873

Seismic liquefaction, lateral spreading, and flow

slides: a numerical investigation into void
redistribution
Mahmood Seid-Karbasi and Peter M. Byrne

Abstract: Experience from past earthquakes indicates that seismically induced large lateral spreads and flow slides in
alluvial sand deposits have taken place in coastal and river areas in many parts of the world. The ground slope in these
slides was often not very steep, gentler than a few percent. Recent research indicates that the presence of low-permeability
silt or clay sublayers within the sand deposits is responsible for this behaviour. Such layers form a barrier to upward
flow of water associated with earthquake-generated pore pressures. This causes an accumulation of pore water at the
base of the layers, resulting in greatly reduced strength and possible slope instability. This paper uses an effective
stress coupled stress-flow dynamic analyses procedure to demonstrate the effects of a low-permeability barrier layer on
ground deformations from an earthquake event. The analyses show that an expansion zone develops at the base of barrier
layers in stratified soil deposits under seismic loading which can greatly reduce shear strength and result in large
deformations and flow failure. Without such a layer or layers, the slope may undergo significant displacements, but not
a flow slide. Slopes with a barrier layer can be stabilized by drains.
Key words: liquefaction, lateral spreads, stratification, flow failure, dynamic analysis, UBCSAND model, drain.
Résumé : L’expérience antérieure des tremblements de terre indique que de grands étalements latéraux et des coulées
dans les dépôts de sable alluvionnaires se produisent dans des régions côtières et de rivières dans plusieurs parties du
monde. La pente du terrain dans ces glissements n’est souvent pas très abrupte, plus douce que quelques pour cent. La
recherche récente indique que la présence de sous-couches de silt ou d’argile de faible perméabilité dans les dépôts de
sable est responsable de ce comportement. De tels couches forment une barrière pour l’écoulement de l’eau vers le
haut associé avec les pressions interstitielles générées par le tremblement de terre. Ceci cause une accumulation de pression
interstitielle à la base des couches, et résulte en une résistance grandement diminuée et une instabilité potentielle des
pentes. Cet article utilise une procédure d’analyses en contrainte effective couplée à la dynamique contrainte-écoulement
pour montrer les effets d’une couche-barrière de faible perméabilité sur la déformation causée par un événement de
tremblement de terre. Les analyses montrent qu’une zone d’expansion se développe à la base des couches-barrières dans
les dépôts de sol stratifié sous un chargement séismique qui peut réduire grandement la résistance au cisaillement et
résulter en de grandes déformations et en une rupture par écoulement. Sans une telle couche ou de telles couches, la
pente peut subir des déplacements appréciables, mais pas un glissement par écoulement. Les pentes avec des couches-
barrières peuvent être stabilisées par drains.
Mots-clés : liquéfaction, étalements latéraux, stratification, rupture par écoulement, analyse dynamique, modèle
UBCSAND, drain.

[Traduit par la Rédaction] Seid-Karbasi and Byrne 890

Introduction (Kokusho 2003). Submarine slides have been seismically trig-

Experience from past earthquakes indicates that lateral
spreads and flow slides have taken place in liquefied ground in
coastal and river areas in many regions of the world, including
Alaska, Japan (Niigata), and Turkey. Movements may exceed
several metres, even in gentle slopes of less than a few percent
Received 3 November 2004. Accepted 29 January 2007.
Published on the NRC Research Press Web site at cgj.nrc.ca
on 14 August 2007.
M. Seid-Karbasi1,2 and P.M. Byrne. Department of Civil
Engineering, University of British Columbia, 6250 Applied
Science Lane, Vancouver, BC V6T 1Z4, Canada.
1
Corresponding author (e-mail: mseidkarbasi@golder.com).
2
Present address: Golder Associates Ltd., 500-4260 Still
Creek Drive, Burnaby, BC V5C 6C6, Canada.
Can. Geotech. J. 44: 873–890 (2007) doi:10.1139/T07-027
gered in many regions as reported by Scott and Zuckerman
(1972) and Hamada (1992). More interestingly, lateral spreads
or flow slides have occurred not only during earthquake shak-
ing, but also after earthquake shaking. These large movements
are mainly driven by gravity, although they are initially trig-
gered to liquefy by seismic stresses.
In the Hakusan District in Niigata City, Japan, an area
250 m × 150 m bordering the Shinano River moved toward the
river during the 1964 Niigata earthquake (Kawakami and
Asada 1966). Girders of the Showa Ohashi Bridge fell down
due to flow failure of the liquefied riverbed, which moved the
supporting pile foundations. Eyewitnesses reported that the
girders began to fall a few minutes after the earthquake motion
had ceased (Hamada 1992). At the lower San Fernando dam, a
hydraulic fill in California, flow failure due to liquefaction in
the upstream slope occurred about a minute after the strong
© 2007 NRC Canada
874
Fig. 1. Cross section of a sand volcano formed on the surface of
the disposal pond (modified from Ishihara 1984).
shaking in the 1971 earthquake (Seed 1987). Mochikoshi tail-
ings dams in Japan failed as a result of the 1978 Izu-Ohshim-
Kinkai earthquake due to liquefaction-induced flow slides,
causing release of the tailings (Ishihara 1984). One dam failed
during the shaking, and a second dam failed 24 h later. Figure 1
shows a sand boil formed in the Mochikoshi disposal pond
comprising sandy silt and silt layers. More recently, Harder and
Stewart (1996) reported that in the 1994 Northridge, California,
earthquake, liquefaction-induced flow failure of the 24 m high
Tapo Canyon tailings dam occurred about 10 min after the
main shock. Yoshida et al. (2005) reported large lateral
displacements in the order of several metres that occurred in
near-level ground conditions in the 1983 Nihonkai–Chubu
earthquake, with the movements obtained by comparing aerial
photographs taken before and after the earthquake.
Flow failures in liquefied soil deposits have also occurred
in submarine slopes triggered by earthquakes. The slopes in
these slides were gentle, normally less than 5° and some-
times less than 1° (Hampton and Lee 1996). During the
1964 Alaska earthquake, the port cities of Valdez and
Seward suffered great loss of life and property from large-
scale submarine slides (Coulter and Migliaccio 1966; Lemke
1967). In the 1999 Kocaeli earthquake, Turkey, widespread
large lateral spreads are reported by Kokusho (2003). Coastal
areas along Izmit Bay suffered submarine slides triggered by
the earthquake. The 1995 Aegion earthquake of magnitude
(Ms) 6.2 caused intense ground failures in Eratini, Greece
(Bouckovalas et al. 1999). Geotechnical exploration revealed
that the soil profile at the failure sites is characterized by a
continuous interchange of silty sand and clay layers. The
ground inclination in these cases was less than 7°. During
the 1980 Mammoth lakes earthquake in California, a 2 km ×
20 km area of seafloor, consisting of sand and mud on a
slope of 0.25°, suffered a flow slide. There was evidence of
liquefaction in the form of sand boils on the seafloor as
described by Field et al. (1982).
Although lateral flow failures have been reported in past
earthquakes causing damage to structures (e.g., Berrill et al.
1997), the mechanism leading to large lateral displacements
is still poorly understood. Sand and silt deposits often com-
3
Effects of void redistribution on liquefaction-induced ground deformations in earthquakes: a numerical investigation. Ph.D. dissertation. In
preparation.
Can. Geotech. J. Vol. 44, 2007
prise many sublayers as a result of the sedimentation pro-
cess. A number of researchers have examined the effect of
layering on post-liquefaction sliding, including Scott and
Zuckerman (1972), Huishan and Taiping (1984), Liu and
Qiao (1984), Elgamal et al. (1989), Adalier and Elgamal
(1992), Fiegel and Kutter (1992), Kokusho (1999, 2000),
Kulasingam et al. (2001, 2004), Malvick et al. (2002, 2005,
2006), Yang and Elgamal (2002), Kulasingam (2003), Seid-
Karbasi and Byrne (2004a), Malvick (2005), Byrne et al.
(2006), and Seid-Karbasi3. Based on physical model tests
and site investigations, Kokusho (1999) and Kokusho and
Kojima (2002) concluded that failure can be caused by the
formation of a water film at the base of a sublayer leading to
a zone of essentially zero strength. The water-film effect is
associated with upward flow of water arising from liquefac-
tion. When such flow is impeded by a lower permeability
layer, it results in a void redistribution and an accumulation
of pore water that can lead to the formation of a water film
or a thin water-rich zone.
Kulasingam et al. (2004) using centrifuge testing demon-
strated that water films are an extreme case of void redistri-
bution but are not a requisite for large ground deformation to
occur. Under postseismic conditions, the void expansion
may cause the strength to drop significantly and become
smaller than the driving stress, thereby resulting in large
ground deformations. The permeability contrast to cause
such an effect can be associated with silt and (or) clay layers
sandwiched within sand and (or) gravel layers. Layering and
associated void ratio change during and or after an earth-
quake can explain why the steady-state strength approach
based on an undrained condition (Poulos et al. 1985) can
lead to significantly higher values of residual strength than
those estimated from back analysis of field case studies
(e.g., Seed and Harder 1990; Stark and Mesri 1992; Olson
and Stark 2002; among others). This has already been
pointed out by a number of researchers (Seed 1987; Castro
et al. 1989; Byrne and Beaty 1997; Kokusho 2003; Seid-
Karbasi and Byrne 2004a; Byrne et al. 2006).
Our understanding of the behaviour of liquefiable soils
has increased dramatically in the past 30 years because of
(i) observations from field case histories, (ii) extensive labo-
ratory testing of soil elements under monotonic and cyclic
loading, (iii) model testing of earth structures under simu-
lated earthquake loading, and (iv) development of numerical
modeling procedures.
Laboratory model testing suggests that slopes comprised
of clean sands and having relative densities, Dr, as loose as
20% are unlikely to suffer a flow slide (Kulasingam et al.
2004). These slopes can be readily triggered to liquefy and
may undergo large displacements, but their strengths are
generally adequate to prevent a flow slide. However, if the
sands contain low-permeability silt layers that impede drain-
age, expansion of the sand skeleton beneath the silt occurs
and can lead to the formation of a water film and complete
loss of strength and a flow slide.
Traditionally, drainage conditions in laboratory investiga-
tions to characterize liquefiable soils are considered as either
undrained or (fully) drained, but these conditions do not rep-
© 2007 NRC Canada
Seid-Karbasi and Byrne 875

resent the real situation for stratified deposits in the field be- Fig. 2. Partially drained instability of loose Fraser River sand
cause during and after shaking water migrates from zones (modified from Vaid and Eliadorani 1998): (a) stress paths;
with higher excess pressure towards zones with lower excess ′ , minor
(b) strain paths. Drc, relative density at consolidation; σ3c
pressure. principal stress at consolidation.
In this paper a dynamic coupled stress-flow analysis pro-
cedure is utilized to study the mechanisms involved in large
deformations leading to lateral spreads and flow failures in
gentle slopes in liquefiable ground. Based on such analyses,
implications for design of mitigation methods to resist seismic
loading are examined.

Sand liquefaction and failure

Seismic liquefaction refers to a sudden loss in stiffness
and strength of soil due to cyclic loading effects of an earth-
quake. The loss arises from a tendency for soil to contract
under cyclic loading, and if such contraction is prevented or
curtailed by the presence of water in the pores that cannot
escape, it leads to a rise in pore-water pressure and a resulting
drop in effective stress (Silver and Seed 1971; Martin et al.
1975). If the effective stress drops to zero (100% pore-water
pressure rise), the strength and stiffness also drop to zero
and the soil behaves as a heavy liquid. Unless the soil is
very loose (e.g., Dr ≤ 20%), however, it will dilate and regain
some stiffness and strength as it strains. If this strength is
sufficient, it will prevent a flow slide from occurring but
may still result in excessive displacements commonly re-
ferred to as lateral spreading.
The excess pore water generated by seismic loading gen-
erally drains upwards, and if no barrier layers are present,
dissipation results in a decrease in void ratio and an increase
in strength and stability. The potential for lateral spreading
and flow slides can greatly increase if low-permeability layers,
e.g., silt layers within a soil deposit, impede drainage, form-
ing a barrier to flow that can result in an expansion of the
sand skeleton (void redistribution) and accumulation of pore
water at the base of the layers.
The majority of the laboratory element tests conducted to
assess liquefaction resistance of sands have been carried out
under undrained conditions. Pore-pressure rise in these ele-
ments occurs under applied shear stress, causing a reduction
in effective stresses and strength. In recent experimental
studies, Vaid and Eliadorani (1998) and Eliadorani (2000) dem-
onstrated that a small net flow of water into an element
(injection) causing it to expand results in additional pore-
pressure generation and further reduction in strength. They
examined this phenomenon by injecting or removing small
volumes of water from the sample as it was being sheared
and referred to this as a partially drained condition.
The results of inflow tests on Fraser River sand indicate a
potential for triggering liquefaction at constant shear stress.
Small imposed expansive volumetric strains resulted in effec-
tive stress reduction and flow failure of samples of sand con-
solidated to an initial state of stress corresponding to
Rc = σ1c′ / σ′3c = 2, as shown in Fig. 2a, where Rc is the
consolidation effective stress ratio, and σ′1c and σ′3c are the
major and minor principle effective stresses, respectively. Vaid
and Eliadorani (1998) defined this condition as instability that
occurs when a soil element subjected to small effective stress
perturbation cannot sustain the current stress state and results in
runaway deformation as seen in Fig. 2b, or liquefaction flow.
In these tests, expansive volumetric strains (ε v ) were
imposed by injection of water into the samples at a constant
rate dε v / dε1 = − 0.4, where ε1 is the axial strain. The samples
were stable under the initial stress state. The stress paths fol-
lowed during injection indicate a reduction in effective
stresses at a constant shear stress. At each initial confining
stress, the stress changes leading to instability occurred with
little change in shear stress and void ratio and at very small
axial strains of the order of 0.1%. Positive pore pressures con-
tinued to develop even beyond the phase-transformation line.
This occurs because the rate of imposed expansive volumetric
strain is greater than the tendency for the skeleton to expand.
Chu (1991), Bobei and Lo (2003), Sento et al. (2004), and
Yoshimine et al. (2006) have reported similar results for sands
and silty sand materials. As a result, soil elements can liquefy
due to expansive volumetric strains that cannot be predicted
from analyses based on the results of undrained tests.
To investigate the effects of low-permeability layers on
ground deformations due to earthquakes, it is necessary to
predict the generation, redistribution, and dissipation of
excess pore pressures during and after earthquake shaking.
A fundamental approach requires a dynamic coupled stress-
flow analysis. In such an analysis, the volumetric strains are
controlled by the compressibility of the pore fluid and flow
of water through the soil elements. To predict the instability
and liquefaction flow, an effective stress approach based on
an elastic–plastic constitutive model (UBCSAND) was used.
The model is calibrated against laboratory element data and
centrifuge data and is described in the next section.

© 2007 NRC Canada

876 Can. Geotech. J. Vol. 44, 2007

Stress–strain model for sand Fig. 3. Moving of yield loci with stress ratio and plastic strain
increment vectors. φcv , friction angle at constant volume; εpv,
The UBCSAND constitutive model is based on the elastic– plastic volumetric strain.
plastic stress–strain model proposed by Byrne et al. (1995) and
has been further developed by Beaty and Byrne (1998) and
Puebla (1999). The model has been successfully used in ana-
lyzing the CANLEX liquefaction embankments (Puebla et al.
1997) and predicting the failure of the Mochikoshi tailings dam
(Seid-Karbasi and Byrne 2004b). It has been used to examine
dynamic centrifuge test data (e.g., Byrne et al. 2004; Seid-
Karbasi et al. 2005) and also partially saturation effects (Seid-
Karbasi and Byrne 2006). It is an incremental elastic–plastic
model in which the yield loci are lines of constant stress ratio
(η = τ / σ′, where η is the shear stress ratio, τ is the shear stress,
and σ′ is the effective normal stress on the plane of maximum
shear stress). The flow rule relating the plastic strain increment
directions is nonassociated (see Fig. 3) and leads to a plastic
potential defined in terms of the dilation angle.
Fig. 4. Plastic shear strain increment and shear modulus.
Elastic properties
The elastic component of response is assumed to be iso-
tropic and specified by a shear modulus, G e , and a bulk
modulus, B e , as follows:
ne
 σ′ 
[1] G e
= K Ge Pa  

 Pa 

[2] Be = α G e

where K Ge is an elastic shear modulus number, Pa is the

atmospheric pressure, σ′ = ( σx′ + σ′y ) / 2, n e is an elastic expo-
nent approximately equal to 0.5, and α ( = 2 (1 + ν) / 3 (1 − 2ν))
depends on Poisson’s ratio, ν, and ranges from 2/3 to 4/3.

Plastic properties
The plastic shear strain increment d γ p is related to the
stress ratio, dη, where η = τ / σ′, as shown in Fig. 4 and can
be expressed as follows:
whereas no compaction is predicted at stress ratios corre-
dη sponding to ϕ cv (plastic potential vector is vertical). For
[3] dγ p =
G p / σ′ stress ratios greater than ϕ cv , shear-induced plastic expan-
sion or dilation is predicted (where plastic potential vectors
where G p is the plastic shear modulus and is given by a slope to the left).
hyperbolic function as follows: This simple flow rule is in close agreement with the char-
2
acteristic behaviour of sand observed in drained laboratory
 η   element testing. The response of sand is controlled by the
= G ip 1 −  Rf 
p
[4] G  skeleton behaviour outlined previously. The presence of a
 η f   fluid (air–water mix) in the pores of the sand acts as a volu-
where G ip is the plastic shear modulus at a low stress ratio metric constraint on the skeleton if drainage is fully or par-
level (η = 0); η f is the stress ratio at failure and equals sin tially curtailed. It is this constraint that causes the pore
ϕ f , where ϕ f is the peak friction angle; and Rf is the failure pressure rise that can lead to liquefaction. Provided the skel-
ratio. G ip in turn is related to G e and the relative density of eton or drained behaviour is appropriately modeled under
the sand. The associated increment of plastic volumetric monotonic and cyclic loading conditions and the stiffness of
strain, dε pv , is related to the increment of plastic shear strain, the pore fluid and drainage are accounted for, the liquefac-
d γ p , through the flow rule as follows: tion response can be predicted.
This model was incorporated in the commercially avail-
[5] dε v = dγ p (sin ϕ cv − η ) able computer code FLAC (fast lagrangian analysis of con-
tinua), version 4.0 (Itasca Consulting Group Inc. 2005). This
where ϕ cv is the friction angle at constant volume or phase program models the soil mass as a collection of grid zones
transformation. It can be seen from Fig. 3 that significant or elements and solves the coupled stress flow problem using
shear-induced plastic compaction is occurring (where plastic an explicit time stepping approach. The program has a number
potential vectors slope to the right) at low stress ratios, of built-in stress–strain models including an elastic–plastic
© 2007 NRC Canada
Seid-Karbasi and Byrne
Fig. 5. Comparison of predicted and measured response for Fraser
River sand: (a) stress–strain; (b) excess pore pressure ratio, Ru,
versus number of cycles; (c) CSR versus number of cycles (tests
data from Sriskandakumar 2004).
Mohr–Coulomb model, and UBCSAND is a variation of this
model in which friction and dilation angles are varied to
incorporate the yield loci and flow rule described previ-
ously. Pore fluid stiffness and Darcy hydraulic flow are ba-
sic to the FLAC program, so only the skeleton stress–strain
relation is needed to simulate liquefaction, and that is what
the UBCSAND model does. Drainage conditions are built
into FLAC, and drained, undrained, or coupled stress flow
conditions are specified by the user.
The key elastic and plastic parameters can be expressed in
terms of relative density, Dr , or normalized standard pene-
tration test values, ( N 1 ) 60 . Initial estimates of these parame-
ters have been approximated from published data and model
calibrations. The response of sand elements under mono-
tonic and cyclic loading can then be predicted and the results
compared with laboratory data. In this way, the model can
be made to match the observed response over the range of
relative density or (N 1)60 values. The model has also been
877
Fig. 6. Predictions of element response in undrained and partially
drained (inflow) triaxial tests for Fraser River sand: (a) stress–
strain; (b) volumetric strains; (c) stress paths (modified from
Atigh and Byrne 2004).
calibrated to reproduce the National Center for Earthquake
Engineering Research (NCEER) 1997 (Youd et al. 2001)
triggering chart, which in turn is based on field experience
during past earthquakes and is expressed in terms of the
standard penetration test resistance value, ( N 1 ) 60 . The model
properties to obtain such an agreement are therefore ex-
pressed in terms of (N1)60.
Model simulation of laboratory element tests
The model was applied to simulate cyclic simple shear
tests under undrained conditions. Figure 5 shows model pre-
dictions along with test results on Fraser River sand. The
test had an initial vertical consolidation stress σ′v = 100 kPa
and Dr = 40%. The results in terms of stress–strain and
excess pore pressure ratio, R u , compare reasonably well with
the laboratory data. A comparison of model predictions with
tests results in terms of required number of cycles to trigger
liquefaction for different cyclic stress ratios, CSR, is shown
in Fig. 5c and shows good agreement.
The model was also used to predict the effect of both the
undrained condition and partial drainage as observed in
triaxial monotonic tests. The partial drainage involved inject-
ing the sample with water to expand its volume as it was
sheared. The injection causes a drastic reduction in strength.
In the numerical model, the same volumetric expansion was
© 2007 NRC Canada
878 Can. Geotech. J. Vol. 44, 2007

Fig. 7. Ground conditions used in this study: (a) case I, uniform profile without low-permeability sublayer; (b) case II, profile with
low-permeability sublayer; (c) case III, profile treated with drain column.

(a) Groundwater table (b) Material: (c)

(m)
10 Silt barrier
Loose sand
9 Drain column

5
Element:
(1,13)
4 (1,10)
(1,5)
3 (1,3)

-1
Firm impervious ground
-1 0 1 2 3 4 (m)

Table 1. Properties of the materials used in the analyses.

Dry density, ρd Porosity, UBCSAND Permeability,
Material (×1000 kg/m3) n (N1)60 k (m/s)
Loose soil, Dr = 40% 1.50 0.448 6.2 8.81×10–4
Silt barrier 1.50 0.448 — 8.81×10–7

applied and the results are shown in Figs. 6a, 6b, and 6c in Fig. 8. Acceleration–time history for base input harmonic motion.
terms of stress–strain, volumetric strain versus mean stress,
and stress path, respectively (model prediction with the solid
line). The predictions are in remarkably good agreement
with the measured data.
The previous simulations illustrate that the model can gen-
erate the appropriate pore pressures and stress–strain response
to undrained loading and can account for the effect of volu-
metric expansion caused by inflow of water into an element.

Soil profile used in the analyses

The soil profile used in this study is a 10 m thick deposit
representing a submerged ground condition as shown in
Fig. 7a. It comprises a loose sand deposit resting on an
impermeable rigid foundation. The effect of a low-
permeability layer within the loose sand at a depth of 4 m is
examined (see Fig. 7b). Fraser River sand with a relative
density of Dr = 40% is considered to represent the loose
sand. Materials properties are listed in Table 1, in which ρ d ,
n, and k are the material dry density, porosity, and perme-
ability, respectively. The UBCSAND model was applied to
the loose sand layer with a corresponding equivalent
UBCSAND ( N 1 ) 60 value. The low-permeability silt layer
barrier is simulated with a Mohr–Coulomb model having a
friction angle ϕ = 30° and permeability k one thousand times
lower than that of the loose sand layer. Its stiffness in terms
of bulk modulus and shear modulus was modeled as 1.0 ×
104 and 0.5 × 104 kPa, respectively. It is not considered to
generate excess pore pressure.
Input base motion in terms of an acceleration–time history
is shown in Fig. 8. It is a harmonic (sinusoid) excitation
© 2007 NRC Canada
Seid-Karbasi and Byrne
Fig. 9. Excess pore pressure ratio Ru versus time at selected points (elements (1, 13), (1, 10), and (1, 3), see Fig. 7) with increasing
depth: (a) case I; (b) case II.
applied at the base of the soil layer. It ramps up to 2.5 m/s2
within 1 s and dies out in 2 s, lasting for 7 s in total.
Analyses were conducted for three cases: (I) sloping
ground without a low-permeability layer, (II) sloping ground
with a low-permeability layer, and (III) sloping ground with
a low-permeability layer treated with a drain column.
Analyses and results
To model the free-field condition, a mesh with 9 × 22
zones as illustrated in Fig. 7 was used. Material types are
recognized with different permeability values as shown in
the figure. The nodes on the left and right boundaries were
linked to force the soil column to deform as a shear beam.
The earthquake motion was applied as a time history of
acceleration at the base of the mesh.
Ground behaviour in cases I and II
A uniform sloping ground condition with 1° inclination
without a low-permeability layer was analyzed as a bench-
mark condition (case I), and then the same soil profile with a
low-permeability layer (case II) was analyzed. The barrier
879
layer was located at a depth of 4 m and had a permeability
1000 times lower than that of the loose sand. The predicted
behaviour for these cases in terms of excess pore pressure,
surface lateral displacement, and deformation pattern are
presented and compared in this section.
Figure 9 shows the time histories of excess pore pressure
ratio, Ru, for selected depths for the two cases (for positions
of the points refer to Fig. 7). The predicted patterns of
excess pore pressure in case I (see Fig. 9a) indicate that
essentially high excess pore pressures build up within the
soil profile during the strong shaking. Small dilation spikes
are predicted when R u ≈ 1 and are less pronounced in the
upper parts as a result of upward inflow. It may be seen that
excess pore pressures dissipate somewhat more rapidly at
depth, e.g., Ru at 10 s is 55% and 75% for elements (1, 3)
and (1, 13), respectively. A similar trend has been observed
in centrifuge tests conducted for level and sloping ground
(Taboada and Dobry 1993a, 1993b, 1998).
The corresponding Ru time histories for case II at different
depths are shown in Fig. 9b. Again, pore pressures increase
very rapidly; however, beneath the barrier layer they remain
high (Ru ≈ 100%) after the end of shaking, and dissipation
© 2007 NRC Canada
880
Fig. 10. Deformation pattern of soil profile: (a) without barrier, case I (with maximum lateral displacement of 0.95 m after 14 s);
(b) with barrier (darker area), case II (with maximum lateral displacement of 1.75 m after 30 s).
Fig. 11. Surface lateral displacement (X-dis) versus time for pro-
files with and without a barrier.
occurs at greater depths. This indicates that water flows
from the greater depths towards the layer beneath the bar-
rier, causing higher excess pore pressures to last for a signif-
icantly longer time compared with the case without a
barrier. The injected flow causes an expansion of the layer
beneath the barrier to occur at essentially zero effective
stress and leads to large deformation.
Figure 10 shows the predicted deformed meshes for the
two cases. As seen from Fig. 10a, distortion in case I is pro-
nounced at the base and tapers off towards the surface,
resulting in a maximum displacement of 0.95 m at the top
surface. This pattern compares well with dynamic centrifuge
data reported by Sharp et al. (2003), who modeled a 10 m
uniform liquefiable layer with a 5.2° inclination shaken with
Can. Geotech. J. Vol. 44, 2007
22 cycles of harmonic motion with peak ground acceleration
(PGA) = 0.23g at 2 Hz using viscose pore fluid. The soil used
in their tests was Nevada sand 120, deposited at Dr = 45%.
The deformation pattern after 30 s for case II is shown in
Fig. 10b. Comparing the pattern and magnitude of lateral
displacements for these cases (with and without a barrier
layer shown in Figs. 10a and 10b), it can be seen that the
magnitude has increased from 0.95 to 1.75 m and the pattern
is quite different, with a large slippage occurring at the base
of the barrier layer.
Time histories of horizontal displacement of the top sur-
face for the two cases (with and without a barrier layer) are
compared in Fig. 11, which indicates that for case I, without
the barrier, displacements occur during shaking and cease
shortly after the end of shaking. It can be seen that the sur-
face displacements are much larger when a barrier layer is
present. They are larger during shaking and continue to
increase after shaking ceases.
Figure 12 shows a profile of volumetric strain beneath
the barrier for case II and indicates that the lower two
thirds of the soil profile contracts while expansion occurs
in the upper one third of the profile, with the highest rate
of expansion directly beneath the barrier. Therefore, un-
drained conditions do not exist, especially locally within
the soil deposit beneath the barrier. This also suggests that
undisturbed samples taken prior to the earthquake will not
be representative of conditions during and shortly after the
earthquake due to void redistribution resulting from the up-
ward flow of water.
© 2007 NRC Canada
Seid-Karbasi and Byrne
Fig. 12. Profile of volumetric strain beneath the barrier layer
(after 30 s).
The increasing rate of expansion predicted in Fig. 12 as
the barrier is approached suggests that very high strains may
be occurring in the zone at or near the boundary. This is
referred to as strain localization.
Similar analyses were also carried out for the same soil
profile including a barrier layer with zero inclination (level
ground). In this case, no significant lateral displacements are
predicted because of zero static driving force or shear stress
bias for level ground conditions.
Flow failure
The results from the previous section revealed that delayed
large deformations in liquefiable gently sloping grounds can
occur when a sublayer with low permeability is present.
However, two questions arise in this regard. (1) In view of
the predicted localization in the element beneath the barrier,
to what extent is the predicted deformation pattern (i.e., con-
traction in the lower section and expansion in the upper sec-
tion) related to the model configuration (mesh size effects)?
(2) What are the requirements for a flow failure?
The rest of this section discusses the results of analyses
carried out to answer these questions.
The predicted strain localization implies that the com-
puted results can be mesh-size dependent, as also noted by
Yang and Elgamal (2002) and Uzuoka et al. (2003). To
investigate scale effects on the predicted characteristic
behaviour of liquefiable grounds with a sublayer barrier, a
separate series of analyses was conducted for the same soil
profile with a barrier but modeled in two different ways:
(i) as a 10 m soil profile in a 1g field (prototype scale), and
881
(ii) as a 0.1 m soil profile in a 100g field (model scale) fol-
lowing appropriate conversion laws for modeling described
by Schofield (1981) and Kutter (1995). It is noted that the
mesh size used in the latter way of modeling was one-
hundredth that used in the former.
The base input motion and mechanical properties of materials
are the same as those used previously. The displacement pat-
terns were identical for both models when examined in
prototype scale and are the same as those shown in Fig. 10b
(the profiles of volumetric strain beneath the barrier layer for
both models were also identical and the same as those
depicted in Fig. 12). This suggests that for a soil profile
comprising a barrier layer, contraction at the lower parts and
expansion at the upper parts are characteristic behaviours
that occur due to pore pressure migration and result in void
redistribution regardless of layer size. These findings indi-
cate that the phenomenon of pore-water redistribution can be
captured in centrifuge tests and that the actual physical size
of the layers is not important. What matters from a numerical
modeling point of view is the size of the mesh in relation to
the size of the liquefiable layer.
Further studies were undertaken to analyze the problem
under the same conditions but using meshes of decreased
size for the element beneath the barrier. The results showed
that a decrease in element size results in greater volumetric
expansion for the very first element at the base of the barrier
layer. Figure 13a shows maximum volumetric strain of the
element beneath the barrier versus the element thickness
normalized with respect to soil layer thickness beneath the
barrier. The figure suggests that void redistribution leads to a
very thin, water-rich zone at the base of the barrier. If
enough water flows into the soil element, it can expand until
the steady or critical state at zero effective stress is reached.
This corresponds approximately to the maximum void ratio
state. At this state the skeleton can undergo no further
expansion, and this is simulated by setting the dilation angle
to zero. Additional inflow will result in the formation of a
water film at the interface and zero shear strength.
The predicted increasing trend in volumetric strain of the
barrier base element with a decrease in its size (shown in
Fig. 13a) suggests that volumetric strain of the base element
becomes infinity as the size of the base element approaches
zero. This implies that for a liquefiable layer with a sublayer
barrier, flow failure occurs regardless of other involved
factors, i.e., liquefied soil layer thickness, density, shaking
level and duration, and ground slope.
Figure 13b shows time histories of volumetric strain for
the base element with various element thickness ratios
(ETRs) for the analyzed sloping ground using the
UBCSAND model. It can be seen that the amount of volu-
metric expansion in the base element increases with time
and with a decrease in element thickness. More detailed
analyses indicate that the volumetric expansion does not go
to infinity at the boundary. Further analyses to examine flow
and expansion issues were carried out and are presented as
follows.
Figure 14 shows the analyses results in terms of the iso-
chrones of Y-Flow, the vertical specific discharge velocity
(flow rate per unit area) at different time intervals for the
mesh shown in Fig. 7b using the UBCSAND model. The
© 2007 NRC Canada
882
Fig. 13. Volumetric strain of the base element with various thicknesses: (a) maximum volumetric strain versus normalized element
thickness; (b) time histories of volumetric strain.
figure indicates that water flows upward from the lower
parts that have the highest excess pore pressures towards the
upper low-permeability layer where the excess pore pressure
is lowest. Maximum flow occurs near the mid-height of the
soil layer with a descending trend as it approaches the
boundaries both top and bottom. Flows initially increase
quickly with time and then drop off as pore pressures dissi-
pate. Figure 15 shows the time history of volumetric strain
rate for the base element of the mesh shown in Fig. 10b. It
can be seen that the rate of expansion tends to essentially
zero some time after shaking ceases. Similar results were
obtained for the meshes with smaller base zone thicknesses.
Figure 16 shows the maximum expansion rate versus nor-
malized zone thickness. It clearly shows no indication of an
infinite value for the expansion rate as the grid size
approaches zero. The results indicate that, although the
expansion increases with a reduction in element thickness,
its maximum value is finite.
From a practical point of view, localized flow failure occurs
in a thin zone of soil when a shear band of limited thickness
Can. Geotech. J. Vol. 44, 2007
forms within the soil mass. In the next section an approach
for numerical modeling of flow failures is presented.
Approach to modeling localized flow failure
The effect of mesh size can be approximately accounted
for in the analysis by specifying a dilation cut-off on skeleton
expansion that is related to the initial relative density and
mesh size. Denser material would have more dilation capacity
to reach its maximum void ratio, at which point dilation
would be set to zero. For a coarser mesh, less expansion is
predicted for the element and the amount of expansion
required to trigger dilation cut-off would be less.
The mesh size should also be related to the particle size.
Roscoe (1970) noted that a shear band may occur within a
thin zone with a thickness 10–20 times that of the mean
grain size of the soil, D50 . Similar finding have been reported
from further experimental investigations (e.g., Alshibli and
Sture 1999; Desrues and Viggiani 2004; Lu et al. 2004) and
numerical investigations (e.g., district element method,
DEM) by a number of researchers. A reasonable lower limit
© 2007 NRC Canada
Seid-Karbasi and Byrne
Fig. 14. Vertical specific discharge (Y-Flow) isochrones at differ-
ent times.
Fig. 15. Time history of volumetric strain rate of base element.
on mesh size could be based on the size of the shear band.
For a soil profile comprising Fraser River sand of Dr = 40%
with D50 = 0.3 mm, a shear band could form within a zone
3–6 mm in height. In Fig. 17 the data shown in Fig. 13a are
plotted in semilogarithmic format with the initial portion
extrapolated to reach a value corresponding to an element
thickness of 3 mm (ETR = 3/6000 = 0.0005). The maximum
volumetric strain predicted for that ETR is about 5.2%. If
the shear band zone can expand by 5.2% before reaching its
critical state, then it can retain some strength and prevent a
flow slide. A key question then is as follows: What is the
maximum volumetric expansion (threshold strain) that an
element can tolerate before a water film forms?
Potential maximum volumetric (threshold) strain can be
estimated based on data from various soils available in the
literature. Table 2 lists a number of soils with their minimum
and maximum void ratios compiled by Cubrinovski and Ishihara
883
Fig. 16. Maximum expansion rate versus base element thickness
ratio. Isochrones at different time intervals for UBCSAND model
with 0.5 m element height.
(2002) and Olson and Stark (2003). The threshold volumetric
strain, ( ε v ) T , can be calculated as follows:
e max − e min
[6] ( ε v ) T = Dr
1 + ei
where Dr is the soil relative density, e max is the maximum
void ratio, e min is the minimum void ratio, and e i is the initial
void ratio. The numerator of eq. [6] should be e ss − e min
(where e ss is the steady-state void ratio at zero effective
stress), but it can be approximated by e max − e min for the sake
of simplicity. A similar approach was adopted by Tsukamoto
et al. (2004) and Sawada et al. (2006) to estimate potential
maximum liquefaction-induced settlement. In Table 2, the
potential maximum expansion is also provided for a typical
medium-dense state (Dr = 50%). Figure 18 shows the upper
and lower bounds for potential maximum expansion for a
complete range of initial Dr calculated based on eq. [6]. A
range of relative densities from 25% to 80% is of interest
from a practical point of view. For Fraser River sand at Dr
of 40%, (ε v )T = 6.9%.
For the case involved in Fig. 17 with Fraser River sand,
the predicted maximum volumetric strain in the shear band
zone was 5.2%. Since this is less than the threshold value,
i.e., ( ε v ) T = 6.9%, a flow slide is not predicted. If the expan-
sion limit is reached, the soil element deforms in a steady-
state condition with essentially zero shear strength. Not all
elements along a potential slide surface will necessarily
reach this state, and the low residual strengths back-
calculated from field experience are a reflection of the aver-
age strength on the failure surface.
When the element thickness is chosen equal to the shear
band thickness, there is no correction for element size. It is
generally not practical to make the element height this small,
and a correction for element size is then needed. To examine
this approach, the sloping-ground problem was analyzed using
the mesh shown in Fig. 7b. This mesh has an ETR of 0.042
compared with a localized failure or shear band zone ratio of
0.0005. Based on Fig. 13, the predicted volumetric strain in the
element beneath the barrier would be 0.79% compared with
5.10% in the shear band. To account for this, the strain in the
element must be modified to ( ε v )* = ε v (5.10/0.79) = 6.4ε v ,
where 6.4 is a mesh correction factor.
© 2007 NRC Canada
884 Can. Geotech. J. Vol. 44, 2007

Fig. 17. Estimate of maximum volumetric strain, ( εv )max , for a base element with 10D50 (0.003 m) thickness for Fraser River sand
using extended trend of ( εv )max with element thickness ratio.

Table 2. Minimum and maximum void ratio and potential maximum expansion of sands at Dr = 50%.
Void ratio
Expansion potential,
Soil emin emax (εv)50 (%) Reference
Fraser River sand 0.596 0.900 8.8 Vaid and Eliadorani 1998
Cambria sand 0.538 0.767 7.0 Lade et al. 1998
Nevada sand 50/80 0.581 0.858 8.0 Lade et al. 1998
Nevada sand 80/200 0.617 0.940 9.1 Lade et al. 1998
Nevada fines 0.754 1.178 10.8 Lade et al. 1998
Ottawa sand 50/200 0.550 0.805 7.6 Lade and Yamamuro 1997
Ottawa sand F-95 0.580 0.865 8.3 Lade and Yamamuro 1997
Host sand A2 0.60 0.98 10.6 Thevanayagam 1998
Toyoura sand 0.616 0.988 10.3 Zlatovic 1994
Ottawa sand 0.48 0.78 9.2 Salgado et al. 2000
Ottawa sand C-109 0.50 0.83 9.9 Pitman et al. 1994
Quiou sand 0.78 1.20 10.6 Pestana and Whittle 1995
Mine tailings sand 0.69 1.06 9.9 Vaid et al. 1985
Brasted sand 0.48 0.79 9.5 Cornforth 1974
Dune sand 0.54 0.91 10.7 Konrad 1990
Well-rounded silica sand 0.67 1.06 10.5 Konrad 1990
Nerlerk 0%–2% 0.62 0.94 9.0 Sladen et al. 1985
Tottri sand 0.938 1.008 10.2 Takeshita et al. 1995
Monterey No. 9 sand 0.53 0.86 9.7 Riemer et al. 1990
Massey Tunnel sand 0.712 1.102 10.2 Konrad and Pouliot 1997
Quebec sand 0.54 0.79 7.5 Konrad 1998
Sand B 0.50 0.84 10.2 Castro 1969
Sand C 0.66 0.99 9.0 Castro 1969
Sand A 1.23 1.88 12.7 Castro 1969

© 2007 NRC Canada

Seid-Karbasi and Byrne
Fig. 18. Maximum expansion potential versus initial relative density
for sands (based on Table 2 data and eq. [6]).
Fig. 19. Time histories of surface lateral displacement (X-dis) for
two events.
The analysis showed that the ( ε v )* so computed was less
than the threshold value of 6.9%, and the results were
identical to those predicted in Figs. 11 and 12. When the
duration of base motion was increased from 7 s to 12 s, the
computed ( ε v )* was greater than 6.9%, at which time dila-
tion was suppressed and a flow slide was predicted.
Figure 19 shows the surface displacement – time histories
for both cases. It can be seen that surface displacement for
12 s motion is ever increasing, whereas displacements stop
some time after the end of shaking in the 7 s case. Ground
surface velocity is a good indicator of flow failure. Figure 20
shows ground surface velocity versus time for the two
events. For the 7 s event, the surface velocity decreases and
then becomes zero after the end of shaking, whereas for the
12 s event it is essentially constant over that period. This re-
veals that flow failure takes place in the second case when
the dilation threshold is reached.
Countermeasure for flow failure
Drains have the potential to nullify the barrier effect and
curtail or prevent the occurrence of lateral spread or flow
slides in the event of an earthquake. They can facilitate dis-
sipation of excess pore pressure and mitigate the impedance
effects of low-permeability layers. This is examined for the
same soil profile as that used in the previous analyses with
885
Fig. 20. Time histories of surface lateral velocity (X-vel) for two
events.
1.0° inclination treated with gravel drains penetrating
through the low-permeability sublayer down to the loose
sand layer at the bottom (case III).
For the study conducted in plane strain conditions, the
drains have been accounted for with a soil column having a
permeability 100 times greater than that of the surrounding
liquefiable soils. In a real prototype problem, the three-
dimensional effects of the drain columns pattern can be
treated in plane strain analysis using an equivalent drain area
approach (e.g., Indraratna and Redana 1997, 2000). Fig-
ure 21a shows the model (case III) with a drain. Figure 21b
shows distributions of maximum excess pore pressure ratio,
Rumax , within the model along with flow vectors after 3.5 s
of shaking. It is seen that Rumax increases with an increase in
distance from the drain column. The drain exhibits some
excess pore pressure during shaking. The maximum pre-
dicted lateral displacement in this case was negligible, even
compared with that of the model without a low-permeability
layer (case I). This demonstrates that the drain column is a
promising measure to mitigate liquefaction-induced defor-
mations.
A comparison of predicted time histories of excess pore
pressure buildup for a point at mid-depth of loose sand for
cases I, II, and III is shown in Fig. 22, which indicates that
drains can reduce generated excess pore pressures and signi-
ficantly speed up the dissipation of excess pore pressure.
The effect of drains in decreasing excess pore pressure during
shaking is also reported by Chang et al. (2004) and Brennan
and Madabhushi (2005) in field model tests and centrifuge
tests, respectively. Figure 23 shows excess pore pressure ratios
measured in field tests within a liquefiable soil for a deposit
without and with vertical drains subjected to a harmonic
excitation. These data show that the excess pore pressure
build-up rate is significantly slower with drains and confirms
the effectiveness of drains as a remediation measure.
Practical implications
It was demonstrated that the applied dynamic procedure
can reproduce the seismic response of liquefiable soil when
a low-permeability sublayer is present. The presence of such
a sublayer impedes upward drainage that can lead to the for-
mation of water films having essentially zero shear strength
and result in large ground deformation and flow failure. The
© 2007 NRC Canada
886
Fig. 21. Treated model, case III: (a) central vertical drain; (b) Rumax and flow vectors at 3.5 s.
k (m/s)
8.81 x 10-7
8.81 x 10-4
8.81 x 10-2
Vertical drain
(a)
analyses also show that drains can be very effective in alle-
viating the destructive effects of sublayer barriers.
Important aspects for design are as follows:
(1) Appropriate site investigation techniques should be applied
to detect low-permeability, thin layers within sand and
(or) gravel layers.
(2) If a low-permeability sublayer is present and significant
liquefaction is predicted for the design earthquake, either
low residual shear strengths consistent with field back
analysis should be used or drainage–densification should
be considered. The design of drains–densification can
be optimized from a coupled stress-flow dynamic analysis.
(3) The perfect undisturbed sample obtained prior to an
earthquake will not be representative of conditions during
and shortly after the earthquake due to expansion result-
ing from the upward flow of water in stratified deposits.
Dissipation of excess pore pressure some time after the
earthquake will reconsolidate soil elements beneath the
barrier such that the perfect sample taken some time
after the earthquake will also not be representative of
the critical conditions.
Conclusions
Many failures of earth structures and submarine slides
have been reported during past earthquakes worldwide. A
number of civil structures and soil deposits in coastal or
river areas have suffered large deformations during past
earthquakes as a result of soil liquefaction. The deformations
may exceed several metres, even in gentle slopes of less than
a few percent. Deformations occur not only during but also
after earthquake shaking.
Clean, loose (e.g., Dr ≥ 20%) sands are unlikely to suffer
a flow slide. Although they can be triggered to liquefy and
undergo large strains and displacements, their undrained
strengths are generally adequate for stability. However, if the
Can. Geotech. J. Vol. 44, 2007
R umax
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
(b)
sands contain low-permeability layers that impede drainage,
a void expansion resulting in a water-rich thin zone or water
film may cause a near complete loss of strength of the soil
directly beneath that layer and result in a flow slide. The low
residual strengths based on back analyses of field case histo-
ries and reported by Seed and Harder (1990) are likely a result
of void expansion related to the presence of barrier layers.
A numerical approach is used in this paper that captures
element sand behaviour in monotonic and cyclic loading under
different drainage conditions; undrained and inflow were
utilized to study the effects of low-permeability layers on
the sloping ground response during earthquake loading. The
following conclusions are made based on the analysis results:
(1) The computed results show that for a homogeneous
sand layer that is triggered to liquefy, upward flow
resulting from excess pore pressure causes dissipation
and a reduction in void ratio within the sand layer. How-
ever, if a low-permeability layer(s) impedes drainage,
then contraction occurs at the base of the sand layer but
expansion at the top where flow is impeded, and this
expansion is responsible for flow failures. The occur-
rence of contraction and expansion zones, respectively,
at the lower and upper zones due to pore pressure redis-
tribution is a characteristic behaviour of a liquefiable
deposit with sublayer barriers.
(2) Void redistribution results in a thin, water-rich zone at
the base of a barrier layer that could ultimately form a
water film when enough water is available for injection.
The thickness of this shear band zone is related to particle
diameter; if and when the predicted expansion exceeds
the threshold expansion, a water film will form, and this
is simulated in the analyses by setting the dilation equal
to zero.
(3) It is generally not practical to have the mesh size as
small as the shear band thickness, in which case a cor-
rection to the computed expansion is required.
© 2007 NRC Canada
Seid-Karbasi and Byrne
Fig. 22. Predicted time history of Ru at mid-depth of loose sand:
(a) without barrier layer; (b) with barrier layer; (c) with barrier
layer treated with drain (Seid-Karbasi3).
Fig. 23. Measured Ru in field liquefaction test with and without
a drain (data from Chang et al. 2004).
887
(4) Most of the relative movements occur at the base of the
low-permeability sublayer. Maximum displacements occur
at or near the surface above the barrier layer.
(5) A large part of the deformations may occur some time
after shaking has ceased, depending on the time needed
for migration of water from zones with higher excess
pore pressures.
(6) Installation of vertical drains that penetrate the barrier
layers can mitigate the destructive effects of low-
permeability layers. This has been demonstrated in these
analyses and has also been observed from field model
tests.
(7) The design of these drains can be assessed from
dynamic coupled flow – effective stress analyses using
appropriate modeling parameters and design input
motions. The dimensions and location of remediation
measures can be optimized from dynamic analyses.
Acknowledgments
The authors acknowledge the financial support from BC
Hydro through the Professional Partnership program and the
support of the Natural Sciences and Engineering Research
Council of Canada (NSERC) through Strategic Liquefaction
Grant NSERC 246394 and NSERC COSTA Grant 03608-
CG068625. The authors are also grateful to Professor
D.L. Anderson and to Ernest Naesgaard for helpful discus-
sions during the course of this research. In addition, the au-
thors express their appreciation to the reviewers for their
constructive comments, which led to many changes from the
original manuscript.
References
Adalier, K., and Elgamal, A.W. 1992. Post-liquefaction behavior of
soil systems. Report, Department of Civil Engineering,
Rensselaer Polytechnic Institute, Troy, N.Y.
Alshibli, K., and Sture, S. 1999. Sand shear band thickness mea-
surements by digital imaging techniques. Computing in Civil
Engineering, ASCE, 13: 103–109.
Atigh, E., and Byrne, P.M. 2004. Liquefaction flow of submarine
slopes under partially undrained conditions: an effective stress
approach. Canadian Geotechnical Journal, 41(1): 154–165.
Beaty, M.H., and Byrne, P.M. 1998. An effective stress model for
predicting liquefaction behavior of sand. In Geotechnical Earth-
quake Engineering and Soil Dynamics III: Proceedings of a
Specialty Conference, Seattle, Wash., 3–6 August 1998. Edited
by P. Dakoulas, M. Yegian, and R.D. Holtz. ASCE Geotechnical
Special Publication 75, Vol. 1, pp. 766–777.
Berrill, S.A., Christensen, S.A., Keenan, R.J., Okada, W., and
Pettinga, J.R. 1997. Lateral-spreading loads on a piled bridge
foundation. In Proceedings of the 14th International Conference
on Soil Mechanics and Foundation Engineering, Discussion
Session, Seismic Behavior of Ground and Geotechnical Struc-
tures, Hamburg, Germany, 6–12 September 1997. A.A.
Balkema, Rotterdam. pp. 173–183.
Bobei, D.C., and Lo, S.-C. 2003. Strain path influence on the
behavior of sand with fines. In Soil and Rock America 2003:
Proceedings of the 12th Panamerican Conference on Soil
Mechanics and Geotechnical Engineering, Cambridge, Mass., 22–
26 June 3003. Edited by P.J. Culligan, H.H. Einstein, and A.J.
Whittle. Verlag Gluckauf GMBH, Essen, Germany. pp. 583–588.
© 2007 NRC Canada
888
Bouckovalas, G.D., Gazetas, G., and Papadimitriou, A.G. 1999.
Geotechnical aspects of the 1995 Aegion (Greece) earthquake.
In Proceedings of the 2nd International Conference on Earth-
quake Geotechnical Engineering, Lisbon, 21–25 June. Edited by
P. Seco e Pinto. A.A. Balkema, Rotterdam. pp. 739–748.
Brennan, A.J., and Madabhushi, S.P. 2005. Liquefaction and drainage
in stratified soil. Journal of Geotechnical and Geoenvironmental
Engineering, ASCE, 131: 876–885.
Byrne, P.M., and Beaty, M.H. 1997. Post-liquefaction shear
strength of granular soils: theoretical/conceptual issues. In Pro-
ceedings of the Workshop on Post-Liquefaction Shear Strength
of Granular Soils, Urbana–Champion, Ill., 17–18 April 1997.
pp. 16–45.
Byrne, P.M., Roy, D., Campanella, R.G., and Hughes, J. 1995.
Predicting liquefaction response of granular soils from pressure-
meter tests. In Static and Dynamic Properties of Gravelly Soils:
Proceedings of Sessions of the ASCE National Convention, San
Diego, Calif., 23–24 October 1995. Edited by M.D. Evans and
R.J. Fragaszy. ASCE Geotechnical Special Publication 56,
pp. 122–135.
Byrne, P.M., Park, S., Beaty, M., Sharp, M., Gonzalez, L., and
Abdoun, T. 2004. Numerical modeling of liquefaction and com-
parison with centrifuge tests. Canadian Geotechnical Journal,
41(1): 193–211.
Byrne, P.M., Naesgaard, E., and Seid-Karbasi, M. 2006. Analysis
and design of earth structures to resist seismic soil liquefaction.
In Proceedings of the 59th Canadian Geotechnical Engineering
Conference, Hardy Lecture, Vancouver, B.C., 1–4 October 2006.
Canadian Geotechnical Society, Alliston, Ont. pp. 1–24.
Castro, G. 1969. Liquefaction of sands. Ph.D. thesis, Harvard Uni-
versity, Cambridge, Mass.
Castro, G., Keller, T.O., and Boynton, S.S. 1989. Re-evaluation of
the Lower San Fernando Dam. Contract Report GL-89-2, US Army
Waterways Experiment Station, Vicksburg, Miss. Vols. 1 and 2.
Chang, W.J., Rathje, E., Stokoe, K.H., II, and Cox, B.R. 2004.
Direct evaluation of effectiveness of prefabricated vertical drain
in liquefiable sand. Soil Dynamics and Earthquake Engineering
Journal, 24: 723–731.
Chu, J. 1991. Strain softening behavior of granular soils under
strain path testing. Ph.D. thesis, University of New South Wales,
Kensington, NSW, Australia.
Cornforth, D.H. 1974. One-dimensional consolidation curves of a
medium sand. Géotechnique, 24: 678–683.
Coulter, H.W., and Migliaccio, R.R. 1966. Effects of the earth-
quake of March 27, 1964 at Valdez, Alaska. US Geological
Survey, Professional Paper 542-C. 36 pp.
Cubrinovski, M., and Ishihara, K. 2002. Maximum and minimum
void ratio characteristics of sand. Soils and Foundations, 42: 65–78.
Desrues, J., and Viggiani, G. 2004. Strain localization in sand: an
overview of the experimental results obtained in Grenoble using
stereophotogrammetry. International Journal for Numerical and
Analytical Methods in Geomechanics, 28: 279–321.
Elgamal, A.W., Dobry, R., and Adalier, K. 1989. Small scale
shaking table tests of saturated layered sand–silt deposits. In
Proceedings of the 2nd US–Japan Workshop on Soil Liquefac-
tion, Large Ground Deformation, and Their Effects on Life-
lines, Grand Island and Ithaca, N.Y., 26–29 September 1989.
Edited by T.D. O’Rourke and M. Hamada. NCEER Technical
Report NCEER-89-0032, National Center for Earthquake Engi-
neering Research (NCEER), Buffalo, N.Y. pp. 233–245.
Eliadorani, A.A. 2000. The response of sands under partially
drained states with emphasis on liquefaction. Ph.D. thesis, Civil
Engineering Department, The University of British Columbia,
Vancouver, B.C.
Can. Geotech. J. Vol. 44, 2007
Fiegel, G.L., and Kutter, B.L. 1992. Liquefaction mechanism for
layered soil. Journal of Geotechnical Engineering, ASCE, 120:
737–755.
Field, M.E., Gardner, J.V., Jennings, A.E., and Edwards, B.D.
1982. Earthquake induced sediment failures on a 0.258 slope,
Klamath River Delta. Journal of California Geology, 10: 542–546.
Hamada, M. 1992. Large ground deformations and their effects on
lifelines: 1964 Niigata Earthquake. In Case studies of liquefac-
tion and lifeline performance during past earthquakes. Vol. 1.
Japanese case studies. Edited by M. Hamada and T.D.
O’Rourke. NCEER Technical Report NCEER-92-0001, National
Center for Earthquake Engineering Research (NCEER), Buffalo,
N.Y. pp. 3/1–3/123.
Hampton, M.A., and Lee, H.J. 1996. Submarine landslides. Reviews
of Geophysics, 34: 33–59.
Harder, L., and Stewart, J. 1996. Failure of Tapo Canyon tailings
dam. Journal of Performance of Constructed Facilities, ASCE,
10(3): 109–114.
Huishan, L., and Taiping, Q. 1984. Liquefaction potential of sat-
urated sand deposits underlying foundation of structure. In
Proceedings of the 8th World Conference on Earthquake En-
gineering, San Francisco, Calif., 21–28 July 1984. Prentice–
Hall Inc., Englewood Cliffs, N.J. Vol. 3, pp. 199–206.
Indraratna, B., and Redana, I.W. 1997. Plane strain modeling
smear effects associated with vertical drains. Journal of
Geotechnical Engineering, ASCE, 119: 1321–1329.
Indraratna, B., and Redana, I.W. 2000. Numerical modeling of
vertical drains with smear and well resistance installed in soft
clay. Canadian Geotechnical Journal, 37(1): 132–145.
Ishihara, K. 1984. Post-earthquake failure of a tailings dam due to
liquefaction of the pond deposit. In Proceedings of the Interna-
tional Conference of Case Histories in Geotechnical Engi-
neering, University of Missouri-Rolla, Rolla, Mo., May 1984.
Vol. 3, pp. 1129–1143.
Itasca Consulting Group, Inc. 2005. Fast lagrangian analysis of
continua (FLAC), version 5, user’s guide. Itasca Consulting
Group, Inc., Minneapolis, Minn.
Kawakami, F., and Asada, A. 1966. Damage to the ground and
earth-structures by the Niigata earthquake of June 16, 1964.
Soils and Foundations, 1: 14–30.
Kokusho, T. 1999. Water film in liquefied sand and its effect on
lateral spread. Journal of. Geotechnical and Geoenvironmental
Engineering, ASCE, 125: 817–826.
Kokusho, T. 2000. Mechanism for water film generation and lat-
eral flow in liquefied sand layer. Soils and Foundations, 40:
99–111.
Kokusho, T. 2003. Current state of research on flow failure consid-
ering void redistribution in liquefied deposits. Journal of Soil
Dynamics and Earthquake Engineering, 23: 585–603.
Kokusho, T., and Kojima, T. 2002. Mechanism for post-
liquefaction water film generation in layered sand. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 128:
129–137.
Konrad, J.-M. 1990. Minimum undrained strength of two sands.
Journal of Geotechnical Engineering, ASCE, 116: 932–947.
Konrad, J.-M. 1998. Sand state from cone penetration tests. Géo-
technique, 48: 201–215.
Konrad, J.-M., and Pouliot, N. 1997. Ultimate state of reconstituted
and intact samples of deltaic sand. Canadian Geotechnical Journal,
34(5): 737–748.
Kulasingam, R. 2003. Effects of void redistribution on
liquefaction-induced deformations. Ph.D. thesis, Civil and En-
vironmental Engineering Department, University of California,
Davis, Calif.
© 2007 NRC Canada
Seid-Karbasi and Byrne
Kulasingam, R., Malvick, E.J., Boulanger, R.W., and Kutter, B.L.
2001. Void redistribution and localization of shear strains in
model sand slopes with silt seams. Report on first year activities.
In Proceedings of the US–Japan Cooperative Research in Urban
Earthquake Disaster Mitigation Workshop, Seattle, Wash., 16–
18 August 2001. MCEER, N.Y. pp. 117–128.
Kulasingam, R., Malvick, E.J., Boulanger, R.W., and Kutter, B.L.
2004. Strength loss and localization at silt interlayers in slopes
of liquefied sand. Journal of Geotechnical and Geo-
environmental Engineering, ASCE, 130(11): 1192–1202.
Kutter, B.L. 1995. Recent advances in centrifuge modeling of seis-
mic shaking. In Proceedings of the 3rd International Conference
of Recent Advances in Geotechnical Earthquake Engineering and
Soil Dynamics, St. Louis, Mo., 2–7 April 1995. University of
Missouri-Rolla, Rolla. Vol. 2, pp. 927–942.
Lade, P.V., and Yamamuro, J.A. 1997. Effects of nonplastic fines
on static liquefaction of sands. Canadian Geotechnical Journal,
34(6): 918–928.
Lade, P.V., Liggio, C.D., and Yamamuro, J.A. 1998. Effects of
non-plastic fines on minimum and maximum void ratios of sand.
ASTM Journal of Geotechnical Testing, 21: 336–347.
Lemke, R.W. 1967. Effects of the earthquake of March 27, 1964 at
Seward, Alaska. US Geological Survey, Professional Paper 542-E.
Liu, H., and Qiao, T. 1984. Liquefaction potential of saturated sand
deposits underlaying foundation of structure. In Proceedings of
the 8th World Conference on Earthquake Engineering, San Fran-
cisco, Calif., 21–28 July 1984. Prentice–Hall Inc., Englewood
Cliffs, N.J. pp. 199–206.
Lu, X., Shuyun, W., Yihua, W., and Cui, P. 2004. An approximate
method for evaluating the shear band thickness in saturated
sand. International Journal of Numerical and Analytical Methods
in Geomechanics, 28: 1533–1541.
Malvick, E.J. 2005. Void redistribution-induced shear localization
and deformation in slopes. Ph.D. dissertation, University of Cal-
ifornia at Davis, Davis, Calif. 285 pp.
Malvick, E.J., Kulasingam, R., Kutter, B.L., and Boulanger,
R.W. 2002. Void redistribution and localized shear strains in
slopes during liquefaction. In Physical Modeling in
Geotechnics, ICPMG‘02: Proceedings of the International
Conference, St. John’s, Nfld., 10–12 July 2002. Edited by R.
Phillips, P.J. Guo, and R. Popescu. A.A. Balkema, Rotterdam.
pp. 495–500.
Malvick, E.J., Kutter, B.L., Boulanger, R.W., Kabasawa, K., and
Kokusho, T. 2005. Void redistribution research with 1-g and
centrifuge modeling. In Proceedings of the 16th International
Conference on Soil Mechanics and Geotechnical Engineering
(ICSMGE), Osaka, 12–16 September 2005. A.A. Balkema, Rot-
terdam. Vol. 4, pp. 2543–2546.
Malvick, E.J., Kutter, B.L., Boulanger, R.W., and Kulasingam, R.
2006. Shear localization due to liquefaction-induced void redis-
tribution in a layered infinite slope. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 132: 1293–1303.
Martin, G.R., Finn, W.D.L., and Seed, H.B. 1975. Fundamentals of
liquefaction under cyclic loading. Journal of the Geotechnical
Engineering Division, ASCE, 101: 423–438.
Olson, S.M., and Stark, T.D. 2002. Liquefied strength ratio from
liquefaction flow failure case histories. Canadian Geotechnical
Journal, 39(3): 629–647.
Olson, S.M., and Stark, T.D. 2003. Use of laboratory data to confirm
yield and liquefied strength ratio concepts. Canadian
Geotechnical Journal, 40(6): 1164–1184.
Pestana, J.M., and Whittle, A.J. 1995. Compression model for
cohesionless soils. Géotechnique, 45: 611–631.
889
Pitman, T.D., Robertson, P.K., and Sego, D.C. 1994. Influence of
fines on the collapse of loose sands. Canadian Geotechnical
Journal, 31(5): 728–739.
Poulos, S.J., Castro, G., and France, J.W. 1985. Liquefaction evalu-
ation procedure. Journal of Geotechnical Engineering., ASCE,
111: 772–792.
Puebla, H. 1999. A constitutive model for sand analysis of the
CANLEX embankment. Ph.D. thesis, Civil Engineering Depart-
ment, University of British Columbia, Vancouver, B.C.
Puebla, H., Byrne, P.M., and Phillips, R. 1997. Analysis of
CANLEX liquefaction embankments: prototype and centrifuge
models. Canadian Geotechnical Journal, 34(5): 641–654.
Riemer, M.F., Seed, R.B., Nicholson, P.G., and Jong, H.L. 1990.
Steady state testing of loose sands: limiting minimum density.
Journal of Geotechnical Engineering, ASCE, 116: 332–337.
Roscoe, K.H. 1970. The influence of strains in soil mechanics.
Géotechnique, 20: 129–170.
Salgado, R., Bandini, P., and Karim, A. 2000. Shear strength and
stiffness of silty sand. Journal of Geotechnical and Geo-
environmental Engineering, ASCE, 126: 451–462.
Sawada, S., Tsukamoto, Y., and Ishihara, K. 2006. Residual defor-
mation characteristics of partially saturated sandy soils sub-
jected to seismic excitation. Journal of Soil Dynamics and
Earthquake Engineering, 26: 175–182.
Schofield, A.N. 1981. Dynamics and earthquake geotechnical
centrifuge modeling. In Proceedings of the International Confer-
ence on Recent Advances in Geotechnical Earthquake Engi-
neering and Soil Dynamics, St. Louis, Mo., 26 April – 2 May
1981. Edited by S. Prakash. University of Missouri-Rolla, Rolla.
Vol. 3, pp. 1081–1100.
Scott, R.F., and Zuckerman, K.A. 1972. Sandblows and liquefac-
tion. In The Great Alaskan earthquake of 1964. Engineering
Publication 1606, National Academy of Sciences, Washington,
D.C., pp. 170–189.
Seed, H.B. 1987. Design problems in soil liquefaction. Journal of
Geotechnical Engineering, ASCE, 113: 827–845.
Seed, R.B., and Harder, L.F. 1990. SPT-based analysis of cyclic
pore pressure generation and undrained residual strength. In
Proceedings of the H.B. Seed Memorial Symposium, May 1990.
Edited by J.M. Duncan. BiTech Publishing Ltd., Richmond,
B.C. Vol. 2, pp. 351–376.
Seid-Karbasi, M., and Byrne, P.M. 2004a. Liquefaction, lateral
spreading and flow slides. In Proceedings of the 57th Canadian
Geotechnical Engineering Conference, Québec City, Que., 24–
27 October 2004. Canadian Geotechnical Society, Alliston, Ont.
p. G13.529.
Seid-Karbasi, M., and Byrne, P.M. 2004b. Embankment dams and
earthquakes. International Journal on Hydropower and Dams,
11(2): 96–102.
Seid-Karbasi, M., and Byrne, M.P. 2006. Effects of partial satura-
tion on liquefiable ground response. In GeoCongress 2006:
Geotechnical Engineering in the Information Technology Age,
Atlanta, Ga., 26 February – 2 March 2006. Edited by D.J.
DeGroot, J.T. DeJong, D. Frost, and L.G. Baise. ASCE, Reston,
Va. Paper 11803.
Seid-Karbasi, M., Byrne, P.M., Naesgaard, E., Park, S.,
Wijewickreme, D., and Phillips, R. 2005. Response of sloping
ground with liquefiable materials during earthquake: a class A
prediction. In International Association of Computer Methods
and Advances in Geomechanics, IACMAG: Proceedings of the
11th International Conference, Turin, Italy, 19–21 June 2005.
Edited by G. Barla and M. Barla. Patron Editore, Bologna, Italy.
Vol. 3, pp. 313–320.
© 2007 NRC Canada
890
Sento, N., Kazama, M., Uzuoka, R., Ohmur, H., and Ishimaru, M.
2004. Possibility of post liquefaction flow failure due to seepage.
Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 130: 707–716.
Sharp, M., Dobry, R., and Abdoun, T. 2003. Liquefaction centri-
fuge modeling of sands of different permeability. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 129:
1083–1091.
Silver, M.L., and Seed, H.B. 1971. Volume change in sands during
cyclic loading. Journal of the Soil Mechanics and Foundations
Division, ASCE, 97: 1171–1182.
Sladen, J.A., D’Hollander, R.D., Krahn, J., and Mitchell, D.E.
1985. Back analysis of the Nerlerk berm liquefaction slides.
Canadian Geotechnical Journal, 22: 579–588.
Sriskandakumar, S. 2004. Cyclic loading response of Fraser River
Sand for numerical models simulating centrifuge tests. M.A.Sc.
thesis, Civil Engineering Department, University of British
Columbia, Vancouver, B.C.
Stark, T.D., and Mesri, G. 1992. Undrained shear strength of lique-
fied sands for stability analysis. Journal of Geotechnical Engi-
neering, ASCE, 118: 1727–1747.
Taboada, V., and Dobry, R. 1993a. Experimental results of model
No. 1 at RPI. In Verification of Numerical Procedures for the
Analysis of Soil Liquefaction Problems (VELACS): Proceedings
of the International Conference, Davis, Calif., 17–20 October
1993. Edited by K. Arulandan, R.F. Scott, and X. Zeng. A.A.
Balkema, Rotterdam. Vol. 1, pp. 3–17.
Taboada, V., and Dobry, R. 1993b. Experimental results of model
No. 2 at RPI. In Verification of Numerical Procedures for the
Analysis of Soil Liquefaction Problems (VELACS): Proceedings
of the International Conference, Davis, Calif., 17–20 October
1993. Edited by K. Arulandan, R.F. Scott, and X. Zeng. A.A.
Balkema, Rotterdam. pp. 277–294.
Taboada, V., and Dobry, R. 1998. Centrifuge modeling of earth-
quake-induced lateral spreading in sand. Journal of Geotechnical
Engineering, ASCE, 124: 195–206.
Takeshita, S., Takeishi, M., and Tamada, K. 1995. Static liquefaction
of sands and its liquefaction index. In Proceedings of the 1st
International Conference on Earthquake Geotechnical Engi-
neering, Tokyo, Japan, 14–16 November 1995. Edited by K.
Ishihara. A.A. Balkema, Rotterdam. Vol. 1, pp. 177–182.
Thevanayagam, S. 1998. Effect of fines and confining stress on
undrained shear strength of silty sands. Journal of Geotechnical
and Geoenvironmental Engineering, ASCE, 124: 479–491.
Can. Geotech. J. Vol. 44, 2007
Tsukamoto, Y., Ishihara, K., and Sawada, S. 2004. Settlement of
silty sand deposits following liquefaction during earthquakes.
Soils and Foundations, 44: 135–148.
Uzuoka, R., Sento, N., and Kazama, M. 2003. Numerical analysis
on liquefaction-induced progressive deformation with pore water
pressure migration. In Proceedings of the 3rd International Sym-
posium on Deformation Characteristics of Geomechanics, Lyon,
22–24 September 2003. Edited by H. Di Benedetto, T. Doanh,
H. Geoffroy, and C. Sauzét. A.A. Balkema, Rotterdam.
pp. 1095–1101.
Vaid, Y.P., and Eliadorani, A. 1998. Instability and liquefaction of
granular soils under undrained and partially drained states.
Canadian Geotechnical Journal, 35(6): 1053–1062.
Vaid, Y.P., Chern, J.C., and Tumi, H. 1985. Confining pressure,
grain angularity, and liquefaction. Journal of Geotechnical Engi-
neering, ASCE, 111: 1229–1235.
Yang, Z., and Elgamal, A. 2002. Influence of permeability on
liquefaction-induced shear deformation. Journal of Engineering
Mechanics, ASCE, 128: 720–729.
Yoshida, N., Yasuda, S., and Ohya, Y. 2005. Two criteria for
liquefaction-induced flow. In Proceedings of the 16th Interna-
tional Conference on Soil Mechanics and Geotechnical Engi-
neering (ICSMGE), Performance Based Design in Earthquake
Geotechnical Engineering: Concepts and Research Satellite
Conference, Osaka, Japan, 10–15 September 2005. A.A. Balkema,
Rotterdam. pp. 109–116.
Yoshimine, M., Nishizaki, H., Amano, K., and Hosono, Y. 2006.
Flow deformation of liquefied sand under constant shear load
and its application to analysis of flow slide of infinite slope.
Soil Dynamics and Earthquake Engineering Journal, 26: 253–
264.
Youd, T.L., Idriss, I.M., Andrus, R., Arango, I., Castro, G., Christian,
J., Dobry, J., Finn, L., Harder, L., Jr., Hynes, H.M., Ishihara, K.,
Koester, J., Liao, S.S., Marcuson, W.F., III, Martin, G., Mitchell,
J.K., Moriwaki, Y., Power, M.S., Robertson, P.K., Seed, R.B.,
and Stokoe, K.H., II. 2001. Liquefaction resistance of soils:
summary report from the 1996 NCEER and 1998 NCEER/NSF
Workshops on Evaluation of Liquefaction Resistance of Soils.
Journal of Geotechnical and Geoenvironmental Engineering,
ASCE, 127: 817–833.
Zlatovic, S. 1994. Residual strength of silty soils. D.Eng. thesis,
University of Tokyo, Tokyo.
© 2007 NRC Canada