Urban site investigation

C. R. I. Clayton

School of Civil Engineering and the Environment, University of Southampton, Southampton,

SO17 1BJ, UK
(e-mail: [email protected])

Abstract: This paper provides a written version of a Keynote Presentation made at the 10th International Congress of the
Association for Engineering Geology and the Environment (IAEG), held in Nottingham in September 2006. It identifies the
objectives of site investigation, as stated in the British Standard. It then reviews the difficulties that face the investigator,
and identifies factors that have in the UK been found to have an influence on the success of site investigations. Developments
in general UK ground investigation strategy are outlined. The challenges of urban site investigations are identified, as is the
vulnerability of urban development. Recent advances in investigation techniques are described, and conclusions drawn.

As noted in the introduction to BS5930 (BSE 1999), site (2) Design. To allow an adequate and economic design to
investigations are widely seen as ‘an essential preliminary be prepared, including the design of temporary works.
to the construction of all civil engineering and building (3) Construction. To plan the best method of construction;
Works’. The unforeseen ground conditions that frequently to foresee and provide against difficulties and delays that may
occur during construction, and that can lead to very large arise during construction as a result of ground, groundwater
cost and construction time over-runs, have frequently been and other local conditions; in appropriate cases, to explore
blamed on inadequate site investigations (e.g. Rowe 1972). sources of indigenous materials for use in construction; and
In the first part of this paper, recent developments in UK to select sites for the disposal of waste or surplus materials.
site investigation practice are reviewed. In the second part, (4) Effect of changes. To determine the changes that may
some of the special challenges associated with site investi- arise in the ground and environmental conditions, either
gation for urban redevelopment are discussed. naturally or as a result of the works, and the effect of such
changes on the works, on adjacent works, and on the
environment in general.
(5) Choice of site. Where alternatives exist, to advise on
Recent developments of site investigation the relative suitability of different sites, or different parts
in the UK of the same site.
(6) Existing works. Unless the contrary can be demon-
As a result of the perceived problems with UK site investi- strated, it should be assumed that site investigations are
gation, there have been a number of attempts to improve necessary in reporting upon the existing works, and for
the effectiveness of UK site investigation, including work investigating cases where failure has occurred.
by the Construction Industry Research and Information These complex requirements cover most foreseeable uses
Association (CIRIA) (Uff & Clayton 1986, 1991), the UK of site investigations. However, in almost all cases, routine
Site Investigation Steering Group (ICE 1993a–d), and the site investigations are carried out on land that is already in
Association of Geotechnical and Geo-environmental the ownership, or control, of the project developer, before
Specialists (e.g. AGS 2006). planning, design and construction. The objectives of site
investigation are necessarily much more restricted. There
Site investigation objectives is generally no choice between different sites, and little or
no debate about the suitability of the given site, as the con-
The British Standard Code of Practice for Site Investigations struction professionals involved in the project are assumed
(BS5930: BSI 1999) identifies the primary objectives in to be capable of producing a satisfactory design on any
making site investigations as follows. given ground conditions. Given the lack of site investigation
(1) Suitability. To assess the general suitability of the site data at this stage, key geotechnical details of construction
and environs for the proposed works, including, where appli- (e.g. pile types, diameters and depths, dewatering methods
cable, the implications of any previous use or contamination and other temporary works) will probably remain unknown
of the site. until the main construction works are tendered.

From: CULSHAW , M. G., REEVES , H. J., JEFFERSON , I. & SPINK , T. W. (eds) Engineering Geology for Tomorrow’s Cities.
Geological Society, London, Engineering Geology Special Publications, 22, 125–141.
DOI: 10.1144/EGSP22.9 0267-9914/09/$15.00 # The Geological Society of London 2009.
126 C. R. I. CLAYTON

In the author’s view, many of the primary objectives

stated in BS5930 are unlikely to be achieved and it is unrea-
listic to expect that they will be. Site investigation remains,
however, the key to good geotechnical design. The debate
should not be whether and how much it should be done but
how it can be most effective.

Site investigation philosophy in the UK

The traditional view in the UK has been that the ineffective-
ness of site investigations is the client’s fault. To obtain good
site investigations clients should, according to this view,
only employ competent professionals, spend sufficient to
provide a good amount of data, and use methods of procure-
ment that encourage quality.
Many countries around the world, and particularly in
southern Europe, appear to have legislation that restricts
those who can provide professional services to the public.
In the UK, very few professions enjoy this privilege. Prac-
titioners of civil engineering and engineering geology are
not restricted by law and there is little expectation that
they will be in the foreseeable future. In the absence of
legally restricted practice, one view is that we should Fig. 1. Impact of expenditure on ground investigations on highway
educate our clients, provide them information on levels of contract cost over-runs in the UK (after Mott MacDonald & Soil
competence needed for their work, and produce registers Mechanics Ltd. 1994). SI, site investigation.
of competent persons (AGS 2005a). A survey carried out
by the AGS has suggested that support for the concept for However, it should be noted that current rates of expendi-
or against a registration scheme is balanced amongst their ture are considerably less than 1% of construction cost. It is
members but that, if a scheme were introduced, it should unthinkable that clients should increase expenditure on site
include all the disciplines involved in ground engineering investigation by an order of magnitude but, even if they
and geoenvironmental issues. did, the data in Figure 1 suggest that potential cost over-runs
The need to spend more on site investigation has been a would remain unacceptable, given that either the client or the
constant theme in the UK for many years (e.g. Williams & main works contractor will normally take the loss and
Mettam 1971; Rowe 1972). If only clients would spend the level of profit margins in most businesses and in the
more money on site investigation, runs the argument, then construction industry.
site investigation would achieve ‘good’ results. There is An analysis of the causes of cost over-runs on 12 major
some evidence to support the idea that more expenditure on (and relatively troublesome) highway contracts (Tyrrell
site investigation results in more certainty of outcome for et al. 1983) is shown in Figure 2. These projects had, on
the client. Figure 1 shows some statistics obtained from a average, a 35% cost over-run. The data strongly support
survey of UK highway projects (Mott MacDonald & Soil the idea that highway cost over-runs are associated with
Mechanics Ltd. 1994). The x-axis gives the percentage of site investigation; on average, about 50% of the cost over-
the tender price spent on ground investigation, and the runs were attributed to only two sources. However, it
y-axis shows the cost over-run as a percentage of the tender should be noted that it was the planning and interpretation
price. It can be seen that the maximum cost over-run is less of site investigation, rather than the quality or amount of dril-
than 25% of tender price when expenditure on ground inves- ling, sampling and testing that was considered the cause of
tigation is 5% or more of the tender price. In contrast, expen- the problems. These data suggest that: (1) effective risk
diture of less than 1% would appear to expose the client (in and hazard identification is the key, because only once geo-
this case the government) to up to 100% of cost over-run. technical hazards are identified and understood can site
The lessons to be learnt from Figure 1 are, first, that cost investigation be properly planned; (2) knowledgeable
over-runs can be very high on certain types of contract, and realistic interpretation is very important; (3) the quantity
particularly those with a high percentage of ground-related or quality of routine ground investigation drilling, sampling,
work, such as highway construction and tunnelling. There- etc., may well be a secondary issue. The nature of ground-
fore, not all types of construction are equally risky. For related risks and hazards will be discussed below.
example, there will be lower levels of geotechnical risk The issue of procurement has been thoroughly debated in
when constructing in an area with familiar ground the UK (Uff & Clayton 1986, 1991). Procurement methods
conditions, using well-tried geotechnical components and that focus on minimizing cost through competitive tender
processes, in geologically uniform ground. have been widely condemned because, as has been noted,
 URBAN SITE INVESTIGATION 127

‘engineers imagined that the future science of foundations

would consist in carrying out the following program: drill
a hole into the ground. Send the soil samples obtained
from the hole through a laboratory with standardized appar-
atus served by conscientious human automatons. Collect the
figures, introduce them into equations, and compute the
result’ seems as relevant today as in 1936. He complained
that ‘the major part of the college training of civil engineers
consists in the absorption of the laws and rules which apply
to relatively simple and well-defined materials, such as steel
and concrete. This type of education breeds the illusion that
everything connected with engineering should and can be
computed on the basis of a priori assumptions.’
The main problem still faced by engineering geologists
and geotechnical engineers is that much of engineering is
deterministic and, as geotechnical engineering is a branch
of engineering, there is an expectation of certainty from
our clients. However, experienced engineering geologists
Fig. 2. Analysis of the causes of cost over-runs on 12 highway con- and geotechnical engineers will recognize many factors
tracts (after Tyrrell et al. 1983). that can lead to uncertainty of outcome when dealing with
the ground, some examples of which are listed in Table 1.
Figure 3 gives an example of the uncertainty with which
the cost of site investigation is small compared with the cost the settlement of a foundation on sand can be calculated.
of construction but the effects of unforeseen ground con- The graph shows (on the y-axis) the ratio of the observed
ditions can be very great (Fig. 1). Based on Uff & Clayton settlement divided by the predicted settlement. The x-axis
(1986, 1991) it is suggested that a site investigation procure- gives the predicted settlement for each case. All predictions
ment system should have the following aims. used data from the standard penetration test (SPT) which,
(1) To ensure that a competent geotechnical adviser is although known to be problematical (e.g. Clayton 1995) is
retained by the project promoter or developer at an early probably the most widely used test in the world for the engin-
stage during the conceptualization of the project, to ensure eering characterization of granular materials. The data were
that key elements such as an initial desk study and hazard produced by collecting well-documented case records,
identification are carried out. where the settlements of foundations had been observed
(2) As far as possible to give overall responsibility for all and SPT data were available, and then using a range of
geotechnical matters, at all stages of the project, to a single methods to make predictions of settlements. It can be seen
company (or, for smaller projects, individual). that, approximately, actual (observed) settlements may be
(3) To select geotechnical advisers and contractors on the up to five times larger or smaller than are predicted.
basis of their resources (staff, equipment, etc.) and experi- The second example (Wheeler 1999) involves the predic-
ence with similar forms of construction and ground con- tion of the behaviour of piles under load. Figure 4 shows the
ditions, and not primarily on the basis of their fee level results of a pile prediction competition. Given the ground
and unit rates. In the UK, two systems of procurement have conditions and pile construction, different teams of engin-
been recommended, as detailed in CIRIA Special Publi- eers (marked A to P in Fig. 4) predicted the shaft and base
cation SP45 (Uff & Clayton 1986). The first uses a geotech- capacities of a pile that was subsequently load tested. The
nical advisor (an experienced and appropriately qualified results of load testing are shown at the top of the graph.
geotechnical specialist) to carry out the desk study and walk-
over survey, plan and supervise the ground investigation work,
and separately employs a specialist contractor to carry out Table 1. Some uncertainties associated with
fieldwork, in situ and laboratory testing and factual reporting geotechnical work
of these. The second system, more often used for smaller pro-
jects, involves a package deal where all geotechnical activities Examples of geotechnical uncertainties
are carried out by a single organization.
Ground model Software performance
Contaminant model Constitutive framework
Site investigation in an uncertain and Spatial and temporal variability Construction management
complex world Parameters needed Whole-life performance
Groundwater model Engineering competence
Required volume of data Engineering knowledge
Terzaghi’s critical comments (Terzaghi 1936) on the Method of calculation Engineering experience
early and unrealistic expectations of civil engineers, that
128 C. R. I. CLAYTON

the many things that can go wrong during construction. As

Donald Rumsfeld said in February 2002: ‘There are known
knowns. These are things that we know that we know.
There are known unknowns. That is to say, there are things
that we now know we don’t know. But there are also
unknown unknowns. There are things we don’t know we
don’t know.’ To carry out successful site investigation we
need to accept the uncertainty of our situation and systema-
tically defend ourselves against it. We should also remember
that the upside of risk is opportunity. If we can work more
effectively with uncertainty than our competitors, we and
our clients can have an increased opportunity to cut costs
and make profits.

Recent site investigation philosophy in the UK

Geotechnical uncertainties represent only one of a number of
Fig. 3. Comparison of predicted and observed settlements of spread
foundations on sand (after Clayton et al. 1988). types of uncertainty that must be dealt with by the promoter
of a construction project. Other obvious examples include
financial risks and operational risks (e.g. Internal Control
The various predictions are shown in order of total pile Working Party of the Institute of Chartered Accountants in
capacity and the lack of accuracy in almost all predictions England & Wales 1999). Therefore, it is important not
can be judged. Fortunately, in this case, most predictions only that the riskiness of the ground-related phase of con-
of performance were conservative. struction is accepted by the client but that the risks that are
In both the cases described above, the major uncertainty identified are incorporated in the overall project risk
arose from the effectiveness of the method of calculation analysis.
of a value (settlement, bearing capacity). Not only was the For the engineering geologist, it is obvious that the ground
problem well defined ab initio but ground geometry and par- is often heterogeneous and that the scale and degree of its
ameters were also well described, meaning that there were heterogeneity, as well as its structure, are, to a great
far fewer uncertainties than would typically be experienced extent, the product of geological processes. For the engineer,
in routine practice (see Table 1). In fact, it is hard for the historian and archaeologist it is clear that geologically con-
inexperienced engineer or engineering geologist to envisage trolled ground conditions are modified by subsequent land
use. Therefore, it is growing practice in the UK to develop
a number of conceptual ground models for the site under
investigation: (1) a geological model, which uses geological
knowledge to create a conceptual 3D model of the ground;
(2) a geotechnical model, which uses the geological
model and knowledge of the likely physical properties of
the ground and groundwater to identify key issues for
geotechnical design and construction; (3) a land-use model,
which uses historical, archaeological and other data to ident-
ify land contamination, the existence of archaeological heri-
tage, etc. These models must be developed very early in the
investigation process if direct methods of investigation, such
as drilling and in situ testing, are to be planned effectively.
Therefore, a thorough desk study is required, and this needs
to be supplemented (see the Donald Rumsfeld quote above)
by experience captured from senior specialists in the ground
conditions and the expected forms of construction. Figure 5
shows a (relatively simple) example of a geological model
from the first Glossop Lecture by Fookes (1997). Examples
of geotechnical and contamination models, provided in the
AGS ‘Client’s Guide to Desk Studies’ (AGS 2006), are
given in Figures 6 and 7.
It is obvious, given the variability in the ground created by
Fig. 4. Predicted and observed pile load capacity (after Wheeler geological processes and subsequent land use, that a huge
1999). amount of investigation would be necessary to completely
 URBAN SITE INVESTIGATION 129

Fig. 5. A geological model (after Fookes 1997).

Fig. 6. A geotechnical model (after AGS 2006). mAOD, metres above Ordnance Datum.
130 C. R. I. CLAYTON

Fig. 7. A contaminated land model (after AGS 2006).

remove uncertainties of geometry and material properties. details of some construction may not be available until after
On a typical geotechnical project the amount of ground the main works tender period. This, coupled with the need
sampled will be less than 1 part in 106 of that affected by for early information on ground conditions, suggests that it
construction (Broms 1980). It is impossible to eliminate will be difficult to avoid up to four different phases of geotech-
these types of uncertainty. The author’s experience has nical site investigation: (1) desk study, walk-over survey and
been that, even when exceptional amounts of drilling, capture of prior experience for risk analysis; (2) direct ground
boring and testing have been carried out, the ground can investigation (geophysics, boring, drilling, in situ and labora-
still produce major surprises during construction. So, as tory testing of soils and rocks); (3) supplementary direct
von Clausewitz (1832) said, ‘Decisions that are made with investigations to obtain parameters for temporary works,
the awareness of doubt are in most cases better than ground works and geotechnical processes; (4) observation
decisions in the case in which the doubt is pushed into the and monitoring of ground conditions during construction.
background. Pushing doubt into the background will back Communication of data on the risks perceived at different
fire; correctly fostered doubt on the other hand can be used stages of the project, by different people and organizations is
to support well thought through actions.’ vital, and is increasingly being done through the use of risk
In addition to the problems created by complex subsoil geo- registers (Godfrey 1996; Clayton 2001; HR Wallingford
metry, a ground investigation is increasingly hampered by a 2001) (see below).
lack of knowledge of exactly what will be constructed and Desk studies are best carried out systematically, not only
how. The use of specialist geotechnical knowledge through by people experienced in engineering geology, geotechnics
subcontracted design and build (e.g. of piling, dewatering, and contaminated land but also by those with knowledge
anchoring, grouting, temporary works) means that the and experience of the special needs and weaknesses of the
 URBAN SITE INVESTIGATION 131

types of construction that are envisaged, if the detection of Table 2. Survey of types of problem encountered during
hazard and all possible risks is to be achieved. Because geo- tunnelling and their costs (US National Committee on
technical engineering practice is regionalized, it would seem Tunneling Technology 1984)
sensible to search for hazards using lists based upon aggre-
gated experience, under the headings geological hazards, Type of problem Proportion of Claims
geomaterial hazards, geotechnical engineering hazards, projects on which (%)
and land-use hazards. Because risks are related to hazard encountered (%)
through vulnerability (Fig. 8), it can be seen that early use
Running ground 27 9
of engineering geological and geotechnical expertise, during Flowing ground 5 4
pre-project planning and conceptual design, can be used to Squeezing ground 19 8
avoid many issues. For example, buildings may be relocated Groundwater inflow 33 6
away from areas of made ground or pre-existing slope Noxious fluid 6 4
instability. Methane gas 7 2
The use of experience during desk studies and walk-over Existing utilities 1 0
surveys is vital. Some experience is necessarily tied up in Pressure binding of tunnel 4 4
individuals and can be best obtained by employing senior boring machine (TBM)
Mucking 5 2
staff (whether within the company of from outside it) Surface subsidence 9 2
during the desk study, for geotechnical design reviews and Face instability; soil 11 5
during construction. However, other types of experience Obstructions (boulders, etc.) 12 11
are in the public domain. As examples, Table 2 shows the Steering problems 4 0
results of a survey of the frequency of various types of
problem encountered during tunnelling, and their costs,
carried out by the US National Committee on Tunneling
Technology. Figure 9 shows similar data obtained for These are clearly the areas that it might pay to get right
general UK construction projects. during site investigation but, alternatively, identifying des-
The results of this type of study are invaluable, not only igns that are relatively insensitive to such changes (e.g. of
much because they warn us about what may go wrong but ground geometry) may be a more economical way forward.
also because they show how frequently the same problems We are lucky that there are a great many ready sources of
occur. In the UK a similar survey was carried out as part desk study information in the UK. Examples of desk study
of the ICE/DETR project on ‘Managing Geotechnical information and advice on undertaking walk-over surveys
Risk’, to identify ground-related problems on more routine can be found in textbooks (e.g. Clayton et al. 1995) and else-
civil engineering construction sites (University of Surrey where (BRE 1987, 1989; DoE 1994a, b; BSI 1999; AGS
1999; Clayton 2001). The results of a survey of 28 projects 2006). Figure 10 gives some examples of on-line topographi-
(Fig. 9) showed that just five types of problem, associated cal maps, aerial photographs, geological maps, old topogra-
with soil boundaries, soil properties, groundwater, contami- phical maps, existing borehole records and the British
nated ground and obstructions produced 75% of the unfore- Geological Survey’s borehole database available through
seen ground-related problems during routine construction. their GeoIndex (NERC 2008). Increasingly, sources of infor-
mation and desk study services (e.g. for land contamination
and other environmental factors) are Web-based and their
use in the UK has become routine. It is now easy to
produce a basic desk study at low cost.
The key to good hazard and risk identification remains the
use of high-level expertise and experience in the interpret-
ation of the desk study and walk-over survey data during
pre-project planning. The UK’s Highways Agency (1996)
suggested the use of risk workshops, focusing on: risk identi-
fication (brainstorming), based upon the desk study; risk
ranking for impact (based upon group experience); establish-
ment of the risk register; association of risks with the phases
of the project when they may be encountered; estimation of
costs and variances, and probability (based upon group
experience). Figure 11 shows an example of a risk register.
These should be simple documents based on relatively few
levels (four or five) of impact and probability of occurrence.
A simple approach to the creation of geotechnical risk regis-
ters is given in Appendix A of the ICE/DETR Managing
Fig. 8. Classification of hazards. Geotechnical Risk document (Clayton 2001). An Excel
132 C. R. I. CLAYTON

Fig. 9. Causes of unforeseen problems on UK construction sites (after Clayton 2001).

add-on is available from the UK Construction Industry leave chemical or biological contaminants in the ground,
Research and Information Association (HR Wallingford which may pose a hazard for drilling rig operators (ICE
2001). It is widely agreed that these documents should be 1993d) and laboratory technicians, and is likely to have
updated continuously throughout both the design and con- left obstructions, thus limiting the volume of ground that
struction of the project and should be made available to all can be satisfactorily be investigated.
organizations connected with ground work. However, the importance of the geological model, and of
local geological knowledge, must not be forgotten, and
should continue to be the starting point for site investigation.
The urban environment The 1:50 000 geological map shown in Figure 10 is for an
area near Bristol, which was some years ago developed for
a superstore. The map shows the presence of thin coal
At the beginning of the 20th century, 7% of the world’s seams beneath the site and, as a result of the recognition
population could be considered ‘urban’. Most urban dwellers that coal mining might have taken place in the 18th or
lived in developed countries. By 1950 30% of the population 19th century, an additional site investigation was required
lived in urban areas and by 2000 47% was urban. By 2030, it to assess the possibility that cavities might be present.
is expected that more than 60% of the world’s population Figure 12 shows the geometry of steep-sided ‘scour
will live in towns and cities. Most of the new urban areas hollows’ to be found beneath Battersea Power Station,
will be in developing countries. near Central London (Buro Happold Consulting Engineers,
The urban environment is more difficult for site investi- pers. comm.). These features, which are probably inter-
gators than the rural one. To the normal challenges presented glacial (Berry 1979) in origin, penetrate the Pleistocene
to new construction by the geology are added the vulner- Thames Terrace Gravels (Bridgland 1994) and, at this
ability of the proposed project to more difficult shallow location, can also cut right through the London Clay
ground conditions, the effects of development on existing Formation. Because of their steep-sided 3D form, they are
buildings and infrastructure, and the special difficulties of extremely difficult to locate with routine quantities of
carrying out site investigations in urban areas. site investigation.
Previous land use produces increased complexity of
Urban ground ground and groundwater conditions. Examples from the
London area include: (1) the existence of significant thick-
Ground conditions in urban areas, and particularly in old or nesses of artificial deposits (termed ‘fill’ or ‘made ground’
ancient cities, can be even more demanding than those in a in the UK); (2) lowered groundwater as a result of local
green-field or recently developed areas. The geometry of but widespread abstraction (e.g. Amos 1860; see also
the ground is made more complex, and its properties more Mather 2004); (3) significant undiscovered ordnance (such
variable, by previous land use. Previous land use can also as bombs) in the ground, as cities have often been important
 URBAN SITE INVESTIGATION 133

Fig. 10. Examples of desk study information available in the UK.

134 C. R. I. CLAYTON

Fig. 11. Example of a risk register (courtesy of Arup; after Clayton 2001).

targets in 20th-century wars; (4) both shallow and deep has archaeological significance in itself (see below).
foundations, left in the ground by previous use; (5) services, Figure 15 shows street plans from the 18th and 20th centu-
tunnels and cellars in close proximity. ries. Realignment of streets since the 18th century means
Made ground will typically be loose-tipped and will be that 19th- and 20th-century services may pass through old
variable both in composition and mechanical properties. It basements in this area.
can be associated with chemical or biological contaminants; Since the 1920s, repetitive reconstruction of areas of
for example, derived from previous industry on the site, or central London, sometimes as a result of wartime bombing
simply from previous sanitation arrangements. but also simply because of the needs to modernize office
Figure 13 shows an archaeological reconstruction of the buildings, has resulted in the construction of several
extent of the River Thames and the islands within it in phases and types of foundation. The introduction of concrete
Roman times. Before it was confined by the construction of and steel piling in the early 20th century has allowed pro-
the embankments in Victorian times, about 150 years ago, gressively taller buildings to be constructed. Because, gener-
the Thames was much wider. As can be seen, in this area ally, it is impossible to remove existing foundations (and
its right bank was over 1 km south of its present position. particularly piled foundations), a build-up of obstructions
The large area to the north of the old river bank is underlain is occurring. The simplest approach is to locate and avoid
by penecontemporaneous (relatively recent) alluvium as existing foundations, tunnels and services, but in areas
well as the (relatively ancient) made ground described below. such as central London this is likely to become impossible
Most cities have been redeveloped very significantly over in the near future. Obstructions are clearly a major target
the centuries. Figure 14 shows a record of an excavation for a for urban site investigations.
hotel lift shaft in the London Bridge area, which is one of the One option is to reuse these foundations, but this requires
older areas of development in London. About 3 m of geome- even greater investigation of the site. Figure 16 shows a
trically complex medieval (.500 years old) and Roman sketch of the existing foundations within Battersea Power
(.1600 years old) backfill exists, all of which is potentially Station (Buro Happold Consulting Engineers, pers.
rich in archaeological material and the geometry of which comm.), a large derelict power plant in central London that
 URBAN SITE INVESTIGATION 135

Fig. 12. Scour hollow geometry, Battersea Power Station (courtesy of Buro Happold Consulting Engineers).

Fig. 13. Archaeological reconstruction of the extent of the Thames in central London, in Roman times (after Yule 1988).
136 C. R. I. CLAYTON

Ground-related vulnerability of urban development

Variable, compressible and metastable ground, such as
made ground, can have a significant effect on the behaviour
of shallow foundations and services during construction.
It is obvious that building new foundations or services on
such ground is undesirable. It is perhaps less obvious that
new construction can have a significant effect on existing
and, perhaps, very old buildings and pipelines, which
may have been poorly constructed (these issues were not
well understood even in the first half of the 20th century)
and may now be in poor condition. Such infrastructure
is very common in many cities. Depending upon whether
you are the developer of new buildings or construction, or
the owner of existing buildings or infrastructure, damage
caused by new construction is either an ‘accident waiting
to happen’ or the ‘straw that breaks the camel’s back’.
Fig. 14. Record of an excavation for a hotel lift shaft in the London The cost-effectiveness of urban redevelopment can be
Bridge area of central London. significantly affected by the possible presence of archaeo-
logical remains. There is a growing need to recognize the
is being redeveloped. An immense amount of investigation importance of archaeological heritage and obtain site inves-
and testing is required to determine not only the location tigation data in a way that will help archaeologists; this is
of all foundations but, also, their likely load-carrying even more important in urban areas. Failure to recognize
characteristics. the presence of potentially important archaeological

Fig. 15. Eighteenth- and 20th-century street plans in the London Bridge area, showing realignment.
 URBAN SITE INVESTIGATION 137

Fig. 16. Three-dimensional visualization of existing foundations inside Battersea Power Station, London (Buro Happold
Consulting Engineers, pers. comm).

evidence and artefacts can lead to long delays during rede- prospecting and extensive hand-dug trenching. Methods of
velopment, as the discovery of the remains of the obtaining greater archaeological information from routine
16th-century Shakespearean Rose Theatre, during the rede- ground investigation are under development.
velopment of an office block in London, in 1989, shows Given that urban and inner-city sites have often been the
(Rose Theatre Trust 2008). Not only did the archaeological subject of a number of phases of land use and redevelop-
excavation delay construction but redesign of the substruc- ment, the opportunities for the land to become contaminated
ture was necessary to ensure the preservation of archaeolo- are greatly increased. Contaminated land is a risk not only
gical heritage. for the public and those who may ultimately work or live
The UK Department of the Environment’s Planning in the development. Site investigation personnel are in the
Policy Guidance Note PPG 16 (DoE 1990) required local front line and need to be protected (ICE 1993d).
planning authorities to make archaeological concerns a
‘material consideration’ for all planning applications. Site investigation challenges
There is now a demand on many sites that the archaeological
resource, if significant, be preserved in situ by ‘good As discussed above, urban site investigation needs more skill
engineering practices’. Sites must be thoroughly evaluated and care because of the following factors.
archaeologically and often require a separate archaeological (1) The underlying geology can be expected to be no less
investigation. This investigation may involve geophysical complex than elsewhere.
138 C. R. I. CLAYTON

(2) There will be limited (if any) exposures of the geology potential for land contamination, and as a basis for planning
in an urban area and its surface expression will often be direct further ground investigations.
suppressed. (2) As a result of this, national and company borehole data-
(3) Previous land use will have made near-surface bases and urban geological models have been developed.
materials more geometrically complex, whilst adding the (3) Specialist drilling and sampling apparatus is used to
chance of contamination and the presence of ordnance. deal with shallow and potentially contaminated ground.
(4) Regional dewatering (e.g. for water supply; Mather (4) In uniform ground, sophisticated drilling, sampling,
2004) and future movements of the groundwater table need in situ and laboratory testing methods are used to obtain
to be understood. advanced geotechnical design parameters for complex
(5) Archaeological features and artefacts may be present, numerical (finite-element or finite-difference) modelling.
which need to be identified, perhaps excavated, and at least (5) At the same time, index testing and back-analysed par-
avoided during construction. ameters from monitoring of previous construction in similar
(6) Property owners and developers will have paid a ground conditions, are used as a check on site investigation
premium for an urban site and, generally, will wish to maxi- data and for use in more routine design work.
mize the use of it, perhaps by using deep basements and (6) Methods of adapting geotechnical investigations to
construction up to the property line. meet archaeological needs are under current consideration.
(7) Nearby buildings, tunnels and services must be taken Site investigation workers can be at risk if they work on
into account, their condition assessed and the effects of con- contaminated sites. The first and most cost-effective phase
struction on them estimated. In London, as in most cities, the of any assessment of likely contamination will rely heavily
presence of very close pre-existing buildings and critical on desk study records (BS10175, BSE 2001). In urban site
infrastructure such as tunnels is the norm, as described in investigation this must be carried out very early during
many academic papers (e.g. Clayton et al. 1991; Mair & pre-project planning, so that the nature and extent of
Harris 2001). possible contamination hazards can be identified and con-
(8) Access for drilling rigs will be restricted and may trolled. In the UK, the British Drilling Association has pre-
be difficult. pared a classification system that provides a framework
The consequences of a geotechnical failure in a built-up within which the investigation of urban and brownfield
environment are generally far more severe than in a rural sites can take place (ICE 1993d) and this is currently
setting, as failures of the Heathrow Express tunnels in under revision.
London (Health and Safety Executive 2000), the Merrie- Site investigation databases are in common use in the UK.
spruit tailings dam in South Africa (Strydom & Williams Major and local specialist geotechnical consultants and con-
1999) and the Po Shan slope in Hong Kong (Au 1998) all tractors and local authorities have traditionally maintained
show. Even if fatalities are avoided, the financial effects collections of site investigation reports but for a very long
can be very large. As has been described above, ground con- period there were no publicly available records. The decision
ditions in urban areas are likely to be more challenging but of the British Geological Survey to build a borehole data col-
because the cost of urban development is, generally, lection (NERC 2008) and the widespread use of the AGS
greater and every piece of development is likely to affect electronic data transfer format (AGS 2005b) in the UK
numbers of adjacent structures and services, higher levels (and increasingly elsewhere) has meant that paper borehole
of certainty are likely to be needed before urban develop- records have become easily accessible for desk study pur-
ment can proceed. poses and are likely to be replaced with downloadable elec-
tronic records (and possibly additional data) in the future.
Although borehole records can rarely be used to avoid a
Responses new, bespoke, ground investigation, they are invaluable
when used in conjunction with an engineering geological
Despite the potential difficulties inherent in urban site inves- interpretation, as they can give information on the local
tigations, there are likely to be some compensating factors. thickness of formations, the composition of the material
High property values and the severe consequences of geo- (through the engineering descriptions they contain), the
technical failures can encourage clients to spend more on approximate value of properties of different strata (such as
direct methods of site investigation, if this can be justified SPT N value and undrained shear strength), and the level
in terms of greater exploitation of the site (e.g. deeper base- of groundwater strikes.
ments, or building closer to existing buildings or infrastruc- Specialist compact drilling equipment is increasingly
ture) and reduced risk. Written records and maps relating to being used for the investigation of shallow, often contami-
previous land use are likely to exist for both the site itself and nated, urban ground. An example of a drilling rig and
its environs. Borehole and other data may be available from some samplers is shown in Figure 17. Such machines are
previous site investigations. typically track mounted and have an automatic hammer
The following strategies have developed in the UK. that can be used to drive sampler tubes into the ground, or
(1) There is heavy reliance on desk studies; for example, carry out dynamic penetration testing. Additionally, they
to investigate existing services, old foundations and the may have some limited hydraulic rotary drilling capability.
 URBAN SITE INVESTIGATION 139

establishment of precedent practice but, also, to case his-

tories of the performance of different structures. In the stiff
London Clay, for example, conventional laboratory
measurements of stiffness (using, for example, the oedo-
meter test) can be very much (about an order of magnitude)
lower than estimates made on the basis of movements around
deep basement excavations. The difference results, primar-
ily, from bedding during testing and, if good estimates of
stiffness are required, it may be better to use back-analysed
data than carry out routine testing (e.g. Clayton et al. 1991).
Thus, one approach is to restrict site investigation to simple
(but numerous) index tests (in the London Clay this might
involve moisture content, plasticity and undrained shear
strength and SPT tests). These can then be used to identify
monitored sites in similar ground for which back-analysed
data are available.
As noted above, urban site investigation is often special
because the land being developed is expensive, structures
are larger, with deeper basements, and are surrounded by
existing structures and significant services and tunnels,
often in poor condition. Responses to this seem to differ,
depending upon the complexity of the subsoil geometry. In
the London Clay, complex, high-cost, site investigation
techniques, involving pushed thin-walled tube sampling,
wire-line core drilling, self-boring pressuremeter testing,
seismic cone penetration testing (CPT) and laboratory
stress path triaxial testing with local strain and mid-plane
pore pressure measurement have become reasonably
Fig. 17. Example of compact drilling rig and associated equipment common. The results of such site investigation are typically
(courtesy of Dando and Archway Engineering). used in advanced (finite-element or finite-difference) num-
erical modelling of the ground movements around exca-
vations and tunnels, which is justified because the simple
ground geometry can be modelled in a simple way.
A conventional soft ground drilling rig operates by advan- However, in many other situations (e.g. Manhattan and
cing the borehole, cased or uncased, and then carrying out Amsterdam, where the ground geometry is far more
in situ tests or sampling from the base of the borehole. A complex), a precise knowledge of soil parameters is of sec-
dynamic probe rig takes tube samples as a means of advan- ondary importance. In these areas much more effort is put
cing the hole or, alternatively, can carry out continuous into profiling, for example using the CPT.
dynamic penetration testing, typically in the UK using a
trip hammer with the same weight and height of drop as
for the SPT but at a considerably faster rate. Samples can
be taken using a window sampler, essentially a tube with Conclusions
an opening in its side, allowing recovery of disturbed
material or, preferably, with a composite, thick-walled, Site investigation philosophy is changing from one where
open-drive sampler (termed a ‘windowless sampler’), engineers imagined that they had a reasonable chance
which has a inner, clear, plastic liner, allowing identification (given enough skill and expenditure) of determining the
of the material in the field, and provides a container in entire subsoil geometry and the properties of the ground,
which the material can be sent to the laboratory for further to one where the geometric complexity of the ground is
description and testing. recognized, its hazards and risks are identified, and the
It is often the case in geotechnical engineering that soil uncertainty that can be involved in determining even rela-
parameters obtained from back analysis of monitored struc- tively simple outcomes, such as the settlement of a foun-
tures (e.g. of foundation settlements, movement around dation, are recognized. In this context, engineering
excavations and tunnels, seepage into excavations) are geology plays an important role in helping to construct the
more reliable and accurate than data obtained from in situ geological model for a site, in aiding an understanding of
tests or laboratory tests on soil samples. Urban geotechnical the likely variability of ground and groundwater conditions,
design should be more reliable as a result of the intensity and in identifying the full range of hazards that might be
and frequency of construction, leading not only to the present. Risk analysis and risk registers are vital as means
140 C. R. I. CLAYTON

of communicating information on ground-related risks to all AGS 2006. A Client’s Guide to Desk Studies. Association of
those involved in a project. Geotechnical and Geoenvironmental Specialists, Beckenham.
The urban environment provides particular challenges World Wide Web Address: http://www.ags.org.uk/publi-
for ground investigations. The geometry of the deeper cations/Deskstudies.pdf.
A MOS , C. E. 1860. On the government waterworks in Trafalgar
ground will be no less complex than that on an equivalent Square. Proceedings of the Institution of Civil Engineers, 19,
green-field site but, at shallow depth, the presence of uncom- 21–52.
pacted and, perhaps, organic made ground, contamination A U , S. W. C. 1998. Rain-induced slope instability in Hong Kong.
and obstructions can make the work of site investigation con- Engineering Geology, 51, 1– 36.
siderably more challenging. Gathering information to assist B ERRY , F. G. 1979. Late Quaternary scour-hollows and related
in identifying archaeological heritage has also become features in central London. Quarterly Journal of Engineering
important. Geology, 12, 9– 29.
In an urban area, site investigation also has a wider range BRE 1987. Site investigation for low-rise building: desk studies.
of targets. Providing data to assess the effects of new con- BRE Digest, 318.
BRE 1989. Site Investigation for low-rise buildings: the walk-over
struction on existing buildings and on surrounding infra- survey. BRE Digest, 348.
structure such as services and tunnels can be more of a B RIDGLAND , D. R. 1994. The Pleistocene of the Thames. In:
challenge than determining the likely performance of the B RIDGLAND , D. R. (ed.) Quaternary of the Thames. Chapman
new construction, particularly as the geometry, location & Hall, London, Geological Conservation Review Series, 7,
and condition of existing services, foundations and tunnels 1– 30.
may all be uncertain. The consequences of a geotechnical B ROMS , B. B. 1980. Soil sampling in Europe: state-of-the-art. Pro-
failure are likely to be much more severe in an urban ceedings of the American Society of Civil Engineers, Journal of
environment, so that the standards of certainty required are the Geotechnical Engineering Division, 106, 65–98.
likely to be greater before planning permission can be BSI 1999. BS5930. Code of practice for site investigations. British
Standards Institution, London.
obtained. This applies as much to off-site features (such as BSI 2001. BS10175. Code of practice for the investigation of
surrounding slopes and structures) as to the project itself. potentially contaminated sites. British Standards Institution,
There have been a number of responses to these chal- London.
lenges. First, repetitive construction in a relatively small C LAYTON , C. R. I. 1995. The Standard Penetration Test (SPT):
area means that desk studies are likely to be even more methods and use. CIRIA Special Publication Report, R143.
fruitful than in a non-urban area, and that experience and C LAYTON , C. R. I. 2001. Managing geotechnical risk: improving
reports of problems and successes in previous construction productivity in UK building and construction. Institute of Civil
(precedent practice) can be very useful. The drilling and Engineers/Department of the Environment, Transport and the
probing of large numbers of exploratory holes in a relatively Regions Report. Thomas Telford, London.
C LAYTON , C. R. I., S IMONS , N. E. & I NSTONE , S. J. 1988. Research
small area gives the opportunity to build databases and to on dynamic penetration testing of sands. In: D E R UITER , J. (ed.)
incorporate these in 3D geological models (Neber et al. Proceedings of the 1st International Conference on Penetration
2009), helping to extrapolate the data in the context of an Testing, Florida, Vol 1. Balkema, Rotterdam, 415– 422.
engineering geological interpretation. Second, it has been C LAYTON , C. R. I., E DWARDS , A. & W EBB , M. J. 1991. Displace-
recognized that under favourable ground conditions, back ments within the London Clay during construction. In: A SSO-
analysis of monitored data from previous construction may CIAZONE G EOTECNICA I TALIANA (ed.) Proceedings of the
provide a better measure of ground parameters such as stiff- 10th European Conference on Soil Mechanics and Foundation
ness and permeability than the techniques commonly used Engineering, Florence, Vol 2. Balkema, Rotterdam, 791–796.
during routine site investigation. Third, specialist drilling, C LAYTON , C. R. I., M ATTHEWS , M. C. & S IMONS , N. E. 1995. Site
Investigation. Blackwell, Oxford.
sampling and testing equipment has been developed to D O E 1990. Planning Policy Guidance 16: Archaeology and
deal with shallow depths of contaminated ground and planning. PPG16. Department of the Environment, London.
with greater depths of more uniform soil. Techniques are D O E 1994a. Guidance on preliminary site inspection of contami-
evolving for other specialist urban investigation issues, nated land. CLR Report, 2.
such as the investigation of existing foundations with a D O E 1994b. Documentary research on industrial sites. CLR Report, 3.
view to reuse. F OOKES , P. G. 1997. The geological model, prediction and perform-
ance. 1st Glossop Lecture. Quarterly Journal of Engineering
References Geology, 30, 293–424.
G ODFREY , P. S. 1996. Control of risk: a guide to the systematic
AGS 2005a. Position Paper—Registration. Association of Geotech- management of risk from construction. CIRIA Special Publi-
nical and Geoenvironmental Specialists, Beckenham. World cation Report, SP125.
Wide Web Address: http://www.ags.org.uk/aboutus/Registra- H EALTH AND S AFETY E XECUTIVE 2000. The collapse of NATM
tionGeotechnicalEngineers.pdf. tunnels at Heathrow Airport. Health and Safety Executive
AGS 2005b. Electronic transfer of geotechnical and geoenviron- Books, London.
mental data using XML data format. Progress Report. Associ- H IGHWAYS A GENCY 1996. Value for money manual. Her
ation of Geotechnical and Geoenvironmental Specialists, Majesty’s Stationery Office, London.
Beckenham. World Wide Web Address: http://www.ags.org. HR W ALLINGFORD 2001. RiskCom. Software tool for managing
uk/agsml/AGSMLAugust2005Report.pdf. and communicating risks. Software C561 [CD].
 URBAN SITE INVESTIGATION 141

ICE 1993a. Site investigation in construction. Volume 1: Without R OWE , P. W. 1972. The relevance of soil fabric to site investigation
site investigation ground is a hazard. Site Investigation Steering practice. 12th Rankine Lecture. Geotéchnique, 22, 195– 300.
Group. Thomas Telford, London. S TRYDOM , J. H. & W ILLIAMS , A. A. B. 1999. A review of impor-
ICE 1993b. Site investigation in construction. Volume 2: Planning, tant and interesting technical findings regarding the tailings dam
procurement and quality management. Site Investigation Steer- failure at Merriespruit. Journal of the South African Institution
ing Group. Thomas Telford, London. of Civil Engineers, 41, 1 –9.
ICE 1993c. Site investigation in construction. Volume 3: Specifica- T ERZAGHI , K. 1936. Relation between soil mechanics and foun-
tion for ground investigation. Site Investigation Steering Group. dation engineering. Presidential address. In: Proceedings of
Thomas Telford, London. the 1st International Conference on Soil Mechanics and Foun-
ICE 1993d. Site investigation in construction. Volume 3: Guidelines dation Engineering, Harvard, Vol 3. Harvard University Press,
for the safe investigation by drilling of landfills and contami- Cambridge, Massachusetts, USA, 13–18.
nated land. Site Investigation Steering Group. Thomas T YRRELL , A. P., L AKE , L. M. & P ARSONS , A. W. 1983. An inves-
Telford, London. tigation of the extra costs arising on highway contracts. Trans-
I NTERNAL C ONTROL W ORKING P ARTY OF THE I NSTITUTE OF port and Road Research Laboratory (TRRL) Supplementary
C HARTERED A CCOUNTANTS IN E NGLAND & W ALES 1999. Report, SR814.
Internal control—guidance for Directors on the Combined U FF , J. F. & C LAYTON , C. R. I. 1986. Recommendations for the pro-
Code. Chaired by Nigel Turnbull. (‘The Turnbull Report’). curement of ground investigation. CIRIA Special Publication, 45.
World Wide Web Address: http://www.icaew.co.uk/. U FF , J. F. & C LAYTON , C. R. I. 1991. Role and responsibility in site
M AIR , R. J. & H ARRIS , D. 2001. Innovative engineering to control investigation. CIRIA Special Publication, 73.
Big Ben’s tilt. Ingenia, 9, August, 23– 27. U NIVERSITY OF S URREY 1999. Task 5 Report. Case histories and
M ATHER , J. 2004. British hydrogeology—a brief history. World their interpretation. Unpublished report for DETR/PIT Project
Wide Web Address: http://www.groundwateruk.org. on: Managing geotechnical risk: improving productivity in UK
M OTT M AC D ONALD & S OIL M ECHANICS L TD . 1994. Study of the building and construction.
efficiency of site investigation practices. Transport Research US N ATIONAL C OMMITTEE ON T UNNELING T ECHNOLOGY 1984.
Laboratory Project Report, PR60. Geotechnical site investigations for underground projects.
N EBER , A., A UBEL , J., C LASSON , F., H OEFER , S., K UNZ , A. & National Academy of Sciences, Washington, DC.
S OBISCH , H. G. 2009. From the Devonian to the present: V ON C LAUSEWITZ , K. 1832. Vom Kriege. Dümmlers, Berlin.
landscape and technogenic relief evolution in an urban W HEELER , P. 1999. Scattering predictions—Imperial College pre-
environment. In: C ULSHAW , M. G., R EEVES , H. J., J EFFERSON , dictions competition shows pile design remains a big uncer-
I. & S PINK , T. W. (eds) Engineering Geology for Tomorrow’s tainty. New Civil Engineer, December 2, 34.
Cities. Geological Society, London, Engineering Geology W ILLIAMS , J. T. & M ETTAM , J. D. 1971. Summary of lecture
Special Publications, 22 [on CD-ROM insert, Paper 517]. and discussion. National Ports Council—Port Structures
NERC 2008. GeoIndex. Natural Environment Research Council. Report. Proceedings of the Institution of Civil Engineers,
British Geological Survey, Keyworth, Nottingham. World 48, 475.
Wide Web Address: http://www.bgs.ac.uk/geoindex/index. Y ULE , B. 1988. The natural topography of north Southwark. In:
htm. H INTON , P. (ed.) Examinations in Southwark 1973–6, Lambeth
R OSE T HEATRE T RUST 2008. The Rose. World Wide Web 1973–9. London and Middlesex Archaeological Society and
Address: http://www.rosetheatre.org.uk/. Surrey Archaeological Society Joint Publication, 3, 13–18.