1471

Effect of static shear stress on the cyclic

resistance of sands in simple shear loading
Siva Sivathayalan and Da Ha
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23

Abstract: An experimental study of the effect of (initial) static shear stress on the undrained simple shear response of sands
is presented. Liquefaction resistance of two sands, one generally contractive and the other dilative, over a wide range of den-
sity and effective stress levels was evaluated at various density and stress states. At a given density, the effect· of static shear
on the cyclic resistance of sands is dependent on the type of stress–strain response that ensues during loading. The general
notion that static shear increases the cyclic resistance of sands at relatively high densities is not valid if the sand is contrac-
tive at the denser state. The Ka correction factor for sand at 80% relative density may be as low as about 0.6 for strain-soft-
ening sand depending on the initial stress state. The routine industry practice of disregarding the Ka factor in dense sands
may thus lead to unsafe designs in such materials. The Ka values measured under cyclic simple shear loading are compared
to the Ka values reported in the literature based on cyclic triaxial tests. Test results clearly indicate that Ka is also dependent
on the loading mode. Ka correction factors proposed in the literature, in general, have been derived either from empirical
data or from laboratory tests under triaxial loading mode. Extra attention is required when using those correction factors, as
they may not appropriately account for the loading mode effects.
Key words: cyclic resistance, static shear, liquefaction, strain softening, experimental soil mechanics, geotechnical earthquake
engineering.
Résumé : Cet article présente une étude expérimentale de l’effet de la contrainte de cisaillement statique (initiale) sur le
For personal use only.

comportement non drainé en cisaillement simple de sables. La résistance à la liquéfaction de deux sables, un généralement
contractant et l’autre dilatant sur une gamme de densité et de niveaux de contrainte effective, a été évaluée à plusieurs densi-
tés et états de contraintes. Pour une densité donnée, l’effet du cisaillement statique sur la résistance cyclique des sables est
dépendant du type de comportement en contrainte-déformation qui est produit durant le chargement. La notion générale qui
stipule que le cisaillement statique augmente la résistance cyclique de sables à densités relativement élevées n’est pas valide
si le sable est contractant à l’état plus dense. Le factor de corrélation Ka pour un sable à une densité relative de 80 % peut
être aussi faible qu’environ 0,6 pour un sable anti-écrouissage, dépendant de l’état des contraintes initial. La pratique usuelle
de l’industrie qui ne considère pas le facteur Ka pour des sables denses peut ainsi entraîner des conceptions non sécuritaires
dans ces matériaux. Les valeurs de Ka mesurées sous chargement en cisaillement statique simple sont comparées aux valeurs
de Ka reportées dans la littérature basées sur des essais triaxiaux cycliques. Les résultats des essais indiquent clairement que
Ka est aussi dépendant du mode de chargement. Les facteurs de corrélation Ka proposés dans la littérature sont généralement
dérivés soit de données empiriques ou d’essais en laboratoire sous chargement triaxial. Une attention particulière est néces-
saire lorsque ces facteurs de correction sont utilisés puisqu’ils peuvent ne pas bien tenir compte des effets du mode de char-
gement.
Mots‐clés : résistance cyclique, cisaillement statique, liquéfaction, anti-écrouissage, mécanique des sols expérimentale, génie
de la géotechnique sismique.
[Traduit par la Rédaction]

Introduction
Undrained behaviour of soils is dependent on the initial ef-
fective stress state, which is normally characterized by the
confining and static shear stress levels. The term static shear
Received 10 March 2010. Accepted 19 July 2011. Published at
www.nrcresearchpress.com/cgj on 28 September 2011.
S. Sivathayalan. Department of Civil and Environmental
Engineering, Carleton University, 3432 Mackenzie Building,
1125 Colonel By Drive, Ottawa, ON K1S 5B6, Canada.
D. Ha. Geocomp Corporation, 1145 Massachusetts Avenue,
Boxborough, MA 01719, USA.
Corresponding author: S. Sivathayalan (e-mail: siva_sivathaya-
lan@carleton.ca).
stress is used herein to represent the initial shear stress prior
to earthquake loading. Under the simple shear loading mode,
the confining stress is generally represented by the vertical
0
effective stress, s vc , and static shear is represented by the
shear stress on the horizontal plane, th. The influence of the
level of the confining stress on the response is fairly well
understood both under cyclic and monotonic loading condi-
tions. Increasing confining stresses have been shown to in-
crease the contractive tendency of the sands at a given initial
density (Vaid and Chern 1985; Vaid and Thomas 1995). Such
an increase in contractive tendency due to higher stress levels is
more prominent in angular sands than in rounded sands.
Cyclic resistance of sands is often represented by the cy-
clic resistance ratio (CRR), which is defined as the ratio of
the cyclic shear stress, tcy, causing liquefaction in a specified

Can. Geotech. J. 48: 1471–1484 (2011) doi:10.1139/T11-056 Published by NRC Research Press
 1472
number of load cycles (typically 10 or 15 cycles) to the ef-
0
fective confining stress, s vc . At a given density, the CRR of
a hydrostatically consolidated sand decreases with increasing
confining stress level. This reduction in cyclic resistance due
to increasing confining stresses is accounted for by the cor-
rection factor Ks defined as the ratio between the cyclic re-
sistance at the specified confining stress level to that at
100 kPa (Seed 1983). Extensive research has been carried
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
out on the effect of initial confining stress level on cyclic re-
sistance (Seed and Harder 1990; Vaid and Thomas 1995;
Vaid and Sivathayalan 1996; Hynes and Olsen 1998; Bou-
langer 2003a; Idriss and Boulanger 2003; Sivathayalan and
Ha 2004; Boulanger and Idriss 2007), and the Ks correction
factors proposed in the literature are considered fairly reliable
and have been widely accepted.
Lee and Seed (1967) pioneered research on cyclic resist-
ance of soils subjected to static shear using anisotropically
consolidated triaxial specimens, and Seed (1983) proposed a
correction factor Ka, which is defined as the ratio of the cy-
clic resistance of the soil with static shear to that without
static shear, to account for the effects of static shear on cyclic
resistance. Several attempts have been made over the years to
quantify the cyclic resistance of soils subjected to static shear
stresses (Yoshimi and Oh-Oka 1975; Vaid and Finn 1979;
Vaid and Chern 1985; Seed and Harder 1990; Harder and
Boulanger 1997; Vaid et al. 2001). However, significant un-
For personal use only.
certainties and apparent contradictions exist in the literature,
and the role of static shear on liquefaction resistance has not
been fully understood yet.
Vaid and Chern (1985) have demonstrated that cyclic re-
sistance can either increase or decrease depending on the
confining stress level and density. They further showed that
at a given density, the cyclic resistance depends on the con-
fining and static shear stress levels. It is now well known that
an increase in static shear may either increase or decrease the
cyclic resistance depending on the density as well as the lev-
els of confining and static shear stresses. However, only lim-
ited attention has been paid in the literature to the mechanism
that is responsible for strain development under cyclic load-
ing. The data presented by Seed and Harder (1990) show sig-
nificant reductions in the cyclic resistance of loose sands
with increasing static shear stress, but indicate that cyclic re-
sistance of medium-dense to dense sands increases with in-
creasing static shear stress. The modifications proposed by
Harder and Boulanger (1997) represent marginally further re-
ductions in the Ka values reported by Seed and Harder. In
contrast, experimental data (Vaid et al. 2001) shows that the
cyclic resistance of even the loosest deposited sand increases
with increasing static shear up to moderate levels of static
shear. The attempt by Idriss and Boulanger (2003) to corre-
late Ka with the relative state parameter index, xR (Boulanger
2003b), and not with density is consistent with the findings
of Vaid and co-workers (Vaid and Chern 1985; Vaid et al.
2001) that Ka is dependent on both density and confining
stress level, in addition to the level of static shear a, which
is defined as the ratio of initial shear stress to initial confin-
ing stress. The variation of Ka with a reported by various re-
searchers over the years is shown in Fig. 1. The differences
noted in the figure are reflective of the significant uncertain-
ties that still exist in accounting for static shear using Ka cor-
rection factors. The summary report of the National Center
Can. Geotech. J. Vol. 48, 2011
for Earthquake Engineering Research (NCEER) workshop,
which noted that “although curves relating Ka to a have
been published, these curves should not be used by nonspe-
cialists in geotechnical earthquake engineering or in routine
engineering practice,” attests to this uncertainty (Youd et al.
2001). The Greater Vancouver Seismic Geotechnical Design
Task Force (GVSGDTF) set up in Vancouver, British Colum-
bia, to study the implications of the new National Building
Code of Canada (NRCC 2005) has also recommended not
using a Ka correction factor on account of the uncertainties
associated with the determination of Ka (GVSGDTF 2007).
These recommendations clearly highlight the need for further
insights into the effects of static shear on the cyclic resistance
of sands.
Cyclic simple shear versus cyclic triaxial tests
The stress and strain conditions during cyclic loading dif-
fer significantly between simple shear and triaxial tests. The
initial stress condition in simple shear is K0 (i.e., no lateral
strain during consolidation). The Static shear stress ratio a in
simple shear is defined as the ratio of the initial (static) shear
stress on the horizontal plane to the initial vertical effective
consolidation stress. Triaxial specimens could be isotropically
or anisotropically consolidated. Static shear stresses in triax-
ial tests are introduced on account of anisotropic consolida-
tion, and a is based on the shear and normal stresses on the
45° plane.
Plane strain conditions prevail in cyclic simple shear, and
axisymmetric conditions in triaxial testing. Simple shear
loading causes continuous rotation of principal stress axes,
whereas triaxial loading may involve a jump rotation of 90°
if stress reversal occurs. Shear stress reversal will not occur
if the static shear stress is higher than the cyclic shear stress.
The cyclic shear stress is applied on the horizontal plane in
simple shear, whereas the maximum shear stresses in the tri-
axial test occur at planes inclined at 45° to vertical. Further,
the cyclic stress ratio (CSR) is generally defined as the ratio of
the peak cyclic shear stress to the vertical effective consolida-
0
tion stress (tcy/s vc ) in simple shear, but as the ratio of the peak
0
shear stress to the lateral effective confining stress (sd,cy/2s 3c )
in triaxial testing. The cyclic resistance ratio in field prob-
lems is normally obtained by normalizing the cyclic shear
stress by the vertical effective stress. These definitions have
limitations from a mechanics point of view because the cal-
culated stress ratios correspond to different planes, but they
reflect the current state of practice.
Undrained behaviour of sands is stress path dependent.
Under identical initial conditions a sand may strain-harden
under a given loading mode (e.g., triaxial compression), but
strain-soften and be highly susceptible to liquefaction under
another (e.g., triaxial extension). Therefore, the most credible
assessment of liquefaction susceptibility would be obtained if
the soil element in the laboratory would be loaded along the
stress path anticipated in the field. The loading conditions in
a cyclic simple shear test closely duplicate the stress condi-
tions in soil elements in situ due to vertically propagating
shear waves during an earthquake. Even though cyclic simple
shear is a more realistic representation of sand behaviour
under earthquake loading, liquefaction resistance in the labo-
ratory is often determined using cyclic triaxial tests. The
measured cyclic triaxial resistance is then corrected using
Published by NRC Research Press
 Sivathayalan and Ha
Fig. 1. Range of Ka values reported in the literature: (a) Seed and Harder (1990); (b), Harder and Boulanger (1997); (c) Vaid and Chern
1985; (d) Vaid et al. 2001. Dr, relative density; s 00 , effective overburden stress; N1,60, normalized standard penetration blow count.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
For personal use only.
empirical correction factors to obtain equivalent simple shear,
or in situ response. Several researchers have suggested corre-
lations between the cyclic resistance in simple shear to that in
triaxial for specimens with no static shear (Finn et al. 1971;
Seed and Peacock 1971; Vaid and Sivathayalan 1996). The
influence of state variables on cyclic resistance has been gen-
erally obtained from the measured cyclic triaxial response.
Importantly, attempts to assess the effects of static shear on
cyclic resistance, and establish Ka factors, have been primar-
ily based on cyclic triaxial tests. The effect of loading mode,
if any, on the Ka correction factor has not been addressed in
the literature to date.
Experimental aspects
Monotonic and cyclic simple shear tests were carried out
using an Norwegian Geotechnical Institute (NGI)-type simple
shear device (Bjerrum and Landva 1966) in the geotechnical
research laboratory at Carleton University. The simple shear
specimen was about 70 mm in diameter and 20 mm in
height, and was confined using a steel-wire reinforced rubber
membrane. The specimen was normally consolidated to the
desired effective stress state, and the cyclic shear stress was
1473
applied under constant volume conditions. The pore pressure
in constant volume simple shear tests is always atmospheric,
and thus the change in total vertical stress during shear
equals the excess pore pressure generated in an equivalent
undrained test (Dyvik et al. 1987). All cyclic tests reported
herein were carried out under stress-controlled loading
mode. A limited number of cyclic triaxial tests were carried
out using a stress–strain path capable triaxial device on speci-
mens of about 63 mm diameter × 125 mm height. The sim-
ple shear and triaxial devices used in the study are shown in
Fig. 2. All tests were conducted on normally consolidated
sands and no attempt was made to assess the potential effects
of overconsolidation.
Stress state in simple shear
The true stress state in a simple shear specimen cannot be
determined based on the stress measurements. Several re-
searchers have attempted to determine the stress state in sim-
ple shear, some by using highly instrumented simple shear
devices (Roscoe 1970; Airey et al. 1985; Airey and Wood
1987), and others using hollow cylinder torsional shear devi-
ces to simulate simple shear deformation (Shibuya and Hight
Published by NRC Research Press
 1474
Fig. 2. (a) Triaxial and (b) simple shear devices at Carleton University.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
1987; Pradhan et al. 1988; Ampadu and Tatsuoka 1993).
These findings suggest that interpretation of the test data,
based on the presumption that the horizontal plane is the
plane of maximum shear stress, yields acceptable results.
This assumption permits the construction of the Mohr circle,
and thus fully defines the stress state.
For personal use only.
The overall stress conditions in simple shear are not very
uniform, but it has been demonstrated that the stress condi-
tions in the middle third of the specimen are fairly uniform,
and equal to average stresses measured (Airey and Wood
1987). While the simple shear device is not well suited to in-
vestigate the constitutive behaviour of soils on account of the
stress state uncertainties, it still provides a preferred means to
assess the in situ response of soils subjected to cyclic load-
ing. Finn and Vaid (1977) have previously demonstrated that
constant volume simple shear tests can be used to evaluate
the potential for liquefaction.
Test material No. 1: silica sand
Cyclic simple shear tests were carried out on a subrounded
silica sand that was dredged from the deposits of the Illinois
River in the USA. This sand is quite similar to the widely
used Ottawa sand in mineral composition, but was sourced
through a secondary supplier and has somewhat different gra-
dation that leads to different maximum and minimum density
states. It conforms to the American Society for Testing and
Materials (ASTM) specifications for standard test sand
(ASTM 2006). The soil fabric that ensues upon pluviation
normally has dilative tendencies, and as a result many sands
strain-harden even at the loosest deposited state attainable by
pluviation. This silica sand was chosen for this study due to
its strain-softening tendency over a range of density states
following deposition by pluviation. This was required to as-
sess the effects of strain-softening (versus strain-hardening)
on the measured Ka. This sand is very poorly graded with an
average particle size D50 of about 0.42 mm, and a uniformity
coefficient of 1.6. The maximum and minimum void ratios of
the sand determined according to ASTM standard tests meth-
ods (ASTM 2001a, 2001b) are 0.723 and 0.478, respectively,
and the specific gravity of the sand is 2.66.
Can. Geotech. J. Vol. 48, 2011
Test material No. 2: Fraser River sand
A limited number of simple shear tests were also carried
out on semi-angular Fraser River sand that was dredged
from Abbotsford, British Columbia. This sand underlies large
portions of the heavily populated Fraser Delta in British Co-
lumbia, Canada, and hence forms the strata of interest for
liquefaction susceptibility. Deltaic deposits like this are typi-
cal of many heavily populated regions of the world. The nat-
ural material was processed to remove the fine particles
passing the No. 200 sieve and those retained in the No. 20
sieve. This provides a fairly uniform sand with a mean diam-
eter of 0.30 mm and uniformity coefficient of 2.7. Similar
material has been used in several past studies reported in the
literature (Vaid and Thomas 1995; Vaid and Sivathayalan
1996; Sivathayalan and Vaid 2002; Wijewickreme et al.
2005). The maximum and minimum void ratio of this batch
of Fraser River sand determined according to the ASTM
(2001a, 2001b) standard test methods are 0.806 and 0.509,
respectively. While the mineral composition of this sand is
similar to the various batches of Fraser River sand discussed
in the literature, the differences in gradation and geographical
origin cause fairly significant changes in the maximum and
minimum void ratio. Such changes for different batches of
Fraser Delta sands can be noted in the literature. It can be
observed from the uniformity coefficients that both tested
sands are very poorly graded materials. Such uniformity is
essential for fundamental laboratory studies that require sev-
eral repeatable, homogeneous specimens to be reconstituted
in the laboratory.
Sample preparation
Simple shear specimens were reconstituted using the air
pluviation technique, and triaxial specimens were reconsti-
tuted using the water pluviation technique. The air-pluviation
technique is much simpler, and yields a convenient means of
assessing the liquefaction potential in simple shear due to the
imposed constant volume conditions. The water-pluviation
technique was used to constitute the triaxial specimens be-
cause of the saturation requirement.
Even though these two techniques yield different fabrics in
sands with fines (Collins and McGown 1974), they produce a
Published by NRC Research Press
 Sivathayalan and Ha
similar fabric in clean sands. It has also been reported in the
literature that the differences in the mechanical behaviour (in
stress-strain, pore pressure space) of dry- and wet-pluviated
specimens decrease as the sands get cleaner (Vaid et al.
1999). As the silica sand, which was tested in simple shear
and triaxial loading modes, is very uniform with no measura-
ble amount of fines (0% finer than 0.075 mm; 0.2% finer
than 0.125 mm), the different media in which the soil was
pluviated is not expected to affect either the ensuing fabric
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
or the response.
In addition, there is direct evidence in the literature to in-
dicate that comparisons of dry- and water-pluviated speci-
mens of the silica sand is rational. Finn et al. (1978) made
direct comparisons of the behaviour of air- and water-pluviated
Ottawa sand, and concluded that the response is not de-
pendent on the medium of pluviation. As noted earlier, the
silica sand tested herein is fairly similar to Ottawa sand,
and hence their finding is direct evidence that both of these
preparation techniques yield identical fabric in this sand,
and it is reasonable to compare the responses of air- and
water-pluviated sand to assess the effect of loading modes.
Both of these specimen reconstitution techniques are well
known to yield uniform and repeatable specimens. Specimens
pluviated in water have been shown to mimic the behaviour
of natural fluvial and hydraulic fill sands, both under static
and cyclic loading (Vaid et al. 1999). Thus, the results of
For personal use only.
this study are expected to be applicable to water deposited in
situ sand masses.
All specimens were initially pluviated at the loosest possi-
ble state. Denser specimens were obtained by applying low-
energy, high-frequency vertical vibrations under a small seat-
ing load. Void ratios of the specimens were confidently deter-
mined using the volume of the cavity and mass of the solids.
The load cells are mounted inside the frame, and the linear
variable displacement transducers on the outside in this de-
vice (Fig. 2). Cyclic shear stresses were applied by an elec-
tropneumatic transducer at a frequency of 0.1 Hz, and the
data recorded using a high-resolution analog to digital (A/D)
card at a rate of 64 data points per cycle. This permits an ex-
amination of the mechanism responsible for the development
of strains within the loading cycle. All stresses were meas-
ured with resolutions of better than 0.2 kPa and strains with
resolutions of better than 10–4 in simple shear. The resolution
of stress measurements is better than 0.1 kPa, and strain
measurements about 10–5 in triaxial tests.
The cyclic shear stress ratio, CSR, defined as the ratio of
the cyclic shear stress amplitude to that of the initial vertical
effective confining stress, is used to quantify the level of the
cyclic loading. Several specimens were loaded at different
levels of CSR until liquefaction to yield a cyclic resistance
curve (number of cycles to liquefaction versus CSR) for a
given density and stress level. The cyclic resistance ratio,
CRR, of the sand at a given density, confining stress, and
static shear stress was then obtained from this data.
Liquefaction criteria
The term liquefaction is used herein to refer to all forms of
large deformation without regard to the actual strain develop-
ment mechanism. Simple shear specimens were deemed to
have liquefied when the single amplitude shear strain ex-
ceeded 3.75%, and triaxial specimens were deemed to have
1475
liquefied when the single amplitude axial strain exceeded
2.5%. This definition of liquefaction is often used in labora-
tory testing following the recommendations of NRC (1985).
A pore-pressure based liquefaction criterion is not suitable to
define the onset of liquefaction in laboratory tests because it
would not be applicable to sands that exhibit a flow failure
type of deformation, even though such materials represent
the most critical risk. The CRR reported herein corresponds
to liquefaction in 10 cycles, which represents a magnitude
M = 6.75 earthquake (Seed et al. 1975).
Test results and discussion
Several monotonic and cyclic simple shear tests were car-
ried out at different initial stress and density states. The mon-
otonic behaviour of the loosest deposited silica sand was
assessed at confining stress levels of 100, 200, and 400 kPa,
and at various levels of static shear a ranging from 0 to 0.3.
In addition, the behaviour was assessed at denser initial states
at various values of a, but at a fixed confining stress level of
100 kPa. A value of a = 0 represents a traditional simple
shear test (with no static shear). Cyclic resistance of the
sands was characterized over a range of density and initial
stress states.
The relative density, Drc, of the loosest deposited silica
sand following consolidation to 100, 200, and 400 kPa effec-
tive vertical stress in simple shear was about 25%, 28%, and
35%, respectively. All tests on Fraser River sand were con-
ducted at a confining stress level of 100 kPa, and the loosest
density attainable at this stress level was about 20% under tri-
axial conditions. The loosest attainable density is somewhat
dependent on the level of static shear as well, but the var-
iance is not significant. All triaxial specimens were fully sa-
turated prior to shear, and this was confirmed by ensuring a
Skempton’s B-value of 0.99 or better in all triaxial tests. The
degree of saturation is not applicable in dry-pluviated sands
(technically zero), nor is there a concern in simple shear due
to the enforcement of constant volume conditions during
shear.
Undrained monotonic behaviour
Monotonic response of the sands is assessed to provide a
baseline for evaluating the characteristics of the cyclic re-
sponse. Unlike the Fraser Delta sand, which is known to
strain-harden in triaxial compression and only marginally
strain-soften in simple shear loading even at the loosest de-
posited state attainable by pluviation, the loosest deposited
silica sand exhibited significant strain-softening regardless of
the initial effective stress state. It exhibited marginal strain-
softening at denser states depending on the static shear stress
state as shown in Fig. 3. Increasing static shear stresses ap-
pear to promote further strain-softening at a given density
and confining stress level.
Figure 4 compares the effect of shear stress reversal during
monotonic loading. Sand specimens were consolidated to an
identical initial stress state (with a static shear stress of
10 kPa). In one test the specimen was sheared in the same
direction as the static shear stress, and in a companion test
the shear loading was applied in the opposite direction. Nor-
malised peak shear strength is about 0.17 regardless of the
relative density when there is no stress reversal. It varied be-
Published by NRC Research Press
 1476
Fig. 3. Monotonic undrained behaviour of silica sand at different density (Drc) and initial static shear stress (a) levels.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
For personal use only.
tween 0.12 to 0.15 when the shear loading caused stress re-
versal. Somewhat smaller peak strengths mobilize when the
material is sheared in a direction opposing the static shear
compared to shearing in the same direction, even though
much larger principal axes rotation takes place under such
loading conditions. The influence of relative density on peak
shear strength is relatively insignificant compared to its influ-
ence on minimum strength at this initial state. Normalised
minimum undrained strength increases with increasing rela-
tive density of the sand. However, it is not influenced by the
direction of the undrained loading with respect to the direc-
tion of the initial static shear. Essentially similar minimum
strength values were noted at a given density state, when un-
drained shear simply increased the initial static shear, or
when it reduced the initial static shear to zero and loaded the
sand in the opposite direction.
The modified brittleness index, IB′, which is defined as the
ratio of the difference between the peak and minimum
strength to the difference between the peak and static shear
stress, is an indicator of the degree of strain-softening (Siva-
thayalan and Vaid 2002). IB′ is considered to be equal to zero
for soils that do not strain-soften. An IB′ value of greater than
Can. Geotech. J. Vol. 48, 2011
1.0 indicates that the minimum strength following the peak is
smaller than the static shear stress, and implies high damage
potential. The variation of the modified brittleness index for
silica sand at various initial stress and density states is shown
in Fig. 5. It can be seen that sand strain-softened over a large
range of initial density states, and that increasing static shear
increased the brittleness of the material. The sand with no
static shear does not strain-soften at a relative density of
70%. This is consistent with the expected response of the ma-
terial at such high relative density. However, the sand exhib-
its strain-softening response even at 70% relative density,
when the initial stress state includes a small static bias of
a = 0.1. The degree of strain-softening is clearly dependent
on the level of static bias, and IB′ of about 0.3 can be noted
at 70% relative density when a = 0.10. The sand is very brit-
tle at higher a values of 0.2 and 0.3. It realises much smaller
minimum undrained strength compared to the static shear
stress in these cases (IB′ varying from about 3 to 6.5 for a
values of 0.2 and 0.3 at the loosest state). Such an increase
in strain-softening tendency has also been noted in hollow
cylinder tests on anisotropically consolidated sands (Siva-
thayalan and Vaid 2002).
Published by NRC Research Press
 Sivathayalan and Ha
Fig. 4. Monotonic undrained behaviour of silica sand subjected to
shearing with and without stress reversal. Du, excess pore-water
pressure; s′, vertical effective stress; fPT , phase transformation fric-
tion angle.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
For personal use only.
It is well recognized that the deposition process adopted
leads to a repeatable, but inherently anisotropic, fabric in
these sands. This initially consistent fabric of the sand is
somewhat altered due to the application of initial static shear
(induced anisotropy). In addition, the initial state of sands
subjected to static shear is closer to the instability zone (or
within the instability zone depending on the level of static
shear) compared to specimens not subjected to static shear. As
a result, a small undrained perturbation might initiate a sudden
flow deformation in the sand. This is critical in slope stability
problems, as a catastrophic failure could be triggered due to a
relatively minor undrained disturbance. The risk of such failure
increases with increasing static shear stress or slope angle.
Figure 6a shows that at a given initial density state, the
friction angle mobilized at the peak shear strength state is de-
pendent on the initial stress state. This friction angle is gener-
ally termed fcsr in the literature to signify that it corresponds
to a critical stress ratio at which the material would collapse.
This state represents the onset of instability and a limit of dif-
fuse bifurcation where the incremental second order work
(ds × d3) becomes negative (Vardoulakis and Sulem 1995),
1477
Fig. 5. Variation of modified brittleness index showing the effect of
static bias on the degree of strain-softening. Smin, minimum peak
shear strength; Speak, peak shear strength; Sstatic, static shear stress.
which leads to a loss of control in stress-controlled tests.
The dependence of fcsr on the relative density and loading
mode has been noted in several publications in the literature
(Vaid and Thomas 1995; Vaid and Sivathayalan 1996). The
results reported herein show that it is dependent on the ini-
tial stress state as well.
In contrast, the friction angle mobilized at the state of min-
imum strength can be noted to be essentially a constant re-
gardless of the initial density or stress state in Fig. 6b. In
addition to the data points corresponding to the state of mini-
mum strength in strain-softening sand, data corresponding to
the peak pore-pressure state in strain-hardening sand are also
shown in Fig. 6b. The test results support the contention that
both quasi-steady state (the state at minimum strength in
strain-softening sand) and phase transformation (state of
peak pore pressure in strain-hardening sand) can be treated
together in the stress space. However, such treatment cannot
be extended to void ratio – stress space because the mobi-
lized shear strength at a given void ratio is dependent on the
nature of the response and the initial stress state. The data
presented support the conclusion drawn in previous studies
that the friction angle at this state is essentially a unique
property of the material.
The static behaviour of Fraser River sands has been well
established in the literature (Vaid and Thomas 1995; Vaid
and Sivathayalan 1996; Vaid et al. 2001). Unlike the tested
silica sand, Fraser Delta sand is only marginally strain-softening
at the loosest state in simple shear loading, and is dilative
at higher initial densities. Reported values of the brittleness
index of Fraser River sand in simple shear loading have
been smaller than 0.2 for a values of up to 0.1, which
clearly indicates that strength degradation is fairly small in
Fraser Delta sands even if they strain-soften (Vaid and Siva-
thayalan 1996). In contrast, much larger IB′ values (up to
about 1.1 for a = 0.1, and much higher when a equals 0.2
or 0.3) can be noted in the data presented in Fig. 5 for
silica sand.
Published by NRC Research Press
 1478
Fig. 6. The variation of (a) mobilized friction angle at peak state with a and (b) effective stress states at the instant of the peak pore pressure.
fCSR , friction angle at critical stress ratio.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
Undrained cyclic behaviour
Different mechanisms of strain development have been
found to be responsible for the triggering of liquefaction
under cyclic loading. The initial stress and density state, and
the relative values of the static and cyclic shear stress ampli-
For personal use only.
tudes, dictate the mechanism responsible for strain develop-
ment. Strain-softening was observed in silica sand under
most initial conditions, except at the densest states. However,
Fraser River sand was mostly dilative and exhibited only
marginal strain-softening even at the loosest state. Typical
stress–strain and the stress path under cyclic loading where
the cyclic shear stress amplitude is smaller than the static
shear bias are shown in Fig. 7 for “loose” (Drc ≈ 30%) and
“dense” (Drc ≈ 70%) specimens of silica sand. The effective
stress state reaches the critical stress ratio line during the
sixth cycle in the loose sand (Fig. 7c). Even though only
negligible levels of strains develop up to this instant, shear
strains accumulate progressively with the cycles of loading
(Fig. 7b). Sudden flow deformation, associated with strain-
softening, ensues at this state and very large strains develop
as a consequence. As cyclic loading was applied under a
stress-controlled mode, the control of the tests was lost until
the material strain-hardened upon reaching the quasi-steady
state or an unloading pulse was applied. Such flow deforma-
tion takes place within a short time, typically less than 1 s.
Subsequent loading cycles cause progressively larger cyclic
strains. The excess pore pressure generated does not reach
100% of the initial effective stress, but oscillates between
about 75%–85% during the loading cycles following liquefac-
tion. It should be noted that shear stress reversal is a prereq-
uisite for the development of 100% excess pore pressure, and
therefore specimens with static shear bias higher than the cy-
clic shear stresses cannot realize a state of zero effective
stress. Nonetheless, the specimens can develop very large
shear strains leading to flow deformation. Given the defini-
tion of liquefaction adopted herein following the NRC
(1985) recommendation, this specimen was deemed to have
liquefied in six cycles.
The denser sand shown in Fig. 7 marginally strain-softens
during the second cycle, but the resulting deformation is
Can. Geotech. J. Vol. 48, 2011
somewhat limited due to the unloading pulse. It realises a
shear strain larger than 3.75% during the third cycle, and
hence was deemed to have liquefied in three cycles even
though it had only developed about 70% excess pore pres-
sure. As noted in Fig. 7d, strain development following lique-
faction occurs fairly gradually, and each load cycle after
reaching the phase transformation (PT) line accumulates pro-
gressively smaller cyclic shear strain. As the material strad-
dles the PT line it goes through contractive and dilative
phases of deformation, which leads to a quasi-stable state of
strain development. The material is subjected to cyclic mobi-
lity without going through a state of zero effective stress in
both loose and dense sands. The looser sand reaches the crit-
ical stress ratio or collapse line during the sixth cycle and
suddenly mobilizes a relatively larger amount (∼7%) of
strain, whereas strain mobilization is fairly gradual in the
dense sand. However, the response mechanism following ini-
tial liquefaction is similar in both cases, but each stress cycle
causes much larger strain accumulation in the looser sand
(about 2% per cycle in loose sand compared to about 0.3%
per cycle in the dense sand). Cyclic mobility in scenarios in-
volving stress reversal (cyclic shear stress amplitude larger
than static bias) led to transient states of zero effective stress,
and hence progressively increasing levels of cyclic shear
strain in a manner similar to that noted in Vaid et al. (2001).
Several such tests were carried out at different density and
cyclic stress ratio levels to yield a void ratio versus number
of cycles to liquefaction plot at constant cyclic stress ratio
contours. Figure 8 shows the cyclic resistance of silica sand
consolidated to an effective confining stress of 100 kPa, and
a static shear stress of 20 kPa in the form of void ratio ec (or
relative density Drc) versus number of cycles to liquefaction.
Each contour in this figure represents a constant amplitude of
cyclic stress ratio. Strain development in specimens with rel-
ative density less than about 75% was due to contractive
0
strain-softening for this initial condition (s vc = 100 kPa, a =
0.20), and only a few specimens did not undergo strain-soft-
ening deformation during cyclic loading. These specimens
are identified by the shaded zone in the figure. As noted
under monotonic loading, the contractive tendency of this
Published by NRC Research Press
 Sivathayalan and Ha
Fig. 7. Typical response of sand subjected to cyclic shear stress smaller than the static shear stress: (a–c), loose sand; (d–f), dense sand. ec,
void ratio; PT, phase transformation; tst, static shear stress.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
For personal use only.
subrounded sand at initial states defined by high static shear
stresses is fairly significant, and as a result the sand strain-
softens even at relative density states of about 75%–80%.
The static shear stress amplitude is larger than the cyclic
shear stress amplitude for all the tests shown in Fig. 8, and
hence cyclic mobility without transient states of zero effec-
tive stress was responsible for strain development in dilative
(non strain-softening) sands.
It can also be noted in Fig. 8 that silica sand could be
liquefied by subjecting it to fairly small levels of cyclic shear
stresses (as low as CSR = 0.02) at loose states. Such small
undrained perturbations from the initial state trigger strain-
softening deformation leading to flow failure. This occurs be-
cause the highly contractive sand is already subjected to a
1479
relatively large static shear under drained conditions during
consolidation, and when undrained conditions are imposed
the strength of the material may become smaller than the
static shear.
Ka correction factor
The data shown in Fig. 8 is used to estimate the consoli-
dated relative density (or void ratio) at which the sand would
liquefy in 10 cycles of loading for a specified cyclic stress
ratio. This would relate the relative density and cyclic resist-
ance ratio CRR of the sand for liquefaction induced in 10
0
cycles to the specific s vc and a values. Data similar to that
in Fig. 8, but at other a values, were established to yield the
cyclic resistance curves of the sand at a given effective con-
Published by NRC Research Press
 1480
Fig. 8. Dependence of number of cycles to liquefaction on density at
a given initial stress state.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
0
Fig. 9. Cyclic resistance curves for silica sand at s vc = 200 kPa and
different a levels for an M = 6.75 earthquake. NL, number of cycles
to liquefaction.
For personal use only.
0
solidation stress level s vc , but at different density and a lev-
els. Figure 9 shows the cyclic resistance curves for silica sand
consolidated to 200 kPa consolidation stress, but to different
levels of static shear. The data clearly shows a profound in-
fluence of a on the cyclic resistance of the sand except at
very dense states. An essentially constant CRR of about 0.15
was noted in specimens consolidated to 200 kPa regardless
of a, but only at a relative density state of about 85%. In
practical terms, the cyclic resistance curves given in Fig. 9
specify the density required to withstand a specific “equiva-
lent” cyclic stress ratio induced at a site due to a magnitude
M = 6.75 earthquake, which is expected to produce 10 stress
cycles at the design load level. The equivalent cyclic stress
ratio is often related to the peak ground acceleration using
Seed’s simplified procedure (Seed et al. 1975) in practice.
Similar cyclic resistance curves were established at confining
stress levels of 100 and 400 kPa as well. This permitted the
Can. Geotech. J. Vol. 48, 2011
determination of the Ka correction factor at various levels of
confining stress, densit,y and static shear stress (Ha 2004).
The variation of the Ka correction factor with a is shown in
Fig. 10a at different density levels for samples consolidated
to 100 kPa. Significant reduction in the cyclic resistance of
the silica sand can be noted at looser density states. At the
loosest deposited state, the sand subjected to a = 0.20 pos-
sesses only about 22% of the cyclic resistance compared to
the sand that is not subjected to any static shear stresses.
The reduction in Ka factor is less pronounced at higher con-
solidated density states, but the presence of static shear al-
most always decreases the cyclic resistance in this sand
regardless of the density and the level of static shear. The Ka
factor is marginally greater than 1 only at the relative density
state of 80%. Sands at such high density states are generally
considered less prone to liquefaction.
Similar plots obtained for consolidation stress levels of 200
and 400 kPa shown in Figs. 10b and 10c, respectively, show
that the Ka factor at a given density and a level decreases
with increasing confining stress level. Significant reductions
in cyclic resistance due to increasing static shear can be
noted even at 80% relative density when the sand is consoli-
dated to 400 kPa. At a given relative density level, the CRR
versus a relationship is profoundly influenced by the confin-
ing stress level. The reduction in Ka with increasing stress
levels at a given density and a level is postulated to be a re-
flection of the reduced dilatancy at the higher stresses. The
measured Ka is marginally larger than 1, only for the densest
state at the lowest confining stress level. Vaid et al. (2001)
noted similar dependence of the Ka factor on confining stress
level, but under cyclic triaxial loading, and recommended
that a combined correction factor Kas be used instead of two
single factors Ka and Ks, as adopted in current practice.
Similar cyclic tests were carried out on Fraser Delta sands
under a simple shear loading mode at 100 kPa effective con-
solidation stress and at three different density levels. The Ka
values obtained for Fraser Delta sands were generally much
higher than those obtained for the silica sand under similar
initial states. The Ka values in Fraser Delta sands were often
larger than 1, which indicates that the presence of static shear
increases the cyclic resistance of Fraser Delta sands, espe-
cially at medium-dense or dense states. A similar response
of Fraser Delta sands has been reported by Vaid et al.
(2001) based on cyclic triaxial tests, where Ka values were
close to or larger than 1. The relatively high angularity in
Fraser River sand particles (compared to silica sand) leads to
higher initial anisotropy, and a more dilative fabric. The ini-
tial anisotropy is further altered by the anisotropy induced
during the application of static shear (Wong and Arthur
1985) and the degree of change for a given stress path would
be dependent on the initial fabric.
The variation of Ka in silica and Fraser Delta sands (con-
solidated to the same stress level) is compared in Fig. 11a
under simple shear loading. Dramatic differences can be
seen in the Ka values determined for the two sands. The pres-
ence of static shear stress significantly reduces the cyclic re-
sistance in silica sand at all density states. On the other hand,
the presence of static shear only marginally decreases the cy-
clic resistance of the loosest deposited Fraser Delta sand.
Further, static shear increases the cyclic resistance by as
much as 50% at denser states in Fraser River sand. It appears
Published by NRC Research Press
 Sivathayalan and Ha
0
Fig. 10. Variation of Ka with a and relative density at s vc =
(a) 100 kPa, (b) 200, and (c) 400 kPa.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
For personal use only.
that contractive soils would see a significant reduction in cy-
clic strength and dilative soils would see a marginal reduction
or an increase in cyclic resistance depending on the density
level. These results clearly indicate that the role played by
static shear stresses on cyclic resistance is dependent on the
material itself, in addition to density and confining stress lev-
els. The results suggest that Ka could be much smaller than
1e in contractive strain-softening sands, but is often greater
than 1 in dilative strain-hardening sands. Differences in the
level of dilatancy, occurring on account of different soil fab-
ric, control excess pore-pressure generation, and thus the cy-
clic resistance. The fabric of the material is dependent on the
particle gradations and shapes for a given deposition mecha-
nism. As angular sands are often much more dilative, it can
be argued that increasing static shear would decrease the cy-
clic resistance of subrounded to rounded sands, but will in-
crease the cyclic resistance of angular sands.
1481
Fig. 11. Variation of Ka with soil type in (a) simple shear loading
0
and (b) triaxial loading. s 1c , major principal consolidation stress;
0
s 3c , minor principal consolidation stress.
The fact that the Ka value at a given relative density could
be drastically different depending on the type of response (i.e.,
strain-softening versus strain-hardening) strongly favours the
argument that the attempts to quantify Ka should not be re-
lated to the relative density, but rather to the type of re-
sponse of the soil at the given relative density. Vaid et al.
(2001) suggested the use of a combined Kas factor, while
Boulanger (2003b) suggested relating Ka to the relative
state parameter index to account for such dependence.
All data discussed previously were obtained from cyclic
simple shear tests. However, a significant amount of the re-
search on Ka values in the literature has been conducted us-
ing cyclic triaxial tests. A limited number of cyclic triaxial
tests were conducted on saturated silica sand in this study. In-
itial static bias in triaxial loading is simulated by subjecting
the specimen to nonhydrostatic consolidation stresses. As
noted earlier, this will produce the maximum shear stresses on
a 45° plane in contrast to the horizontal plane in simple shear.
Published by NRC Research Press
 1482 Can. Geotech. J. Vol. 48, 2011

Comparative values of Ka under cyclic triaxial loading for Fig. 12. Dependence of Ka on loading mode in semi-angular Fraser
the two sands are shown in Fig. 11b. The data shown corre- River sand. SS, simple shear; Triax, triaxial; s c0 , effective consolida-
sponds to a confining stress of 100 kPa and a range of a val- tion stress.
ues at two different density states. Ka values for Fraser River
sand under triaxial loading conditions have been obtained
from the literature (Vaid et al. 2001). Even though available
data is limited, Fig. 11b clearly shows that dilative Fraser
River sand will yield higher Ka values compared to silica
sand under triaxial loading conditions as well.
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23

Data shown in Figs. 11a and 11b can also be used to com-
pare the effect of loading mode on Ka for a given sand at a
given density and initial stress state. Figure 12 compares the
Ka values obtained using simple shear tests on Fraser Delta
sand to those obtained using cyclic triaxial tests. All speci-
mens were consolidated to 100 kPa effective normal stress
on the plane of maximum shear stress. The loading mode in-
fluences the determined Ka values regardless of the density
or static shear stress levels. On average the Ka values deter-
mined using cyclic triaxial tests is about 50% higher than
those determined using cyclic simple shear tests in Fraser
River sand. A comparison of the Ka values measured in tri-
axial and simple shear loading modes for the strain-softening
silica sand also show similar dependency of Ka on loading Fig. 13. Dependence of Ka on loading mode in subrounded silica
mode (Fig. 13). Even though the available data is limited, sand.
the Ka values determined from the axisymmetric cyclic triax-
ial tests are clearly higher compared to those determined
For personal use only.

from the plane-strain simple shear tests at comparable density

states for the contractive silica sand. These results show that
regardless of whether a sand is strain-hardening or strain-
softening, the Ka values obtained from cyclic triaxial loading
are larger than those obtained from cyclic simple shear. As
noted earlier,the cyclic simple shear test simulates very
closely the loading conditions in situ due to vertically propa-
gating shear waves caused by an earthquake. As a result, the
Ka values determined under simple shear loading are ex-
pected to be more relevant and applicable to actual field
problems.
Cyclic triaxial tests are generally recognized to overesti-
mate the cyclic resistance, and correction factors (Cr) are ap-
plied to the triaxial resistance to represent the differences in
loading modes between triaxial and field conditions. These
Cr factors have been based on tests with no static shear
(Finn et al. 1971; Seed and Peacock 1971; Vaid and Siva-
thayalan 1996). The overestimations of CRR in triaxial tests cantly reduced cyclic resistance in strain-softening silica sand
appear to become more serious with increasing static shear, regardless of the density. Not accounting for the presence of
and hence the previously proposed Cr factors may not be ap- static shear in such materials may lead to potentially unsafe
propriate for soils subjected to initial static shear. The inher- designs. These results, while limited to two sands, suggest
ent anisotropy in natural and pluviated soils, which results in that the risks are much higher in strain-softening sands com-
higher stiffness in the vertical direction, is potentially one of pared to strain-hardening sands. In this light, the recommen-
the causes of the increased resistance in cyclic triaxial load- dation of the Vancouver Task Force (GVSGDTF 2007) to use
ing. In this regard, the worst discrepancies would be expected Ka = 1 can be regarded as a safe–conservative option be-
to occur when the static shear stress is larger than the cyclic cause the sand that underlies the heavily populated Fraser
shear stress, in which case the entire cyclic loading would be Delta is known to be generally dilative except at the loosest
fully confined to the stronger triaxial compression loading states. However, not using the Ka correction factor may lead
mode. These results suggest that cyclic triaxial tests on aniso- to significant and potentially unsafe errors in the estimated
tropically consolidated sands should not be used to evaluate cyclic resistance ratios in strain-softening sands. Practising
the cyclic resistance of soils subjected to static shear. engineers must recognize that it is imperative to consider the
The results presented herein demonstrate that the com- effects of static shear in rounded to subrounded sands regard-
monly held notion that static shear will increase the cyclic re- less of the relative density. The NCEER recommendation
sistance of “dense” sands is not always valid. It has been (Youd et al. 2001) that “published Ka values should not be
shown that the presence of static shear stress leads to signifi- used in routine engineering practice” should be interpreted

Published by NRC Research Press

 Sivathayalan and Ha
cautiously. Not using Ka values in strain-softening sands is
justified, only if the effects of static shear on cyclic resistance
are taken into consideration by other means, such as a direct
site-specific assessment using undisturbed specimens or a
comprehensive effective stress analysis of the undrained cy-
clic shear loading. While these test results suggest a link be-
tween Ka values and particle shape, it is acknowledged that the
scope of this testing program is limited to just two different
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
sands. Additional research on different sands is needed to con-
fidently establish a general dependency of Ka on particle shape.
Summary and conclusions
Experimental results of a series of cyclic simple shear tests
on a silica sand and Fraser River sand have been presented.
Cyclic resistance of the sands has been assessed over a range
of initial density and stress states to obtain the Ka correction
factors. Ka values are rarely greater than 1 in the subrounded
silica sand, which strain-softens over a large range of relative
density states at the tested stress levels. But Ka values are al-
most always greater than 1 in the semi-angular and strain-
hardening Fraser River sand. Drastically different Ka values
are obtained at the same relative density in both sands re-
gardless of the loading mode. It is postulated that the nature
of response (i.e., strain-softening versus strain-hardening),
dictated by dilatancy, soil fabric, and particle shape is the
For personal use only.
cause of such dramatic differences in the response of the
sands. These results suggest a possible reason for the large
scatter noted in current Ka charts, and indicate that relative
density should not be used to characterize the variation of Ka.
In addition, it has been demonstrated that loading condi-
tions play a key role in the determined Ka value. Significant
differences have been noted in the measured Ka values de-
pending on the laboratory cyclic shear test (triaxial versus
simple shear). As in situ loading conditions during earth-
quakes are well simulated by cyclic simple shear tests, Ka
values determined using cyclic triaxial tests should not be
used to assess the effect of static shear.
Estimation of Ka value in practice should not be based on
the relative density of the sand, but on the anticipated defor-
mation characteristics (strain softening versus strain harden-
ing). The use of Ka factors determined from cyclic triaxial
tests, or not using a Ka factor at all, may lead to unsafe de-
signs, especially in strain-softening sands. It appears that the
large scatter in the reported Ka values is a reflection of the
effects of various parameters on the cyclic resistance of soil
under initial static shear. The effect of increasing static shear
is much more serious compared to that of increasing confin-
ing stress at liquefaction-prone strain-softening sands. The ef-
fect of static shear on cyclic resistance should not be ignored
in geotechnical design if the material is strain-softening. The
tests were conducted on onlytwo sands, and further studies
are required to derive reliable generalized conclusions.
Acknowledgements
The research reported herein has been supported by grants
from the Natural Sciences and Engineering Research Council
of Canada, Canada Foundation for Innovation, and the On-
tario Innovation Trust. Financial assistance provided to the
second author by Carleton University; the technical assis-
tance of Ken McMartin, Stanley Conley, and Jim White-
1483
horne; and the thoughtful comments of the reviewers are
gratefully acknowledged.
References
Airey, D.W., and Wood, D.M. 1987. An evaluation of direct simple
shear tests on clay. Géotechnique, 37(1): 25–35. doi:10.1680/geot.
1987.37.1.25.
Airey, D.W., Budhu, M., and Wood, D.M. 1985, Some aspects of the
behaviour of soils in simple shear. In Developments in soil
mechanics and foundation engineering 2. Edited by P.K. Banerjee
and R. Butterfield. Elsevier Applied Science Publishers, New
York. pp. 185–213.
Ampadu, S.K., and Tatsuoka, F. 1993. A hollow cylinder torsional
simple shear apparatus capable of a wide range of shear strain
measurement. Geotechnical Testing Journal, 16(1): 3–17. doi:10.
1520/GTJ10262J.
ASTM. 2001a. Standard test methods for maximum index density and
unit weight of soils using a vibratory table. ASTM standard D-4253-
00. In Annual Book of ASTM Standards, Vol. 04-08. American
Society for Testing and Materials, West Conshohocken, Pa.
ASTM. 2001b. Standard test methods for minimum index density and
unit weight of soils and calculation of relative density. ASTM
standard D-4254-00. In Annual Book of ASTM Standards,
Vol. 04-08. American Society for Testing and Materials, West
Conshohocken, Pa.
ASTM. 2006. Standard specification for standard sand. ASTM standard
C-778. In Annual Book of ASTM Standards, Vol. 04.01. American
Society for Testing and Materials, West Conshohocken, Pa.
Bjerrum, L., and Landva, A. 1966. Direct simple-shear tests on a
Norwegian quick clay. Géotechnique, 16(1): 1–20. doi:10.1680/
geot.1966.16.1.1.
Boulanger, R.W. 2003a. High overburden stress effects in liquefac-
tion analyses. Journal of Geotechnical and Geoenvironmental
Engineering, 129(12): 1071–1082. doi:10.1061/(ASCE)1090-
0241(2003)129:12(1071).
Boulanger, R.W. 2003b. Relating Ka to relative state parameter index.
Journal of Geotechnical and Geoenvironmental Engineering, 129(8):
770–773. doi:10.1061/(ASCE)1090-0241(2003)129:8(770).
Boulanger, R.W., and Idriss, I.M. 2007. Evaluation of cyclic
softening in silts and clays. Journal of Geotechnical and
Geoenvironmental Engineering, 133(6): 641–652. doi:10.1061/
(ASCE)1090-0241(2007)133:6(641).
Collins, K., and McGown, A. 1974. The form and function of micro
fabric features in a variety of natural soils. Géotechnique, 24(2):
223–254. doi:10.1680/geot.1974.24.2.223.
Dyvik, R., Berre, T., Lacasse, S., and Raadim, B. 1987. Comparison
of truly undrained and constant volume direct simple shear tests.
Géotechnique, 37(1): 3–10. doi:10.1680/geot.1987.37.1.3.
Finn, W.D.L., and Vaid, Y.P. 1977. Liquefaction potential from
drained constant volume cyclic simple shear tests. In Proceedings
of the 6th World Conference on Earthquake Engineering, New
Delhi, India, 10–14 January 1977. Sarita Prakashan Publishers,
Meerut, India. Vol. 3, pp. 2157–2162.
Finn, W.D.L., Pickering, D.J., and Bransby, P.L. 1971. Sand
liquefaction in triaxial and simple shear tests. Journal of the Soil
Mechanics and Foundation Engineering Division, ASCE, 97(4):
639–659.
Finn, W.D.L., Vaid, Y.P., and Bhatia, S.K. 1978. Constant volume
cyclic simple shear testing. In Proceedings of the Second
International Conference on Microzonation for Safer Construction:
Research and Applications, San Francisco, Calif., 26 November –
1 December 1978. National Science Foundation, Washington, D.C.
Vol. 2, pp. 839–851.
Published by NRC Research Press
 1484
GVSGDTF. 2007. Geotechnical design guidelines for buildings on
liquefiable sites in Greater Vancouver in accordance with NBCC
2005. Task Force Report. Greater Vancouver Seismic Geotechni-
cal Design Task Force (GVSGDTF), Vancouver, B.C.
Ha, D. 2004. Effect of initial stress state on the undrained cyclic
behaviour of sands. MASc thesis, Carleton University, Ottawa, Ont.
Harder, L.F., Jr., and Boulanger, R.W. 1997. Application of Ks and Ka
correction factors, In Proceedings of the NCEER Workshop on
Evaluation of Liquefaction Resistance of Soils, Salt Lake City,
Can. Geotech. J. Downloaded from cdnsciencepub.com by MIDDLE EAST TECHNICAL UNIVERSITY on 08/30/23
Utah, 5–6 January 1996. Edited by T.L. Youd and I.M. Idriss.
National Center for Earthquake Engineering Research, State
University of New York at Buffalo, Buffalo, N.Y. pp. 167–190.
Hynes, M.E., and Olsen, R.S. 1998. Influence of confining stress on
liquefaction resistance. In Proceedings of the International Work-
shop on the Physics and Mechanics of Liquefaction, Baltimore, Md.,
10–11 September 1998. Edited by P.V. Lade and J.A. Yamamuro. A.
A. Balkema, Rotterdam, the Netherlands. pp. 145–152.
Idriss, I., and Boulanger, R.W. 2003. Estimating Ka for use in
evaluating cyclic resistance of sloping ground. In Proceedings of
the 8th U.S.–Japan Workshop on Earthquake Resistant Design of
Lifeline Facilities, and Counter Measures against Liquefaction,
Tokyo, Japan, 16–18 December 2002. Edited by M. Hamada, J.
Bardet, and T.D. O’Rourke. Multidisciplinary Center for Earth-
quake Engineering Research, State University of New York at
Buffalo, Bufalo, N.Y. pp. 449–468.
Lee, K.L., and Seed, H.B. 1967. Dynamic strength of anisotropically
consolidated sand. Journal of the Soil Mechanics and Foundations
For personal use only.
Division, ASCE, 93(5): 169–190.
NRC. 1985. Liquefaction of soils during earthquakes. National
Academic Press, Washington, D.C. Report No. CETS-EE-001.
NRCC. 2005. National building code of Canada. 12th ed. Canadian
Commission on Building and Fire Codes, National Research
Council, Ottawa, Ont.
Pradhan, B.S., Tatsuoka, F., and Horii, N. 1988. Simple shear testing on
sand in a torsional shear apparatus. Soils and Foundations, 28(2):
95–112.
Roscoe, K.H. 1970. The influence of strains in soil mechanics.
Géotechnique, 20(2): 129–170. doi:10.1680/geot.1970.20.2.129.
Seed, H.B. 1983. Earthquake-resistant design of earth dams. In
Proceedings of the Symposium on Seismic Design of Earth Dams
and Caverns, Philadelphia, Pa., 16–20 May 1983. Edited by T.R.
Howard. American Society of Civil Engineers, New York. pp. 41–64.
Seed, R.B., and Harder, L.F. 1990. SPT-based analysis of cyclic pore
pressure generation and undrained residual strength. In Proceed-
ings of H.B. Seed Memorial Symposium, University of California
Berkeley, Berkeley, Calif. Edited by J.M. Duncan. BiTech
Publishers, Vancouver, B.C. Vol. 2, pp. 351–376.
Seed, H.B., and Peacock, W.H. 1971. Test procedures for measuring
soil liquefactions characteristics. Journal of the Soil Mechanics
and Foundation Engineering Division, ASCE, 97(8): 1099–1119.
Seed, H.B., Idriss, I.M., Makdisi, F., and Banerjee, N. 1975.
Representation of irregular stress time histories by equivalent
uniform stress series in liquefaction analysis. Earthquake En-
gineering Research Centre, University of California Berkeley,
Berkeley, Calif. Report No. EERC75-29.
Can. Geotech. J. Vol. 48, 2011
Shibuya, S., and Hight, D.W. 1987. On the stress path in simple
shear. Géotechnique, 37(4): 511–515. doi:10.1680/geot.1987.37.4.
511.
Sivathayalan, S., and Ha, D. 2004. Effect of initial stress state on the
cyclic simple shear behaviour of sands. In Proceedings of the
International Workshop on Cyclic Behaviour of Soils and
Liquefaction Phenomena. Edited by T. Triantafyllidis. Taylor &
Francis Group, London. pp. 207–214.
Sivathayalan, S., and Vaid, Y.P. 2002. Influence of generalized initial
state and principal stress rotation on the undrained response of
sands. Canadian Geotechnical Journal, 39(1): 63–76. doi:10.1139/
t01-078.
Vaid, Y.P., and Chern, J.C. 1985. Cyclic and monotonic undrained
response of saturated sands. In Proceedings of Advances in Testing
Soils Under Cyclic Conditions, Detroit, Mich., 24 October 1985.
Edited by V. Khosla. American Society of Civil Engineers, New
York. pp. 120–147.
Vaid, Y.P., and Finn, W.D.L. 1979. Static shear and liquefaction
potential. Journal of the Geotechnical Engineering Division, 105(10):
1233–1246.
Vaid, Y.P., and Sivathayalan, S. 1996. Static and cyclic liquefaction
potential of Fraser Delta sand in simple shear and triaxial tests.
Canadian Geotechnical Journal, 33(2): 281–289. doi:10.1139/t96-
007.
Vaid, Y.P., and Thomas, J. 1995. Liquefaction and post liquefaction
behaviour of sand. Journal of Geotechnical Engineering, 121(2):
163–173. doi:10.1061/(ASCE)0733-9410(1995)121:2(163).
Vaid, Y.P., Sivathayalan, S., and Stedman, D. 1999. Influence of
specimen-reconstituting method on the undrained response of
sand. Geotechnical Testing Journal, 22(3): 187–195. doi:10.1520/
GTJ11110J.
Vaid, Y.P., Stedman, J.D., and Sivathayalan, S. 2001. Confining stress
and static shear effects in cyclic liquefaction. Canadian Geotech-
nical Journal, 38(3): 580–591. doi:10.1139/t00-120.
Vardoulakis, I.G., and Sulem, J. 1995. Bifurcation analysis in
geomechanics. Blackie Academic & Professional Publishers,
Bishopbriggs, Glasgow, UK.
Wijewickreme, D., Sriskandakumar, S., and Byrne, P. 2005. Cyclic
loading response of loose air-pluviated Fraser River sand for
validation of numerical models simulating centrifuge tests.
Canadian Geotechnical Journal, 42(2): 550–561. doi:10.1139/
t04-119.
Wong, R.K.S., and Arthur, J.R.F. 1985. Induced and inherent
anisotropy in sand. Géotechnique, 35(4): 471–481. doi:10.1680/
geot.1985.35.4.471.
Yoshimi, Y., and Oh-Oka, H. 1975. Influence of degree of shear
stress reversal on the liquefaction potential of saturated sand. Soils
and Foundations, 15(3): 27–40.
Youd, T.L., Idriss, I.M., Andrus, R.D., Arango, I., Castro, G.,
Christian, J.T., et al. 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, 127
(10): 817–833. doi:10.1061/(ASCE)1090-0241(2001)127:10(817).
Published by NRC Research Press