
CCR 10% + FA 15% + cement 5% 0.074 0.049 0.003 0.002
CCR 5% + FA 20% + cement 5% 0.293 0.218 0.007 0.006
0.0000 0.0000
1 10 100 1000
1 10 100 1000 CCR 15% + FA 15% 0.192 0.139 0.006 0.0003
log ( )/ (kN/m2) log ( ) (kN/m2)
C 0.038 0.001 0.003 0.0003
(a) mv during loading (b) mv during reloading
0.3000
0.3000 0.2000
30% CCR C Untreated peat
reference line
0.2000 0.1500 CCR20+FA5+Cement5
0.2000
CCR15+FA10+Cement5
C
C
0.1000
0.1000 0.1000 CCR10+FA15+Cement5
C
0.0500 CCR5+FA20+Cement5
0.0000 0.0000
1 10 100 1000 1 10 100 1000 Cement30
log ( ) (kN/m2) log ( ) (kN/m2) 0.0000
1 10 100 1000
(c during loading (d during reloading log ( ) (kN/m2)
Figure 9 reloading for treated peaty clay with CCR,
5.00
FA and Cement
4.50
4.00
3.4 Changes in the microstructure
Void Ratio
3.50
3.00 Microstructure of the stabilized samples were investigated
2.50 through scanning electron microscope (SEM) and are shown in
2.00
Figure 10. Irregular shape of peaty clay particles and the voids
1.50
among particles are clearly depicted in the figure. When the
1 10
Log( ) /(kN/m2)
100 1000 cement or CCR is mixed with soil, a rod-shaped structure
Untreated peat CCR15 CCR30 (Ettringite) can be observed which makes a more aggregated and
CCR45 Cement30
bonded structure hence increases the shear strength. The
spherical shape of FA allows peaty clay to densify more than the
Figure 8. Compressibility behaviour of CCR treated peaty clay CCR with proper packing. Due to the high specific gravity of FA,
it allows peaty clay to be well compact and thus in-crease the
shear strength. However, the soil matrix with 30% cement is
3.3.2 Compressibility of peaty clay with CCR, FA and more densified and has higher ettringite formations compared to
cement other mixtures. Hence, a higher cement percentage than 5%
Table 4 summarises the compression and recompression index should be added for better improvement.
obtained after 28 days and 90 days curing of treated peaty clay
samples. As observed before, 30% cement provides the best
improvement among all the combinations. However, when equal (a) (b)
percentages of FA with CCR and cement is used (15% FA+10%
CCR+5% cement), the improvement is higher than the
improvement by the preloading on natural peaty clay.
Nevertheless, all the additives could improve the compressibility
characteristics of peaty clay and this performance turns better
with time for all cases.
lower (all were less than 0.01) than the values reported for (c) (d)
naturally occurring peaty clay layers in Sri Lanka (which are in
the order of 0.02 – 0.06, Karunawardena (2000)) (refer to Figure
9). Further, like strength behaviour, higher improvement in
compressibility can be observed when equal amounts of CCR
and cement with FA are mixed compared to the compressibility
of soft peaty clay with optimum CCR content (Figure 9).
Comparatively, more significant improvements are visible for
secondary consolidations and is important for long-term stability Figure 10. SEM images of (a) peaty clay, (b) peaty clay with
of highway embankments built on soft peaty clays (refer to 10% CCR 15% FA and 5% cement, (c) 30% cement and (d) 30%
Figure 8 and 9). CCR (magnification of 3k and voltage of 10 kV)
629
4 COMPARISON WITH PREVIOUS STUDIES For better improvements, using similar proportions of FA
with CCR and cement mixtures is suggested.
Results of the present study were compared with Vitharana
(2019) who have used CCR, FA and cement as stabilizers with a
surcharge of 10 kN/m2 and are shown in Table 5. Both studies 3.2
have similar organic content whereas, moisture content of 3.0
present study’s soil is twice as for Vitharana (2019)’s study. peat +
2.8 CCR15% +
When peaty clay is treated only with cement, present study has
void ratio
FA10% +
around 34% lesser strength improvement despite using 10% 2.6 cement 5% -
current study
higher cement for mixing. Thus, it further confirms the previous 2.4 peat +
observation of the moisture content playing a vital role in the CCR15% +
2.2 FA10% -
increment of the shear strength of peaty clay. Furthermore, CCR previous study
fixation point follows a similar pattern as for moisture content 2.0
1 10 100 1000
where the value is doubled in the present research. 15% CCR
Log( ) /(kN/m2)
addition provides similar strength for both soils in short-term.
Whereas the long-term improvement of strength is hindered in Figure 11. Comparison of the effect of cement on soft peaty clay
present study as the Ca+2 ions absorption levels are still halfway improvement with present study and Elpitiya (2019)
(optimum CCR content is 30%) and/or lack of aluminosilicates
in soil compared to Vitharana (2019)’s study. 6 ACKNOWLEDGEMENTS
However, present study shows a higher improvement for
longer durations when cement is added compared to the previous Authors acknowledges T.H. Vitharana, L.B.D.L. Jayathilake,
study. For example, case (i) & case (ii) shows around 12% higher K.D. Elpitiya, T.G.T. Nanjitha, A.K.G.M. Jayamal and R.M.S.U.
increment in strength from 28 days to 90 days compared to the Rathnayake for their contribution in conducting the laboratory
previous study. We can conclude that adding cement provides experiments. The assistance given by the Geotechnical
more pozzolanic materials to continue the strengthening in the Engineering and Materials Science and Engineering laboratory
long run. Further, when comparing the effect of cement on staff of University of Moratuwa are acknowledged.
compressibility characteristics of peaty clay, Figure 11 reveals
the positive influence of cement when added to the same CCR 7 REFERENCES
and FA mixture by improving the compressibility characteristics
of soil. Hence, partial replacement of cement has more Elpitiya, K.D. (2019), Compressibility Behaviour in Soft Peaty Clays
advantages than for the full replacement with secondary raw Stabilized with Calcium Carbide Residues and Fly Ash, Bachelor’s
Dissertation, University of Moratuwa, Sri Lanka.
materials for peaty clay. However, the strength improvement Hobbs, N. B. (1986). Mire morphology and the properties and behaviour
observed here are still lower than some of the records in literature of some British and foreign peats. Quarterly Journal of Engineering
(Jelisic 2002, Kulathilaka 2015) with cement/lime. Hence, a Geology and Hydrogeology, 19(1), 7-80.
reasonable overburden stress should be provided with the Horpibulsuk, S., Phetchuay, C., & Chinkulkijniwat, A. (2012). Soil
Pozzolanic reaction for higher improvements in strength and stabilization by calcium carbide residue and fly ash. Journal of
stiffness. materials in civil engineering, 24(2), 184-193.
Kampala, A., & Horpibulsuk, S. (2013). Engineering Properties of Silty
Table 5. Comparison with previous studies Clay Stabilized with Calcium Carbide Residue. Journal of Materials
in Civil Engineering, 25(5), 632–644.
Reference Vitharana (2019) Present Study Karunawardena, W. A. (2000). A study of consolidation characteristics
Moisture% 300% 614% of Colombo Peat. In Proceedings of 1st International Young
Properties
of Peaty
Geotechnical Engineering Conference, Southampton, UK.
clay
Organic% 40.5% 39.6% Kulathilaka, S. A. S. (2015). Experiences on Improvement of Soft Peaty
Gs 1.8 1.5 Clays in Sri Lanka. ACEPS 2015, 10.
4.5 Menard Jelisic (2002), Stability and Settlement in Mass Stabilized Peat,
Soil only 8
Doctoral Thesis at Lund University, Sweden.
Cement 20 30 Mesri, G., Stark, T. D., Ajlouni, M. A., & Chen, C. S. (1997). Secondary
Shear Strength (kN/m2)
Additive combination (%)
(C) S-30 L-32 S-20 L-28.2 compression of peat with or without surcharging. Journal of
Geotechnical and Geoenvironmental Engineering, 123(5), 411-421.
CCR15+FA10 CCR15+FA10+C5 O’Kelly, B. C. (2015). Atterberg limits are not appropriate for peat soils.
(i)
S-15 L-17 S-12 L-15 Geotechnical Research, 2(3), 123-134.
Porbha A, Tanaka H and Kobayaska M, (1998), State of art in Deep
CCR CCR10+FA20 CCR10+FA15+C5 mixing Technology: Part II Applications, Journal of Ground
(ii)
/FA S-15 L-16 S-13.5 L-16 Improvement Techniques pp 125-139.
Vitharana, T.H. (2019), Investigation of Strength Behaviour in Soft Peaty
CCR15 CCR15 Clays Stabilized with Calcium Carbide Residues and Fly Ash,
(iii)
S-11 L-17 S-11.5 L-12 Bachelor’s Dissertation, University of Moratuwa, Sri Lanka.
CCR Fixation point 15% 30%
S – Short-term (28 days); L – Long-term (90 days)
5 CONCLUSIONS & RECOMMENDATIONS
Partial replacement of cement with CCR and FA
combination has shown considerable development in
strength and stiffness and can be increased by application of
higher overburden stress on peaty clay.
The strength is increased with time, whereas CCR fixation
point is independent of time.
Variation of moisture content plays a main role in the
stabilisation of peaty clay.
630
Authors Index
Abdelaal, F.B. 187 Kim, S. 95
Aftab, A. 95 Kortbawi, M.E. 405
Asan-Mangold, V. 567 Kulathilaka, A. 625
Banerdt, W.B. 567 Kwon, T.-H. 405
Benson, C.H. 187 Lacasse, S. 543
Bjelkevik, A. 121 Lange, L. 567
Bouazza, A. 187 Lee, J.Y. 95
Boulanger, R.W. 351 Lee, M. 405
Bradley, B.A. 619 Legge, K. 187
Briaud, J.-L. 511 Lei, L. 95
Brinkman, N. 567 Lemmon, M. 567
Caicedo-Hormaza, B. 567 Li, J. 535
Castillo-Betancourt, J.P. 567 Liu, H. 561
Chen, L. 535 Liu, Z. 543
Correia, A.G. 41 Lognonné, P. 567
Day, P. 511 Manassero, M. 187
DeJong, J.T. 405 Marinho, F.A.M. 297
Delage, P. 567 Marteau, E. 567
Dominijanni, A. 187 Montoya, B. 405
Dusseault, M. 95 Mueller, N. 567
Edme, P. 567 Murdoch, N. 567
Emidio, G.D. 187 Narsilio, G. 95
Espinoza, D.N. 95 Nelson, D.C. 405
Evans, J.C. 187 Ng, C.W.W. 535
Fan, X. 585 Onodera, K. 567
Fourie, A. 121 Palmeira, E.M. 187
Fratalocchi, E. 187 Parra, D. 187
Gardoni, M.G. 187 Pereira, J.-M. 95
Gens, A. 3, 95 Piqueux, S. 567
Golombek, M.P. 567 Powrie, W. 41
Gomez, M.G. 405 Ranathunga, A. 625
Graddy, C.M.R. 405 Robertsson, J. 567
Grott, M. 567 Romero, E. 297
Hoteit, H. 95 Rowe, R.K. 187
Hurst, K. 567 San Pablo, A.C.M. 405
Iai, S. 381 Sanchez, M. 95
Indraratna, B. 41 Santamarina, J.C. 95
Jefferis, S. 187 Scalia IV, J. 187
Katsumi, T. 187 Scheirs, J. 187
Kedar, S. 567 Schmelbach, C. 567
Kelly, R. 41 Shackelford, C.D. 187
Khalili, N. 297 Shen, Y. 535
Kikuchi, Y. 41 Shimizu, N. 597
631
Soga, K. 95 Vrettos, C. 567
Sollberger, D. 567 Wang, J. 535
Spohn, T. 567 Warner, N. 567
Stähler, S. 567 Widmer-Schnidrig, R. 567
Tobita, T. 381 Williams, N.R. 567
Torres-Cruz, L.A. 121 Xie, G. 561
Tutumluer, E. 41 Zapata, C. 41
Verdier, N. 567 Zhan, T.L.T. 187
Verdugo, R. 121 Zhang, W. 561
Villar, M. V. 95 Ziotopoulou, K. 405
Violay, M. 95 Znidarcic, D. 121
632
You might also like
- Discrete-Time Markov Chains: He ShuangchiNo ratings yetDiscrete-Time Markov Chains: He Shuangchi61 pages
- Pile Load Test Frequently Asked Questions100% (1)Pile Load Test Frequently Asked Questions15 pages
- Konstanta Pegas Dan Interaksi Struktur TanahNo ratings yetKonstanta Pegas Dan Interaksi Struktur Tanah10 pages
- Konstanta Pegas Dan Interaksi Struktur TanahNo ratings yetKonstanta Pegas Dan Interaksi Struktur Tanah10 pages
- Ground Improvment On Reclaimed Land - Indonesian ExperiencesNo ratings yetGround Improvment On Reclaimed Land - Indonesian Experiences22 pages
- GTL SoftSoils Bandung FindingSuitableModelNo ratings yetGTL SoftSoils Bandung FindingSuitableModel7 pages
- 2018 Book ProceedingsOfGeoShanghai2018InNo ratings yet2018 Book ProceedingsOfGeoShanghai2018In790 pages
- Plaxis-Triaxial Test With Hardening Soil Model100% (1)Plaxis-Triaxial Test With Hardening Soil Model4 pages
- Terowongan Galian Dalam Tanah Lunak: Aplikasi Untuk Dan DiNo ratings yetTerowongan Galian Dalam Tanah Lunak: Aplikasi Untuk Dan Di24 pages
- ASCE7 16ground motionScalingProcedureForNonlinearAnalysisOfBuildingsInComparisonWithASCE7 10AndASCE7 0No ratings yetASCE7 16ground motionScalingProcedureForNonlinearAnalysisOfBuildingsInComparisonWithASCE7 10AndASCE7 034 pages
- Evaluasi Status Degradasi Lahan Dataran Tinggi Akibat Produksi Biomasa Di Kabupaten Probolinggo, Jawa TimurNo ratings yetEvaluasi Status Degradasi Lahan Dataran Tinggi Akibat Produksi Biomasa Di Kabupaten Probolinggo, Jawa Timur9 pages
- Sahu & Mahanty, 2021. Soil Health and SustainabilityNo ratings yetSahu & Mahanty, 2021. Soil Health and Sustainability22 pages
- 35 Years of (AI) in Geotechnical Engineering State of The ArtNo ratings yet35 Years of (AI) in Geotechnical Engineering State of The Art54 pages
- Review Article: Impacts of Land Use and Land Cover Change On Soil Erosion and Hydrological Responses in EthiopiaNo ratings yetReview Article: Impacts of Land Use and Land Cover Change On Soil Erosion and Hydrological Responses in Ethiopia10 pages
- Investigations of Pile-Soil Behaviour, With Special Reference To The Foundations of Offshore Structures. RichardJardine-1986-PhD-ThesisNo ratings yetInvestigations of Pile-Soil Behaviour, With Special Reference To The Foundations of Offshore Structures. RichardJardine-1986-PhD-Thesis393 pages
- Section 3.9 Governing Flow Equation For Consolidation AnalysisNo ratings yetSection 3.9 Governing Flow Equation For Consolidation Analysis15 pages
- Section 2.1 Dry Vs Saturated Soils (General Modeling Considerations)No ratings yetSection 2.1 Dry Vs Saturated Soils (General Modeling Considerations)9 pages
- 2 - Ce6101 An Introduction To Plasticity-20092020No ratings yet2 - Ce6101 An Introduction To Plasticity-2009202088 pages
- Shallow Foundations - : Bearing Capacity and Limit AnalysisNo ratings yetShallow Foundations - : Bearing Capacity and Limit Analysis40 pages
- Solution Pipes and Focused Vertical Water Flow - Geomorphology and Modelling - CompressedNo ratings yetSolution Pipes and Focused Vertical Water Flow - Geomorphology and Modelling - Compressed17 pages
- Feasibility For Nemakonde Water Conveyence-For Tsuro - v10No ratings yetFeasibility For Nemakonde Water Conveyence-For Tsuro - v1028 pages
- International Society For Soil Mechanics and Geotechnical EngineeringNo ratings yetInternational Society For Soil Mechanics and Geotechnical Engineering9 pages
- Numerical Analysis of The Reinforced StoNo ratings yetNumerical Analysis of The Reinforced Sto8 pages
- Brinkgreve Et Al 2023 - Automated CPT Interpretation and Modelling in A BIM-Digital Twin EnvironmentNo ratings yetBrinkgreve Et Al 2023 - Automated CPT Interpretation and Modelling in A BIM-Digital Twin Environment7 pages
- Application of Finite Element and Constitutive Models100% (1)Application of Finite Element and Constitutive Models116 pages
- Current and Future Trends of Geotechnical Engineering in MalaysiaNo ratings yetCurrent and Future Trends of Geotechnical Engineering in Malaysia11 pages
- Numerical Analysis in Geotechnical Engineering 1No ratings yetNumerical Analysis in Geotechnical Engineering 120 pages
- 8 - Design and Construction of The New NLB BuildingNo ratings yet8 - Design and Construction of The New NLB Building15 pages
- Hulme 2023 Win T316 Tunnelling in AirportNo ratings yetHulme 2023 Win T316 Tunnelling in Airport13 pages
- 366-Mass Modulus of Rock For Use in The DesignNo ratings yet366-Mass Modulus of Rock For Use in The Design6 pages
- CBSE Class 6 Geography Notes Chapter 6 Major Landforms of The EarthNo ratings yetCBSE Class 6 Geography Notes Chapter 6 Major Landforms of The Earth3 pages
- Burning Environmental Issues of Assam or North-East100% (6)Burning Environmental Issues of Assam or North-East15 pages
- Probabilistic Assessment of Stability of A Cut Slo100% (1)Probabilistic Assessment of Stability of A Cut Slo9 pages
- NG Leung 2007 Wall and Ground Movements Associated With Deep Excavations Supported by Cast in Situ Wall in Mixed GroundNo ratings yetNG Leung 2007 Wall and Ground Movements Associated With Deep Excavations Supported by Cast in Situ Wall in Mixed Ground15 pages
- Dickson Et Al. - 2016 - Coastal Cliff LandslideNo ratings yetDickson Et Al. - 2016 - Coastal Cliff Landslide11 pages
- 142 Piling 2020 Paper Ciantia Et Al RevisedNo ratings yet142 Piling 2020 Paper Ciantia Et Al Revised7 pages
- PGC Orientation Briefing As of 8 Aug 2023No ratings yetPGC Orientation Briefing As of 8 Aug 202315 pages
- 9-4 Keynote 5 - Ir. Dr. Toh Cheng Teik - Ground Movements Due To Bored Piles and Basement Construction - 1No ratings yet9-4 Keynote 5 - Ir. Dr. Toh Cheng Teik - Ground Movements Due To Bored Piles and Basement Construction - 16 pages
- Application of Finite Element Analysis in Geotechnical Engineering by Sallam (2009)No ratings yetApplication of Finite Element Analysis in Geotechnical Engineering by Sallam (2009)3 pages
