Brand (1985) - Predicting The Performance of Residual Soil Slopes

Predicting the Performance of Residual Soil Slopes
E. W. Brand
Geotechnical Control Office, Engineering Development Department, Hong Kong
Brand, E.W., Predicting the performance of residual soil slopes, Proceedings of the 11th International Conference on
Synopsis: Residual soil slopes are common in many parts of the world, but little attention has been paid to
appropriate methods of analysis and design. The basic approaches available for dealing with these heterogeneous
materials are reviewed in this Paper and details are given of some recent developments in design methods, with
particular reference to the design and performance of residual soil cut slopes in Hong Kong. Examination of
the relationship between rainfall and landslides for Hong Kong shows that short-term intensity is almost solely
responsible for the many slope failures, antecedent rainfall being of little account. The state-of-the-art with
respect to the application of soil mechanics methods to the analysis of residual soil slopes is reviewed, and the
main difficulties are examined with reference to a number of Hong Kong case histories. It is concluded that our
soil mechanics predictive tools are far from adequate for analysing residual soil slopes, largely because of the
difficulties of predicting pore pressures and of modelling geological detail which often controls the mode of
slope failure.
Soil Mechanics and Foundation Engineering, vol. 5, pp 2541-2578 © CRC Press.
1. INTRODUCTION (c) they are nearly always unsaturated, which poses

considerable difficulties for shear strength
This Theme Lecture was originally conceived as measurements, and
covering the broad field encompassed by the title (d) they invariably have high permeabilities, which
“Comparison of Prediction and Performance of Earth makes them subject to rapid changes in material
Structures”. During the preparatory stages, however, properties because of external hydraulic influences.
the Author became aware that this was too large an Experience has shown that engineering problems
undertaking for either the oral or written version of the in residual soils are not usually amenable to the
Lecture. The scope was therefore dramatically reduced principles of soil mechanics alone, but that these
to enable concentration on a single aspect of earth must be combined with the appropriate elements of
structures which is of great importance and which ‘geology’, ‘geomorpholopy’ and ‘hydrology’ in much
has been much neglected by geotechnical engineers, greater measure than is necessary for dealing with
namely the performance of residual soil slopes. sedimented soils. Geotechnical engineering in residual
Residual soils are widespread, and they represent soils therefore spans the narrowly separated fields
the most common type of material in those parts of of soil mechanics, rock mechanics and engineering
the world where slope failures are very prevalent. This geology, and the engineering geological approach is
is particularly so in tropical countries, where deep generally the most satisfactory one for these materials.
weathering of the bedrock has invariably resulted in From the point of view of analysis and design,
thick residual soil mantles, and where heavy rainfall residual soils undoubtedly represent the ‘difficult’
is usually the main agent of slope failures. Apart from end of the broad spectrum of engineering soils, as
their economic importance, rain-induced landslides opposed to saturated soft clays, which represent the
often present a significant hazard to life and property. ‘easy’ end. Whereas a great deal has been published
Residual soils have hitherto received very little about our ability to predict the behaviour of soft clays,
attention from the soil mechanics community. This a negligible amount has been written to enable an
neglect, and the resulting paucity of relevant literature, evaluation of soil mechanics methods of prediction as
stems largely from the fact that the very nature of applied to residual soils.
residual soils makes the application of soil mechanics Mention must be made of the well-known writings
principles problematical and renders the materials of Lambe (1973, 1975), which have given us a
difficult to model for the purposes of engineering framework by which to judge our soil mechanics
analysis and design. The main characteristics of predictions, and which have done so much to focus
residual soils responsible for this are: attention on the need for more good measurements
(a) they cannot usually be considered in isolation from of performance. Lambe’s predictive efforts have for
the complete weathered rock profile, of which they the most part, however, been concentrated on the
are one component part, ‘easy’ soils, and even for these, it has been found that
(b) they are generally very heterogeneous, which the careful application of soil mechanics methods
makes them difficult to sample and test, does not always lead to accurate predictions of field
Finish.indb 119 2007/8/17 11:03:28 AM

behaviour. A great deal of experience and engineering cover over weathered rock profiles, it is sometimes
judgement is therefore often necessary for making difficult to distinguish between colluvium and the
economically viable and acceptably safe predictions insitu material, particularly if only drillhole samples
where less predictable soils are involved. In this are available for examination. For geotechnical
context, the ‘observational’ philosophy continually engineering purposes, colluvium can therefore be
expounded by Peck (1969, 1975, 1980) is central to grouped with residual soils.
good geotechnical engineering. A recent publication contained a summary of the
The material in this Paper is based on the Author’s types of material regarded as ‘residual soil’ in eighteen
experiences with the performance of residual soil countries, as revealed by a collection of review papers
slopes in Hong Kong, which is prone to landslides that (Brand & Phillipson, 1985). No universally accepted
occur in the steep granitic and volcanic residual soil definition of ‘residual soil’ emerged, but the following
slopes during times of heavy rainfall. These failures general statements can be made :
are of major social and economic significance. Some (a) residual soils can result from the weathering of any
severe rainfall events have in the past resulted in many parent rock type,
failures which have caused large numbers of casualties (b) under the category of ‘residual soil’ are included
and widespread damage. In recent years, therefore, a materials which are not completely weathered and
great deal of effort has been devoted to the prediction which retain the original structure of the parent
and prevention of landslides. rock (saprolite), and
In part, this Paper draws heavily on two previous (c) colluvium is often categorised as ‘residual soil’ for
publications by the Author (Brand, 1982, 1985a) engineering purposes.
which reviewed the various approaches to analysis and For the purpose of this Paper, therefore, ‘residual
design in residual soils with particular emphasis on soil’ will be defined as that broad group of materials,
slope stability. formed by the insitu weathering of any rock type,

which exhibits engineering behaviour that is
considered to be governed mainly by the principles of
2. RESIDUAL SOILS soil mechanics. In terms of the six-grade weathering
classification system given below (Table 1), these
2.1 Definition materials are grades IV, V & VI. Colluvium is included
No universally accepted definition of ‘residual with these.
soils’ exists. These materials are products of insitu
weathering of rocks, the degree of weathering and 2.2 Published Literature
extent to which the original structure of the rock mass There is only a small amount of published literature on
is destroyed varying with depth from the ground the engineering properties and behaviour of residual
surface. This process gives rise to weathering profiles soils. As a starting point, reference should be made to
which contain material ‘grades’ from fresh rock to the proceedings of the Specialty Session on Lateritic
completely weathered material, the latter usually being Soils held in Mexico City (Moh, 1969) and to the book
described by geotechnical engineers as ‘soil’. For by Gidigasu (1975), which relies heavily on these
engineering purposes, it is difficult to separate the ‘soil’ proceedings. Also of importance are the proceedings
from the rest of the weathering profile, and the whole of the conferences on residual soils held in Hawaii
profile is therefore best treated as a single entity. (ASCE, 1982) and Brazil (ABMS, 1985a).
For the purposes of tropically weathered profiles, The proceedings of the ten International
the earth materials are sometimes categorised simply Conferences on Soil Mechanics and Foundation
as ‘laterite’, ‘saprolite’ and ‘rock’. The engineering Engineering each contain some papers directly
behaviours of laterite and saprolite are usually relevant to residual soils, as do the proceedings of
considered to be governed by the principles of soil the many ISSMFE Regional Conferences held to-
mechanics, and they are therefore the materials of date; the session onslope stability in residual soils at
most importance here. the Fourth Panamerican Conference is particularly
Large deposits of colluvium often exist in noteworthy (ASCE, 1971). To a lesser extent, there are
conjunction with residual materials, particularly as also specifically relevant papers in the proceedings of
colluvial fans on footslopes of hillsides. Colluvium the four Congresses of the International Association of
is material derived from the weathering of any parent Engineering Geology and the five Congresses of the
rock which has been transported downhill by the International Society for Rock Mechanics. There is a
agencies of gravity and water. It can range in general wealth of residual soil literature in the proceedings of
composition from a collection of matrixless boulders at the seven Brazilian Conferences on Soil Mechanics
one extreme to a fine slopewash material at the other. It and Foundation Engineering, but unfortunately this is
possesses many of the same general characteristics as virtually all in Portuguese.
residual soil, particularly in the context of engineering A few key publications address themselves to the
behaviour. Because it is commonly found as slope general philosophy and approach to engineering in
Finish.indb 120 2007/8/17 11:03:28 AM

residual materials. The first of these is the state-of-the- application to weathered rock profiles. Classification
art report on slope stability in residual soils by Deere in terms of weathering ‘zones’ and weathering ‘grades’
& Patton (1971). The others are the review papers by is essential for engineering design, and there have
de Mello (1972) and Brand (1982, 1985a), which all been several major attempts to provide a satisfactory
deal with the subject of residual materials in fairly description and classification system for engineering
broad terms. In addition, the comprehensive report purposes. Deere & Patton (1971) gave a valuable
produced very recently by the ISSMFE Technical comparative summary of the classification systems
Committee on Tropical Soils (ABMS, 1985b) is of available at that time, but there have been some
importance. important subsequent developments. The Geological
Especially worthwhile contributions to the Society of London (1977) produced a Working Party
literature on the engineering properties of residual Report on the subject of the description of weathered
soils have been made by Vargas (1953), Lumb (1962a, rock masses for engineering purposes, the main
1965), Sowers (1963), Little (1967, 1969), De Graft- elements of which have now been incorporated into
Johnson & Bhatia (1969) and Dearman et al (1976). the British Code of Practice for Site Investigation
Only Vaughan & Kwan (1984) and Vaughan (1985a), (British Standards Institution, 1981). Other important
however, appear to have attempted to provide a papers on this general topic are those by Little (1967,
theoretical framework for residual soil behaviour. 1969), Dearman (1974), IAEG (1979, 1981), Hencher
Noteworthy publications specifically on the slope & Martin (1982) and Martin & Hencher (1984).
stability aspects of residual soils have been written by Any weathering description and classification
Vargas (1967), Patton & Hendron (1974), Morgenstern system must be suitable for the particular geological
& de Matos (1975) and Blight (1977). Good reviews conditions and engineering purpose to which it is
of the landslide problems in these materials have applied. In Hong Kong, where site formation and
been made by Vargas & Pichler (1957), Da Costa slope stability are the main geotechnical engineering
Nunes (1969) and Jones (1973) for Brazil, Page & problems (see Section 3 below), the Geotechnical
James (1981) for Colombia, Sowers (1971) for Puerto Control Office (1984) has adopted a system for the
Rico, Lumb (1975) and Brand (1985b) for Hong granites and volcanic rocks which is based on the
Kong, Oyagi (1984) for Japan, and Brand (1984) for original work by Moye (1955) and Ruxton & Berry
Southeast Asia. (1957), but which has several important additions
Colluvium features hardly at all in the technical suggested by Hencher & Martin (1982). A profile is
literature, even though it is a fairly common logged according to the six material grades given in
engineering soil type. Good geomorphological Table 1 and the four profile zones described in Table 2.
textbooks, such as that by Young (1972), provide a It should be noted that weathering zones A and B
general description of the formation of colluvium, but comprise those materials referred to in some countries
few attempts appear to have been made to devise an as ‘laterite’ and ‘saprolite’ respectively.
engineering classification system, nor to investigate its For the purposes of geotechnical analysis and
wide range of material properties. Although colluvium design, the following should be noted :
features fairly prominently in some published (a) grades I to III material are usually treated as ‘rock’,
descriptions of mass movements, there are only a few and grades IV to VI material as , ‘soil’, and
papers which describe engineering designs in this (b) the engineering behaviour of weathering zones A
material, the most useful of which is probably that by and B is broadly considered to be governed by the
D’Appolonia et al (1966). principles of soil mechanics.
There exists no engineering description and
2.3 Weathering Profiles classification system for colluvium, although one is
The accurate logging of weathering profiles is badly needed. An attempt has been made in Hong
fundamental to successful design and construction in Kong to provide a framework for such a system
residual profiles. These often contain a whole range (Huntley & Randall, 1981), but this is entirely
of materials from an engineering point of view from descriptive in character and requires a great deal of
‘soil’ to ‘rock’. The weathering profile is therefore of further development.
great importance for the stability of slopes, because it
2.4 Investigation
usually controls :
(a) the potential failure surface, and therefore the For projects in residual soils, site investigation must
‘mode’ of failure for analysis and design, and generally be more extensive and more expansive than
(b) the groundwater hydrology, and therefore the for more homogeneous earth materials. The emphasis
critical pore pressure distribution in the slope. must be on the engineering geological approach, for
There is no universally accepted system for which the following are the main elements :
describing and classifying the component parts of (a) execution of adequate surface and subsurface
a weathering profile. The classification systems exploration to establish the site engineering
commonly used in soil mechanics have very limited geology, to define the ‘engineering’ materials
Finish.indb 121 2007/8/17 11:03:29 AM

involved, and to retrieve good quality samples for of Hong Kong, an introduction to the Territory is
laboratory testing, appropriate at this stage.
(b) identification of especially significant geological,
geotechnical and hydrological features, 3.1 General Description
(c) study of existing local and other ‘case histories’ of The Territory of Hong Kong has a land area of only
similar projects and of sites with similar geology, 1 050 sq. km and a population of nearly six million. It
(d) routine insitu and laboratory measurement of consists of Hong Kong Island, a piece of the Chinese
selected engineering properties of materials to mainland and more than 200 small islands, the largest
establish lower bound values, and comparison of
of which is Lantau (Figure 1). The population is

these with generalised parameters developed from
existing data, and Table 1. Weathering Grade Classification System
(e) continuous reappraisal of the site investigation recommended by the Geotechnical Control Office
results throughout the period of construction. (1984) for use in Hong Kong
The importance of the study of good case histories
cannot be overemphasized. Particularly important
also is the continuous reappraisal process, since site
investigations carried out in residual soils before
construction commences can in most cases only be
regarded as ‘initial’ investigations, and the geotechnical
engineer must be prepared to modify his design as
excavations reveal much fuller subsurface information
than was available prior to the construction phase. This

is the basis of the ‘observational method’ of design
expounded by Peck (1969).
Whereas the use of generalised soil properties
is not unsatisfactory where the soils are such that
these can be established within a sufficiently sensible
range, this is not so for residual soils, because of their
variability and inhomogeneity. Generalised parameters
for residual soils are difficult to obtain, and insufficient
basic research has been carried out on most residual
materials to provide even approximate values. This is
an unsatisfactory situation which confronts everyone
who works with residual soils, and it enhances the
importance of the non-analytical methods of design
described below.
Drilling and sampling in residual soil profiles Table 2. Weathering Zone Classification System
can be difficult because of the vertical variability recommended by the Geotechnical Control Office
encountered, particularly since it is essential to obtain (1984) for use in Hong Kong
a high recovery for profile description purposes.
A recent international review of the methods used
worldwide for the sampling and testing of residual
soils (Brand & Phillipson, 1985) revealed that the
sampling practice in most countries leaves much to be
desired, heavy reliance often being placed on methods
developed for sedimentary soils. In this context, the
drilling and sampling methods used for good quality
site investigations in the weathered granite and
volcanic profiles of Hong Kong may be of interest
elsewhere (Brand & Phillipson, 1984); triple-tube
rotary core barrels are used for sampling, and air-foam
is sometimes employed as the flushing medium.
3. INTRODUCTION TO HONG KONG

Since there will be continual reference in this Paper
to the slope stability problems in the residual soils
Finish.indb 122 2007/8/17 11:03:29 AM

developable land.
The rainfall in Hong Kong averages 2 225 mm
annually, and more than 80% of this falls during the
period May to September (Figure 4). Intensities can be
high, with 50 mm per hour and 200 mm in 24 hours
being not uncommon.
Figure 1. Map of the territory of Hong Kong showing

the geology
concentrated in a number of distinct geographical

locations dictated to a great extent by the terrain. High
concentrations of building development and population
exist all along the north side of Hong Kong Island
(Figure 2) and over the entire Kowloon peninsula,
as well as in the five new towns which are currently
under construction north of Kowloon. Apart from
Hong Kong Island, the islands are largely undeveloped
and sparsely populated. Figure 2. View looking east along the north coast of
The terrain of Hong Kong is very hilly. The Hong Kong Island showing the central district and the
land rises from sea level to 550 m on Hong Kong mid-levels area (Note Po Shan Landslide Scar)
Island in a distance of only 1.5 km. Most of the
Kowloon peninsula has now been levelled, along with
reclamation of adjacent sea areas, but isolated hills of
up to 100 m still exist. Immediately behind Kowloon,
hills rise to over 450 m. In the New Territories, very
little lowlying land exists, and peaks of over 400 m
are common. Natural slopes throughout the Territory
are steep, typically with upper slopes steeper than 35
degrees, midslopes of 25 to 30 degrees and footslopes
of 15 degrees. Cut slopes formed for roads and site
development works are commonly 40 to 80 degrees,
and fill slopes are 30 to 35 degrees. Figure 3 gives the
distribution of terrain angles for Hong Kong Island
and Kowloon. Figure 3. Distribution of terrain angles over Hong
About two-thirds of the Territory’s small land Kong Island and Kowloon
area is considered undevelopable. The usable land
is therefore an exceptionally valuable commodity. 3.2 Geology
In the absence of flat land, highrise buildings and The geology of Hong Kong, which is summarised in
other structures are increasingly being built on the Figure 1, has been described by Ruxton (1960) and
mid-slopes and upper slopes of natural hillsides. In Allen & Stephens (1971). The main rock types are
these circumstances, very intensive use is made of all granite and acid volcanic rocks, which together cover
Finish.indb 123 2007/8/17 11:03:31 AM

Safety (after Li and Whit
the major portion of the Territory and are by far the accumulation of debris from old landslips and mass
most important from an engineering point of view. The movements very often carpets the lower slopes of most
small amounts of sedimentary and metamorphic rocks of the hills. It varies in composition from a boulder
are of much less importance, although some specific field to a fine slope-wash, but it most commonly
landslide problems have been associated with these. consists of boulders, cobbles and gravel in a matrix of
Granite predominates in those areas of the Territory sand, silt and clay. The colluvium is up to 30 m thick
where building development is densest. in places. It is sometimes in a loose state, with a high
permeability, and it frequently gives rise to perched
water-table conditions. It is also prone to the formation
of ‘pipes’ or ‘tunnels’ as a result of internal erosion

(Pierson, 1983; Nash & Dale, 1983; Brand et al, 1986),
and these features can be of major significance to the
hydrogeology of an area (Leach & Herbert, 1982;
Premchitt et al, 1985).
Some engineering properties of the fresh and
decomposed granite and volcanic rocks of Hong Kong
have been reported by Lumb (1962a, 1962b, 1965,
1975, 1983).
Figure 4. Average monthly rainfalls in Hong Kong
3.3 Slope Stability
The granite varies somewhat in colour and Geotechnical engineering practice in Hong Kong
composition, but the fresh rock is fairly uniform in tends to be dominated by the slope stability problems
its engineering properties. Of major significance for brought about by the combination of the high intensity
slope stability are the joint patterns and the extent of rainfall and the steep terrain. Several hundred failures
weathering. Joint spacings are typically 0.5 to 1 m, may occur in a year, most of which are not of great
but they can be as wide as 3 m. In a few locations, consequence, but severe effects are felt from some
sheeting joints occur almost parallel to the natural of these in terms of casualties and damage. In fact, a
ground surface. The granite is extensively weathered significant landslide event, in which a large number
almost everywhere, with depths of up to 60 m of silty- of failures occur in one day causing considerable
sandy residual soil, often with large corestones in the disruption and damage, can statistically be expected to
matrix or exposed on the surface. Below the residual take place in Hong Kong about once every two years
soil mantle, a considerable depth of differentially (Lumb, 1975; Brand et al, 1984).
decomposed material extends to great depth in a The whole range of slope ‘features’ is prone to
weathered granite profile. The weathering process in landslides in Hong Kong, including natural slopes,
Hong Kong granite has been described by Ruxton & soil cut slopes, rock cut slopes, earth fill slopes,
Berry (1957). retaining walls and boulders. The majority of failures,
The volcanic rocks consist mainly of coarse and usually those with the most severe consequences,
tuffs, fine tuffs and rhyolite, but some ignimbrites take place in man-made features or are triggered by
and breccias also exist. These rocks are usually fine man-made features, particularly cut slopes in soil
grained and have a blocky structure, with close joint (weathering grades IV to VI).
spacings of generally less than about 0.25 m. They Cut slope failures in soil, or in mixed soil and
are more resistant to weathering than the granite, the rock, now constitute by far the most common form
residual soil mantles being only up to 20 m thick. A of landslide. The volcanic rock profiles are more
volcanic profile usually exhibits a steadily decreasing susceptible to failure than the granite profiles, and
degree of decomposition with depth, the unweathered colluvium is frequently involved. The failures nearly
rock being reached at a much shallower depth than in a always occur suddenly during intense rain without
granite profile. prior warning (Brand, 1984, 1985b), and most slip
Steeply dipping dolerite dykes have been intruded surfaces are shallow, the thickness of the failed zone
in many places into the granite and volcanic rocks. usually being less than 3 m. A typical example is
These dykes, which vary in thickness between about shown in Figure 5.
150 mm and 1.5 m, greatly complicate the engineering
geology of many sites and are often a significant
feature in landslides (Hencher et al, 1984; Hudson & 4. METHODS OF SLOPE FAILURE
Hencher, 1984). Granodiorite intrusions also exist in PREDICTION
some places.
Of particular engineering geological significance The stability analysis of a slope is ipso facto related
in Hong Kong are the extensive bodies of colluvium directly to the prediction of the conditions under
which cover about 20% of the total land area. This which the slope could fail. Soil mechanics methods
Finish.indb 124 2007/8/17 11:03:31 AM

y (after Li and White 1987b)
empirical methods are closely related, in that both

are based on an explicit assumption that the stability
characteristics of a slope can be assessed on the basis
of observations of the performance of others with
similar characteristics.
Whereas the vast majority of cut slopes in residual
soils were not designed on the basis of rigorous soil
mechanics methods, the soil mechanics approach
is being increasingly adopted even in this difficult

soil type. This is certainly true in Hong Kong,
where design practice is governed largely by the
Geotechnical Manual for Slopes (Geotechnical Control
Office, 1984). The degree of safety of a slope can only
be quantified on the basis of an analytical method,
whereas methods (a), (b) and (c) provide no such
quantification.
5. RELATIONSHIP BETWEEN RAINFALL AND

LANDSLIDES IN HONG KONG
5.1 Background
The vast majority of failures in residual soil slopes
are caused by rainfall. In several countries of the
world, studies have been carried out to correlate slope
failures with the pattern of rainfall. At their best,
such correlations provide a broad basis for predicting
widespread slope failures, which can lead to the
establishment of a warning system for those whose
lives might be endangered. Rainfall-failure correlations
Figure 5. Typical Hong Kong soil cut slope failure might therefore be regarded as ‘bulk’ predictions.
caused by heavy rainfall Simple direct rainfall-failure correlations have
been made for Brazil (Barata, 1969; Guidicini &
of analysis have long been used for this purpose, but Iwasa, 1977), Italy (Rossetti & Ottone, 1979), Japan
it must be realised that the majority of the world’s (Onodera et al, 1974; Fukuoka, 1980) and the United
man-made slopes were in fact formed on the basis of States (Campbell, 1975; Nilsen et al, 1976), whereas
experience and precedent. The vast majority of residual more sophisticated correlation attempts have been
soil slopes certainly fall into the latter category. undertaken in New Zealand (Crozier, 1969; Eyles et al,
Apart from the application of sound judgement 1978; Eyles, 1979; Crozier & Eyles, 1980) and Hong
and experience alone, there are four basic methods Kong (Lumb, 1975, 1979; Brand et al, 1984).
available for the prediction of rain-induced failures in It is very rare for detailed information to be
residual soil slopes. These are: available anywhere in the world on the geographical
(a) correlations between slope failures and pattern of distribution and short-term intensities of rainfall,
rainfall, measurements of rainfall commonly being made on a
(b) terrain evaluation, mainly on the basis of daily basis at stations which are far apart. It is also rare
geomorphological mapping, for the precise times to be known for the occurrence
(c) semi-empirical (or modified precedent) approach, of landslides. All previously published analyses of
which is based on an examination of the landslide data have therefore been based upon regional
geomorphology and geology of stable and unstable correlations of landslides with one-day or longer
slopes, and duration rainfalls, the more sophisticated correlations
(d) soil mechanics analytical methods, usually in the taking account of the antecedent rainfalls for periods
form of limit equilibrium analysis. of up to several weeks.
The first of these two approaches can be considered Relationships between rainfall and landslides were
to be directly related, since they apply to the stability first established for Hong Kong by Lumb (1975, 1979),
of a land area in one particular location. The last who classified landslide events into four categories,
three methods can be regarded as methods of analysis defined as follows :
and design; all three have been used extensively (a) ‘Disastrous event’ - Territory-wide damage, with
in residual soils. The terrain evaluation and semi- more than 50 individual failures recorded in one day,
Finish.indb 125 2007/8/17 11:03:32 AM

(b) ‘Severe event’ - widespread damage, with between elsewhere in the Territory for many years.
10 and 50 failures in one day, Since 1978, a sophisticated system of automatically
(c) ‘Minor event’ - localised damage, with less than 10 recording raingauges has come into use to provide data
failures in one day, and for the Geotechnical Control Office’s rainfall-landslide
(d) ‘Isolated event’ - a single individual failure. correlation programme. At present, 46 automatic
On this basis, Lumb correlated the landslide events gauges transmit rainfall measurements continuously
for the period 1950 to 1973 with the one day rainfalls through telephone lines to a central micro-computer.
and the 15-day antecedent rainfalls to obtain the The locations of these, and other continuously reading
interesting results shown in Figure 6. gauges, are shown in Figure 7.
Figure 6. Relationship between rainfall and landslides

established by Lumb (1975)
In incorporating antecedent rainfall into his Figure 7. Locations of continually recording

correlations, Lumb (1975) in effect took account of raingauges in Hong Kong, July 1985
the well-known phenomenon of soil moisture deficit
and its relationship to shear strength. His one-day Large geographical variations in rainfall occur
rainfalls, however, were taken as being the calendar- over the Territory during any rainstorm event, as
day rainfalls measured at the Royal Observatory (see dramatically illustrated by rainfall records shown in
below), and it is now known that better correlations are Figure 8 for three automatic gauges during a very
possible if account is taken of the localised variations recent rainstorm. These variations are attributable to
in rainfall. A further correlation study was therefore the effects of topography. A common feature of a Hong
undertaken recently in Hong Kong (Brand et al, 1984) Kong rainstorm is that the highest rainfalls occur near
which had the benefit of extremely detailed data not the tops of large hills, and that the rainfall isohyets
previously available. The results of this have greatly tend to follow the ground elevation contour lines.
aided our understanding of rain-induced slope failures, These features are illustrated in Figure 9, which shows
and may have application elsewhere. the contours of two separate three-hour rainfalls during
The rainfall-landslide correlation programme, a severe storm in May 1982.
carried out by the Geotechnical Control Office, was
based on the six main parameters:
(a) one-hour rainfall,
(b) 24-hour rainfall,
(c) antecedent rainfall for periods of up to 30 days,
(d) number of landslides which caused casualties,
(e) number of casualties, and
(f) times of landslide occurrence.
5.2 Rainstorm and Landslide Data

Excellent rainfall data is available for Hong Kong.
Rainfall records have been kept by the Royal
Observatory (RO) since 1884. Rainfall statistics are
referenced to the ‘principal’ gauge situated at the
Observatory on the Kowloon peninsula (Figure 7), at
which hourly measurements are made. In addition, Figure 8. Example of large variations in rainfall over
several autographic gauges have been installed short distances in Hong Kong
Finish.indb 126 2007/8/17 11:03:33 AM

If meaningful rainfall-landslide correlations are their human consequences. Seven of the rainstorms
to be made for any geographical location, accurate were associated with troughs of low pressure and six
information is needed on the number of landslides were caused by tropical cyclones. For the majority of
and their times of occurrence. In formation on precise these rainstorms, there are no adequate records of the
times of failure, however, is invariably difficult to actual numbers of landslides, but the small numbers of
obtain. In Hong Kong, there is often a delay of several Fire Services reports are indicative of the numbers of
hours between a failure occurring and an engineer particularly serious incidents.
being called out to inspect the failure; landslides in However, detailed landslide information was
rural areas are sometimes not inspected for several obtained by the Geotechnical Control Office for the
days. Information on times of failure is therefore often May and August 1982 rainstorms, which caused
not reliable from this source. over 1 500 slope failures (observed from aerial
In collecting data for its landslide correlation photographs), and the June 1983 rainstorm, which
study, the Geotechnical Control Office found the Fire caused more than 150 failures. While the numbers of
Services Department to be an extremely reliable source landslides reported in newspapers are well below the
of information on the occurrence of landslides in Hong actual figures, they give some relative measure of the
Kong over the past twenty years. Although the FSD number of failures that occurred during each storm.
is involved with only a small proportion of the total It should be noted from Table 3 that the number
number of landslides that occur, these comprise all of people killed and injured was not necessarily
the more serious incidents, and their excellent records proportional to the severity of a particular rainfall
can be considered to be representative of the landslide event, since a few major individual landslides
events in Hong Kong as a whole. Their data which were responsible for large numbers of casualties.
relates to times of call-out is particularly good, and the In particular, the June 1972 storm resulted in 224
fact that they are usually called to a serious incident casualties from the two failures that occurred at Po
immediately after its occurrence enables the time of Shan Road on Hong Kong Island and Sau Mau Ping in
failure to be fixed fairly accurately. The correlation Kowloon (Government of Hong Kong, 1972a, 1972b;
studies were therefore undertaken largely on the basis Vail, 1984). Also, the vast majority of the people
of Fire Services data. who were permanently evacuated from their homes
The Geotechnical Control Office’s rainfall-landslide in the 20-year period as a result of rainstorms were
correlation study was based on data available for the people who occupied temporary squatter dwellings on
20-year period 1963 to 1983, for which the 13 severest unformed hillside sites.
rainstorm events are listed in Table 3, together with
Table 3. The thirteen major rainstorm events in Hong Kong and their consequences during the period 1963-1983
Maximum Rainfall, mm Landslide Consequences
Date Type Observatory Other Location No. Failures No. Fire No. People No. People
of of Reported in Services Killed or Permanently
Rainstorm Storm 24-hour 1-hour 24-hour 1-hour Newspapers* Reports Injured Evacuated
24-25 August 1976 STS Ellen 416 52 500 82 314 23 57 2400
12 June 1966 trough 401 108 525 157 100 30 35 8500
29 May 1982 trough 394 44 430 111 498 15 48 8000
16-17 October 1978 STS Nina 380 37 380 38 15 1 1 no record
16 August 1982 STS Dot 362 68 370 95 62 6 9 1500
17 June 1983 trough 347 69 460 101 114 5 2 600
27 September 1965 TS Agnes 333 47 333 47 9 4 4 200
17 August 1971 Typhoon Rose 328 63 328 63 10 5 7 no record
12-13 October 1964 Typhoon Dot 304 60 375 94 8 10 39 8000
12-13 June 1968 trough 287 100 343 143 10 7 27 200
16-17 June 1972 trough 280 36 560 71 >15 15 21
17-18 June 1972 trough 275 99 300 98 “dozens” 14 229 } 7800
17 May 1972 trough 271 79 377 92 “dozens” 2 0 6000
TS - Tropical Storm *The numbers of landslides reported in newspapers are for comparison only - they
STS - Severe Tropical Storm represent the lower limits of the numbers of landslides that actually occurred.
Finish.indb 127 2007/8/17 11:03:33 AM

Figure 11. Occurrence of landslides in relation to

Figure 9 Locations of landslides in relation to the rainfall at the Observatory and at any location during
3-hour isohyets in the May 1982 rainstorm the June 1966 rainstorm
5.3 Effect of Rainfall Intensity

The available rainfall and landslide data was examined

in detail for all thirteen rainstorm events listed in
Table 3. Figure 10 shows the patterns of rainfall
for five of the most serious rainstorms (in terms
of consequence). It will be seen that these storms
generally lasted for a few days, but the majority of the
rain fell in a few hours in each case. All these storms
were therefore associated with short duration, high
intensity rainfalls. Where sufficient data was available,
a definite pattern emerged from a close examination of
this fact for each of the thirteen cases.
The storms of June 1966, October 1978, May Figure 12. Occurrence of landslides in relation to
1982 and August 1982 will be used as examples. rainfall at the Observatory and at any location during
The last two are of particular importance, because the May 1982 rainstorm
the Geotechnical Control Office obtained detailed
information on these two severe storms, which resulted
in the loss of 27 lives, and which ranked third and
fifth of the last 20 years’ rainstorm events in terms of
24-hour rainfall (Table 3).
Figure 13. Occurrence of landslides in relation to

rainfall at the Observatory and at any location during
the August 1982 rainstorm
In June 1966, the heaviest rainfall in memory fell

on Hong Kong Island, and the highest rainfall to-
Figure 10. Hong Kong’s five severest rainstorms in date was recorded at the Observatory. Widespread
1963-83 landslides and flooding occurred throughout the
Territory. For this rainstorm, Figure 11 shows the
10
Finish.indb 128 2007/8/17 11:03:34 AM

hourly cumulative rainfall at the Observatory and at
the location of measured maximum rainfall in each
case, compared with the occurrence of landslides as
recorded by the Fire Services Department. The pattern
of cumulative rainfall measured at the Observatory
is quite similar to that measured at the location of
maximum rainfall, although it is inevitably somewhat
lower. In contrast, the hourly maximum rainfalls are
much higher than the corresponding ones measured at

the Observatory. Landslide timings for the rainstorm
show very good correlations with hourly maximum
rainfalls, which suggests that high intensity rainfall
of short duration is a major triggering factor for
landslides.
For the May 1982 rainstorm, Figure 12 shows the
rainfall and Fire Services landslide information. It
can be seen that virtually all the recorded landslides Figure 14. Occurrence of landslides in relation to
occurred at the time of the maximum rainfall intensity. rainfall at the Observatory at any location during the
Also, although the maximum hourly rainfalls are October 1978 rainstorm
dramatically higher than those measured at the with increasing maximum hourly intensity. The actual
Observatory, the pattern of cumulative rainfall is very
number of casualties caused by landslides shows a

similar in each case. similar but even more dramatic correlation with hourly
For the August 1982 rainstorm, the data plotted in intensity.
Figure 13 shows essentially the same pattern depicted
in Figures 11 & 12.
In Figure 14 is shown the data for the October
1978 rainstorm. Although the 24-hour rainfall of
380 mm was the fourth highest in the 20-year period
(and is similar to that for the May 1982 rainstorm,
Figure 12), only one landslide and one casualty were
reported for this event, for which the maximum hourly
rainfall was only 38 mm. This lends confirmation to
the dependence of landslide occurrence on the short
duration rainfall intensity.
During 1982, 3 248 mm of rain fell in Hong Kong,
the highest annual rainfall on record. The distributions
of all rainy days and of days with three or more
landslides are shown in Figure 15, plotted against the
maximum hourly rainfall anywhere on each day. When Figure 15. Correlation of landslide occurrence during
the maximum hourly rainfall was less than 20 mm, no 1982 with the maximum hourly rainfall intensity
slope failures usually occurred. In contrast, failures anywhere
nearly always occurred on the few days on which the
maximum hourly rainfall intensity exceeded 40 mm
per hour. The failures that occurred were also found
to be in the vicinity of the area that experienced the
highest rainfall intensity.
For the 20-year period from 1963 to 1983, an
analysis of all the times of occurrence of landslides
was carried out. The results presented in Figure 16
show that the majority of the reported landslides
occurred within four hours of peak rainfall intensity,
with only 10% taking more than 16 hours to occur.
The dependence of landslide occurrence and
casualties on hourly rainfall intensity is clearly shown
in Figure 17. Throughout the twenty-year period, the
number of landslides that caused casualties was very
small unless the maximum hourly rainfall anywhere in Figure 16. Times of occurrence of major landslides
the Territory approached 70 mm. Above this figure, the as related to the maximum hourly rainfall intensity,
number of landslides with casualties increased sharply during the period 1963-83
11
Finish.indb 129 2007/8/17 11:03:35 AM

Correlations between landslides in Hong Kong
and the Observatory 24-hour rainfall are shown in
Figure 19 on the basis of two categories of event. A
‘minor’ event is one for which there are less than ten
recorded landslides in one day, and a ‘major’ event is
one for which there are more than ten landslides in one
day. The ‘minor’ category is identical to that used by
Lumb (1975), while the ‘major’ category corresponds
to his combined ‘severe’ and ‘disastrous’ categories.

Figure 19 demonstrates clearly that, where the 24-hour
rainfall at the Observatory is below 100 mm, only a
few minor events and no major events occur. With
Figure 17. Correlation of landslides causing casualties increasing 24-hour rainfall, the proportion of landslide
and number of casualties during 1963-83 with days increases, with major events increasingly
maximum hourly rainfall anywhere dominating at higher rainfalls. When more than
270 mm of rain falls in 24 hours, every rainstorm
5.4 Correlations with 24-hour Rainfalls event results in a major landslide event.
Although clear evidence has been presented that Hong
Kong landslides are dependent on short duration
rainfall intensity, it is worth examining whether any
correlations exist between landslides and 24-hour

rainfalls.
For the May 1982 rainstorm, Figure 18 shows
the daily rainfalls at the Observatory for the event,
together with the occurrence of all landslides for
which the dates of failure are known. This shows that
the large majority of landslides occurred on the 29th
May, the day of the heaviest rainfall. The few failures
that occurred up to a week later were probably caused
by delayed groundwater build up (Hudson & Hencher,
1984).
Figure 19. Correlation of number of landslides with
the 24-hour rainfall at the Observatory during the
period 1963-83
The relationship shown in Figure 19 is not unlike

the distribution in Figure 15 of landslide days in
relation to maximum hourly rainfall for 1982. This is
to be expected, since the heavier the 24-hour rainfall,
the more likely it is that it will include a short duration
rainfall of high intensity. The main importance of
the correlation between landslides and the 24-hour
rainfall, however, is its usefulness as a landslide
warning criterion, since the accumulation of rainfall to
Figure 18. Relationship between the daily rainfall approach a ‘trigger’ value can be sensibly anticipated
at the Observatory and the occurrence of landslides a few hours in advance, whereas the short duration
during the May 1982 rainstorm intensity cannot possibly be predicted in advance with
any precision. The use of the 24-hour rainfall also has
The often large short-term differences in rainfall the advantage that measurements from the Observatory
from place to place throughout Hong Kong tend gauge alone can be used to represent the situation
to even out over the longer rainfall periods. This is throughout the whole Territory.
demonstrated in Figures 11 to 14, where the 24-hour
rainfalls at the Observatory are generally not greatly 5.5 Effect of Antecedent Rainfall
different from the maximum values measured The evidence presented here suggests that the
anywhere. It is therefore useful to examine the antecedent rainfall is not a significant factor in the
relationship between landslides in Hong Kong and the occurrence of major landslide events in Hong Kong,
Observatory 24 hour rainfall. despite the fact that it has been found to be of major
12
Finish.indb 130 2007/8/17 11:03:35 AM

significance elsewhere. This was confirmed in the
study by the poor correlations obtained between
landslides and antecedent rainfalls for periods of
up to 30 days. In Hong Kong, failures appear to be
directly related to short duration rainfall intensity. It
is common, however, for such high intensity rainfalls
of short duration to occur during prolonged rainfall
events. Where limited information is available on
rainfall intensities and times of landslides, it is

therefore not surprising that a relationship can be
found to exist between antecedent rainfall and the
occurrence of a large number of landslides.
Of interest is the fact that the three- or four-day Figure 20. Approximate relationship between rainfall
antecedent rainfall appears to influence the occurrence intensity and landslide events in Hong Kong
of minor landslide events in Hong Kong when short
duration intensities do not dominate. This indicates engineers, who could make much more use of the
that soil moisture deficit might well be a factor in rain- technique for planning and development studies, and
induced landslides in soil slopes, but that its effects are to provide input for land use management purposes.
secondary in situations of high intensity rainfall. The evaluation of terrain is based on a suitable
‘terrain classification’ system, which provides a
5.6 Conclusions means for classifying terrain into homogeneous land
From the rainfall-landslide correlation study in Hong units, largely on the basis of surface features. For this
Kong, the following main conclusions can be drawn : purpose, suitable geomorphological ‘attributes’ must
(a) The large majority of landslides are induced by be chosen. Attributes for stability purposes commonly
localised short duration rainfalls of high intensity, include slope angle, land form, vegetation, surface
and these landslides take place at about the same hydrology and erosion, each of which is divided into
time as the peak hourly rainfall. a number of ‘classes’. Once classified in this way, the
(b) Antecedent rainfall is not a major factor in terrain evaluation is carried out by sorting the land
landslide occurrence, except in cases of minor units into categories on the basis of predetermined
landslide events which take place under relatively combinations of the attribute classes.
low intensity rainfalls of short duration. In these An excellent example of the development of
circumstances, only a few days antecedent rainfall a comprehensive system of terrain evaluation for
appears to be significant. engineering purposes is the Australian PUCE system
(c) A rainfall intensity of about 70 mm/hour appears described by Grant (1975a, 1975b). This has been
to be the threshold value above which landslides developed continuously over 20 years into a highly
occur. The number of landslides and the severity of complex system which requires the use of a large
the consequences increase dramatically as hourly storage capacity computer (Grant & Finlayson, 1978;
intensity increases above this level. Finlayson, 1984).
(d) The 24-hour rainfall usually reflects short duration In addition to general engineering geological
rainfalls of high intensity, and this can therefore be assessment maps, ‘risk’ or ‘hazard’ maps can be
used as an indicator of the likelihood of landslides. constructed from terrain evaluation data. This has been
A 24-hour rainfall of less than 100 mm is very done in some landslide-prone areas of a few countries,
unlikely to result in a major landslide event. notably France (Porcher & Guillope, 1979), Italy
The results of the rainfall-landslide correlation (Carrara et al, 1978; Carrara, 1983) and the United
study are summarised broadly in Figure 20, which States (Nilsen & Brabb, 1977; Nilsen et al, 1979).
shows the approximate frequency of landslide events Good reviews of the various systems used have been
(as defined by Lumb, 1975) in terms of one-hour and made recently by Brabb (1984), Varnes & Keaton
24-hour rainfalls. This forms the basis of a landslide (1984) and Varnes (1985).
warning system currently used in Hong Kong for The most advanced form of the landslide hazard
hillside squatters. map is the digitized form. The methodology used
is no different in principle from that for manually-
compiled maps, but the terrain classification is
6. TERRAIN EVALUATION digitized to facilitate ease of storage, retrieval and
sorting. The computerstored data also acts as a data
6.1 Background bank of information that can be called upon at will.
Hazard maps have been prepared in this way in a few
Terrain evaluation is a powerful technique for
countries, the most advanced system probably being
categorising terrain for stability purposes. It has
that developed by Carrara et al (1978) and Carrara
hitherto been somewhat neglected by geotechnical
(1983) in Italy.
13
Finish.indb 131 2007/8/17 11:03:36 AM

6.2 The GASP System in Hong Kong Study, an essential part of which is a base map and a
series of transparent overlay maps.
The Geotechnical Control Office in Hong Kong has
For a Regional Study (1:20 000), the overlay maps
for several years placed considerable emphasis on
consist of:
its Geotechnical Area Studies Programme (GASP)
(a) Terrain Classification Map,
to provide systematic geotechnical input for land use
(b) Landform Map,
management and development planning purposes. This
(c) Erosion Map,
programme and some aspects of the system of terrain
(d) Physical Constraints Map,
evaluation employed have been described in several
(e) Engineering Geology Map, and
publications (Brand et al, 1982a, 1982b; Burnett &

(f) Geotechnical Land Use Map (GLUM).
Styles, 1982; Burnett et al, 1985).
GASP was designed to be carried out in the
For a District Study (1:2 500), the overlay maps are :
following three phases :
(a) Terrain Classification Map,
(a) Regional Studies - Initial geotechnical assessments
(b) Surface Hydrology Map,
(at a scale of 1:20 000), based entirely on aerial
(c) Vegetation Map,
photograph interpretation, site reconnaissance and
(d) Engineering Data Sheet,
existing geotechnical information.
(e) Engineering Geology Map, and
(b) District Studies: Stage 1 - Initial geotechnical
(f) Geotechnical Land Use Map (GLUM).
assessments (at a scale of 1:2 500), based
The Engineering Geology Map indicates the
entirely on aerial photograph interpretation,
broad pattern of the geological materials and their
site reconnaissance and existing geotechnical
general engineering characteristics, and it identifies
information, to give more detailed assessments of
features which are of engineering relevance. One
specific areas identified from Regional Studies.

of the major purposes of this map is to present
(c) District Studies: Stage 2 - Expanded geotechnical
the geomorphological and engineering geological
assessments, based on the results of Stage 1
constraints that influence the allocations of GLUM
Studies together with data obtained from planned
classes (see below). It is designed for use by engineers
programmes of site investigation.
and engineering geologists who require explanations
Regional Studies, each of which covers an area
for the nature of the geotechnical limitations which
of 50 to 100 sq. km, provide geotechnical input for
affect the terrain.
outline and strategic planning within the Territory. In
For a Regional Study, the Physical Constraints
addition, they are designed to provide a comprehensive
Map summarises the physical constraints to regional
physical land resource inventory through a computer-
development planning. For a District Study, every
based Geotechnical Terrain Classification System
item of available engineering geological information
(GEOTECS). All eleven Regional Studies have now
(including every borehole location) is plotted and
been completed or are in their final stages.
referenced on the Engineering Data Sheet to avoid
District Studies are typically concerned with areas
confusion on the Engineering Geology Map; this
of 2 to 4 sq. km. They provide information suitable for
information is shown in a much more general way on
local planning needs and are of assistance in the basic
the Engineering Geology Map of a Regional Study.
layout planning of large sites. For most localised areas,
The Engineering Geology Map and Engineering Data
Stage 1 District Studies provide adequate information,
Sheet have been described in some detail by Burnett &
a Stage 2 Study being required only if very significant
Styles (1982).
geotechnical constraints are revealed by Stage 1. Nine
The main findings of a Geotechnical Area Study
Stage 1 Studies have so far been completed. As yet, no
are engendered in the Geotechnical Land Use Map
Stage 2 Studies have been undertaken as an integral
(GLUM). In this interpretative map, land units are
part of GASP, but a related Stage 2 area study was
classified into four GLUM ‘classes’ on the basis
completed several years ago (Geotechnical Control
of combinations of attributes from the terrain
Office, 1982; Rodin et al, 1982) .
classification, with due account being taken of the
The Geotechnical Control Office realised at an early
other data collected during the Study. E x a m p l e s o f
stage that no existing system of terrain evaluation was
a GASP terrain classification map and a GLUM are
ideally suited to Hong Kong’s peculiar conditions. The
given in Figures 21 & 22. The GLUM class is assigned
GASP approach differs in several important aspects
from any other known terrain evaluation or hazard to a unit on the basis of the likely geotechnical
mapping system. It combines elements of both of these limitations on the land unit posed by the combination
approaches, but the zonation framework employed is of its categories of the six attributes - slope gradient,
based on the overall geotechnical assessment of the terrain component, terrain morphology, erosion (which
land units, not only on the identification of ‘hazards’ includes instability), slope condition and hydrology.
from a stability viewpoint. As might be expected, The four GLUM classes are summarised in Table 4.
however, considerations of stability are dominant. The results of GASP are intended for use largely
A comprehensive report is produced for each Area by planners, but they also provide valuable data
for engineering appraisal and feasibility studies. A
14
Finish.indb 132 2007/8/17 11:03:36 AM

Table 4. The four-class classification system used for the Geotechnical Land Use Map (GLUM) in Hong Kong
Characteristics of
Class I Class II Class III Class IV
GLUM Classes
Geotechnical
Low Moderate High Extreme
Limitations
Suitability for
High Moderate Low Probably unsuitable
Development
Engineering Cost for
Low Normal High Very high
Development
Intensity of Site
Normal Normal Intensive Very intensive
Investigation Required
Typical Terrain Insitu terrain with Insitu terrain with Insitu terrain with Very steep insitu
Characteristics gentle slopes slopes between slopes between slopes (>60° and
(Some, but not (0- 15° without 15° & 30° without 30° & 60° without cliffs.
necessarily all, of the severe erosion or instability. severe erosion or Steep to very steep
stated characteristics instability. Insitu terrain of instability. insitu and colluvial
will occur in the Cut platforms in gentle slopes Insitu terrain slopes with history
respective GLUM insitu terrain. associated with less than 15° of instability.
class) drainage but no with history of Colluvial terrain with
instability. landslips. gentle slopes, but
Colluvial terrain Colluvial terrain associated with
with gentle slopes less than 15° instability and

(0-15°) without with evidence of drainage.
severe erosion or instability.
instability. High to moderate fill
slopes.
Figure 21. Example of a Terrain Classification Map

(Alpha-numerics denote attribute classes)
recent study carried out to compare the GLUM class Figure 22. Example of a Geotechnical Land Use Map
assessments in the Mid-levels area with the results
(GLUM) for the area shown in Figure 21
of engineering stability analyses (Styles et al, 1984)
indicated clearly that the GASP system provides a
essentially on the basis of common sense and were
reliable means of rapidly delineating localised areas of
then gradually modified in the light of experience with
potential instability.
the performance of the slopes. In a few instances, this
‘precedent’ approach has evolved into a formulated
semiempirical design method in the form of a set of
7. SEMI-EMPIRICAL DESIGN METHODS design rules.
7.1 Background A particular set of slope design rules can obviously
only be applied to a specific geological formation at
The majority of existing cut slopes in residual soils one geographical location, because these rules have
were not designed in the engineering sense but were evolved from the topography, geology and climate
formed on the basis of judgement or precedent. of that location. It is true, however, that design rules
In most cases, the precedents were established developed for one location may provide a good guide
15
Finish.indb 133 2007/8/17 11:03:37 AM

for the development of rules for a similar geological
formation elsewhere.
Semi-empirical design rules must be based on
the examination of stable and unstable slopes of
similar geology. This can in some cases result in the
establishment of simple relationships between slope
height and slope angle. At the other extreme, very
complicated statistical relationships can sometimes
be devised among a large number of geological and

geomorphological parameters.
An example of the simplest type of semi-empirical
design rules is illustrated in Figure 23, which shows
the recommended slope angles for different heights
of cut slopes on the basis of the degree of weathering
of Wellington greywacke (Grant-Taylor, 1964; Taylor
et al, 1977). Earlier, Lane (1961) had correlated slope Figure 24. Design guidelines developed by Lane
height and slope angle for slopes in clay shales in (1961) on the basis of the performance of clay shale
Montana, North Dakota and Kansas, to produce the slopes in Northern USA
plot shown in Figure 24. The work by Shuk (1965,
1968), summarised in Figure 25, is of particular note,
since he combined a probability of failure in his log-
log plot of height versus slope angle for natural shale

slopes near Bogota, Colombia. Shuk’s approach was
later used by McMahon (1976) for data obtained on
233 slopes, for which he found that the log-log straight
line relationship applied for any rock type.
Morphometric models have been used to establish
the attributes common to landslides. The work by
Blong (1973) and Crozier (1973) are good examples
of this approach in its more descriptive form. A more
complex model was devised by Neuland (1976)
using 31 variables, which included morphometric,
geological, stratification and soil mechanics attributes.
Equally sophisticated are the terrain models being used
in parts of Italy (Carrara et al, 1977, 1978; Carrara,
1983).
7.2 The CHASE Study in Hong Kong

A major Study (CHASE) was carried out in Hong
Kong a few years ago (Brand, 1982; Brand & Hudson,
1982) to examine the possibility of establishing semi-
empirical guidelines for the design of cut slopes in
the residual soil profiles. Detailed examinations were
Figure 25. Design guidelines developed by Shuk

(1965, 1968) for clay shale slopes near Bogota
conducted on 177 stable and failed slopes, and the
data was analysed statistically in an attempt to obtain
simple correlations among the geological, hydrological
and geometrical factors which control cut slope
stability. More than 200 items of data were collected
for each slope to ensure that any factors which
influenced its stability were recorded.
A typical CHASE slope is shown in Figure 26.
This is 45 m high and is cut at an angle of about
Figure 23. Simple empirical design rules developed for 60° into a steep natural hillside covered with dense
Wellington greywacke by Grant-Taylor (1964) vegetation. The material of the slope comprises 5.5 m
16
Finish.indb 134 2007/8/17 11:03:37 AM

Figure 26. Typical stable Hong Kong cut slope

included in the CHASE Study (Brand & Hudson,
1982)
of colluvium overlying decomposed granite, which

varies from grade V at the top to grade IV at road
level. The surface is covered with a protective coating
of ‘chunam’ (lime stabilised soil), which is commonly Figure 27. Relationship between height and angle for
used in Hong Kong to prevent surface erosion and to the 177 Hong Kong cut slopes included in the CHASE
minimise infiltration. Study (Brand & Hudson, 1982)
Despite the vast amount of data collected, and the
extensive statistical analyses undertaken, the CHASE
Study did not reveal the simple relationships hoped
for, but it did result in the establishment of some lower
bound envelopes for the relationships between height
and slope angle of the kind shown in Figures 27 & 28.
These have proved to be useful for the initial design of
new cut slopes, and for assessments of the likelihood
of failure of existing slopes.
Figure 27 is of interest for the information it
provides about empirical design rules adopted at
various times in Hong Kong’s history. It can be seen
that none of these was meaningful in the context of
the results of the CHASE Study. It is also evident
that classical stability analyses are probably very
conservative for Hong Kong conditions. The same
conclusion is well supported by Figure 28, where
the changes in slope height and angle are shown for
a large number of granitic soil cut slopes which had
previously failed.
8. SOIL MECHANICS PREDICTIONS

8.1 Summary
Slope stability analysis is usually based on the limit
equilibrium approach, for which the equilibrium of a Figure 28. Relationship between height and angle for
sliding mass is examined (Figure 29). The degree of the 117 Hong Kong granite soil cut slopes included in
stability is quantified in terms of a ‘factor of safety’, the CHASE Study (Brand & Hudson, 1982)
17
Finish.indb 135 2007/8/17 11:03:40 AM

which is most commonly defined as the ratio between Table 5 summarises, in very broad terms, the
the average shear resistance and the average shear Author’s assessment of the overall state-of-the-art of
stress along the most critical slip surface, i.e.: stability predictions in residual soil slopes in terms
of the five components of prediction. The accurate
Sa determination of loads is thought to present very
F= (1) little difficulty, the dead weight of the sliding mass
τa
(Figure 29) being readily computed on the basis of
data from borehole samples. Neither are there reasons
The determination of the F-value for a particular to believe that the available methods of analysis are
slope therefore requires : anything but very good, as long as these are restricted
(a) prediction of the correct mode of failure (i.e. to appropriate methods for the particular mode of
selection of the critical slip surface), failure. Our ability to select a correct mode of failure
(b) prediction of the distribution of shear stress over varies fairly widely depending on the geology and
the critical slip surface, and weathering profile for the slope. Because of the small
(c) prediction of the distribution of shear resistance amount of work that has been done to investigate
over the critical slip surface. the shear strength properties of residual soils, our
The distribution of shear stress along the critical knowledge of this component is only fair. However,
slip surface is dependent on the loading and the by far the major difficulty with stability predictions
method of analysis employed. The shear resistance in residual soil slopes is the poor state-of-the-art with
along the slip surface is governed by the effective respect to the prediction of pore pressure distribution
shear strengths of the materials of the slope and at failure.
the normal effective stress distribution. The normal

effective stress distribution is, in turn, a function of the 8.2 Mode of Failure
pore pressure distribution at failure and the method of The selection of a suitable mode of failure is crucial
analysis. The five ‘components’ of stability prediction to accurate stability predictions in all soil types. It
are therefore: is particularly crucial for slopes in residual profiles,
(a) mode of failure, where the potential failure surface is often governed
(b) loading, by geological detail. Some guidance on the accurate
(c) method of analysis, assessment of failure modes for geologically complex
(d) shear strength, and conditions was given many years ago by Terzaghi
(e) pore pressure distribution at failure (1950). More recent important publications in this
context are those by Skempton & Hutchinson (1969),
Deere & Patton (1971), Patton & Hendron (1974) and
Coates (1977).
Whereas the pre-failure geometry of a slope is
easily defined, it is sometimes difficult to decide
upon the critical potential failure surface for design.
Occasionally, even a post-failure surface cannot
readily be determined because of the multiple-failure
nature of some landslides in residual materials.
As a general rule, the shear strength of a residual
soil increases with depth, and slope failures can
therefore be expected to occur on relatively shallow
slip surfaces. These surfaces are largely controlled by
Figure 29. Basis of the limit equilibrium method of the weathering profile. Failures most frequently occur
stability analysis for slopes along surfaces dictated largely by relict joints or by
boundaries between weathering zones. By the very
Table 5. Assessed state-of-the-art of stability nature of residual weathering profiles, non-circular
predictions for residual soil slopes failure surfaces are the most common, and these are
often almost planar over a major proportion of their
Component State-of-the-Art length.
Mode of failure Good to poor Geological detail is often crucial to the location of
Loading Very good the critical slip surface. A complex weathering profile
Method of analysis Very good can rarely be adequately investigated at the design stage.
Shear strength Fair In those few instances where the profile is determined in
considerable detail by extensive surface and subsurface
Pore pressure Very poor
investigations, the true geological situation usually
Overall Fair to poor does not lend itself to being properly modelled for the
purposes of meaningful engineering analysis.
18
Finish.indb 136 2007/8/17 11:03:40 AM

The importance of soil fabric to the mode of slope methods of analysis for non-circular slip surfaces.
failure cannot be over-emphasized. In particular, Many incorrect answers were received, and this
relict joints, especially when slickensided, can be reflected the fact that many of the computer programs
instrumental in the initiation of a failure, and their being used had never been properly checked or
presence can appreciably reduce the mass strength of documented. In addition, an alarming number of errors
the soil (see Section 8.4). had resulted from careless data input.
Some examples of the importance of geological
detail to slope failure modes in residual soils are given 8.4 Shear Strength Measurement
later in Section 9. Despite the obvious objections to the measurement of

shear strengths on residual soils by means of laboratory
8.3 Methods of Analysis
tests, these still comprise the most satisfactory means
Although the state-of-the-art with respect to of establishing the likely range of shear strengths on
methods of stability analysis is thought to be very the softer materials (grades V & VI). The effects of
good, it is well-known that the available methods corestones and other large-sized particles, however,
produce different theoretical F-values because of the cannot be determined, and there is no doubt that
different assumptions made in their formulation. Many laboratory strength tests carried out on the ‘matrix’
papers have been published which compare the various material of residual soils and colluvium will usually
assumptions and which make comparisons of the underestimate the shear strength of the insitu material
numerical factors of safety obtained for certain specific because of the neglect of the boulder content.
slope situations. For the up-to-date position, reference Residual soils and colluvium are invariably
should be made to the recent publications by Fredlund unsaturated and are of relatively high permeability.
et al (1981), Fredlund (1984) and Ching & Fredlund Stability computations must therefore always be made
(1984). in terms of effective stresses; analysis on the basis of
For residual soil slopes, methods of stability undrained strengths has no relevance.
analysis which employ circular slip surfaces are Because of their high permeabilities (usually 10-4
usually inappropriate, and methods which apply to to 10 -6 m/sec.), rainwater infiltrates with ease into
any shape of surface must be used. In this category, most residual soils and colluvium, and it is thought
are the well-known methods by Janbu (1954, 1973), likely that saturation conditions will be approached at
Morgenstern & Price (1965) and Sarma (1973, 1979). shallow depths in the field during the life of a slope. It
The first two of these are in very common usage. is therefore generally felt to be appropriate to measure
Although Morgenstern & Price made more satisfactory strength parameters on the basis of shear tests carried
assumptions than Janbu, the latter’s method is much out on saturated soil specimens. Although this may
easier to programme and requires much smaller often be a more severe condition than that experienced
computer capacity. For these reasons, it is more widely by the soil insitu, it remains the only certain means
used for routine stability calculations. Sarma’s method by which a ‘base’ shear strength envelope can be
is as rigorous as that of Morgenstern & Price, but it established.
requires less iterations with consequent less computer The triaxial test is the most widely used method for
time. shear strength measurement on residual soils (Brand
All the above methods of slope stability analysis & Phillipson, 1985). Test specimens should be as large
are based on a two-dimensional failure mode. In as possible, full-diameter lengths of drillhole sample
practice, however, slope failures are nearly always being ideal for routine work, and they are usually
three-dimensional, but it is rare for account to be taken saturated by the application of a sufficiently high
of this in design or stability assessment, or even during back pressure prior to shear. Either drained tests (CD)
the back analyses of failures that have occurred. Three- or consolidated undrained tests with pore pressure
dimensional effects on slope stability analysis have measurement (CU) can be used, but the latter are
been examined by Baligh & Azzouz (1975), Azzouz et much to be preferred, because they are quicker and
al (1981), Lovell (1984) and Leshchinsky et al (1985), provide much more information about the stress-strain
but more work on this needs to be done. behaviour of the soil.
A great deal of effort has been put into refining Cell pressures used for triaxial testing must relate
the details of methods of stability analysis over the to the correct insitu stress range if the measured
years, without sufficient attention being paid to the strengths are to be meaningful. Critical slip surfaces
way in which these methods are used. A small study in residual soil slopes are most commonly shallow,
carried out in Hong Kong (Lumsdaine & Tang, 1981, and the effective stresses on these are therefore low
1982) is revealing in this respect. A large number of (typically 30 to 200 kPa). At such low effective cell
organisations intimately involved with slope stability pressures, however, triaxial tests are difficult to control
assessment in Hong Kong on a regular basis were satisfactorily, and this stress range is not recommended
asked to compute the factors of safety for a number for routine use. There is some evidence to suggest
of problem situations by means of the well-known that the strength envelopes for some residual soils
19
Finish.indb 137 2007/8/17 11:03:41 AM

are curved at low effective stresses, and that the
straightline projection of strengths measured at high
stresses underestimates the strengths in the low stress
range (Figure 30). It is almost certain that this is part
of the explanation for why many stable residual soil
slopes have theoretical factors of safety of less than 1.0.
As discussed in Section 8.5 below, failures of
residual soil slopes are invariably caused solely by
pore pressure increases. In its simplest form, the

mechanism of failure is therefore that the soil suction
(i.e. negative pore pressure) decreases (i.e. pore
pressure increases), thereby reducing the shear strength
of the soil, as dictated by the effective stress principle.
Failure will occur when the average shear resistance
on the critical slip surface has decreased to the value of
the average shear stress on that surface. Rain-induced Figure 31. Comparison between stress paths for
slope failures, therefore, take place under conditions rain-induced slope failure and those for triaxial tests
of almost constant total stress and increasing pore (Brand, 1981)
pressure (Brand, 1981).
for stability assessment (Figure 30). It is possible to
follow the correct stress path in the laboratory simply

by decreasing the cell pressure, but this does not
simulate the correct mechanism of failure in the field.
This can only be done by means of a constant load test
in which the pore pressure is increased from an initial
negative value until failure occurs. Tests of this kind
have been carried out by the Geotechnical Control
Office for some time, but the experimental difficulties
are considerable, and this test procedure cannot be
Figure 30. Underestimate of shear strengths because

of incorrect strength envelope deduced from data
obtained at too high normal stresses
The triaxial test is most commonly conducted

by increasing the axial stress, σl, to failure while the
cell pressure σ3 is kept constant. This is carried out
either as a drained test (∆u = 0), or as an undrained
test (or constant water content test) with pore pressure
measured throughout the shearing process. In contrast,
for rain-induced failure, σl and σ3 are almost constant,
and the pore pressure increases to failure. The
comparative stress paths for the common triaxial tests
and for the real field situation are shown in Figure 31,
in which p' & q are defined as :
σ'1 + σ'3 σ1 - σ3
p' = ;q= (2)
2 2
Figure 31 illustrates that the stress paths commonly

followed in the triaxial test are quite different from
that which pertains in the field, and that the stress
ranges over which triaxial tests are usually conducted
are generally not appropriate to the field stress
conditions. Where the strength envelope of a soil
is markedly curved, this disparity will result in an Figure 32. Insitu direct shear test apparatus used on
appreciable underestimate of the correct shear strength Hong Kong residual soil (Brand et al, 1983b)
20
Finish.indb 138 2007/8/17 11:03:42 AM

recommended for routine use. to extremely low relict joint strengths. One such well-
Direct shear (shear box) testing is deservedly documented Hong Kong case (Hunt, 1982), which was
becoming more commonly used for the strength of major economic significance, led to an extensive
assessment of residual soils. Apart from its relative investigation into the shear strength of jointed
simplicity, this method has obvious merits over the sedimentary volcanic material (Koo, 1982a, 1982b).
triaxial test when determinations of strengths along Persistent joint surfaces were coated with a thin black-
relict joints are required, because orientated specimens brown deposit believed to be the precipitation of iron
can be prepared. However, there is no certain method and manganese oxide products that filled the joints
of saturation, and pore pressures cannot be measured in the parent rock during the course of weathering.
during shear. Despite its lack of theoretical ‘purity’, the The results of laboratory direct shear and triaxial tests
direct shear test is seen by some as providing a means showed that the effective strength parameters on the
of obtaining shear strength data under conditions joints were much lower than those measured on the
which model those in the field more closely than intact material.
the triaxial test. It is also readily adaptable for insitu
measurements of shear strength on relatively large 8.5 Pore Pressure Prediction
masses of material. The equipment shown in Figure 32
For a residual soil slope, the prediction of the pore
has been used successfully for testing 300 x 300 mm
pressure distribution is by far the most critical factor
soil and soft rock specimens in Hong Kong (Brand et
for stability analysis. This is particularly so since most
al, 1983b). A typical test specimen of grade V granitic
failures in residual soil slopes are caused by rainfall.
soil is shown in Figure 33.
The hydrological effects of rainfall on a permeable
slope are depicted in Figure 34. Some of the water

runs off the slope and may cause surface erosion if
there is inadequate surface protection. Because of the
high soil permeability, however, the majority of the
water infiltrates. This causes the water table to rise,
or it may cause a perched water table to be formed at
some less permeable boundary, usually dictated by the
weathering profile. Above the water table, the degree
of saturation of the soil increases, and the soil suction
(i.e. negative pore pressure) therefore decreases.
Failures in residual cut slopes are thought to be
caused mostly by the ‘wetting-up' process by which the
soil suction (and hence the soil strength) is decreased,
but there is some evidence to suggest that transient
rises in groundwater tables are responsible for some
rain-induced landslides (Premchitt et al, 1985). Rain-
induced slope stability failures thus occur as a direct
result of pore pressure increases (see Section 8.4), and
pore pressure distribution is therefore the variable of
most concern.
Figure 33. Block sample of grade V Hong Kong

granite used in the apparatus shown in Figure 32
Relict joints and other such discontinuities are

worthy of special mention, because they frequently
play a major part in slope instability (Hencher et al,
1984). These are often slickensided and sometimes
coated with thin deposits of very weak material. The
significance of smooth slickensides in decomposed
igneous and metamorphic rocks was fully discussed
in a paper by St John et al (1969). There have been Figure 34. Diagrammatic representation of the effects
some reported failures of quite flat slopes attributable of rainwater on a slope
21
Finish.indb 139 2007/8/17 11:03:43 AM

Figure 35. Rapid changes in the groundwater table

measured in a Hong Kong slope
The pore pressure distribution in a residual

soil is dependent on the pattern of rainfall and the
hydrogeology of the slope. Pore pressure is therefore
a variable which is independent of soil mechanics
considerations, being imposed upon a slope by external
influences. For this reason, it is extremely difficult to

predict the appropriate pore pressures for slope design
or stability assessment. This fact is vividly illustrated
in Figure 35, which shows the rapid variation in Figure 36. Representation of the advance of the
piezometric head recorded in a Hong Kong slope wetting front as water infiltrates into soil (Lumb,
during a rainstorm in June 1982. 1962b, 1975)
For the infiltration of water into the horizontal
surface of a porous medium, Lumb (1962b) derived
an expression for the advance of the ‘wetting front’ For the analysis of slope stability in residual soils
(Figure 36). On the assumption that diffusion is and colluvium, measured pore pressure data is much to
negligible at the end of an intensive rainfall, Lumb be preferred to the application of uncertain predictive
(1975) later showed that the thickness, h, of the methods. For such data to be meaningful, however,
wetting front after time t could be approximated by: pore pressures must be monitored for a sufficiently
long period of time, and piezometers must be installed
kt at the appropriate depths at sufficient locations on the
h= (3) slope. In addition, an appropriate type of instrument
η(Sf - So)
must be used which can respond rapidly to pore
where k is the saturated permeability, η is the porosity, pressure changes. In Hong Kong, numerical modelling
and S o and S f are the initial and final degrees of techniques have been applied in a few cases (Leach
saturation. & Herbert, 1982), but simple direct correlation and
Despite its shortcomings, equation (3) is frequently extrapolation methods are more commonly resorted to
used for the prediction of design pore pressures (Koo & Lumb, 1981; Endicott, 1982).
in Hong Kong (Beattie & Chau, 1976; Beattie & Because of the rapid changes in pore pressure
Attewill, 1977; Geotechnical Control Office, 1984) that occur with rainfall in many residual soil slopes,
by the superimposition of the depth of wetting front the critical pore pressures are rarely obtained from
onto the groundwater table level at the end of the wet normal piezometer measurements. In order to enable
season. For this purpose, a rainfall return period of ten standpipe piezometers to record the maximum
years is commonly employed. This approach suffers water levels attained during a given period, Halcrow
from the fact that the value of h is directly proportional ‘buckets’ (British Patent No. 1538487)(Figure 37)
to Sf-So, and the initial and final degrees of saturation have become popular in Hong Kong. These simple
are somewhat speculative. In practice, the Hong Kong plastic devices are threaded onto a weighted nylon
wetting band thickness is usually calculated as being string at selected depth intervals above the normal
about two metres. Such predictions will generally base water level in a standpipe. When the string of
lead to overly conservative slope design and stability buckets is withdrawn, the highest transient water level
assessments, especially if it is assumed that no soil is indicated by the upper limit of water-filled buckets.
suction exists anywhere in the soil profile. This system is cheap and simple, and it can be used to
provide important design information when expensive
22
Finish.indb 140 2007/8/17 11:03:44 AM

automatic piezometer systems, of the type described
below, cannot be justified.
Figure 37. Halcrow plastic ‘buckets’ used for

recording the maximum water elevation attained in a Figure 38. Agricultural tensiometer for the
standpipe piezometer measurement of insitu soil suction

During most of the year, suctions exist in many stability purposes. However, some measurements
residual soil slopes. These suctions, which can be made at considerable depth in Hong Kong granite
of high magnitudes, are reduced dramatically by the (Sweeney, 1982; Sweeney & Robertson, 1982) suggest
process of water infiltration during rainfall. There are that not all soil suction is destroyed by infiltration.
almost certainly situations where even the heaviest The extent to which suctions play a significant part in
rainfall does not completely destroy the soil suctions, slope stability in residual soil profiles elsewhere in the
and these continue to contribute to the stability of a world clearly depends on the pattern of rainfall and the
slope. Whereas slope design is commonly based on infiltration characteristics of the material concerned.
predictions of the maximum positive pore pressures Much valuable data on soil suctions can be
that are likely to develop in the slope, no account is obtained for shallow depths by means of simple
usually taken of negative pore pressures that might agricultural tensiometers (Anderson, 1984; Greenway
be sustained under the worst rainfall conditions. This et al, 1984). These instruments, which consist of
neglect of the contribution of soil suction to the shear a small porous ceramic tip connected to a Perspex
strength of the soil might be the main reason why tube and a vacuum gauge (Figure 38), have been
stable residual soil slopes often have theoretical factors found to perform satisfactorily. They suffer from the
of safety of less than unity. serious disadvantage, however, that they must be read
Programmes of laboratory testing carried out over manually, and critical readings during intensive rainfall
many years on Hong Kong residual soils have proved are frequently unable to be obtained.
that suction pressures act as ‘modified’ effective Some sophisticated, automatic reading equipment
stresses, a matric suction of u a - u w increasing the has been developed to measure both positive and
shear strength by (ua - uw)tan φb, where φb is the angle negative pore pressures in Hong Kong conditions
of internal friction with respect to matric suction (Pope et al, 1982). This is based on similar equipment
(Fredlund, 1981; Ho & Fredlund, 1982; Fredlund & designed in the United Kingdom (Anderson & Burt,
Rahardjo, 1985). It has been suggested (Sweeney & 1977; Burt, 1978). Small diameter tensiometer tips
Robertson, 1982; Boonsinsuk & Yong, 1982) that of high air-entry ceramic are each connected by twin
this increase in soil strength should be permissible water-filled flexible nylon tubes to a single transducer
for the design of Hong Kong slopes as long as insitu through a fluid scanning switch (Figure 39). The
measurements are available to substantiate that switch is analogous to an electrical wafer switch, in
suctions continue to exist in the slopes throughout the that it allows sequential measurement of a number of
year. fluid pressure inputs. A rotating metal valve inside the
Much valuable data on soil suctions has been water-filled wafer (Scanivalve) enables the pressure
obtained on some Hong Kong slopes over several transducer to be connected to 24 different inputs in
years. It has been found that the suctions on relatively turn, two of which are used as reference reservoirs for
shallow potential slip surfaces decrease everywhere calibration purposes. The measured pore pressures are
close to zero during a wet season, and that suction recorded on a variety of types of strip-chart recorder,
cannot therefore normally be relied upon for slope the pressure-sensitive paper type now being regarded
23
Finish.indb 141 2007/8/17 11:03:45 AM

as the most reliable. The system is powered by a Control Office, 1984), which specifies factors of
twelve volt battery, and all measuring and recording safety. Table 6 shows the recommended factors against
components are housed in a water-tight metal box. loss of life and against economic risk for new soil
Although installation and maintenance costs are high, and rock cut slopes. These values are for groundwater
the systems have been found to be reliable for periods conditions resulting from a ten-year return period
of up to three years. rainfall. There are three risk categories in each case -
‘negligible’, ‘low’ and ‘high’. The risk-to-life category
reflects the likelihood of loss of life in the event of
failure, while the economic risk category reflects the
likely magnitude of economic loss. Typical of high

risk-to-life slopes are high cut slopes adjacent to
occupied buildings. An example of a negligible risk-
to-life slope is one which threatens only a lightly
trafficked secondary road. Other examples of the two
risk categories for Hong Kong conditions are given in
Tables 7 & 8.
Table 6. Factors of safety specified for new cut slopes

in Hong Kong (Geotechnical Control Office, 1984)
Figure 39. Automatic Scanivalve System used in Hong

Kong for the measurement of positive and negative
pore pressures (Pope et al, 1982)
Note: these factors of safety are for a rainfall return
In the context of pore pressure prediction, mention period of ten years
must be made of soil ‘pipes’ (erosion tunnels) which
frequently occur in residual and colluvial terrain. These Hong Kong’s Geotechnical Manual for Slopes
can vary in size from a few millimetres to perhaps also gives guidance on preventive and remedial
a metre in diameter, and they are capable of very works to existing slopes. When analysing an existing
rapidly conveying large volumes of water downslope slope to determine the extent of any preventive (i.e.
(Figure 35). There is no doubt that they are frequently before failure) or remedial (i.e. after failure) works
an important factor in slope failures (Pierson, 1983). required, the performance history of that slope can be
In Hong Kong, pipes can be of major significance to of considerable assistance to the designer. There is, for
the hydrogeology of an area (Leach & Herbert, 1982; example, an opportunity to examine the geology of the
Nash & Dale, 1983; Premchitt et al, 1985). slope more closely than for an undeveloped site, and to
obtain more realistic information on groundwater. The
8.6 Factors of Safety Manual maintains that the designer is therefore able to
adopt with confidence factors of safety for proposed
Residual soil slopes commonly occur in steep natural preventive or remedial works that are slightly lower
terrain, and it is often necessary to work to very low than those specified in Table 6 for new works, as long
factors of safety in order to avoid the excessive costs as rigorous geological and geotechnical investigations
of adopting higher values. In these circumstances, are conducted (which include a thorough examination
it is logical to relate design safety factors to the of slope maintenance history, groundwater records,
consequences of failure in terms of risk of casualties rainfall records and any slope monitoring records).
and risk of damage to property and services. These reduced factors of safety are shown in Table 9.
Slope design practice in Hong Kong is governed
by the Geotechnical Manual for Slopes (Geotechnical
24
Finish.indb 142 2007/8/17 11:03:45 AM

Table 7. Examples of risk-to-life categories for slope of the parameters remains extraordinarily large even
failures in Hong Kong (Geotechnical Control Office, after completion of exploration and testing that costs
1984) as much as the project can possibly afford.”
Risk to Life Experience with the application of soil mechanics
Example methods of analysis to slope failures in Hong Kong
Neg. Low High
enables a more detailed state-of-the-art assessment to
(1) Failures affecting country be made than that contained in Table 5. This has been
parks and lightly used • done for Hong Kong’s residual soil slopes by Hencher
open-air recreation areas.
et al (1984) (Table 10), who also suggested where

(2) Failures affecting roads advances could be made in the two particularly poorly
•
with low traffic density. understood areas of shear strength and pore pressure
(3) Failures affecting prediction (Table 11).
storage compounds (non- •
dangerous goods). Table 8. Examples of economic risk categories for
(4) Failures affecting densely slope failures in Hong Kong (Geotechnical Control
used open spaces and Office, 1984)
recreation facilities •
(e.g. sitting-out areas, Economic Risk
Example
playgrounds, car parks). Neg. Low High
(5) Failures affecting roads (1) Failures affecting country
•
with high vehicular or • parks.
pedestrian traffic density. (2) Failures affecting rural (B),

(6) Failures affecting public feeder district distributor
waiting areas (e.g. railway and local distributor •
•
platforms, bus stops, petrol roads which are not sole
stations). accesses.
(7) Failures affecting occupied (3) Failures affecting open-air
buildings (e.g. residential, •
• car parks.
educational, commercial, (4) Failures affecting rural
industrial). (A) or primary distributor
(8) Failures affecting buildings •
• roads which are not sole
storing dangerous goods. accesses.
(5) Failure affecting essential
8.7 Conclusions service which could cause
loss of that service for a
It is clear from the preceding brief review that the •
temporary period (e.g.
application of rigorous soil mechanics methods of
power, water and gas
slope stability analysis to residual profiles is extremely
mains).
difficult. It is prudent to adopt this approach only in
conjunction with a thorough engineering geological (6) Failures affecting rural or
assessment and with the liberal application of sound urban trunk roads or roads •
judgement. Analytical solutions alone cannot be relied of strategic importance.
upon, as succinctly stated by Peck (1975), thus : (7) Failures affecting essential
“Analytical procedures have been developed services, which could
for calculating the factors of safety of a slope cause loss of that service •
under various conditions ... The theories and their for an extended period.
applications have met with some successes and more (8) Failures affecting
failures. Most of the failures are a consequence of buildings, which could
oversimplification. •
cause excessive structural
Even the most complex theories are necessarily damage.
oversimplifications of nature; they are general rather
than specific. A theory that would take into account all
the significant variables at a given location would be Table 9. Factors of safety specified for existing cut
far too complex for use. Furthermore the parameters slopes in Hong Kong (Geotechnical Control Office,
that must be evaluated to use the theory are not simple 1984)
invariant quantities, few in number, but are instead Risk to Life High Low Negligible
complex, highly variable, and usually not constant
with respect to time. Frequently the range of variability Factor of Safety 1.2 1.1 >1.0
25
Finish.indb 143 2007/8/17 11:03:45 AM

Table 10. Assessment of the state-of-knowledge of the various aspects of slope stability predictions for Hong
Kong conditions (Updated from Hencher et al, 1984)
Overall
Aspect Current State-of-Knowledge for Hong Kong Conditions Rating of
Knowledge
Methods of Janbu (1954, 1973) method of analysis for non-linear surfaces thought satisfactory.
Stability Recommended factors of safety of 1.2 to 1.4 are satisfactory (GCO, 1984). Very good
Analysis Computational data is often poorly handled (Lumsdaine & Tang, 1981, 1982).
Pre-failure geometry is easily defined.

Geometry Often difficult to decide critical potential failure surface for design, especially Good to
of Failure where geology is complex (Hencher et al, 1984; Hencher & Martin, 1984; Very good
Hudson & Hencher, 1984).
Site investigation procedures are adequate, but descriptions often poor.
Geology Complex weathering profiles are difficult to describe (Hencher & Martin, 1982). Fair
Understanding of influence of geological details on hydrogeology is poor.
Mass strength as distinct from sample strength is poorly understood.
Laboratory tests are commonly used to determine saturated strengths of samples
in terms of effective stress, but doubt exists about applicability of test results
Shear Fair to
(Brand, 1981, 1982).
Strength Poor
Limited amount of insitu strength testing carried out (Brand et al, 1983a).
Weakening effect of relict joints recognised (Koo, 1982a, 1982b).

Effects of boulder and corestone content unknown (Hencher & Martin, 1982).
Useful correlations available between landslides and rainfall (Lumb, 1975; Brand
et al, 1984).
Rapid changes in pore pressure with rainfall are very difficult to predict for design
(Anderson et al, 1983; Premchitt et al, 1985).
Groundwater Only limited attempts made to model groundwater (Leach & Herbert, 1982).
and Pore Extrapolation of insitu measurements seems best design approach (Koo & Lumb, Poor
Pressures 1981; Endicott, 1982).
Some progress made with field instrumentation (Pope et al. 1982; Brand et al,
1983a).
Erosion pipes are important in transmitting water (Nash & Dale, 1983; Brand et al,
1986).
Table 11. Suggested advances which can be made in the state-of-knowledge of slope stability predictions for
Hong Kong conditions (Updated from Hencher et al, 1984)
Methods Field Testing Theoretical Back
Key Areas Observations Field Laboratory Studies Analysis
Shear Advances can be Limited advances Considerable Advances can be This is the only
Strength made by observa- can be made in advances can made in establish- means of checking
tion of factors relating sample be made in our ing models for the validity of
which influence strengths to mass understanding of mass behaviour of relationships
the mass strengths strengths through the shear strength materials between
of materials field testing behaviour of un- properties
saturated residual
soils
Groundwater These are useful Case studies Not applicable Advances are This is the only
and Pore for checking relating subsurface possible, e.g. means for
Pressure theories and profiles to development of checking
methods of hydrogeology more sophisticated that the methods
prediction and infiltration infiltration and of incorporating
characteristics can hydrogeological groundwater into
lead to improved models design are
predictions satisfactory
26
Finish.indb 144 2007/8/17 11:03:46 AM

9. SOME HONG KONG CASES and piezometers were installed to monitor pore
pressures throughout one or two wet seasons. Triaxial
9.1 Introduction tests were carried out in the laboratory to obtain good
In establishing his framework by which to judge soil quality shear strength data.
mechanics predictions, Lambe (1973) recognised The three slopes were all analysed using the Janbu
three main categories of prediction, namely: type A routine method (Janbu, 1954, 1973) applied to a
predictions (made before the event), type B predictions large number of slip surfaces chosen on the basis of
(made during the event), and type C predictions the site geology as revealed by the field studies. The
(made after the event). For obvious reasons, type A minimum calculated factor of safety was in each case
predictions are thought to be the only true predictions, significantly lower than 1.0. In May 1982, August
and these are certainly the most valuable kind for 1982 and June 1983, severe rainstorms occurred
assessing the reliability of our predictive methods. (Table 3) which between them caused well over 2 000
Type C predictions are regrettably the most common slope failures, but the three slopes in question did not
type reported in the published literature, and alone they fail, and they remain stable to this day.
provide unconvincing confirmation of the accuracy Case 1 slope is shown in the simplified cross-
of our predictive tools. However, these ‘post-mortem’ section in Figure 40. It has a height of about 36 m
predictions can be extremely valuable where they are from the flat open space below. Deep excavations
accompanied by full details of the engineering design into the original ground surface left a relatively
carried out before construction. thin layer of granitic soil (grades IV, V & VI) over
Mention has been made in Section 8.7 of the the granite bedrock. The slope surface is covered
detailed back analyses carried out on some slope partly by vegetation and partly by lime-stabilised
failures in Hong Kong (Hencher et al, 1984). These soil (‘chunam’). The water table, as determined
were part of a continuing programme to evaluate the by piezometer installations, is low even during the
engineering design and analysis methods currently wet season, no positive pore pressures having been
in use and specified in the Geotechnical Manual for measured in the soil.
Slopes (Geotechnical Control Office, 1984). This
continuing evaluation process is vitally important in
Hong Kong, where the design factors of safety are
necessarily very low (Table 6).
In order to illustrate the state-of-the-art with respect
to predicting the performance of residual soil slopes,
details will be given here of the performance of six
selected Hong Kong cut slopes. These fall into the two
categories:
(a) slopes which have remained stable even though the
present theoretical factors of safety are less than 1.0
(Cases 1, 2 & 3), and
(b) slopes which have failed, even though the ‘design’
factors of safety were above 1.0 (Cases 4, 5 & 6).
The slopes represented by Cases 1 to 3 were
formed some years ago on the basis of precedent, no Figure 40. Case 1 : Stable cut slope with F <1.0
engineering design methods having been used. Cases 4
to 6, on the other hand, involved the latest analysis and
design procedures. All six cases depict common types
of cut slope in Hong Kong, and they together represent
a small proportion of the examples that could be used
to illustrate the same points.
9.2 Stable Slopes with F <1. 0

The three stable slopes (Cases 1 to 3) were studied as
part of a Territory-wide appraisal of existing cut slopes
carried out between 1979 and 1981. Each slope was
the subject of detailed assessment, which included a
review of the slope’s history, a detailed site survey,
and an appropriate site investigation and testing Figure 41. Detailed weathering profile for the Case 1
programme. Holes were drilled to supplement the slope showing the grades IV, V & VI Soil
surface geological examinations and to obtain samples,
27
Finish.indb 145 2007/8/17 11:03:46 AM

The measured shear strength parameters for the and partly by vegetation. The granitic soil varies from
material of the Case 1 slope vary between c' =14 kPa grade IV to grade VI, the vast majority being of grade
and ø' = 45° for the grade IV material, and c' = 0 and V with strength parameters of c' = 2 kPa and ø' = 39°.
ø' = 42° for the grade VI material. On the basis of these At its highest, the water table is located many metres
parameters, the factor of safety for the steep upper below the top of the slope, but it is close to the surface
part of the slope, above the berm, is calculated to be of the bottom part of the slope. The results of the
0.95, with the critical slip surface being very shallow stability analyses are depicted in Figure 43. Several
and passing largely through the grade V material. The slip surfaces give factors of safety which are less than
detailed cross-section of the ‘unstable’ part of the slope 1.0, the critical surface, which is very shallow, having
(Figure 41) clearly illustrates the importance of the a factor of safety of only 0.80.
weathering profile on stability predictions for slopes of
this kind. 9.3 Unstable Slopes with F >1.0
Case 2 involves a cut slope in granitic Soil which
Examples will be given of three cut slopes which
was formed as part of a combined site formation for
were designed to be adequately stable and which
a housing block and a road (Figure 42). The steep
subsequently failed during the severe rainstorms of
cut slope, which is about 18 m high and coated with
1982 and 1983 (Table 3). These will be referred to as
chunam, is retained at its toe by a 3 m concrete
Cases 4, 5 & 6.
retaining wall. The very thick deposit of granitic
Case 4 cut slope, shown in Figure 44, was
soil is predominantly of grade V, with shear strength
designed and constructed in 1980-81 as part of a new
parameters of c' = 5 kPa and ø' = 35°. The water table
road scheme. What was thought to be adequate site
remains low throughout the wet season. The calculated
investigation was carried out at the time in the form of
factor of safety on the deep critical slip surface shown

drilling, sampling and strength testing. This indicated
in Figure 42 is 0.82
that the granite rock was overlain by a relatively thin
layer of grades IV, V & VI materials. Piezometers
installed in drillholes showed the water table to be
quite close to the ground surface in the upper part
of the slope, and two rows of horizontal drains were
included at mid-slope as part of the design. On the
assumption that the horizontal drains would control
the maximum level of the water table, the critical slip
surface was found to be as shown in the failure. The
measured shear strengths of c' =10 kPa and ø' = 35°
yielded a design factor of safety of 1.24, which was
considered adequate for the risk category concerned.
Figure 42. Case 2: Stable cut slope with F <1.0

Case 3 involves a 30 m high slope (Figure 43) cut
in granitic soil with a thin covering of colluvium over
its lower half. At the toe of the slope is a 5 m high
masonry retaining wall, and above the slope is a road.
The surface of the slope is protected partly by chunam
Figure 44. Case 4: Unstable cut slope previously

predicted as being stable (F = 1.24 ?)
After the heavy rainstorm at the end of May 1982,

failure of a large portion of the Case 4 slope occurred,
as can be seen in the aerial photograph in Figure
45. The failed mass measured approximately 70 m
along the slope and 40 m from top to bottom, the
deduced failure surface being as shown in Figure 44.
Figure 43. Case 3: Stable cut slope with F <1.0
It should be noted that, unlike the vast majority of
28
Finish.indb 146 2007/8/17 11:03:47 AM

Hong Kong landslides, failure did not occur suddenly
and dramatically, but took place over several days,
and the large mass of failed material remained on
the slope. The failure was extensively investigated,
many boreholes being drilled to obtain a much more
detailed geological profile than that on which the
design was based. Reanalysis of the slope on the basis
of the actual failure surface, which was much deeper
than the design critical surface (Figure 44), gave a
range of F-values from 0.82 to 1.0 depending on the

water table level, with the most probable value being
F = 0.89. The crucial design error, therefore, was a
misinterpretation of the weathering profile because of
insufficient borehole data. Figure 46. Case 5: Unstable cut slope previously
predicted as being stable (F = 1.3?)
further study was then carried out to investigate the

cause. The failure surface was accurately determined
to be as shown in Figure 46, for which the calculated
factor of safety was 1.3, unless considerably higher
pore pressures than those measured were assumed
in the calculations. Of major importance was the

presence of almost vertical slickensided relict joints
visible at the failure scarp, which could have resulted
in the joint-controlled failure mechanism depicted
in Figure 47. In addition, soil pipes (see Section 8.5)
discovered in the vicinity could have transmitted large
quantities of water to the slip surface to bring about
Figure 45. Aerial view of Case 4 slope after failure localised transient pore pressures higher than those
measured by the piezometers. It is therefore difficult
Case 5 concerns a very large cut slope formed in to envisage how this slope failure could possibly have
the late 1960s for the construction of a road. At its been predicted on the basis of available soil mechanics
highest point, the crest of the slope is 90 m above the methods.
road level. The slope was originally cut to a ‘standard’ Case 6 is a 70 m high cut slope in volcanic soil
angle without the benefit of investigation or design. A constructed in 1974-77 as part of the Tuen Mun
number of shallow failures occurred at various points Highway, which features in Section 10 below. No
on this long slope during 1973, 1974 and 1976, as a proper investigation or stability analysis was carried
result of which a limited amount of investigation was out before construction. Failure of the slope first
conducted to enable remedial works to be designed occurred in 1975, and this led to remedial works,
and carried out. These took the form of cutting back which included cutting back the slope and the
together with the installation of some horizontal drains installation of some horizontal drains, to leave the
in the failed areas. slope as shown in the cross-section in Figure 48. On
Figure 46 shows a section of the upper portion the basis of site investigation data collected for the
of the Case 5 slope as it was in 1981. Several metres remedial works, a full stability analysis was later
of colluvium overlay about 12 m of grades V & VI carried out in 1978, to give F = 1.3 on the ‘design’
granitic soil. A comprehensive site investigation on critical slip surface shown in the figure. In September
the slope was carried out in 1981-82 to reassess its 1983, an extremely large failure, about 10 000 sq. m in
stability. This included the installation of piezometers area (Figure 49), occurred on the slope.
and a thorough programme of laboratory shear strength A very detailed geotechnical study was carried
testing. The strength parameters were measured as c' out on the Case 6 failure. The many boreholes drilled
= 4 kPa and ø' = 36° for the colluvium, c' = 6 kPa and (Figure 49) indicated that the bedrock (grades III or
ø' = 40° for the grade VI material, and c' = 20 kPa and stronger) was much deeper over most of the slope than
ø' = 40° for the grade V material. On the basis of this assumed in the 1978 analysis. In spite of this, borehole
data and the maximum groundwater elevation recorded data and inclinometer readings determined the failure
during 1981-82, stability analyses gave a ‘design’ surface as being fairly close to that predicted in 1978
factor of safety in excess of 1.3. (Figure 48), although the back scarp was appreciably
A portion of the slope, 30 m high and 15 m long, steeper. This fact appeared to be related to the
failed during the rainstorm of August 1982, and a subvertical relict joints in the decomposed volcanic
29
Finish.indb 147 2007/8/17 11:03:47 AM

soil which were crucial in dictating the mode of
failure. The computed F-value for the actual failure
surface, however, was still about 1.3 on the basis of
strength data obtained from good quality laboratory
triaxial tests. Like Case 5, it is thought that gradual
slope ‘deterioration’ caused by the ingress of water
along the relict joints could have been the main cause
of failure. Piezometers installed for the back analyses
gave measured groundwater levels very close to those
assumed in the original 1978 stability calculations.
9.4 Discussion
The six Hong Kong examples described above
illustrate our apparent inability to accurately predict
the failure of residual soil slopes. On the one hand, the (a) Rotation and displacement of joints
classical soil mechanics methods of stability analysis
would seem to be conservative, as evidenced by the
Cases 1 to 3 slopes which have remained stable with
factors of safety of less than 1.0. On the other hand,
Cases 4 to 6 suggest that the same methods over-
estimate the F-values for some slopes which are in fact
unstable.
There are two possible contributory factors to
the incorrect predictions that the Cases 1 to 3 slopes
should have failed, namely :
(a) the insitu mass shear strengths are higher than
those measured in the laboratory, for the reasons
discussed earlier in Section 8.4, and
(b) soil suctions are sustained in the slopes throughout
the heaviest rainstorms which are sufficient to
render them stable. (b) Ingress of water through open joints
The usually dominant part played in slope stability Figure 47. Possible failure mechanism of Case 5 and
calculations by positive pore pressures cannot in Case 6 slopes
any way be the cause of error in these three cases,
since the pore pressures were zero everywhere on the
determined critical slip surfaces.
For soil suction alone to be the reason why the
three slopes have remained stable (i.e. F >1.0),
suction values greater than about 2 kPa, 16 kPa and
8 kPa would need to be sustained in the three slopes
respectively throughout severe rainstorms. In order
for the slopes to possess factors of safety of 1.4, as
specified in Hong Kong for high risk situations, the
soil suctions required are about 8 kPa, 45 kPa and
28 kPa respectively. As mentioned earlier, it is unlikely
that suctions of these large magnitudes could exist
at such shallow depths in these slopes during severe
rainstorms. However, it is probable that the soil never
becomes fully saturated insitu, in contrast to the full Figure 48. Case 6: Unstable cut slope previously
saturation state usually imposed for laboratory shear predicted as being stable (F = 1.3?)
measurements.
Where slopes fail with ‘design’ factors of safety successfully with hindsight. An example of such
which exceed 1.0 (Cases 4 to 6), the role of geological a situation is shown in Figure 50, for which back
detail appears to be paramount, and this cannot be analyses have been carried out by Hencher & Martin
readily accounted for in our routine methods of slope (1984).
analysis and design. It is not uncommon in weathered It is noteworthy that none of the three slopes
profiles for the geological complexity to be such that represented by Cases 4, 5 & 6 failed dramatically, but
our soil mechanics approach can only be applied all involved slow and relatively small displacements
30
Finish.indb 148 2007/8/17 11:03:48 AM

situations of 1.4, as recommended by the Geotechnical
Manual for Slopes (Geotechnical Control Office, 1984)
(Table 6).
10. PROBABILITY OF SLOPE FAILURE

10.1 Prediction Quality
The calculated factor of safety for a slope is a way of
representing the probability of failure. Engineering

design aims to produce a low probability of failure
combined with a high degree of economy. A perfect
prediction method would result in the failure of all
slopes with F <1.0 and no failures of slopes with F
>1.0. In other words, the probability of failure for
Figure 49. Plan showing the extent of the Case 6 F <1.0 should ideally be 1.0, and the probability of
failure for F >1.0 should be 0. This ideal situation is
failure
depicted in Figure 51(a), which shows the probability
of slope material which in each case remained on the of failure as a function of theoretical factor of safety.
slope surface. This suggests that a calculated value of In practice, the ideal prediction is not achievable, but
1.3 might be the lower bound F-value to ensure long- predictions based on the best slope design methods
could be expected to yield results represented by
term stability in all Hong Kong’s large soil cut slopes.

Of importance is the fact that there are no recorded the probability curves A & B. While prediction A is
failures of Hong Kong slopes for which the F-values clearly superior to prediction B, both are ‘balanced’
predicted on the basis of state-of-the-art soil mechanics predictions, in the sense that each gives a probability
methods have exceeded 1.4. This gives good support of failure of 0.5 for F =1.0.
to the use of a minimum factor of safety for high risk Where our predictive tools are not adequate
Figure 50. Diagram showing the complex geological conditions at the failure site studied by Hencher & Martin
(1984)
31
Finish.indb 149 2007/8/17 11:03:49 AM

to enable a balanced prediction to be made, the construction.
probability of failure for a designed slope could be as The majority of the Tuen Mun Highway slopes
shown in Figure 51(b). The balanced prediction curve are cut in weathered granite, but a few at the western
moves to the left for a conservative prediction, and end are in weathered volcanic rocks. Most of the
to the right for an unsafe prediction. A conservative slopes have exposed rock (grades I to III) at the toe
prediction is one for which the probability of failure with residual soil (grades IV to VI) above. Of the rest,
for a slope with F =1.0 is less than 0.5; an unsafe a few are cut entirely in rock, while the others are
prediction is one for which the probability of failure composed of residual soil throughout. While the slopes
for F =1.0 is greater than 0.5. As our prediction can therefore be expected to vary considerably in their
capabilities improve, the prediction becomes steeper stability, they have all been in existence for the same
and approaches the ideal prediction situation depicted length of time and they have all experienced the same
in Figure 51(a). rainfall history.
The Territory-wide slope survey carried out in
10.2 A Hong Kong Example 1977-78 recorded 78 cut slopes higher than 5 m along
In order to evaluate, in a general way, our ability to the Tuen Mun Highway. Fourteen of these were selected
predict slope failures in residual soils, the probability for detailed geotechnical studies, which included full
of slope failure will be examined for the Tuen Mun stability analyses. The distribution of the calculated
Highway in Hong Kong. This major highway, which factors of safety for these is shown in Figure 52. Half
was constructed over a number of years commencing of the slopes had factors of safety greater than 1.4,
in 1974, runs for about 15 km along the southwest while the others had F-values which varied from 1.1 to
coast of the mainland portion of the Territory. A large 1.4. No stability analyses were carried out on the other
number of cut slopes were required to be formed in 64 slopes along the Highway, but the crude assumption
steep natural hillsides, and this was done on the basis will be made here that the distribution of F-values for
of the standard ‘design’ geometry of 5 vertical to these slopes would be similar to the distribution for the
3 horizontal (about 60°), with 1.5 m wide berms at fourteen shown in Figure 52.
7.5 m vertical intervals. Provision was made for any The Tuen Mun Highway slopes have endured six
slope to be flattened if it proved to be unstable during major rainstorm events since they were constructed
(Table 3). The rainstorms of 1982 and 1983 were
particularly severe tests of their stability. Four
slopes failed in the storm of May 1982 and another
(Case 6, Section 9.3) in September 1983, which all
caused disruption to traffic for many days. During
the highway’s first seven years of life, therefore, five
major failures occurred in the 78 slopes, which gives
a rough annual probability of failure of about 1 in 100
for each slope.
(a) Ideal and actual balanced predictions
Figure 52. Distribution of factors of safety for the 14

Tuen Mun Highway slopes studied in detail
The rough annual probability of slope failure on

the Tuen Mun Highway can be compared with the
(b) Conservative and unsafe predictions probabilities quoted by Whitman (1984) for other types
Figure 51. Reliability of slope stability predictions in of geotechnical failure. Figure 53 shows Whitman’s
terms of probability of failure and theoretical factor of plot of probability of failure against consequence of
safety failure for the situation in the United States. To this has
32
Finish.indb 150 2007/8/17 11:03:50 AM

been added a zone to represent the group of Tuen Mun
Highway slopes. Whitman’s acceptability criteria, as
represented by straight lines in the figure, suggest that
the Tuen Mun Highway situation would represent an
acceptable risk in the USA and would be little different
from the risk accepted for foundation failure.
The Geotechnical Control Office has established
a comprehensive catalogue of approximately 10 000
large cut slopes in Hong Kong, the majority of which
were constructed before engineering design methods

were adopted. If the probability of failure of 1 in 100
for each Tuen Mun Highway slope is representative of
the slope stability situation overall in Hong Kong, then
approximately 100 slope failures can be expected in
an average year. This figure is in fact not far in excess
of the average number of significant failures recorded Figure 53. Annual probabilities and consequences of
annually over the last 20-year period (Table 3). The geotechnical failures in the USA (Whitman, 1984)
situation is improving dramatically, however, because compared with those on the Tuen Mun Highway
of the high standard of geotechnical control which
of failure, with the meaningful measurement of shear
is now applied to the design and construction of all
strengths and, particularly, with the prediction of the
new cut slopes, which are required to comply with the
most unfavourable pore pressures which are likely to
Geotechnical Manual for Slopes (Geotechnical Control
occur during the life of the slope.

Office, 1984). In time, it is envisaged that the overall
There is considerable scope for improvements to
probability of failure will tend towards a reduction by
the application of analytical methods for the prediction
an order of magnitude (from 10-2 to 10-3), in much the
of slope performance. Long-term programmes of
same way as Peck (1980) hoped for a reduction from
research and development are required, into the shear
10-4 to 10-5 in the annual probability of failure for the
strength behaviour of residual soils and into improved
world’s dams.
methods for predicting the pore pressures induced
by rainfall. Most importantly, good case histories
are badly needed which give detailed comparisons
11. CONCLUSIONS of stability analyses and slope performance. There is
It has been shown that there are several basic no doubt that the available analytical methods have
approaches that can be used for the design and stability far out-striped our ability to observe, measure and
assessment of residual soil slopes. For most residual interpret data. As Peck (1975, 1980) has commented,
soil problems, the traditional division between the analytical methods of design are usually too heavily
methods developed separately for soil mechanics, relied upon by the engineer at the expense of the
rock mechanics and engineering geology is not application of engineering judgement based on sound
meaningful, and a unified approach is necessary for experience.
good engineering design. Because of the nature of
residual materials, the application of soil mechanics
methods alone will generally not be adequate. ACKNOWLEDGEMENTS
The selected use of terrain evaluation has much to
Acknowledgement is due to many of the Author’s
recommend it, especially for planning and feasibility
colleagues in the Geotechnical Control Office who
studies of relatively large land areas. In addition, the
have collectively contributed significantly to the
establishment of semi-empirical design guidelines
progress made in the past few years in geotechnical
will often be invaluable where there is adequate
engineering practice in residual soils. He is particularly
information available on the performance of existing
grateful to Dr Jerasak Premchitt, who gave generous
cut slopes over a sufficiently long period of time.
assistance with the preparation of this Paper, and to Dr
The state-of-the-art with respect to the application
S.R. Hencher, who supervised many of the slope
of soil mechanics methods to the analysis of residual
studies described. As usual, Miss Margaret Wong
soil slopes is generally poor. Whereas some simple
produced a high quality manuscript in a very short
situations readily lend themselves to solution by
time.
analytical means, the often complex geological
This Paper is published with the permission of the
conditions which exist in residual soil slopes usually
Director of Engineering Development of the Hong
preclude the sensible application of the well-known
Kong Government.
methods of stability analysis. Major difficulties
invariably arise with the selection of a suitable mode
33
Finish.indb 151 2007/8/17 11:03:50 AM

REFERENCES Brabb, E.E. (1984). Innovative approaches to landslide
hazard and risk mapping. Proc. 4th Int. Symp.
ABMS (1985a). Proceedings of the First International Landslides, Toronto, (1), 307-323.
Conference on Geomechanics in Tropical Lateritic Brand, E.W. (1981). Some thoughts on rain-induced
and Saprolitic Soils, Brasilia, 4 vols, ABMS, slope failures. Proc. 10th Int. Conf. SMFE,
Brazilian Soc. Soil Mech., Sao Paulo, 1758 p. Stockholm, (3), 373-376 .
ABMS (1985b). Peculiarities of Geotechnical Brand, E.W. (1982). Analysis and design in residual
Behaviour of Tropical Lateritic and Saprolitic Soils. soils. Proc. ASCE Spec. Conf. Engineering and
(ISSMFE Technical Committee on Tropical Soils, Construction in Tropical and Residual Soils,
Progress Report 1982-1985). ABMS, Brazilian Honolulu, 89-143.

Soc. Soil Mech., Sao Paulo, 470 p. Brand, E.W. (1984). Landslides in Southeast Asia:
Allen, P.M. & Stephens, E.A. (1971). Report on the a state-of-the-art report. Proc. 4th Int. Symp.
Geological Survey of Hong Kong. Hong Kong Landslides, Toronto, (1), 17-59. (Addendum, (3)
Govt Press, 116 p. & 2 maps. 105-106).
Anderson, M.G. (1984). Prediction of Soil Suction for Brand, E.W. (1985a). Geotechnical engineering
Slopes in Hong Kong. Geotech. Control Office, in tropical residual soils. Proc. 1st Int. Conf.
Hong Kong, 242 p. Geomechanics in Tropical Lateritic and Saprolitic
Anderson, M.G. & Burt, T.P. (1977). Automatic Soils, Brasilia, (3), 23-99.
monitoring of soil mositure conditions in a hillslope Brand, E.W. (1985b). Landslides in Hong Kong. Proc.
spur and hollow. Jour. Hydrol., (33), 27-36. 8th S.E. Asian Geotech. Conf., Kuala Lumpur, (2),
Anderson, M.G., McNicholl, D.P. & Shen, J.M. (1983). in press.
On the effect of topography in controlling soil Brand, E.W., Borrie, G.W. & Shen, J.M. (1983a). Field
water conditions, with specific regard to cut slope measurements in Hong Kong residual soils. Proc.
piezometric levels. Hong Kong Engr, (11), 11, Int. Symp. Field Measurements in Geomechanics,
35-41. Zurich, (1), 639-648.
ASCE (1971). Session 11: Slope Stability in Residual Brand, E.W., Burnett, A.D. & Styles, K.A. (1982a).
Soils. Proc. 4th Panam. Conf. SMFE, San Juan, The Geotechnical Area Studies Programme in
Puerto Rico, (2), 105-167, and (3), 125-146. Hong Kong. Proc. 7th S.E. Asian Geotech. Conf.,
ASCE (1982). Proceedings of the ASCE Specialty Hong Kong, (1), 107-123.
Conference on Engineering and Construction in Brand, E.W., Dale, M.J. & Nash, J.M. (1986). Soil
Tropical and Residual Soils, Honolulu, 746 p. pipes and slope stability in Hong Kong. Qtrly Jour.
Azzouz, A.S., Baligh, M.M. & Ladd, C.C. (1981). Eng. Geol., in press.
Three dimensional stability analyses of four Brand, E.W. & Hudson, R.R. (1982). CHASE - An
embankment failures. Proc. 10th Int. Conf. SMFE, empirical approach to the design of cut slopes in
Stockholm, (3), 343-346. Hong Kong soils. Proc. 7th S.E. Asian Geotech.
Baligh, M.M. & Azzouz, A.S. (1975). End effects on Conf., Hong Kong, (1) 1-16. (Discussion , (2),
stability of cohesive slopes. Jour. Geotech. Eng. 61-72 and 77-79).
Div., ASCE, (101), 1105-1117. Brand, E.W. & Phillipson, H.B. (1984). Site
Barata, F.E. (1969). Landslides in the tropical region of investigation and geotechnical engineering practice
Rio de Janeiro. Proc. 7th Int. Conf. SMFE, Mexico in Hong Kong. Geotech. Eng., (15), 97-153.
City, (2), 507-516 . Brand, E.W. & Phillipson, H.B. (Eds.)(1985).
Beattie, A.A. & Attewill, L.J.S. (1977). A landslide Sampling and Testing of Residual Soils - A Review
study in the Hong Kong residual soils. Proc. 5th S.E. of International Practice. Scorpion Press, Hong
Asian Conf. Soil Eng., Bangkok, 177-188. Kong, 194 p.
Beattie, A.A. & Chau, E.P.Y. (1976). The assessment Brand, E.W., Phillipson, H.B., Borrie, G.W. & Clover,
of landslide potential with recommendations for A.W. (1983b). In-situ direct shear tests on Hong
future research. Hong Kong Engr, (4), 1, 27-44. Kong residual soils. Proc. Int. Symp. Soil and Rock
(Discussion, (4), 2, 55-62). Investigations by In-situ Testing, Paris, (2), 13-17.
Blight, G.E. (1977). Slopes and excavations in residual (Discussion, (3), 55-56).
soils. Proc. 9th Int. Conf. SMFE,Tokyo, (3), Brand, E.W., Premchitt, J. & Phillipson, H.B. (1984).
582-590. Relationship between rainfall and landslides in
Blong, R.J. (1973). Morphometric attributes of Hong Kong. Proc. 4th Int. Symp. Landslides,
landslides. Zeitschrift für Geomorphologie, Suppl. Toronto, (1), 377-384.
Vol. 18, 66-77. Brand, E.W., Styles, K.A. & Burnett, A.D. (1982).
Boonsinsuk, P. & Yong, R.N. (1982). Some aspects of Geotechnical land-use maps for planning in Hong
slope stability in Hong Kong residual soils. Proc. Kong. Proc. 4th Int. Cong. IAEG, New Delhi, (1),
ASCE Spec. Conf. Engineering and Construction 145-153.
in Tropicial and Residual Soils, Honolulu, 463-483. British Standards Institution (1981). Code of Practice
34
Finish.indb 152 2007/8/17 11:03:50 AM

for Site Investigation BS 5930: 1981. Brit. Stan. De Graft-Johnson, J.W.S. & Bhatia, H.S. (1969).
inst., London, 150 p. Engineering characteristics (of lateritic soils). Proc.
Burnett, A.D., Brand, E.W. & Styles, K.A. (1985). Spec. Sess. Engineering Prorities of Lateritic Soils,
Terrain evaluation mapping for a landslide 7th Int. Conf. SMFE, Mexico City, (2), 13- 43.
inventory in Hong Kong. Proc. 4th Int. Conf. & De Mello, V.F.B. (1972). Thoughts on soil engineering
Field Workshop on Landslides, Tokyo, 63-68. applicable to residual soils. Proc. 3rd S.E. Asian
Burnett, A.D. & Styles, K.A. (1982). An approach Conf. Soil Eng., Hong Kong, 5-43.
to urban engineering geological mapping as used Deere, D.U. & Patton, F.D. (1971). Slope stability in
in Hong Kong. Proc. 4th Int. Cong. IAEG, New residual soils. Proc. 4th Panam. Conf. SMFE, San
Delhi, (1), 167-176. Juan, Puerto Rico, (1), 87-170.

Burt, T.P. (1978). An automatic fluid-scanning switch Endicott, L.J. (1982). Analysis of piezometer data
tensiometer system. British Geomorpholog. Res. and rainfall records to determine groundwater
Group., Tech. Bull. No. 21, 1-30. conditions. Hong Kong Engr, (10), 9, 53-56.
Campbell, R.H. (1975). Soil slips, debris flows and Eyles, R.J. (1979). Slip-triggering rainfalls in
rainstorms in Santa Monica mountains and vicinity, Wellington City, New Zealand. New Zealand Jour.
Southern California. US Geol. Survey, Prof. Paper Sc., (22), 117-121.
851, 55 p. Eyles, R.J., Crozier, M.J. & Wheeler, R.H. (1978).
Carrara, A. (1983). Multivariate models for landslide Landslips in Wellington City. New Zealand
hazard evaluation. Math. Geol., (15), 403-426. Geographer, (34), 2, 58-74.
Carrara, A., Catalano, E., Sorriso Valvo, M., Reali, C. Finlayson, A.A. (1984). Land surface evaluation for
& Ossi, I. (1978). Digital terrain analysis for land engineering practice: applications of the Australian
evaluation. Geologia Applicata e Idrogeologia, (13), PUCE System for terrain evaluation. Qtrly Jour.
69-127 & 3 drgs. Eng. Geol., (17), 149-158.

Carrara, A., Catalano, E., Sorriso Valvo, M., Reali, Fredlund, D.G. (1931). The shear strength of
C., Merendi, L. & Rizzo, V. (1977). Landslides unsaturated soils and its relationship to slope
morphometry and topology in two zones, Calabria, stability problems in Hong Kong. Hong Kong
Italy. Bull. IAEG, No. 16, 8-13. Engr, (9), 4, 37-45.
Ching, R.K.H. & Fredlund, D.G. (1984). Fredlund, D.G. (1984). Analytical methods for slope
Quantitative comparison of limit equilibrium stability analysis. Proc. 4th Int. Symp. Landslides,
methods of slices. Proc. 4th Int. Symp. Landslides, Toronto, (1), 229-250 .
Toronto, (2) 373-379. Fredlund, D.G., Krahn, J. & Pufahl, D.E. (1981).
Coates, D.R. (1977). Landslides perspectives. The relationship between limit equilibrium slope
Landslides, ed. D.R. Coates. Geol. Soc. America, stability methods. Proc. 10th Int. Conf. SMFE,
Reviews in Eng. Geol., (3), 3- 28. Stockholm, (3), 409-416.
Crozier, M.J. (1969). Earthflow occurrence during Fredlund, D.G. & Rahardjo, H. (1985). Theoretical
high intensity rainfall in eastern Otago. Eng. Geol., context for understanding unsaturated residual soil
(3), 325-334. behaviour. Proc. 1st Int. Conf. Geomechanics in
C r o z i e r , M . J . ( 1 9 7 3 ) . Te c h n i q u e s f o r t h e Tropical Lateritic and Saprolitic Soils, Brasilia, (1),
morphometric analysis of landslips. Zeitschrift für 295-306.
Geomorphologie, (17), 78-101. Fukuoka, M. (1980). Landslides associated with
Crozier, M.J. & Eyles, R.J. (1980). Assessing the rainfall. Geotech. Eng., (11), 1-29.
probability of rapid mass movement. Proc. 3rd Aus. Geological Society of London (1977a). The logging of
-NZ Conf. Geomechanics, Wellington, (2), 47-51. cores for engineering purposes. (Second Rep. by
Da Costa Nunes, A.J. (1969). Landslides in soils of Geol. Soc. Eng. Group Working Party). Qtrly Jour.
decomposed rock due to intense rainstorms. Proc. Eng. Geol., (10), 45-52.
7th Int. Conf. SMFE, Mexico City, (2), 547-554. Geological Society of London (1977b). The description
D’Appolonia, E.R., Alperstein, R. & D’Appolonia, of rock masses for engineering purposes. (Rep. by
D.J. (1966). Behavior of a colluvial slope. Proc. Geol. Soc. Eng. Group Working Party). Qtrly Jour.
Conf. Stability and Performance of Slopes and Eng. Geol., (10), 355-388.
Embankments, Berkeley, Cal., 489-515. (Reprinted Geotechnical Control Office (1982). Mid-levels Study:
Jour. Soil Mech. Found. Div., ASCE, (93), SM4, Report on Geology, Hydrology and Soil Properties.
447-493, 1976). Geotech. Control Office, Hong Kong, 265 p. & 54
Dearman, W.R. (1974). Weathering classification in the drgs.
characterization of rock for engineering purposes Geotechnical Control Office (1984). Geotechnical
in British practice. Bull. IAEG, No. 9, 33-42. Manual for Slopes (2nd ed.). Geotech. Control
Dearman, W.R., Baynes, F.J. & Irfan, T.Y. (1976). Office, Hong Kong, 295 p.
Engineering properties of weathered granite. Eng. Gidigasu, M.D. (1975). Laterite Soil Engineering.
Geol., (12), 345-374. Elsevier, Amsterdam, 570 p.
35
Finish.indb 153 2007/8/17 11:03:51 AM

Government of Hong Kong (1972a). Interim Report engineering geological mapping. Part I: rock and
of the Commission of Inquiry into the Rainstorm soil materials. Bull. IAEG, No. 19, 364-371.
Disasters, 1972. Hong Kong Govt Printer, 20 p. IAEG (1981). Report of the Commission on Site
Government of Hong Kong (1972b). Final Report Investigation. Bull. IAEG, No. 24, 185-226.
of the Commission of Inquiry into the Rainstorm Janbu, N. (1954). Application of composite slip
Disasters, 1972. Hong Kong Govt Printer, 91 P. surfaces for stability analysis. Proc. Europ. Conf.
Grant, K. (1975a). The PUCE programme for terrain Stability of Earth Slopes, Stockholm, (3), 43-49.
evaluation for engineering purposes: I. Principles. Janbu, N. (1973). Slope stability computations.
CSIRO Australia, Div. App. Geomech., Paper No. Embankment Dam Engineering (Casagrande
11, 209 p. Volume), ed. R.C. Hirschfield & S.J. Poulos,

Grant, K. (1975b). The PUCE programme for terrain 47-107. Wiley, New York.
evaluation for engineering purposes: II. Procedures Jones, F.O. (1973). Landslides of Rio de Janeiro and
for terrain classification. CSIRO, Australia, Div. the Serra das Araras escarpment, Brazil. US Geol.
App. Geomech. Paper No. 19, 66 p. & 2 drgs. Survey Prof. Paper 697, 48 p. & 2 maps.
Grant, K. & Finlayson, A.A. (1978). The assessment Jones, J.A.A. (1981). The Nature of Soil Piping - A
and evaluation of geotechnical resources in urban Review of Research. Geo Books, Norwich, UK,
and regional environments. Eng. Geol., (12), 315 p .
219-293. Koo, Y.C. (1982a). Relict joints in completely
Grant-Taylor, T.L. (1964). Stable angles in Wellington decomposed volcanics in Hong Kong. Canad.
greywacke. New Zealand Engng, April 15, Geotech. Jour., (19), 117-123.
129-130. Koo, Y.C. (1982b). The mass strength of jointed
Greenway, D.R., Anderson, M.G. & Brian-Boys, K.C. residual soils. Canad. Geotech. Jour., (19), 225-231.
(1984). Influence of vegetation on slope stability Koo, Y.C. & Lumb, P. (1981). Correlation between
in Hong Kong. Proc. 4th Int. Symp. Landslides, rainfall intensities and groundwater levels in the
Toronto, (1), 399-404 natural slopes of Hong Kong. Hong Kong Engr, (9),
Guidicini, G. & Iwasa, O.Y. (1977). Tentative 10, 49-52.
correlation between rainfall and landslides in a Lambe, T.W. (1973). Predictions in soil engineering.
humid tropical environment. Bull. IAEG, No. 16, Géotechnique, (23), 149-202.
13-20. Lambe, T.W. (1975). PRIME: Predict - instrument -
Hencher, S.R. & Martin, R.P. (1982). The description measure - evaluate. Proc. 4th S.E. Asian Conf. Soil
and classification of weathered rocks in Hong Kong Eng., Kuala Lumpur, G.1-G.6.
for engineering purposes. Proc. 7th S.E. Asian Lane, K.S. (1961). Field slope charts for stability
Geotech. Conf., Hong Kong, (1) , 125-142. studies. Proc. 5th Int. Conf. SMFE, Paris, (2),
Hencher, S.R. & Martin, R.P. (1984). The failure of a 651- 655.
cut slope on the Tuen Mun Road in Hong Kong. Leach, B. & Herbert, R. (1982). The genesis of a
Proc. Int. Conf. Case Histories in Geotechnical numerical model for the study of the hydrogeology
Engineering, St Louis, Missouri, (2), 683-688. of a steep hillside in Hong Kong. Qtrly Jour. Eng.
Hencher, S.R., Massey, J.B. & Brand, E.W. (1984). Geol., (15), 243-259.
Application of back analysis to some Hong Kong Leroueil, S. & Tavanas, F. (1981). Pitfalls of back
landslides. Proc. 4th Int. Symp. Landslides, analyses. Proc. 10th Int. Conf. SMFE, Stockholm,
Toronto, (1), 631-638 (1), 185-190. (Discussion, (4), 603-604).
Hudson, R.R. & Hencher, S.R. (1984). The delayed Leshchinsky, D., Baker, R. & Silver, M.L. (1985).
failure of a large cutting in Hong Kong. Proc. Int. Three dimensional analysis of slope stability. Int.
Conf. Case Histories in Geotechnical Engineering, Jour. Numeric. & Analyt. Methods in Geomech.,
St Louis, Missouri, (2), 679-682. (9), 199-223.
Ho, D.Y.F. & Fredlund, D.G. (1982). Increase in Little, A.L. (1967). Laterites. Proc. 3rd Asian Reg.
strength due to suction for two Hong Kong Conf. SMFE, Haifa, Israel, (2), 61-71.
soils. Proc. ASCE Spec. Conf. Engineering and Little, A.L. (1969). The engineering classification of
Construction in Tropical and Residual Soils, residual tropical soils. Proc. Spec. Sess. Lateritic
Honolulu, 263-295. Soils, 7th Int. Cont. SMFE, Mexico City, (1), 1-10.
Hunt, T. (1982). Slope failures in colluvium overlying Lovell, C.W. (1984). Three dimensional analysis
weak residual soils in Hong Kong. Proc. ASCE of landslides. Proc. 4th Int: Symp. Landslides,
Spec. Conf. Engineering and Construction in Toronto, (2), 451-455
Tropicai and Residual Soils, Honolulu, 443-462. Lumb, P. (1962a). The properties of decomposed
Huntley, S.L. & Randall, P.A. (1981). Recognition of granite. Géotechnique, (12), 226-243 (& 2 plates).
colluvium in Hong Kong. Hong Kong Engr, (9), Lumb, P. (1962b). Effect of rain storms on slope
12, 13-18. stability. Proc. Symp. Hong Kong Soils, Hong
IAEG (1979). Classification of rocks and soils for Kong , 73 -87.
36
Finish.indb 154 2007/8/17 11:03:51 AM

Lumb, P. (1965). The residual soils of Hong Kong. Nilsen, T.H., Wright, R.H., Vlasic, T. & Spangle,
Géotechnique, (15), 180-194. (Discussion, (16), W.E. (1979). Relative slope stability and land-
1966, 76-81 and 359-360). use planning in the San Francisco Bay region,
Lumb, P. (1972). Landslides in Hong Kong. Proc. 1st California. US Geol. Survey, Prof. Paper 944, 104 p.
Int. Symp. Landslide Control, Kyoto, 91-93. & 3 drgs.
Lumb, P. (1975). Slope failures in Hong Kong. Qtrly Onodera, T., Yoshinaka, R. & Kazama, H. (1974).
Jour. Eng. Geol., (8), 31-65. Slope failures caused by heavy rainfall in Japan.
Lumb, P. (1979). Statistics of natural disasters in Hong Proc. 2nd Int. Cong. IAEG, Sao Paulo, (2), V.II.I-
Kong, 1884-1976. Proc. 3rd Int. Conf. Applications V II.10 .
of Statistics & Probability in Soil & Structural Oyagi, N. (1984). Landslides in weathered rocks and
Engineering, Sydney, (1), 9-22. residual soils in Japan and surrounding areas. Proc.
Lumb, P. (1983). Engineering properties of fresh and 4th Int. Symp. Landslides, Toronto, (3), 1-31.
decomposed igneous rocks from Hong Kong. Eng. Page, W.D. & James, M.E. (1981). Landslides and
Geol., (19), 81-94. mudflows, major contributors to the landscape in
Lumsdaine, R.W. & Tang, K.Y. (1981). On stability Northwest Colombia. Revista CIAF, (6), 455-467.
analysis in Hong Kong. Proc. 10th Int. Conf. Patton, F.D. & Hendron, A.J. (1974). General report on
SMFE, Stockholm, (4), 921-922. “Mass Movements”. Proc. 2nd Int. Cong. IAEG,
Lumsdaine, R.W. & Tang, K.Y. (1982). A comparison Sao Paulo, (2), V-GR.1 -V.GR. 57
of slope stability computations. Proc. 7th S.E. Peck, R.B. (1969). Advantages and limitations of the
Asian Geotech. Conf., Hong Kong, (1), 31-38. observational method in applied soil mechanics.
Martin, R.P. & Hencher, S.R. (1984). Principles for Géotechnique, (19), 171-187.
description and classification of weathered rocks Peck, R.B. (1975). General introduction to the session
for engineering purposes. Proc. 20th Conf. Eng. “Landslides and Their Prevention”. Proc. 4th
Group Geol. Soc. London, Guilford, 304-308. Guelph Symp. Geomorphology, Guelph, Ontario,
McMahon, B.I. (1976). Estimation of upper bounds 133-136.
to rock slopes by analysis of existing slope data. Peck, R.B. (1980). “Where has all the judgement
CANMET Canada Centre Min. & Energy Tech.,, gone?”. The fifth Lauritz Bjerrum memorial lecture.
Rep. 79-14, 65 p. Canad. Geotech. Jour., (17), 584-590. (Reprinted,
Moh, Z.C. (Ed.)(1969). Proceedings of the Specialty Norweg. Geotech. Inst. Pub. No. 134, 1-5).
Session on Engineering Properties of Lateritic Pierson, T.C. (1983). Soil pipes and slope stability.
Soils, 7th lnt. Cont. SMFE, Mexico City, 2 vols., Qtrly Jour. Eng. Geol., (16), 1-11.
419 p. Pope, R.G., Weeks, R.C. & Chipp, P.N. (1982).
Morgenstern, N.R. & de Matos, M. (1975). Stability Automatic recording of standpipe piezometers.
of slopes in residual soils. Proc. 5th Panam. Conf. Proc. 7th S.E. Asian Geotech. Conf., Hong Kong,
SMFE, Buenos Aires, (3), 369-384. (1), 77-89.
Morgenstern, N.R. & Price, V.E. (1965). The Porcher, M. & Guillope, P. (1979). Cartographic
analysis of the stability of general slip surfaces. des risques ZERMOS appliquée à des plans d’
Géotechnique, (15), 79-83. occupation des sols en Normandie. Bull. Liais.
Moye, D.G. (1955). Engineering geology for the Lab. Ponts Chaus., No. 99, 43-54.
Snowy Mountains Scheme. Jour. Inst. Engrs Premchitt, J., Brand, E.W. & Phillipson, H.B. (1985).
Australia, (27), 287-298. Landslides caused by rapid groundwater changes.
Nash, J.M. & Dale, M.J. (1983). Geology and Proc. 21st Conf. Eng. Group Geol. Soc. London,
hydrology of natural tunnel erosion in superficial Sheffield, 31-42
deposits in Hong Kong. Proc. Conf. Geology of Rodin, S., Henkel, D.J. & Brown, R.L. (1982).
Surficial Deposits in Hong Kong, Hong Kong, Geotechnical study of a large hillside area in Hong
61-72. Kong. Hong Kong Engr, (10), 5, 37-45.
Neuland, H. (1976). A prediction of landslips. Catena, Rossetti, R. & Ottone, C. (1979). Esame preliminaire
(3), 215-230. delle condizioni pluviometriche dell’Oltrepò
Nilsen, T.H. & Brabb, E.E. (1977). Slope stability Pavese e dei valori critici delle precipitazioni
studies in the San Francisco Bay region, California. in relazione ai fenomeni di dissesto franoso
Geol. Soc. America, Reviews in Eng. Geol., (3), (Preliminary analysis of rainfall in Oltrepò Pavese
235-243. and the critical rainfall values as related to landslide
Nilsen, T., Taylor, F.A. & Brabb, E.E. (1976). failures). Geologia Applicata e Idrogeologia, (14),
Recent landslides in Alameda County, California 3, 83-99.
(1940-71): An estimate of economic losses and Ruxton, B.P. (1960). The geology of Hong Kong.
correlations with slope, rainfall and ancient Qtrly Jour. Eng. Geol., (115), 233-260 (& 2 plates
landslide deposits. US Geol. Survey, Bull. 1398, & 1 map).
24 p. & 1 drg. Ruxton, B.P. & Berry, L. (1957). Weathering of granite
37
Finish.indb 155 2007/8/17 11:03:52 AM

and associated erosional features in Hong Kong. Sweeney, D.J. & Robertson, P.K. (1982). Slope
Bull. Geol. Soc. America, (68), 1263-1291 (& 1 stability in residual soils in Hong Kong. Canad.
plate). Geotech. Jour., (19), 521-525.
Sarma, S.K. (1973). Stability analysis of embankments Taylor, D.K., Hawley, J.G. & Riddolls, B.W. (1977).
and slopes. Géotechnique, (23), 423-433. Slope stability in Urban Development. New
Sarma, S.K. (1979). Stability analysis of embankments Zealand Geomech. Soc., Wellington, 71 p.
and slopes. Jour. Geotech. Eng. Div., ASCE, (105), Terzaghi, K. (1950). Mechanism of landslides.
1511-1524. Application of Geology to Engineging Practice
Shuk, T. (1965). Discussion on “Some aspects of the (Berkey Volume), 83-123. Geol. Soc. America,
safety factor in soil mechanics” by A. Langejan. Boulder, Col.
Proc. 6th Int. Conf. SMFE, Montreal, (3), 576-577. Vail, A.J. (1984). Two landslide disasters in Hong
Shuk, T. (1968). Unmetodo sencillo de diseno para Kong. Proc. 4th Int. Symp. Landslides, Toronto, (1),
minimizar el costo de taludes en roca (A simple 717-722.
design method to minimize the costs of slopes in Vargas, M. (1953). Some engineering properties of
rock). Proc. 1st Columbian Symp. Eng. Geol. , residual clay soils occurring in southern Brazil.
Bogota, (2), 17 p. (English translation reprinted Proc. 3rd Int. Conf. SMFTE, Zurich, (1), 67-71.
under Appendix B to “Estimation of upper bounds Vargas, M. (1967). Design and construction of large
to rock slopes by analysis of existing slope data” cuttings in residual soils. Proc. 3rd Panam Conf.
by B.K. McMahon. CANMET Canad. Centre Min. SMFE, Caracas, (2), 243-254.
& Energy Tech., Rep. 76-14, 18-34. Vargas, M. & Pichler, E. (1957). Residual soil and
Skempton, A.W. & Hutchinson, J. (1969). Stability of rock slides in Santos (Brazil). Proc. 4th Int. Conf.
natural slopes and embankment foundations. Proc. SMFE, London, (2), 394-398.
7th Int. Conf, SMFE Mexico City, State-of-the-Art Varnes, D.J. (1985). Landslide Hazard Zonation:
vol., 291-340. Review of Principles and Practice. UNESCO,
So, C.L. (1971). Mass movements associated with the Paris.
rainstorm of June 1966 in Hong Kong. Trans. Inst. Varnes, D.J. & Keaton, J.R. (1984). Trends in
British Geographers, No. 53, 55-56. engineering geological and related mapping
Sowers, G.F. (1963). Engineering properties of residual 1972-1983. Bull. Assoc. Eng. Geologists, (21),
soils derived from igneous and metamorphic rocks. 255-267.
Proc. 2nd Panam. Conf. SMFE, Brazil, (1), 39-62. Vaughan, P.R. (1985a). Mechanical and hydraulic
Sowers, G.F. (1971). Landslides in weathered properties of in-situ residual soils, particularly as
volcanics in Puerto Rico. Proc. 4th Panam. Conf. related to their structure and mineral components.
SMFE, San Juan, Puerto Rico, (2),105-115 Proc. 1st Int. Conf. Geomechanics in Tropi Lateritic
St John, B.J., Sowers, G.F. & Weaver, Ch. E. (1969). and Saprolitic Soils, Brasilia, (3), 231-263.
Slickensides in residual soils and their engineering Vaughan, P.R. (1985b). Pore pressures due to
significance. Proc. 7th Int. Conf. SMFE, Mexico infiltration into partially saturated slopes. Proc. 1st
City, (2), 591-597. Int. Conf. Geomechanics in Trogical Lateritic and
Styles, K.A., Hansen, A., Dale, M.J. & Burnett Saprolitic Soils Brasilia, (2), 61-71.
A.D. (1984). Terrain classification methods for Vaughan, P.R. & Kwan, C.W. (1984). Weathering
development planning and geotechnical appraisal: a structure and in situ stress in residual soils.
Hong Kong case. Proc. 4th Int. Symp. Landslides, Géotechnique, (34), 43-59.
Toronto, (2), 561-568. Whitman, R.V. (1984). Evaluating calculated risk in
Sweeney, D.J. (1982). Some in situ soil suction geotechnical engineering (7th Terzaghi Lecture).
measurements in Hong Kong’s residual soil slopes. Jour. Geotech. Eng., ASCE, (110), 145-168.
Proc. 7th S.E. Asian Geotech. Conf., Hong Kong, Young, S. (1972). Slopes. Longman, London, 295 p.
(l), 91-106. (Discussion, (2), 93-96).
38
Finish.indb 156 2007/8/17 11:03:52 AM

Brand (1985) - Predicting The Performance of Residual Soil Slopes

Uploaded by

Copyright:

Available Formats

Brand (1985) - Predicting The Performance of Residual Soil Slopes

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Brand (1985) - Predicting The Performance of Residual Soil Slopes

Uploaded by

Copyright:

Available Formats

Predicting the Performance of Residual Soil Slopes

1. INTRODUCTION (c) they are nearly always unsaturated, which poses

Finish.indb 119 2007/8/17 11:03:28 AM

slope stability. formed by the insitu weathering of any rock type,

Finish.indb 120 2007/8/17 11:03:28 AM

Finish.indb 121 2007/8/17 11:03:29 AM

of which is Lantau (Figure 1). The population is

than was available prior to the construction phase. This

3. INTRODUCTION TO HONG KONG

Finish.indb 122 2007/8/17 11:03:29 AM

Figure 1. Map of the territory of Hong Kong showing

concentrated in a number of distinct geographical

Finish.indb 123 2007/8/17 11:03:31 AM

of ‘pipes’ or ‘tunnels’ as a result of internal erosion

Finish.indb 124 2007/8/17 11:03:31 AM

empirical methods are closely related, in that both

is being increasingly adopted even in this difficult

5. RELATIONSHIP BETWEEN RAINFALL AND

Finish.indb 125 2007/8/17 11:03:32 AM

Figure 6. Relationship between rainfall and landslides

In incorporating antecedent rainfall into his Figure 7. Locations of continually recording

5.2 Rainstorm and Landslide Data

Finish.indb 126 2007/8/17 11:03:33 AM

Finish.indb 127 2007/8/17 11:03:33 AM

Figure 11. Occurrence of landslides in relation to

5.3 Effect of Rainfall Intensity

The available rainfall and landslide data was examined

Figure 13. Occurrence of landslides in relation to

In June 1966, the heaviest rainfall in memory fell

Finish.indb 128 2007/8/17 11:03:34 AM

much higher than the corresponding ones measured at

number of casualties caused by landslides shows a

Finish.indb 129 2007/8/17 11:03:35 AM

to his combined ‘severe’ and ‘disastrous’ categories.

correlations exist between landslides and 24-hour

The relationship shown in Figure 19 is not unlike

Finish.indb 130 2007/8/17 11:03:35 AM

rainfall intensities and times of landslides, it is

Finish.indb 131 2007/8/17 11:03:36 AM

publications (Brand et al, 1982a, 1982b; Burnett &

specific areas identified from Regional Studies.

Finish.indb 132 2007/8/17 11:03:36 AM

with gentle slopes less than 15° instability and

Figure 21. Example of a Terrain Classification Map

Finish.indb 133 2007/8/17 11:03:37 AM

be devised among a large number of geological and

log plot of height versus slope angle for natural shale

7.2 The CHASE Study in Hong Kong

Figure 25. Design guidelines developed by Shuk

Finish.indb 134 2007/8/17 11:03:37 AM

Figure 26. Typical stable Hong Kong cut slope

of colluvium overlying decomposed granite, which

8. SOIL MECHANICS PREDICTIONS

Finish.indb 135 2007/8/17 11:03:40 AM

the normal effective stress distribution. The normal

Finish.indb 136 2007/8/17 11:03:40 AM

later in Section 9. Despite the obvious objections to the measurement of

Finish.indb 137 2007/8/17 11:03:41 AM

pore pressure increases. In its simplest form, the