Geomorphology 49 (2002) 269 – 279

www.elsevier.com/locate/geomorph

A new approach to soil characterisation for hydrology–stability

analysis models
Ioana Ibraim, Malcolm G. Anderson *
School of Geographical Sciences, University of Bristol, University Road, Bristol BS8 1SS, UK

Received 10 January 2002; received in revised form 12 June 2002; accepted 21 June 2002

Abstract

Landslide stability analysis increasingly utilises high-resolution coupled hydrology – slope stability models (CHASM) to
improve stability assessments in areas subject to dynamic pore pressure regimes. In such environments, the estimation of soil
hydraulic conductivity (K) is a key parameter but one which is not always readily available or determined with the required
resolution. By using basic soil particle-size distribution (PSD) data, we evaluate the microscopic composition of the actual soil,
and applying the analytical relations obtain by a Self-Consistent Method (SCM) approach, we determine an appropriate value of
K. This is of importance in that it allows within-soil type variability to be reflected in terms of K and hence within the model
structure. The SCM methodology is briefly reviewed and an illustrative application is undertaken for a slope typical of Hong
Kong. The results show model output sensitivity in terms of moisture content and factor of safety (FOS) when comparing K
values determined using the SCM approach and the conventional field determination. In attempting to determine slope
hydrological processes and attendant stability conditions, we conclude that the application of SCM approach offers a novel
methodology for potentially improving the parameterisation of hydrology – slope stability models.
D 2002 Elsevier Science B.V. All rights reserved.

Keywords: Hydraulic conductivity; Slope stability; Self-Consistent Method

1. Introduction essential for the problems of consolidation (McManus

and Kulhawy, 1991; Anderson et al., 1991), internal
The hydraulic conductivity of natural or artificially dam and embankment erosion (Sherard, 1984; Tedd et
reconstituted soils is an important physical parameter al., 1987; Van Zyl and Harr, 1981) and stability
in many areas of the physical and natural sciences. (Kenney et al., 1984; Skempton and Brogan, 1994).
This parameter is a key element in modelling seepage, Within the field of hydrology and slope stability
and consolidation, in influencing stability analysis and process relationships, hydraulic conductivity is a key
in the accurate prediction of fluid movement through parameter in two specific contexts: in environments
soils. For instance, knowledge of this property is (1) where pore pressure changes are highly dynamic
and (2) where there are additional demands on stand-
ard analyses, such as the desire to include bioengin-
*
Corresponding author. Fax: +44-117-928-7878. eering components in hydrology– stability modelling
E-mail address: [email protected] (M.G. Anderson). schemes (Collison et al., 1995).

0169-555X/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 9 - 5 5 5 X ( 0 2 ) 0 0 1 9 0 - 3
270 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279

Sensitivity analyses of combined hydrology – stabil- Because the direct measurement of hydraulic con-
ity models have shown the importance of hydraulic ductivity is so susceptible to numerous sources of
conductivity (Anderson and Howes, 1985). As the error, has poor repeatability and demands moderately
refinement in terms of both spatial scale and numerical expensive testing, it is of particular interest to attempt
procedures increases, a clear consequence is the a relevant estimation procedure using only standard
requirement to parameterise hydrology– stability mod- geotechnical measurements, such as the percentage of
els more precisely. Central to such parameterisation constituents and the mineralogical structure of the clay
needs is the ability to better estimate hydraulic con- phase (e.g., plasticity index). Assessing the perme-
ductivity. This paper seeks to address this issue by: ability of a given medium in an analytical form from
such constituents is thus of great interest. The majority
of such previous investigations have been concerned

Introducing a new analytical method for determin- with granular soils (Loudon, 1952; Amer and Awad,
ing hydraulic conductivity (K) derived from 1974; Kenney et al., 1984) or fine-grained soils
percentages of the mineralogical structure constit- (Tavenas et al., 1983; Al Tabbaa and Wood, 1987).
uents, However, even if the basic physics of the phenomena

Incorporating this method within a hydrology – are well understood, this problem is complex and the
slope stability model structure, question is still open despite a number of experimental

Developing an application to illustrate the poten- and theoretical studies (Chapuis, 1990; Kenney et al.,
tially restrictive influence that conventional param- 1992; Schuster, 1987).
eterisation of K can have and Taking into account the large number of parame-

Suggesting the possible revision of site investiga- ters involved in the identification and the character-
tion procedures. isation of the hydraulic behaviour of a mixed soil
(size of elements, grain-size distribution and arrange-
ment of elements), the real configuration of the
2. Standard approaches to the characterisation of geometry and flow within a typical specimen of
hydraulic conductivity ground are too complicated to be described at the
microscopic level. Assessing hydraulic conductivity
There have been numerous studies dealing with of a medium given by a mixed soil could, however,
fundamental aspects of soil hydraulic conductivity. be possible by using the principles of the self-con-
Such studies include the validity of Darcy’s law (e.g., sistent homogenisation method (Boutin, 2000; Boutin
Olsen, 1966; Mitchell, 1993; Law and Lee, 1981; and Ibraim, 2001).
Jacquin, 1965), the principal factors controlling This paper uses the initial principles and analytical
hydraulic conductivity (Chapuis, 1990; Daniel, results developed by Boutin (2000) and Boutin and
1984; Haug and Wong, 1992; Elsbury et al., 1990) Ibraim (2001) to define the homogeneous hydraulic
and the measurement of the value itself (Shackelford, conductivity of a heterogeneous medium in theoretical
1994). Hydraulic conductivity can be obtained by terms. The currently reported research extends this
laboratory tests (direct evaluation in a permeameter, analysis to a natural soil (requiring procedures for
indirect evaluation from a consolidation test or use of estimating free water) having three constituents (sand,
a triaxial test). Laboratory tests are small scale and clay and water). This improved analysis is then
constrain the water flow direction to be vertical, applied to a sample 30j Hong Kong hillslope typified
whereas horizontal flow rates may in fact be much by deep colluvium and subjected to intense precip-
greater. Field tests overcome these shortcomings by itation events. The characteristics of the slope are
direct site measurement (such as drill hole and infil- configured from detailed available geotechnical data.
tration tests). However, because the pattern of water The Hong Kong slope environment is an important
flow from a borehole cannot generally be known, analogue for the more general problem of slope
interpretation of the results can be difficult and stability investigations in subtropical areas having
uncertain. Thus, one problem is exchanged for another highly dynamic pore pressure regimes within weath-
(Carter and Bentley, 1991). ered residual soils.
 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279 271

3. Self-Consistent Method (SCM) to estimate . granular particles (index s), having zero hydraulic
hydraulic conductivity conductivity (Ks = 0). This impervious constituent can
be sand, silt, gravel or rock blocks.
The main feature of the SCM is to propose and . water pockets (index w) having infinite hydraulic
develop analytical expressions of the effective coef- conductivity (Kw = l). In fact, this is an approxima-
ficient of a disordered medium (in our case, for the tion because the flow in a water pocket also dissipates
hydraulic conductivity) based on an analysis of sim- energy. However, we consider water pockets to be of a
plified configurations. The purpose of the method is to large size compared to the other elements in the
define a homogenous continuous media, with structure so that the dissipation is negligible and the
hydraulic behaviour equivalent to the actual hetero- approximation is justified.
geneous media. It is convenient for this kind of model to treat soils
The search for an equivalent macroscopic (i.e., containing saturated clay with one or two additional
homogenised) behaviour of a heterogeneous medium constituents. The volume concentration of constitu-
makes sense only if the condition of scale separation ents, defined in the following manner, are used:
is fulfilled (Auriault, 1991). This condition leads to Cs = Vs/(Vc + Vs + Vw) and Cw = Vw/(Vc + Vs + Vw),
two requirements that result in a ‘‘local invariance’’ where Vc, Vs and Vw are volumes of the constituents
for all kinds of homogenisation methods: in the elementary representative volume of the heter-
ogeneous soil.
the first concerns the material, which must be With these definitions in place, there are five
sufficiently regular so that one can define a steps to implementing the SCM to provide a way
characteristic microscopic size l, which can also of estimating the effective macroscopic hydraulic
be termed the elementary representative volume conductivity of heterogeneous media or perfectly
(ERV), and disordered media (Kröner, 1978; for more details,
the second concerns the phenomenon, for which refer to Hashin, 1968, who treats an identical prob-
the characteristic size of its variations (hydraulic lem for thermal conduction in a biconstituents
conductivity to macroscopic pressure relation) media).
must be large in comparison with l. The steps of this method are (Boutin, 2000):

In the case of heterogeneous soils, the scale sepa- (a) assume macroscopic behaviour, the coefficients of
ration required by the SCM and specified above is which have to be determined
implicitly satisfied. (b) choose the topological assemblage of the media to
Heterogeneous soils can be considered to be a define the ‘‘generic heterogeneity,’’ characterised
mixture of two or three constituents such as saturated by a representative microstructure of the soil,
clay, impervious grains (silt, sand and gravel) and embedded in such a medium (for simplicity, we
perfectly permeable water pockets. The reasoning in used spherical particles)
the following analysis is conducted at a macroscopic (c) solve the basic problem under homogeneous
scale so that Darcy’s law is valid for any constituent boundary conditions (applied at the bounds of
and the aim of SCM is to give a pertinent assessment the infinite medium)
for the hydraulic conductivity. (d) express energy equivalence between the whole of
This study is based on the soil constituents having the ‘‘generic heterogeneity’’ and the equivalent
spherical (or spherical shell) shapes and with the medium
following definitions having been made for the con- (e) deduce from this relation the value of the macro-
stituents: scopic coefficient
. saturated clay (index c), having hydraulic con-
ductivity Kc. This constituent is considered at a The SCM thus constitutes a straightforward
sufficiently large scale, so that it appears as a homo- approach for assessing the effective coefficients of
geneous medium, in which the solid clay particles and a disordered medium. However, even if the concen-
microvoids are strongly associated. tration of each ‘‘generic heterogeneity’’ is known,
272 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279

their assemblage is not specified and the micro- Because of the linearity of the problem, the flux for
morphology adapted to the results given by SCM is every phase i depends linearly on the amplitude of the
generally unknown. For instance, the connectedness imposed pressure gradient, so
(or dispersion) of a given phase can sometimes only
be checked a posteriori, whereas it could be an < v >Vi ¼  < Kgradp >Vi ¼ biA < gradp > A
essential feature for the physics (Ibraim et al., in X
and then: K < gradp >¼  CibiA< gradp >A
press).
A simple way to explore the implicit morphology
Taking the scalar product by e and writing bi = bie,
is to make the physical coefficient of the considered i
one obtains K = SCibi (where Cibi expresses the
phase infinite or vanishing to zero, the coefficient of
contribution of the phase i embedded in the other
the other phases remaining finite. For instance, we
phases), from which the permeability is determined
consider the case of a mixed soil of permeable
when the value of bi is known.
constituents. Letting the permeability of the i constit-
The SCM facilitates the determination of bi. The
uent be ki, then when ki = l, the macroscopic per-
value of bi is estimated by replacing the real media
meability K = l, then phase i is connected, but if K
around the representative pattern associated with the
remains finite, phase i is dispersed. Conversely, if
phase i by the researched unknown equivalent homo-
ki = 0, then K = 0, all the other phases considered
geneous medium of hydraulic conductivity K. Con-
together are dispersed in the phase i, but if K remains
sequently, the initial complex interphase interaction
finite, then the other phases, considered together, are
problem is significantly simplified and can be solved.
connected.
bi depends on K, so that the equation allowing the
As we have already commented, if one applies
determination of K is K = SCibi(K). The value bi for
SCM to estimate the hydraulic conductivity of a
different ‘‘generic inclusions’’ (tricomposite, bicom-
mixture of isotropic permeable media, the reasoning
posite and monocomposite) is given in Boutin (2001)
is conducted at the macroscopic scale so that Darcy’s
and Ibraim et al. (in press).
law is valid for any constituent. The initial heteroge-
neous medium is formed of several phases i in volume
concentration Ci (SCi = 1), each of them being char-
acterised by a representative pattern. One thus looks
for a macroscopic Darcy’s law relating the average
flow to the average pressure gradient. For simplicity,
one assumes isotropy of constituents (at the micro-
scopic level) and isotropy of the medium (at the
macroscopic level).
The media is subjected at its boundary to a pressure
corresponding to a uniform macroscopic pressure
gradient in direction e:
X
< gradp >¼ A < gradp > Ae ¼ Ci < gradp >Vi

Considering an isotropic permeability K at a macro-

scopic level, the resulting macroscopic flux is in the
same direction e:

Fig. 1. Topological pattern for two constituents materials modelling:

(a) classical self-consistent modelling, with two simple inclusions,
< v > ¼ A < v > Ae ¼ K < gradp > and (b) Hashin (1968) self-consistent modelling, with bicomposite
X pattern, made of an internal sphere of granular phase, coated with a
¼ Ci < v >Vi concentric external spherical shell of saturated clay phase.
 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279 273

It is possible to develop simple models allowing sistent trimodelling (with two possibilities: clay/sand/
the description of the medium with different morphol- water pocket mixtures and clay/water/granular phase
ogy, as the soil is composed of a mixture of clay, mixtures) (Fig. 2d and e).
granular particles and water. Below we enumerate the All the modelling topologies are built up to ensure
general description of some representative patterns for the connectivity of the clay phase. However, we note
two and three constituent media and we give the that the classical SCM approaches (for both soil
general results in terms of analytical relationships compositions—two or three constituents) cannot
for the homogenised hydraulic conductivity: achieve the connectivity of the clay phase and the
. for two constituent materials modelling, the clas- dispersion of the granular phase and/or water pockets
sical self-consistent modelling (with two simple pure if Cc < 1/3 (see Figs. 1a and 2a).
inclusions) (Fig. 1a) and Hashin self-consistent mod- These different homogenisation schemas can be
elling (with bicomposite pattern, made of an internal used to determine the hydraulic conductivity of the
sphere of granular phase, coated with a concentric ex- mixed soil, in direct connection with the hydraulic
ternal spherical shell of saturated clay phase) (Fig. 1b); conductivity of the saturated clay phase, Kc. However,
. for three constituent materials modelling, classi- the question is which model is suited to the morphol-
cal self-consistent modelling (with three simple pure ogy of the considered medium or, in another words,
inclusions) (Fig. 2a), double homogenisation model- what kind of morphology of the soil is implicitly
ling (with two possibilities: clay/sand mixture coating described by the model?
water pocket and clay/water pocket mixture coating Among all possible topologies of homogenisation,
granular phase) (Fig. 2b and c) or Hashin self-con- the most realistic from a physical point of view,

Fig. 2. Topological pattern for three constituents materials modelling: (a) classical self-consistent modelling, with three simple inclusions, (b)
double homogenisation modelling: clay/sand mixture coating water pocket, (c) double homogenisation modelling: clay/water pocket mixture
coating granular phase, (d) Hashin (1968) self-consistent trimodelling: clay/sand/water pocket mixtures and (e) Hashin (1968) self-consistent
trimodelling: clay/water/granular phase mixtures.
274 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279

considering the complex geometry of the mixed soil,

are those which impose the granular phase as an
inclusion in a matrix of saturated clay (for biconstitu-
ent structure) or a homogeneous mixture of saturated
clay and water (for triconstituent structure).
The biconstituents structure (granular phase sur-
rounded by saturated clay phase) gives independence
on the size of the constituents, in the sense that a given
spherical volume of this mixing can equally be con-
sidered as a unique, representative pattern or a homo-
genised zone and so represents more explicitly the real
interaction between the constituents. An experimental
validation of this structure applied to thermal conduc-
tion of ultralightweight concrete is presented by Bou-
tin (1996). The formation of this soil structure depends
on the electrochemical bonding, which itself depends
on the presence of water. So it is not unusual to
assume, as a first approximation, that all of the water
in the soil is associated with the clay phase (water is Fig. 4. Hashin (1968) double self-consistent modelling, with clay/
strongly attracted to clay mineral surfaces and results water pocket mixture coating the granular phase.
in plasticity) and granular particles act simply as
impervious and dry inclusion in this matrix. sentation of homogenised permeability variation, with
For the triconstituents chosen modelling, the the volumetric concentration constituents (Cs, the
Hashin homogenisation procedure was used twice, granular phase concentration, and Cw, free water
so this corresponding morphology should be adapted concentration).
for its microscale to large domains (in comparison to The problem is then to decide when the soil can be
water pocket size) of uncompacted or oversaturated considered with a biconstituent or triconstituent struc-
clay surrounding impervious grains. In another words, ture because this is necessary to examine the micro-
the granular phase size is bigger (and visible at the scopic structure of the mixed soil. Hence, the granular
macroscopic level) than the other two constituents particles (gravel, sand and silt) are considered infin-
(free water and saturated clay). itely impervious and the clay phase is contained within
For a better comprehension of the chosen topology, voids between granular particles and, in the presence
Figs. 3 and 4 show the two schemas and the repre- of water, hydrates and swells. When the void ratio of
the clay phase is less that its free-swell capacity, clay
completely fills the void space of the granular phase
and thus prevents direct contact between granular
particles. However, if the void space in the granular
phase exceeds the free-swell capacity of the clay, the
spaces will not become completely filled with
hydrated clay, and some will contain free water, form-
ing holes or, if connected to other pocket of water,
channels (Kenney et al., 1992). One chooses to define
the capacity of free swelling of the clay phase as the
volume of absorbed water by clay phase, normalised
by the volume of the solid particles of the clay.
Fig. 3. Hashin (1968) self-consistent modelling, with bicomposite
With the moisture content (or its dry density) and
pattern, made of an internal sphere of granular phase, coated with a the degree of saturation of the soil, it is possible to
concentric external spherical shell of saturated clay phase. directly determine the porosity of the soil. For reasons
 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279 275

of simplicity, we assume the specific gravity of clay not only within standard stability analysis but also in
and the granular particles to be the same. regard to slope surface cover effects. This is especially
Usually, geotechnical site investigation measure- so in terms of large engineered and natural slopes in
ments give soil composition in terms of dry mass the Far East (Anderson et al., 1994, 1996; Collison et
concentrations of constituents (Cc = mass of clay/total al., 1995), where there is considerable interest in the
mass of dry soil). After the estimation of the volume of role of negative pore pressures in controlling slope
water adsorbed by dry clay during the swelling process stability (Anderson, 1990).
and the free water volume (not involved in the satu- It was therefore deemed appropriate to demonstrate
ration of the clay phase), one can estimate the volu- the inclusion of the SCM Ksat within the slope stability
metric concentration of characteristic inclusion method for conditions typical of colluvium slopes. In
(granular phases concentration Cs and free water Hong Kong, substantial geotechnical data are available
concentration Cw). We thereby obtain the following and we based the following application on data typical
analytical relationship for those two parameters, which of such slopes. From a key data source (Hong Kong
are able to completely define the equivalent hydraulic Government, 1982), we derived the initial data in
conductivity of the initial heterogeneous soil: terms of particle-size distribution (PSD) results, initial
moisture content, degree of saturation and initial dry
Cs ¼ ð1  CcÞ=½1 þ n=ð1  nÞ and density free-swelling capacity for the clay phase.
Cw ¼ ½n=ð1  nÞ  IsCc=½1 þ n=ð1  nÞ Utilisation of these data with the self-consistent
method made possible the estimation of the hydraulic
where n is the porosity of the soil and Is is the free- conductivity of the colluvium soil. As we have stressed
swelling capacity of the clay phase. above, in contrast to standard applications, which
In conclusion, a mixed soil is represented by an assume a constant value of Ksat associated with a
assemblage of three constituents, one permeable, one given soil type, the SCM methodology allows estima-
impervious (having concentration Cs) and one per- tion of within-soil class variability.
fectly permeable (having concentration Cw), for
applying the self-consistent homogenisation method. Table 1
These extreme values of the permeability allow a Geometrical, hydrological and vegetation parameters for the test
simple analytical expression for the homogenised slopes
saturated hydraulic conductivity K*, taking into Assigned values
account the physical considerations outlined above: Geometrical parameters
Slope height (m) 15
K* ¼ Kcf1 þ 3Cw=ð1  Cs  CwÞg Slope length (m) 22
Slope angle (j) 30
 f1 þ Cs=½1 þ ð1  CsÞ=3g Hydrological parameters
Initial suction in top cell (m)  2.0
We should further note that even if there is enough Rainfall scenario (48 h simulation: 12.87
saturated clay in the media to fill all the void space of 5 h dry + 24 h rain)
the granular phase (and thus prevent direct contact Rainfall intensity (mm h  1)
Vegetation parameters
between granular particles), the composite system will
Net radiation (W m  2) 700
be formed by two constituents (Cw = 0), the analytical Average daily air temperature (jC) 15
relationship for the hydraulic conductivity shown Average daily humidity (%) 60
above remains valid. Canopy resistance (s m  1) 120
Soil aerodyn. res. (s m  1) 50
Vegetation aerodyn. res. (s m  1) 20
Leaf area index (m2 m  2) 13
4. Application of a hydrology–stability model with Canopy cover (%) 36
K generated by the Self-Consistent Method Maximum transpiration (m s  1) 3  10  7
Root tensile strength (kN m  2) 30
There is increasing interest in being able to assess Root area ratio (%) 25
Additional surcharge (kN m  2) 5
accurately the complex process interactions that exist
276 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279

(1959) procedure. Pore pressures, both negative and

positive, are incorporated directly into the effective
stress determination of the Mohr – Coulomb equation
for soil shear strength. This allows derivation of the
minimum factor of safety (FOS) (using a Janbu non-
circular automated slip search routine) (Wilkinson et
al., 2000), with temporal variations arising from
hydrodynamic responses and changes in the position
of the critical slip surface. To illustrate the importance
of the choice of the soil hydraulic conductivity,
application is made to a hypothetical slope configu-
ration with assumed geometrical, hydrological and
Fig. 5. Three cases of geotechnical characterisation of the
‘‘colluvium’’ soil: (a) homogenous soil, (b) differentiation of soil
vegetation parameters as summarized in Table 1. A
into four sublayers and (c) four sublayers of soil with continuous 1:5 year, 24-h return period storm for Hong Kong
characteristic variation. (Hong Kong Government, 1982) was imposed on the
top boundary (12.87 mm/h) and the modeled slope
For this illustrative application, we used a numer- was assumed covered with trees.
ical model that incorporates surface cover and slope In this application, the 30j colluvium slope is
topography within a fully dynamically coupled considered with three cases of geotechnical character-
hydrology – slope stability scheme (CHASM). The isation (Fig. 5 and Table 2): (a) a homogenous soil,
model has been described elsewhere (Anderson et mirroring the standard approach that would normally
al., 1996; Collison and Anderson, 1996). be taken using the field measured K of 5.09  10  6 m
The hydrological system is a forward explicit finite s  1 (Hong Kong Government, 1982), (b) a soil
difference scheme in which the slope is divided into a differentiated into four sublayers and (c) a soil with
series of rectangular columns, each subdivided into continuous geotechnical characteristic variation.
regular cells. The model simulates detention storage, Fig. 6 shows the results of the model application
infiltration, evapotranspiration and unsaturated and for these three representations. Fig. 6a– c shows the
saturated flow regimes. Rainfall is allowed to infiltrate soil moisture results at the time of failure representa-
the top cells governed by the infiltration capacity and tions of the colluvium as defined respectively in a– c
unsaturated vertical flow through each column is of Fig. 5. Additionally, Fig. 6d shows the dynamic
computed using the Richards’ equation (Richards, factor of safety plot for all three cases (the irregular
1931) solved in explicit form, with the unsaturated nature of the plots is due to different noncircular slip
conductivity defined by the Millington and Quirk surfaces being identified as critical at different time

Table 2
Geotechnical parameters for the test slope and the three geotechnical characterizations (a – c)
Characteristics (a) (b) (c)
Layer 0 Layer 0 Layer 1 Layer 2 Layer 3 Layer 0 Layer 1 Layer 2 Layer 3
Saturated bulk 19.517 19.27 19.26 20 19.59 189.278 19.436 19.595 19.832
density (kN m  2)
Unsaturated bulk 18.233 17.74 18.56 19.34 18.78 17.149 18.149 18.453 18.91
density (kN m  2)
Effective cohesion 4.4 4.4 4.4 4.4 4.4 4.4 4.4 4.4 4.4
(kN m  2)
Effective angle of 34.3 34.3 34.3 34.3 34.3 34.3 34.3 34.3 34.3
internal friction (j)
Saturated hydraulic 5.09E  06 5.04E  06 4.7E  06 4.92E  06 9.5E  06 4.24E  06 4.43E  06 4.75E  06 5.71E  06
conductivity
(m s  1)
 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279
Fig. 6. (a – c) Soil moisture content at the time of failure using hydraulic conductivity representations of the colluvium as defined in Fig. 5a – c, respectively. (d) Dynamic factor of
safety plot for all three cases.

277
278 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279

periods). The results show differences in moisture advanced and developed in this paper, we believe,
content, factor of safety and detailed slip surface has important implications for the collection of appro-
location, all of which are directly attributable to the priate slope stability model parameterisation data.
differing methods adopted for representing hydraulic Many geomorphological investigations in this field
conductivity. In the slip critical zone of the slope, utilise slope stability as a base model from which
using the homogeneous representation of K (Fig. 5) temporal extensions for investigations relating to
results in generally higher moisture content (Fig. 6a) land-use change and climate change are then required
than when using the SCM (Fig. 6b and c). The (e.g., Wu and Thornes, 1996; Van Beek, 2002), whilst
homogeneous representation thus results in the lowest other geomorphological extensions relate to process
factor of safety (Fig. 6d) in this particular application. interactions between vegetation and stability over
The range of the minimum factor of safety from 1.2 recent times (e.g., Sidle, 1992). In all of these cases,
(homogeneous soil) to 1.27 (soil with four sublayers) it is of importance to achieve the highest level of
is an illustration that the characterisation methods do confidence in the parameterisation of the soil hydraulic
have an impact on inferences attributable to stability conductivity. The proposed methodology outlined
analysis procedures. here, we believe, assists in achieving this objective.
Moreover, it does so in an analytical form from data
(PSD) that are generally more widely available and
5. Conclusions more precisely determined than the direct determina-
tion of hydraulic conductivity.
This paper has commented on the limitations gen-
erally associated with the parameterisation of hydraulic
conductivity for hydrology –stability models. We have References
outlined a new methodology based upon the SCM. The
key advantage of this method is that it delivers ana- Al Tabbaa, A., Wood, D.M., 1987. Some measurements of the
lytical expressions of the effective coefficient of a permeability of kaolin. Geotechnique 37 (4), 499 – 503.
Amer, A.M., Awad, A.A., 1974. Permeability of cohesionless soils.
disordered medium (the hydraulic conductivity in our J. Geotech. Eng. Div. 1000 (GT12), 1309 – 1316.
case) based on the analysis of simplified configura- Anderson, M.G., 1990. A feasibility study in mathematical modelling
tions. The topological description of the media for these of slope hydrology and stability. Geotechnical Control Office
derivations relate to the parameters of percentage sand, Civil Engineering Services Department, Hong Kong, Report
CE 23/90.
clay and water. Such PSD data have two key advan-
Anderson, M.G., Howes, S., 1985. Development and application of
tages over hydraulic conductivity—they are more reli- a combined soil water-slope stability model. Q. J. Eng. Geol. 18,
ably determined and more readily available at a given 225 – 236.
site. The increased spatial availability of PSD means Anderson, W.F., Pyrah, I.G., Fryer, S.J., 1991. A clay calibration
that within-soil type variability of hydraulic conductiv- chamber for testing field devices. Geotech. Test. J. 14 (4),
ity can now be determined by the SCM (Fig. 5). It is to 440 – 450.
Anderson, M.G., Lloyd, D.M., Othman, A., 1994. Using a com-
be expected that the significance of this improved bined slope hydrology/slope stability model for cut slope design
model parameterisation may be especially significant in the Tropics. Malays. J. Trop. Geogr. 25, 1 – 10.
in the case of residual soil slopes experiencing highly Anderson, M.G., Collison, A.J.C., Hartshorne, J., Lloyd, D.M.,
dynamic rainfall regimes (illustrated by the sample Park, A., 1996. Developments in slope hydrology – stability
application depicted in Fig. 6). In particular, our results modelling for tropical slopes. In: Anderson, M.G., Brooks,
S.M. (Eds.), Advances in Hillslope Processes, vol. 2. Wiley,
have shown that there can be significant differences in Chichester, pp. 799 – 821.
FOS resulting from different characteristics using PSD. Auriault, J.L., 1991. Heterogeneous medium. Is an equivalent
Fig. 6d shows, for example, that if a homogeneous macroscopic description possible? Int. J. Eng. Sci. 29 (7),
layering is assumed, a lower minimum FOS results 785 – 795.
compared to a circumstance that capitalizes on the Boutin, C., 1996. Conductivité thermique du béton cellulaire auto-
clavé. Modélisation par méthode autocohérente. Mater. Struct.
variation of PSD with depth. Whilst this study must 29, 609 – 615, RILEM.
be viewed as a prototype in terms of illustrating the Boutin, C., 2000. Study of permeability by periodic and self-con-
effects from a single application, the methodology sistent homogenisation. Eur. J. Mech. A, Solids 19, 603 – 632.
 I. Ibraim, M.G. Anderson / Geomorphology 49 (2002) 269–279 279

Boutin, C., Ibraim, I., 2001. On the correspondences between peri- McManus, K.J., Kulhawy, F.H., 1991. A cohesive soil for large size
odic and self-consistent homogenisation methods. The case of laboratory deposits. Geotech. Test. J. 14, 26 – 34.
static and dynamic permeability. First Bilateral TUL/ENTPE Millington, R.J., Quirk, J.P., 1959. Permeability of porous media.
Workshop, Szklarska Poreba, Poland. Technical University of Nature 183, 387 – 388.
Lodź, Lodź, pp. 151 – 177. Mitchell, J.K., 1993. Fundamentals of Soil Behaviour, 2nd ed. Wi-
Carter, M., Bentley, S.P., 1991. Correlations of Soil Properties. Pen- ley, New York, 437 pp.
tech Press, London, 130 pp. Olsen, H.W., 1966. Darcy’s law in saturated kaolinite. Water Re-
Chapuis, R.P., 1990. Sand-bentonite liners: predicting permeability sour. Res. 2 (2), 287 – 295.
from laboratory test. Can. Geotech. J. 27, 47 – 57. Richards, L.A., 1931. Capillary conduction of liquids in porous
Collison, A.J.C., Anderson, M.G., 1996. Using a combined slope mediums. Physics 1, 318 – 333.
hydrology/slope stability model to identify suitable conditions Schuster, P., 1987. The proprieties of sealing materials consisting of
for landslide prevention by vegetation in the humid tropics. silty sands mixed with clay exhibiting swell-capacity. 9th Euro-
Earth Surf. Processes Landf. 21, 737 – 747. pean Conference on Soil Mechanics and Foundation Engineer-
Collison, A.J.C., Anderson, M.G., Lloyd, D.M., 1995. Impact of ing, Dublin. Int. Soc. Soil Mech. and Geotechn. Engr., London,
vegetation on slope stability in a humid tropical environment: a pp. 927 – 935.
modeling approach. Proc. Inst. Civ. Eng., Water Marit. Energy Shackelford, C.D., 1994. Waste – soil interactions that alter hydrau-
112, 168 – 175. lic conductivity. Hydraulic Conductivity and Waste Contami-
Daniel, D.E., 1984. Predicting hydraulic conductivity of clay liners. nant Transport in Soil. ASTM Spec. Tech. Publ., vol. 1142,
J. Geotech. Eng. 110 (20), 285 – 300. Am. Soc. for Testing and Mater., Pennsylvania, pp. 111 – 168.
Elsbury, B.R., Daniel, D.E., Sraders, G.A., Anderson, D.C., 1990. Sherard, J.L., 1984. Hydraulic fracturing in embankment dams. J.
Lessons learned from compacted clay liner. J. Geotech. Eng. Geotech. Eng. 112 (10), 905 – 927.
116 (11), 1641 – 1660. Sidle, R.C., 1992. A theoretical model of the effects of timber har-
Hashin, Z., 1968. Assessment of self-consistent scheme approxima- vesting on slope stability. Water Resour. Res. 28, 1897 – 1910.
tion: conductivity of particulate composites. J. Compos. Mater. Skempton, A.W., Brogan, J.M., 1994. Experiments on piping in
2, 284 – 304. sandy gravels. Geotechnique 44 (3), 449 – 460.
Haug, M.D., Wong, L.C., 1992. Impact of modelling water content Tavenas, F., Jean, P., Leblond, P., Leroueil, S., 1983. The perme-
on hydraulic conductivity of compacted sand-bentonite. Can. ability of natural soft clay: part II. Permeability characteristics.
Geotech. J. 29, 253 – 262. Can. Geotech. J. 20 (4), 645 – 660.
Hong Kong Government, 1982. Mid-Levels Study Geotechnical Tedd, P., Charlles, J.A., Boden, J.B., 1987. Internal seepage erosion
Control Office, Hong Kong, 245 pp. in old embankment dams. 9th European Conference on Soil
Ibraim, I., Boutin, C., Arnaud, L., 2002. Morphology and perme- Mechanics and Foundation Engineering, Dublin 1. Int. Soc. Soil
ability assessment of heterogeneous soils. 2nd Biot Conference Mech. and Geotechn. Engr., London, pp. 507 – 510.
on Poromechanics, Grenoble, France, in press. Van Beek, R., 2002. Assessment of the Influence of Changes in
Jacquin, C., 1965. Etude des écoulements et des équilibres de flu- Land Use and Climate on Landslide Activity in a Mediterranean
ides dans les sables argileux. Rev. Inst. Fr. Pet. XX (4), 1 – 49. Environment. Netherlands Geographical Studies 294, University
Kenney, T.C., Lau, D., Ofoegbu, G.I., 1984. Permeability of com- of Utrecht, 363 pp.
pacted granular materials. Can. Geotech. J. 21, 726 – 729. Van Zyl, D., Harr, M.E., 1981. Seepage erosion analyses of struc-
Kenney, T.C., Van Veen, W.A., Swallow, M.A., Sungaila, M.A., tures. Proc. 10th Int. Conf. on Soil Mechanics and Foundation
1992. Hydraulic conductivity of compacted bentonite-sand mix- Engineering, Stockholm 1. Int. Soc. Soil Mech. and Geotechn.
tures. Can. Geotech. J. 29, 364 – 374. Engr., London, pp. 503 – 509.
Kröner, E., 1978. Self-consistent scheme and graded disorder in Wilkinson, P., Brooks, S.M., Anderson, M.G., 2000. Design and
polycrystal elasticity. J. Phys. F, Met. Phys. 8, 2261 – 2267. application of an automated non-circular slip surface search
Law, K.T., Lee, C.F., 1981. Initial gradient in a dense glacial till. within a Combined Hydrology and Stability Model (CHASM).
Proc. 10th Int. Conf. on Soil Mechanics and Foundation Engi- Hydrol. Process. 14, 2003 – 2017.
neering, Stockholm 1. Int. Soc. Soil Mech. and Geotechn. Engr., Wu, K., Thornes, J.T., 1996. Landslides and stability in the Nepal-
London, pp. 441 – 446. ese middle hills under seasonal and agricultural land-use change
Loudon, A.G., 1952. The computation of permeability from simple scenarios. In: Anderson, M.G., Brooks, S.M. (Eds.), Advances
soil test. Geotechnique 3, 165 – 182. in Hillslope Processes, vol. 2. Wiley, Chichester, pp. 773 – 797.