Bulletin of Earthquake Engineering (2019) 17:2899–2933

https://doi.org/10.1007/s10518-019-00592-6

ORIGINAL RESEARCH

Dynamic soil properties and liquefaction potential

of northeast Indian soil for non‑linear effective stress
analysis

Pradeep Kumar Dammala1   · Shiv Shankar Kumar2 · A. Murali Krishna1 ·

Subhamoy Bhattacharya3

Received: 19 April 2018 / Accepted: 2 March 2019 / Published online: 7 March 2019
© Springer Nature B.V. 2019

Abstract
This article presents a comprehensive study of dynamic soil properties [namely, initial
shear modulus-Gmax; normalized shear modulus reduction (G/Gmax); and damping ratio
(D) variation curves] and pore water pressure parameters of a river bed sand (Brahmaputra
sand), sampled from a highly active seismic region (northeast India). Two independent high
quality apparatus (resonant column-RC and cyclic triaxial-CTX) are adopted in the study.
Resonant column apparatus was used to obtain the small strain properties (up to 0.1%)
while CTX equipment was adopted to obtain the high strain properties along with the pore
water pressure parameters. The results obtained from both the equipment are combined
to provide a comprehensive data of dynamic soil properties over wide range of strains.
A modified hyperbolic formulation was suggested for efficient simulation of G/Gmax and
D variations with shear strain. Based on the CTX results, a pore water pressure genera-
tion model is presented. Furthermore, a nonlinear effective stress ground response study
incorporating the pore water pressure generation, is performed using the recorded earth-
quake motions of varying peak bed rock acceleration (PBRA) in the region, to demonstrate
the applicability of proposed dynamic soil properties and pore pressure parameters. High
amplification for low PBRA ground motions (< 0.10  g) was observed and attenuation of
seismic waves was witnessed beyond a PBRA of 0.10 g near the surficial stratum due to the
induced high strains and the resulting high hysteretic damping of the soil. Also, increased
excess pore pressure generation with increased PBRA of the input motion was observed
and the considered soil stratum is expected to liquefy beyond a PBRA of 0.1 g. The estab-
lished properties can be handy to the design engineers during seismic design of structures
in the northeast Indian region.

Keywords  Dynamic soil properties · Resonant column · Cyclic triaxial · Pore water
pressure · Nonlinear effective stress analysis

* Pradeep Kumar Dammala

dammala@iitg.ac.in; pradeepkumardammala@gmail.com
Extended author information available on the last page of the article

13
Vol.:(0123456789)
 2900 Bulletin of Earthquake Engineering (2019) 17:2899–2933

1 Introduction

Tremendous developments in seismic ground response studies have been initiated follow-
ing the 1985 Mexico earthquake which lead to massive destruction due to the soft soil
amplification and resonance (Esteva 1988). Many computer programs such as SHAKE,
EERA, DEEPSOIL, Cyclic1D etc. were developed and such programs require the user to
define strain dependent dynamic behavior of underlying soils. Dynamic behavior of soils
is traditionally represented using strain dependent dynamic soil properties and liquefaction
potential. Dynamic soil properties include initial shear modulus-Gmax; normalized shear
modulus reduction (G/Gmax); and damping ratio (D) variation. Such properties can either
be obtained by in  situ (field) testing or by laboratory element testing techniques. In-situ
testing can only provide the dynamic soil properties at low strain range and is not always
feasible, while laboratory tests can furnish the required properties over wide strain range
and varying loading conditions. Laboratory element testing apparatus such as resonant
column (RC), bender element (BE), cyclic torsional shear (CTS), cyclic triaxial (CTX),
dynamic simple shear (DSS), etc. have been used to determine the dynamic soil proper-
ties, by various researchers (Kramer 1996; Towhata 2008). Each apparatus is unique in its
application with different advantages and limitations.
Researchers (Hardin and Richart 1963; Chung et al. 1984; Maheshwari and Kirar 2017)
utilized RC apparatus to determine the dynamic behavior of sandy soils up to a shear strain
of 0.1%. Similarly, Kokusho (1980) and Sitharam et al. (2004) considered CTX apparatus in
determining the high strain behavior of sandy soils. However, any earthquake geotechnical
application involving dynamic soil behavior requires the properties over wide range of shear
strains (from very low strain range − 0.0001 to large strains − 5%). In this regard, El Mohtar
et al. (2013) and Chattaraj and Sengupta (2016) adopted two independent equipment (RC
and CTX) to obtain the dynamic soil properties individually and combined finally to provide
comprehensive properties over wide strain range. Therefore, present study considers two
independent apparatus (RC and CTX) to obtain the strain dependent dynamic soil properties
and liquefaction potential of a northeast Indian soil (Brahmaputra sand, BS).
Northeast India is classified as one of the highly active seismic regions of the country
(IS:1893 2002) due to its proximity to the Himalayan seismic belt in the north and the Indo-
Burmese Arc in the east (Kayal and De 1991). The region has witnessed more than 20 large
earthquakes (moment magnitude, Mw ranging from 6.0 to 8.0) and three great earthquakes
(Mw ≥ 8.0) in the last century. The great earthquakes (1897 Shillong earthquake and 1950
Assam earthquake) have caused widespread damages leading to excessive settlements,
structural failures and liquefaction triggers (Oldham 1899; Poddar 1950). Previous studies
pertaining to the dynamic response of soils of this region revealed the possibility of intense
liquefaction, severe amplification and excessive settlements (Raghu Kanth and Dash 2010;
Ayothiraman et  al. 2012). Seismologists (Guha and Bhattacharya 1984; Khattri 1999) are
of the opinion that this region is due to a large impounding earthquake in the near future.
Due to the rapid urbanization of this region and increasing seismic risk, strength and stiffness
properties of soils of this region are of vital importance for seismic design of new structures
or seismic requalification studies of existing structures (Dammala et al. 2017a). In view of
this, Dammala et al. (2015, 2017b) investigated the low strain dynamic behavior of soils of
this region using RC apparatus. Similarly, Govindaraju (2005) and Kumar et al. (2017) uti-
lized CTX apparatus for assessing the liquefaction susceptibility while Dammala et al. (2016)
adopted DSS apparatus to explore the dynamic behavior of soils of this region at high shear
strains. As there are no comprehensive studies covering wide strain dynamic soil properties of

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2901

this region, this article forms a connection to provide the required wide strain range dynamic
soil properties and liquefaction parameters of BS, determined by RC and CTX apparatus.
The article is arranged in the following order to achieve the formulated objectives.

1. The planned experimental programme is presented initially, in which two high quality
laboratory element testing techniques (RC and CTX) adopted are described.
2. Results obtained from the laboratory tests are presented in terms of shear modulus and
damping ratio along with liquefaction potential of BS.
3. Data from both the testing techniques are combined to provide comprehensive dynamic
soil properties over wide strain range. Simplified analytical formulations are proposed
to estimate the dynamic soil properties (Gmax, G/Gmax and D) at varying loading condi-
tions, based on the regression analysis performed on the test data. Pore water pressure
parameters representing the pore pressure generation are proposed.
4. Finally, a nonlinear effective stress ground response analysis is performed on a typical
soil profile to demonstrate the application of evaluated properties.

2 Experimental program

2.1 Index properties of chosen soil

The soil considered for the study is Brahmaputra sand (referred as BS hereafter) which
was collected from the shore of Brahmaputra River near Guwahati city in Assam (one
of the seven north-eastern states of India). Figure 1 presents the grain size distribution
of BS. It can be observed that the BS falls in the highly liquefiable boundaries suggested
by PHRI (1997), based on the gradation properties. Index properties of the soil were
determined according to the ASTM standards (ASTM D2487 2006) and are presented in
Table 1. The soil has been classified as poorly graded sand (SP) according to unified soil
classification system (USCS).

Fig. 1  Particle size distribution BS

of BS compared with liquefiable Boundaries for partial liquefiable zone (PHRI, 1997)
soil zones proposed in PHRI Boundaries for highly liquefiable zone (PHRI, 1997)
(1997) 100

80
Percent passing (%)

60

40

20

0
0.01 0.1 1 10
Particle size (mm)

13
 2902 Bulletin of Earthquake Engineering (2019) 17:2899–2933

Table 1  Index properties of Brahmaputra sand

Gs emax emin D10 (mm) D30 (mm) D50 (mm) D60 (mm) Cu Cc Fines (%) Symbol (USCS)

2.72 0.96 0.62 0.15 0.19 0.21 0.22 1.46 1.09 4.5 SP

Two types of advanced apparatus (resonant column-RC and cyclic triaxial-CTX)

were utilized to achieve the strain dependent dynamic properties and liquefaction poten-
tial of BS. Resonant column apparatus provides the reliable shear modulus and damping
ratio of soils up to a shear strain level of 0.1% (ASTM D4015 2014) while CTX can
furnish the required properties from 0.01 to 5% (ASTM D5311 1996). Table 2 lists the
testing programme using both the RC and CTX apparatus.

2.2 Resonant column (RC) testing

A fixed-free type (bottom fixed-top free) of the RC apparatus has been used for the low
strain testing (0.0001–0.1%). Wave propagation through prismatic rods is the basic prin-
ciple of testing in RC apparatus. Only dry BS samples were tested in the RC apparatus as

Table 2  Tests performed and their loading conditions

Type of test Test ID Relative Cell pres- Gmax (MPa) at Results presented
density, Dr sure, σʹc strains < 0.001%
(± 2%) (kPa)

RC RC1 30 50* 48.49 G-γ, D-γ from 0.001 to 0.1%

RC2 100 66.08
RC3 300 113.96
RC4 50 50 53.90
RC5 100* 76.57
RC6 300 138.57
RC7 70 50 61.28
RC8 100 86.10
RC9 300 166.38
RC10 30 50–600 52–174 Maximum shear modulus (Gmax)
RC11 50 50–600* 60–211
RC12 70 50–600 67–218
CTX CTX1 30 50* G-γ, D-γ from 0.15 to 5% (at 10
CTX2 100 different strain levels)
CTX3 150
CTX4 60 50
CTX5 100
CTX6 150
CTX7 90 50
CTX8 100*
CTX9 150

Tests with symbol * are repeated to check the reliability of the testing methodology; relative density values
are rounded to the nearest % (± 2); each CTX test (CTX1–CTX9) was conducted at different shear strain
with a new saturated sample

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2903

the saturation is not expected to significantly alter the dynamic soil properties at low strains
(Saxena and Reddy 1989). The process of sample preparation and testing was followed
according to the ASTM standards (ASTM D4015 2014) and was described in detail by
Dammala et al. (2017b). Two types of RC testing approaches were adopted to achieve the
desired dynamic properties (Gmax, G/Gmax and damping) at different confining pressures
(σʹc) and relative densities (Dr). The first method involves the application of incremental
shear strains (starting from strains < 0.001%) up to a shear strain of 0.1% at a fixed confin-
ing pressure (test IDs RC1-RC9 in Table 2) to establish the shear modulus and damping
variation. The second approach involves the application of incremental confining pressures
on the sample (keeping the input strain as low as possible, typically < 0.001%) to establish
Gmax over wide range of confining pressures (test IDs RC10-RC12). The second approach
(incremental pressures) was started with a confining pressure of 50 kPa and repeated for
intermediate pressures up to a maximum confining pressure of 600 kPa.

2.3 Cyclic triaxial (CTX) testing

Cyclic triaxial (CTX) apparatus has been utilized for the high strain (0.1–5%) testing of
BS. Fully saturated BS specimens were prepared at different Dr and confining pressures
(Table 2) as saturation would affect the high strain behavior of soils due to the rise in pore
water pressure and subsequent reduction in shear strength (Seed and Lee 1966). Lique-
faction susceptibility of the BS specimens was also recorded by monitoring the induced
pore pressures in the sample. Strain-controlled approach was adopted for the testing. Sam-
ple preparation and the test methodology was followed according to the codal provisions
(ASTM D3999 2003) which was described in detail for the adopted CTX apparatus by
Kumar et al. (2017). Separate samples were prepared for each shear strain at the same rela-
tive density and confining pressure in order to avoid the preloading effects at high shear
strains, counting the total number of CTX tests to 90.

3 Laboratory test results and discussion

3.1 Resonant column tests results

The typical output obtained from RC tests and the corresponding analysis and discussion
are presented in this section. In a RC test, input voltage (V) induces cyclic shear strains (γ)
in the soil sample and the frequency of input voltage leading to peak shear strain amplitude
is termed as resonant frequency or fundamental frequency (fnz). The obtained fnz can then
be related to shear wave velocity (Vs) and shear modulus (ASTM D4015 2014).
The first approach of RC testing (incremental strains) yields fnz at various strain lev-
els. Figure 2 presents the typical output from a RC test (variation of fnz with input volt-
age) and the corresponding dynamic soil properties (G and D). Similarly, the incremen-
tal pressure approach yields fnz at various confining pressures and the resulting Gmax can
be evaluated (Fig. 3). It is to be noted that the resonant frequency values cannot be used
in further studies but only the dynamic properties of soil in terms of Gmax, normalized
shear modulus (G/Gmax) and damping ratio (D) versus shear strain curves only be used
for any dynamic analysis.

13
 2904 Bulletin of Earthquake Engineering (2019) 17:2899–2933

Voltage, V
1E-4 1E-3 0.01 0.1 1
80 80 8
fnz and Voltage
G and Shear strain
70
D and Shear strain
70 6
Resonant frequency, Hz

60 Shear modulus, MPa

Damping ratio, %
50 60 4

40

50 2
30

20 40 0
1E-4 1E-3 0.01 0.1 1

Shear strain, %

Fig. 2  Typical output expected from RC test obtained from incremental strain approach, presented in terms
of variation of fnz with input voltage and the corresponding G and D variation with shear strain at 50% rela-
tive density and 100 kPa confining pressure

fnz and CP 200

90
Gmax and CP
180
Resonant frequency, fnz (Hz)

80
160
Gmax (MPa)

140
70

120

60 100

80
50
60

40 40
0 100 200 300 400 500 600

Confining pressure, CP (kPa)

Fig. 3  Output from RC test using the incremental pressure approach, presented in terms of variation of fnz
and Gmax with confining pressure at 50% relative density

3.1.1 Maximum shear modulus ­(Gmax)

As very well established, soil behaves linearly elastic in the low strain range (γ < 0.001%)
and is expected to possess maximum shear stiffness-Gmax (Drnevich et al. 1967). The val-
ues of Gmax were evaluated from Vs using the density of the specimens. Using both the

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2905

aforementioned approaches, one with the increasing confining pressures and later with
the increasing shear strains (at 50, 100 and 300  kPa confining pressures) as listed in
Table 2, Gmax values were determined. Figure 4 presents the variation of Gmax with confin-
ing pressure using both the approaches (solid data points indicate the incremental pres-
sure approach while the hollow data points indicate incremental strain approach). An
exponential increase of Gmax with confining pressure can be observed at the three chosen
relative densities (30%, 50% and 70%). It is important to note the negligible effect of pre-
confinement (incremental pressure approach) as the Gmax values obtained using both the
approaches fall in the narrow range. Similar negligible effect of pre-confinement for Gmax
was observed for Monterey No 0 sand by Chung et al. (1984). Therefore, it may be con-
cluded that the incremental pressure method of testing would yield reasonable estimates of
Gmax for granular soils.

3.1.2 Shear modulus and damping ratio

As presented (Fig. 2), increase in the voltage increases the induced cyclic shear strains and
reduces the fnz and thereby elasticity of the soil diminishes, extending to nonlinear behav-
iour. Shear modulus (G) at different strain amplitudes can be evaluated using the obtained
fnz, Vs and density of the specimen. The variation of G and D with increasing shear strains
is required for any nonlinear dynamic analysis. The effect of confining pressure and rela-
tive density on the shear modulus with varying shear strains is shown in Fig. 5a, b respec-
tively. It is a known fact that both the confining pressure and relative density increases the
shear stiffness due to the increase in inter-particle arrangement. Figure 6a, b presents the
variation of normalized shear modulus reduction (G/Gmax) and damping ratio with shear
strain for different relative densities and confining pressures. An increase in the confin-
ing pressure (essentially the depth of overburden) increases the modulus ratio. However,
increase in the effective stress on the sample (overburden) decreases the dissipated energy
resulting in reduced damping ratio. An important observation can be made from Fig. 6b—
the negligible effect of relative density on the normalized shear modulus reduction and
damping ratio at any given strain level. This essentially means that the compactness of a

Fig. 4  Variation of Gmax with 250

cell pressure at different relative 30% Dr
densities Data from incremental
50% Dr pressure tests
200
70% Dr

150
Gmax (MPa)

100

30% Dr
Data from incremental
50 50% Dr strain tests
70% Dr

0
0 200 400 600
Confining pressure (kPa)

13
 2906 Bulletin of Earthquake Engineering (2019) 17:2899–2933

(a) (b)

Fig. 5  Variation of shear modulus with shear strain for BS sand at different, a σ′c, b Dr (redrawn after Dam-
mala et al. (2017b))

(a) (b)

Fig. 6  Variation of G/Gmax and D with shear strain for BS sand at different, a σʹc, b Dr (redrawn after Dam-
mala et al. (2017b))

granular soil doesn’t affect the rate of reduction of G and the variation of damping ratio at
any given shear strain. Similar findings were also reported by Saxena and Reddy (1989)
and Bai (2011).

3.2 Cyclic triaxial tests results

Strain controlled cyclic triaxial (CTX) tests were conducted on fully saturated BS speci-
mens at varying strain amplitudes (0.15–5% as listed in Table 2) in order to understand the
high strain dynamic behavior, both in terms of dynamic soil properties and liquefaction
analysis. Typical input to the CTX system is shown in Fig.  7a with a variation of axial
strain over number of loading cycles. Using the axially varying load and the effective cross
section area of the specimen, resulting deviatoric stress (σd) can be evaluated. Figure  7b
presents the variation of σd with loading cycles (N). The stress–strain (σd and axial strain)
response thus obtained is shown in Fig.  7c. In order to evaluate the shear modulus and
damping ratio, deviatoric stress versus axial strain response (Fig. 7c) is converted to shear
stress versus shear strain (Fig. 7d). Shear stress is evaluated as half of the cyclic σd while
the shear strain is calculated based on indirect correlation with axial strain using a Pois-
son’s ratio of 0.5, as recommended by Silvestri (2001) and ASTM D3999 (2003).

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2907

(a) (b)

(c) (d)

Fig. 7  Typical input and output from the CTX system

3.2.1 Shear modulus and damping ratio

Dynamic soil properties of BS at intermediate to high strain level (0.15–5%) are evalu-
ated using the strain-controlled CTX tests. A typical stress strain behavior of a cyclically
loaded soil follows a hysteresis loop and was observed to be asymmetrical at shear strains
greater than 0.15% (Kumar et al. 2017), see Fig. 7d. Dynamic soil properties (G and D)
at a shear strain of interest can be obtained by considering the hysteretic loop, typically
at the first loading/unloading cycle. Traditionally, the shape of hysteretic loop is assumed
as symmetrical (similar in loading and unloading conditions) and the corresponding
dynamic properties are evaluated (Kramer 1996). However, Kumar et al. (2017) concluded
that such assumption would often yield in unreliable results and hence, proposed a modi-
fied approach in estimating stiffness properties using the realistic asymmetric loop. The
same approach is adopted here to determine the dynamic soil properties of BS using CTX
results.
Figure 8a, b presents the variation of G and D with shear strain at different confining
pressures (σ′c) for 60% Dr. It can be observed that the variation in G with σ′c becomes
narrower with increasing shear strain, meaning-shear stiffness attains a minimal value
(almost similar) irrespective of the overburden depth at high shear strains. An interesting
trend of damping ratio can be observed at high strains (Fig.  8b)—increase and decrease
of D with γ. Such untraditional damping trend at high strains is attributed to the dilative

13
 2908 Bulletin of Earthquake Engineering (2019) 17:2899–2933

behavior of sands resulting from S shaping of the stress–strain response (Matasovic and
Vucetic 1994). Similar trends were observed for sands by Matasovic and Vucetic (1994),
Kiku and Yoshida (2000), Chiaradonna et al. (2015), Pagliaroli et al. (2018) and for clays
by Chehat et  al. (2018). Figure  8c depicts the comparison of D at high strains from the
above mentioned literature to that of BS. The magnitude of damping ratio ranges from 20
to 28% at strain range of 0.1–1% and degrades to almost 12–15% at very high strains. How-
ever, for soft clays, the damping values are relatively lower (Chehat et al. 2018). It can be
inferred that the apparent change of damping ratio for sands is prevailing at strains close to
1%-meaning-the dilative behavior of sands resulting from S shaping, happens at 1% strain.
It can also be observed that the D of BS is consistent with the literature, and the change of
trend is seen around 1% shear strain. This means that the critical strain for transformation
of dilative behavior of BS is around 1%. In order to provide the normalized shear modulus
reduction (G/Gmax) at high strains using CTX tests, the required Gmax values were consid-
ered from RC tests. Similar combining technique has been employed by Chattaraj and Sen-
gupta (2016) for sands and El Mohtar et al. (2013) for sands with fines. Figure 8d presents
the variation of G/Gmax of BS at 60% Dr, which shows the negligible effect of σ′c at high
shearing strains.

(a) (b)

(c)
(d)

Fig. 8  Variation of a G, b D, c comparison of D with literature at high strains, d G/Gmax with shear strain at
different confining pressures for 60% relative density

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2909

3.2.2 Liquefaction results

Liquefaction is often represented using pore water pressure ratio (ru), which is the ratio of
excess pore pressures (ΔU) generated in the soil column to the mean effective confining
pressure (σ′c). Figure 9 presents the variation of excess pore water pressure (PWP) and ru
with loading cycles (N) at 100 kPa effective confining pressure for 30% Dr.
Figure 10 depicts the variation of PWP ratio at different γ, σʹc and Dr. Due to the rise of
pore pressures with the increase of N and γ, liquefaction potential of sands decrease (Seed
and Lee 1966). It can be observed that higher the γ, higher is the probability of liquefaction
at a given loading cycle. In other words, for identical N, tendency to liquefy the specimen
increases with higher γ at a particular σʹc (Fig. 10a). Figure 10a also depicts that BS speci-
men attains ru ≥ 0.90 in 1st cycle at γ ≥ 0.75% in loose condition (Dr = 30%), thus revealing
the onset of liquefaction at the very 1st cycle (N = 1). However, it is noted that more N are
required to initiate liquefaction for soil specimen subjected to γ < 0.60%. Some of the tests
were conducted at a γ > 0.75%, though, the data are presented up to 0.75%, because higher
γ shows failure of soil specimens in first loading cycle (N = 1). Similar responses were
observed for the specimens prepared at Dr = 60% and 90% and at σʹc = 50 kPa and 150 kPa.
Figure 10b illustrate the variations of ru at different σʹc. It can be observed that the sand
specimens prepared at a particular Dr (30%) requires larger N to attain initial liquefaction
when subjected to higher σʹc, testifying the high resistance to deeply located sandy soils
(high over burden pressure). Figure  10c illustrates the variation of ru at different Dr i.e.
from 30 to 90% and σʹc = 50 kPa. It reflects that the ru decreases with the increase of Dr
means at higher Dr, higher N are required to liquefy the soil specimens for a constant γ and
σʹc. The ru significantly decreases with the increase of σʹc whereas, the same is marginally
affected, which can be neglected, by relative density for first loading cycle. Hence, it can
be stated that the liquefaction susceptibility of BS decreases with σʹc, whereas negligibly
affected by Dr.

Fig. 9  Typical PWP generation and ru variation

13
 2910 Bulletin of Earthquake Engineering (2019) 17:2899–2933

(a) (b)

(c)

Fig. 10  Variation of ru with N at different a γ, b σʹc, c Dr

Cyclic stress ratio (CSR) obtained from CTX is generally used to represent the strength
of soil against cyclic loads. The developed maximum CSR at different γ amplitudes for
different combinations of Dr and σ′c have been presented in Tables  3 and 4. It has been
observed that the CSR increases with the increase of shear strains because the developed
deviatoric stresses (σd) increased with the increase of γ. It has also been observed that the
CSR values decrease with the increase of σ′c whereas, relative density increases the CSR
values.
Tables 3 and 4 also deal with the N at which specimen attains the initiation of liquefaction
(NL), corresponding to the γ and developed CSR. For both the combinations, it is observed
that the developed σd is much higher than the initial σʹc at shear strain (γ) greater than 0.75%,
thus the evaluated CSR (= σd/2σʹc) becomes equal to or greater than 0.5. The developed CSR
attains a value of 0.5 nearly at γ = 1%, however it is not true for all cases but a critical γ = 1%
for NL can be considered from Tables 3 and 4. Beyond this limiting γ, CSR becomes greater
than 0.5 and in such condition, soil was unable to resist even a single loading cycle and hence
specimen collapse was observed due to the sudden rise of PWP. Based on these observations
(Tables  3,  4 and Fig.  10), a critical value of γ (γcritical) can be defined for each σʹc for BS,
beyond which the PWP ratio reaches unity (liquefaction triggering). Table 5 lists the values of
γcritical obtained.

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2911

Table 3  Parameters of the strain-controlled CT tests at Dr = 30% and f = 1 Hz

σʹc (kPa) γ (%) 0.015 0.045 0.075 0.15 0.30 0.45 0.60 0.75 1.0 1.5 3.0

50 σstd,1cycle 6 15 25 30 41 41 49 48 60 61 102
CSR 0.06 0.15 0.25 0.30 0.41 0.41 0.49 0.48 0.60 0.61 1.02
NL NL NL NL 16 4 4 2 1 1 1 1
100 σstd,1cycle 25 35 57 71 79 82 95 120
CSR 0.13 0.18 0.29 0.36 0.40 0.41 0.48 0.6
NL NL NL NL 16 7 4 1 1
150 σstd,1cycle 31 46 66 96 105 112 138
CSR 0.1 0.15 0.22 0.32 0.35 0.38 0.46
NL NL NL NL 18 9 5 3

NL number of cycles for liquefaction, NL no liquefaction till 40 cycles, σstd,1cycle deviator stress (in kPa) at
1st cycle

Table 4  Parameters of the strain-controlled CT tests at Dr = 90% and f = 1 Hz

σʹc (kPa) γ (%) 0.045 0.075 0.15 0.30 0.45 0.60 0.75 1.0 1.5 2.0 3.0

50 σstd,1cycle 20 32 36 49 66 83 85 133
CSR 0.20 0.32 0.36 0.49 0.66 0.83 0.85 1.33
NL NL NL NL 10 3 1 1 1
100 σstd,1cycle 34 43 58 87 90 103 150 250
CSR 0.17 0.22 0.29 0.44 0.45 0.52 0.75 1.25
NL NL NL NL 28 12 4 1 1
150 σstd,1cycle 40 56 73 106 120 130 170 197 247 372
CSR 0.14 0.19 0.24 0.35 0.40 0.44 0.57 0.66 0.83 1.24
NL NL NL NL 32 20 8 7 5 1 1

NL number of cycles for liquefaction, NL no liquefaction till 40 cycles, σstd,1cycle deviator stress (in kPa) at
1st cycle

Table 5  Values of γcritical of BS at Dr (%) 30 60 90

different Dr and σʹc
σʹc (kPa) 50 100 150 50 100 150 50 100 150

γcritical (%) 0.075 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15

4 Modelling of experimental data

To provide design engineers the required dynamic soil properties (Gmax, G/Gmax and D)
at varying range of confining pressure, analytical formulations are required. Therefore,
multiple regression analysis has been performed on the experimental data to arrive at the
required best fit parameters.

13
 2912 Bulletin of Earthquake Engineering (2019) 17:2899–2933

4.1 Maximum shear modulus (Gmax)

Analytical relationship for Gmax estimation may be required, especially if there is a lack
of laboratory/field data. An analytical expression in the form (Eq. 1) developed by Hardin
(1978) has been considered to provide the best fit parameters for BS based on the RC data.
The same expression has been used by many researchers to better fit the experimental data
(Table 6). As can be observed from Table 6, different sands have different fitting param-
eters rising a need to have the soil-specific best fit parameters
�
A × (Pa )1−m × (𝜎c )m
Gmax = ( ) (1)
0.3 + 0.7e2

where A parameter adjusting the magnitude; m confining pressure coefficient.

As the Gmax values obtained from incremental pressure approach seem reliable, regres-
sion analysis has been performed on such data, to achieve the best fit parameters for BS.
Figure 11 presents regression analysis in terms of normalized Gmax with atmospheric pres-
sure (Pa) and an exponential form was found to model the variation efficiently. The best fit
parameters thus achieved are: A = 589.31 and m = 0.489 (Fig. 11) with a correlation coef-
ficient (R2) of 0.996. Figure 12 depicts the efficiency of the proposed relationship as most
of the experimental data fits within ± 5% of the predicted Gmax values.

4.2 Combined results of RC and CTX

Results obtained (in terms of G/Gmax and D) from both the testing techniques (RC and
CTX) are combined to provide the comprehensive data of dynamic soil properties over
wide strain range (0.001–5%). Figure  13a, b present the variation of G/Gmax and D
with shear strain, compared with the boundaries for sands proposed by Seed and Idriss
(1970) and Darendeli (2001). Traditionally, Ground Response Analysis (GRA) studies
consider these two literature based stiffness curves for sands, see for example, Desai
and Choudhury (2012), Chatterjee and Choudhury (2016) and Kumar et  al. (2016). It
is interesting to note that the G/Gmax of BS falls higher than that of Seed and Idriss
(1970) suggested range and even higher than the limits proposed by Darendeli (2001)
up to intermediate strains which are normally the induced strains in typical earthquake
conditions. It must be noted that Darendeli’s upper limit is for 1600  kPa (σ′c) while

Table 6  List of curve fitting parameters for various sands using Eq. 1

Author/s Soil A m

Hardin (1978) Clean sand 625 0.5

Chung et al. (1984) Monterey no. 0 sand 523 0.48
Saxena and Reddy (1989) Monterey no. 0 sand 428 0.574
Bai (2011) Berlin sand 751 0.504
Bai (2011) Braunschweig fine sand 654 0.484
Bai (2011) Braunschweig coarse sand 523 0.603
Bai (2011) Cuxhaven fine sand 736 0.490
Bai (2011) Cuxhaven medium sand 658 0.476
Chattaraj and Sengupta (2016) Kasai River sand 612 0.468

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2913

Fig. 11  Fitted Gmax results using

Eq. 1

Fig. 12  Predicted and measured

Gmax values

the present experimental data is only up to 300  kPa, still are significantly higher than
Darendeli (2001) limits. Similarly, damping ratio falls in the lower range than the litera-
ture suggested boundaries for sands, meaning-damping of the BS will be overestimated
by adopting the traditional literature suggested data, which may lead to underestimation
of the seismic demands on the structures.

4.3 Normalized shear modulus reduction (G/Gmax)

Hardin and Drnevich (1972) initiated modelling the modulus reduction and damping of
soils using hyperbolic stress–strain relationship. The relationship was further modified by
various researchers (Ishibashi and Zhang 1993; Matasovic and Vucetic 1994; Darendeli

13
 2914 Bulletin of Earthquake Engineering (2019) 17:2899–2933

(a) (b)

Fig. 13  a G/Gmax, b damping ratio variation over wide range of shear strains

2001) to better fit the experimental data. Darendeli’s (2001) model (Eq. 2) was found to be
satisfactorily estimating the stress–strain response of soils (Zhang et al. 2005; Vardanega
and Bolton 2013) and hence the same is considered here

G 1
=
(2)
( )𝛼
Gmax 𝛾
1+ 𝛾ref

where γref is the reference shear strain (shear strain at G/Gmax value of 0.5) and α is the
curve fitting parameter adjusting the shape of the modulus curve (found to be 0.92 using
Bayesian analysis by Darendeli (2001)). As the shape of modulus curve is soil specific and
hence the corresponding α needs to be determined from the regression analysis of experi-
mental data.
Figure  14 presents the variation of G/Gmax of BS at 50% relative density and 50  kPa
confining pressure, with the normalized shear strain (γ/γref) fitted with the Darendeli’s
single parameter (α) model. The γref value was considered from the experimental results
(at G/Gmax of 0.5). Along with the Darendeli’s single parameter model, a two parameter
model (Eq. 3) (referred to as modified Darendeli’s—MD model hereafter) with two curve
fitting coefficients (α and β) is considered to better fit the experimental data. Matasovic
and Vucetic (1994) adopted similar two parameter model to simulate the G/Gmax of soils,
however, the definition of reference shear strain was different. El Mohtar et al. (2013) also
adopted similar two parameter model in modelling the G/Gmax behavior of sand mixed with
fines

G 1
= )𝛼 .
Gmax (3)
(
𝛾
1+𝛽 𝛾ref

It can be observed from Fig. 14 that the modified Darendeli’s (MD) model better fits the
experimental data both at low and high shear strains while Darendeli’s model could only
able to fit the modulus reduction at low shear strains. The correlation coefficient using MD
model was found to be 0.988 while that of Darendeli’s model is 0.946, showcasing the effi-
ciency of the MD model.

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2915

Fig. 14  Efficiency of modified
Darendeli’s model in predicting
G/Gmax

Figure 15 presents G/Gmax boundaries of BS obtained by the regression analysis. It can

be noted that most of the experimental data falls in the proposed boundaries and the best
fit parameters are presented. It must be noted that G/Gmax and the value of γref are pressure
dependent variables. The values of γref were considered from experimental data (at a G/
Gmax of 0.5 according to Stokoe et al. 1995) if available, and from the Stokoe’s relationship
(Zhang et al. 2005) (Eq. 4) if the experimental data is unavailable
( �
)k
𝜎c
𝛾ref = 𝛾ref 1 (4)
Pa

Fig. 15  Normalized shear modulus reduction variation with normalized shear strain

13
 2916 Bulletin of Earthquake Engineering (2019) 17:2899–2933

Table 7  Best fit parameters for G/Gmax and damping ratio

Sand type Confining pres- ϒref G/Gmax Damping
sure (kPa)
α β R2 b p R2

BP 50 0.08 1.085 1.177 0.982 0.65 0.2 0.865

100 0.14 1.116 1.591 0.971 0.60 0.185 0.880
300 0.18 1.361 1.250 0.942 0.55 0.185 0.822

where γref1 refers to the reference shear strain at 100  kPa confining pressure, Pa is the
atmospheric pressure = 101.3 in (kPa) and k is the stress correction exponent, given as 0.42
for residual soils (Zhang et al. 2005). As the effect of relative density was found insignifi-
cant in the G/Gmax (Fig. 6b), only the effect of confining pressure was considered. All the
values of considered γref and the obtained best fit parameters (α and β) at different confin-
ing pressures are presented in Table 7.

4.4 Damping ratio

Similar to the modulus reduction, analytical expressions for evaluating damping ratio at
a given shear strain were also proposed by several researchers. As damping ratio can be
quantified using the stress–strain response of the soil, therefore, D can be conveniently
related to the G/Gmax (Ishibashi and Zhang 1993; Darendeli 2001; Zhang et al. 2005; Chat-
taraj and Sengupta 2016). Darendeli (2001) proposed a damping model (Eq. 5), relating it
to the G/Gmax and Masing damping
( )p
G
D(%) = b × Dmas × + Dmin (5)
Gmax

where b and p are the scaling coefficients adjusting the high strain damping ratio, Dmas is
the Masing damping (function of shear strain and the curvature coefficient (α)) and Dmin is
the minimum damping ratio at the lowest possible shear strain (0.5% based on the experi-
mental results). A regression model is run to find the best fit values of b and p. A b value
of 0.60 and p value of 0.18 were found with an R2 of 0.868 (Fig. 16). Individual values of
b and p for different confining pressures are presented in Table  7. Although, experimen-
tal trend of damping ratio at high strains is simulated using the Masing formulation, an
underestimation of D at high strains can be observed (Fig.  16). Such complexities often
arise while combining the data obtained using independent apparatus (Chattaraj and Sen-
gupta 2016). In such cases, ideal solution is to find the better fit at practical range of strains
(0.01–0.1%). Therefore, the damping model has been cautiously fitted at strains of engi-
neering interest.
Finally, both the normalized shear modulus reduction and damping ratio of BS esti-
mated using the regression coefficients (α, β, b and p) at practical values of σ′c values are
shown in Fig. 17. As the presented data is soil-specific (BS), it is suggested to adopt the
present data in seismic analysis, for the soils of similar properties.

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2917

Fig. 16  Variation of D with shear

strain

4.5 Pore water pressure model parameters

Pore water pressure (PWP) models simulate the generation of excess pore pressures during
cyclic loading. Numerous PWP models, ranging from simple to complex nature, are avail-
able in the literature, see for example—Park et al. (2015), Finn and Bhatia (1982), Dobry
et al. (1985), Ivsic (2006), Seed et al. (1975) and Chiaradonna et al. (2018). However, to
perform non-linear effective stress analysis, commercial programs such as DEEPSOIL
(Hashash et  al. 2016), employ extended Dobry et  al. (1985) PWP model as the required
input parameters can be efficiently obtained by performing cyclic stress/strain controlled
tests on saturated soil samples udner undrained conditions. Based on the strain-controlled
CTX tests on sandy soil, Vucetic and Dobry (1988) extended the basic PWP model of

Fig. 17  Design G/Gmax and D ratio curves for BS

13
 2918 Bulletin of Earthquake Engineering (2019) 17:2899–2933

Dobry et al. (1985) (Eq. 6). The model emphasizes the generation of PWP with number of
cycles and applied cyclic shear strain
p . N . F . (𝛾 − 𝛾t )S
ru,N = (6)
1 + N . F. (𝛾 − 𝛾t )S

where ru,N excess PWP ratio at N number of cycles; γ cyclic shear strain amplitude; p, F
and s are the curve fitting parameters; and γt is the threshold shear strain below which no
significant PWP is generated or shear strain below which no significant permanent vol-
ume changes are observed (also called volumetric threshold shear strain). Traditionally,
the value of γt is established from laboratory tests on saturated samples (Vucetic 1995).
However, as the CTX tests were conducted at high strains and RC tests were performed in
dry conditions, the value of γt was established at a G/Gmax value of 0.80 as suggested by
Vucetic (1995). The magnitude of γt increases with increasing confining pressure which
means that the depth of overburden increases the strain required for the volume changes or
significant pore pressure generation. Therefore, for the three tested confining pressures (50,
100 and 150 kPa), the values of γt established are 0.02%, 0.03% and 0.035% respectively,
based on Fig. 4a. The proposed γt values are consistent with the literature suggested range
for sands − 0.01 to 0.04% (Hsu and Vucetic 2004; Dammala et al. 2015; Heshmati et al.
2015).
The PWP model curve fit parameters (p, F and s) were obtained by performing regres-
sion analysis on the experimental data for the three considered σ′c values (using the Eq. 6).
As described, the effect of relative density is insignificant in the PWP generation (Fig. 10c)
and hence the same is neglected in the analysis. Figure 18 depicts the obtained results of
PWP model with the fitting parameters, at three σ′c values.

5 Example application of established soil properties

A one dimensional nonlinear (NL) effective stress GRA, incorporating the PWP gen-
eration and dissipation, has been performed to showcase the applicability of proposed
dynamic soil properties and liquefaction parameters. The computer program DEEPSOIL
v6.1 (Hashash et al. 2016) has been used for the analysis as it is a widely used nonlinear
time domain site response analysis software which utilizes a discretized multi-degree-
of-freedom lumped parameter model of the 1D soil column. Briefly, in the NL effective
stress analysis, the dynamic equation of motion is solved numerically following the New-
mark β method (Newmark 1959). This approach requires the constitutive behavior of soil
to be modelled using the nonlinear stress strain response, along with the pore pressure
parameters. The hysteretic soil response is captured by a pressure-dependent hyperbolic
model that represents the backbone curve of the soil along with the modified extended
unload–reload Masing rules (Phillips and Hashash 2009). Further details about the theo-
retical background of the methodology can be found in Park and Hashash (2008) and Phil-
lips and Hashash (2009) and the complete description is avoided here for brevity. Similar
NL effective stress GRA studies for typical sites in Guwahati city were conducted by Basu
et al. (2017) and Singhai et al. (2016).

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2919

(a) (b)

(c)

Fig. 18  Pore water pressure ratio variation of BS at a 50 kPa, b 100 kPa, c 150 kPa effective confining pres-
sure

5.1 Soil profile

The considered soil profile for the analysis is located near the Brahmaputra River in Guwa-
hati—a northeastern Indian city, classified as one of the highly active seismic regions of
the country, surrounded by many tectonic faults, see Fig. 19 for the details. As the soil pro-
file is near the center of the Saraighat Bridge in Brahmaputra River, most of the explored
depth (42 m out of 48 m) consists of BS sand with increasing relative density, underlain
by a very hard silty clay of 6 m deep. Guwahati city is spread over plain areas interspersed
with hills all around and the typical geological setting consists of quaternary alluvium
deposited over the undulated and faulted granitic rocks (Raghu Kanth and Dash 2010).
The depth of bedrock ranges from 30 to 60 m at the alluvium deposits-close to the river
(Raghu Kanth and Dash 2010; Basu et al. 2017; Kumar et al. 2018). Therefore, the cho-
sen soil profile is assumed to be underlain by a rigid bedrock beyond the stiff clay layer
(beyond 48  m). The selection of the soil profile at the center of the bridge is motivated
by the recent bridge failures in which the central span of the bridge is most vulnerable to
liquefaction induced effects (Mohanty et al. 2017). Therefore, identification of liquefaction
possibility of the soil profile at the central span may be beneficial for the seismic perfor-
mance assessment of the bridge. An equivalent linear GRA for the same profile has been
performed by Dammala et  al. (2017b). Further details about the seismicity of the region
and borelog data can be found in Kayal and De (1991) and Dammala et al. (2017b) respec-
tively. Table 8 presents the bore log data along with the calculated σ′c values based on the

13
 2920 Bulletin of Earthquake Engineering (2019) 17:2899–2933

Fig. 19  a Seismic zonation map of India (IS:1893 2002), b seismotectonic faults in the northeastern India
(Raghu Kanth and Dash 2010), c location of the chosen soil profile in Guwahati City (star symbol)

unit weight of the strata. The shear wave velocity (Vs) required for the analysis is evaluated
based on the SPT N value, following Imai and Tonouchi (1982) relationship. Predominant
period (Ts) estimated for the profile based on the Vs and the thickness of each layer, cor-
responds to 0.80  s. Efficiency of seismic wave propagation depends on the element size
(maximum thickness of each layer) as it would govern the maximum frequency that can be
propagated through the layer. In this regard, Kuhlemeyer and Lysmer (1973) recommended
that the size of element (thickness of each layer) must be less than one-eighth of the mini-
mum wavelength (λ = Vs/fmax), (where Vs is the shear wave velocity of the layer and fmax is
the predominant frequency of the input motion) for accurate wave transmission. However,
recent study by Ramirez et al. (2018) suggested that the maximum size of element as λ/16
for soils of soft or liquefiable nature. Therefore, for the present study, soil layers have been
discretized following the recommendation of Ramirez et al. (2018). In this way, thickness

13
 Table 8  Design soil profile considered along with the dynamic soil properties and liquefaction parameters [modified after Dammala et al. (2017b)]
Layer no. Soil type (depth) D i, m SPT Navg Vs, m/s γtotal, kN/m3 σ′c−I, kPa σ′c, kPa PWP parameters
p F s

1 Loose fine clean sand (11 m) 1.5 4 149 15.1 5.31 10 1.297 3.837 0.963
2 1.5 4 149 15.1 10.63
3 2 8 178 15.7 19.73 25
4 3 8 178 15.7 31.57
5 3 8 178 15.7 43.4 50
6 Moderate to medium dense fine sand (14 m) 2 12 211 16.2 55.65
7 3 15 226 16.5 67.23
8 2 15 226 16.5 80.68 100 1.205 3.029 0.952
9 3 24 262 17.8 112.41
Bulletin of Earthquake Engineering (2019) 17:2899–2933

10 2 24 262 17.8 123.12

11 2 24 262 17.8 133.83
12 Highly dense deep sand partially mixed with 2 31 284 19.4 173.48 200 1.117 3.698 1.054
13 grey silt (17 m) 3 31 284 19.4 192.75
14 2 36 298 20.6 231.33
15 3 36 298 20.6 253.02 300
16 3 36 298 21.7 302.71
17 2 36 298 21.7 318.65
18 2 36 298 21.7 334.58
19 Very hard deep silty clay (6 m)* 3 47 324 22.0 – – – – –
20 3 47 324 22.0

Di thickness of each layer, Vs shear wave velocity, γtotal total unit weight, σ′c−I mean effective confining pressure of ith layer, σ′c mean effective confining pressure of entire
unit; ground water table (GWT) is 16 m above the ground surface; *clay layer with PI = 85; PWP parameters for bottom most clay according to Carlton (2014) are: S = 0.0464;
r = 0.478; A = 7.45; B = -14.714; C = 6.38; D = 0.6922
2921

13
 2922 Bulletin of Earthquake Engineering (2019) 17:2899–2933

of layers increased towards the depth of the stratum with the minimum thickness adopted at
the surface being 0.5 m.

5.2 Dynamic soil properties and PWP parameters for NL effective stress analysis

The required strain dependent dynamic soil properties (G/Gmax and damping ratio) for each
layer for the profile, at different values of σ′c were considered from the established prop-
erties (Fig.  17). Ideally, each layer should have unique dynamic soil properties based on
the corresponding σ′c, however, defining such small sized pressure dependent properties
is tedious and time consuming. In this view, Stokoe et al. (1995) suggested that the field
effective confining pressure of ith layer (σ′c−i) should be within about 50% of the chosen
σ′c. Therefore, the soil profile is divided into a total of seven major units (Table  8). The
reference shear strain values required were estimated from the Stokoe et  al. (1995) for-
mulation (Eq.  4). The obtained modulus and damping values with varying shear strain
were inputted as discrete data points to the program to act as a target G/Gmax and damp-
ing ratio. The provided data points were then fitted using the available MRDF-UIUC pres-
sure dependent, hyperbolic procedure implemented in DEEPSOIL (Phillips and Hashash
2009). The primary reason for fitting the established soil properties is due to the efficient
hyperbolic representation of the fitted damping ratio as well as for a better representation
of high strain damping ratio. In case of underlying clay stratum, dynamic soil properties
proposed by Vucetic and Dobry (1991) are considered as the target data and the described
fitting procedure was adopted. Pore pressure parameters of BS for carrying out the effec-
tive stress analysis were considered from the established values of p, N and F based on
CTX test results. As the CTX tests were only conducted at three σ′c magnitudes (50, 100
and 150  kPa), the parameters (p, N and F) for the sandy layers are defined accordingly
(see Table 8 for details). In case of bottom most clayey stratum, PWP model proposed by
Matasovic and Vucetic (1995) has been adopted and the required parameters of the model
are obtained by considering the Carlton (2014) formulations based on plasticity index
(PI) and over consolidation ratio (OCR). For the clayey stratum, parameters are deter-
mined using PI of 85 and OCR of 1.0 and the achieved values are listed in the foot notes of
Table 8.

5.3 Considered ground motions

Multiple ways of ground motion selection are available such as recorded motions, scal-
ing the recorded motions, spectrum compatible generated, artificial motions by stochastic
seismological models, and by hybrid approaches. However, recorded accelerograms are a
viable option-being more realistic than both spectrally matched and artificially generated
motions and are also easier to obtain (Bommer and Acevedo 2004). Therefore, the chosen
soil profile is subjected to five recorded ground motions with Peak Bed Rock Accelera-
tion (PBRA) ranging from 0.02 to 0.33 g. The ground motions correspond to 2011 Sikkim
earthquake with 0.02 g and 0.10 g PBRA (SKM 0.02 g and SKM 0.10 g), 2009 Sonitpur
earthquake of 0.03 g PBRA (SONIT 0.03 g), 1988 Indo-Burma earthquake of 0.18 g and
0.33 g PBRA (INDBUR 0.18 g and INDBUR 0.33 g respectively). Such range of PBRA is
selected so as to check the effect of varying PBRA on the liquefaction characteristics of BS.
Indian standard (IS:1893 2002) suggests a PBRA of 0.36 g for Guwahati city and therefore
the maximum value of PBRA chosen is justifiable. All the considered input motions are
expected to contain the source and path characteristics effectively, as they were recordings

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2923

of earthquakes occurred in the northeastern India. Table 9 lists the chosen ground motions
along with their Ground Motion Parameters (GMP). Four out of five ground motions are
recorded and one motion (SKM 0.02 g) is linearly scaled to achieve a targeted PBRA of
0.10 g (SKM 0.1 g). Two ground motions (SKM 0.02 g and SONIT 0.03 g) were recorded
in Guwahati while INDBUR 0.18 g and INDBUR 0.33 g were recorded at Diphu station in
Assam. Figure 20 depicts the considered ground motions. The chosen ground motions are
given as input at the base of the profile (assumed rigid bedrock).

5.4 Results and discussion

5.4.1 Peak ground acceleration (PGA)

Figure 21a–c present the variation of Peak Ground Acceleration (PGA), amplification and
PWP ratio (ru) along the depth respectively, for the considered ground motions. As can
be observed from the PGA profile, ground motions with low PBRA (0.02  g and 0.03  g)
exerted high amplification nearing the surface while the strong ground motions (PBRA
of 0.18  g and 0.33  g) deamplified in the shallow depths. An amplification of 83% and
86% was observed for 0.02 g and 0.03 g PBRA motions respectively, while 25%, − 13%
and − 54% amplification/attenuation was noticed for the 0.1  g, 0.18  g and 0.33  g PBRA
motions respectively. The amplification/attenuation is characterized by both the dynamic
soil properties and the ground motion parameters. With the increase in intensity of a
ground motion (PBRA), an increase in the soil displacement can be expected and this
increased displacement is predominant in the surficial layers as they typically contain loose
deposits with low Vs. The increased strains in the soil deposit leads to increased damp-
ing (Fig. 17) which attenuates the incoming waves and reduces the magnitude on reaching
the surface. Such deamplification is most likely to prevail in loose and soft soil deposits
as observed during the 1985 Mexico earthquake and 1989 Loma Prieta earthquake (EPRI
1993; Romero and Rix 2005). More recently, Kumar et  al. (2015) also concluded that
loose/soft soil deposits of northeastern India may experience deamplification for ground
motions with PBRA > 0.10  g. In the present study as well, deamplification started with
PBRA > 0.10 g inferring that BS sand in shallow depths would experience high amplifica-
tion for low PBRA motions. As the northeast Indian region suffers frequently from mild to
moderate earthquakes (PBRA ranging from 0.02 to 0.1 g) and the region is currently under
high development mode with huge infrastructural development such as railway and road
bridges over the Brahmaputra river, it is imperative to estimate such amplifications caused
by seismic events.

5.4.2 Peak spectral acceleration (PSA)

Figure 22 presents the comparison of Peak Spectral Acceleration (PSA) at surface and at
base, for all the considered ground motions. Similar to the PGA amplification/deamplifica-
tion pattern, high surface PSA was observed for low PBRA ground motions (Fig. 22a, b)
and with increasing PBRA, the surface PSA gradually reduced relative to the PSA of input
motion (Fig.  22c–e). Table  10 illustrates the magnitude of decrease in surface PSA with
PBRA of the ground motion. An increase in PSA of 89% was observed for 0.02 g PBRA
while the same is reduced to − 54% for 0.33 g PBRA motion. It is interesting to note that
the predominant period (Tinput = 0.44 s) of the INDBUR 0.33 g motion varied significantly

13
 2924

13
Table 9  Strong motion parameters of the considered ground motions
Earthquake parameters Sikkim 2011 earthquake Sonitpur 2009 earthquake Indo-Burma 1988 earthquake Indo-Burma 1988 earthquake

Short name SKM 0.02 g SONIT 0.03 g INDBUR 0.22 g INDBUR 0.33 g
Date of occurrence 18-09-2011 19-08-2009 06-08-1988 06-08-1988
Moment magnitude 6.8 4.9 7.2 7.2
Epicenter 27.52°N 26.60°N 25.15°N 25.15°N
88.04°E 92.50°E 95.13°E 95.13°E
Recording station IIT Guwahati Guwahati Diphu Diphu
Component E–W N–S N–E S–W
Distance from source (km) 378 424 91 193
PBRA (g) 0.02 0.031 0.18 0.330
Depth (km) 20 20 91 91
Predominant period, (s) 0.38 0.20 0.10 0.44
Mean period (s) 0.633 0.211 0.172 0.310
Bracketed duration (s) – – 77.74 75.32
Significant duration (s) 34.89 7.975 34.66 16.78
Bulletin of Earthquake Engineering (2019) 17:2899–2933
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2925

0.02
SKM 0.02g
0.01

0.00

-0.01

-0.02
0 30 60 90 120
0.030

0.015 SONIT 0.03g

0.000

-0.015

-0.030
Acceleration (g)

0 20 40 60
0.15

0.10 SKM 0.1g

0.05

0.00

-0.05

-0.10

-0.15
0 30 60 90 120
0.2

0.1 INDBUR 0.18g

0.0

-0.1

-0.2
0 20 40 60 80
0.4

0.2
INDBUR 0.33g

0.0

-0.2

-0.4
0 20 40 60 80

Time (s)

Fig. 20  Considered recorded ground motions

upon reaching the surface (Tsurface = 0.12 s) while the input and surface period is consistent
for the rest of the motions.

5.4.3 Liquefaction assessment

Figure  21c presents the PWP ratio (ru) profile for the considered ground motions. It can
be observed that the intensity of shaking increased the vulnerability to liquefaction. Low
PBRA motions (SKM 0.02  g and SONIT 0.031  g) did not liquefy the stratum while the
strong ground motions (SKM 0.10 g, INDBUR 0.18 g and INDBUR 0.33 g) liquefied the
surficial deposits. A liquefiable depth of 11 m and 15 m were observed for INDBUR 0.18 g
and INDBUR 0.33 g events respectively. For the purpose of comparison, liquefaction sus-
ceptibility of the profile was also evaluated following the semi-empirical procedure of
Idriss and Boulanger (2006). A liquefiable depths of 4 m, 10 m and 14 m were observed for
0.10 g, 0.18 g and 0.33 g events respectively while no traces of liquefaction were obtained
for low intensity ground motions (SKM 0.02 g and SONIT 0.03 g).
The PWP ratio (ru) variation with time at 10  m depth for all the considered ground
motions is illustrated in Fig.  23. Low PBRA motions (SKM 0.02  g and SONIT 0.03  g)

13
 2926 Bulletin of Earthquake Engineering (2019) 17:2899–2933

PGA (g) Deamplification (%) Amplification (%)

PWP ratio
0.0 0.1 0.2 0.3 0.4 -100 -50 0 50 100 150 0.00 0.25 0.50 0.75 1.00
0

10
Depth from surface, m

20

30

40

50
(a) (b) (c)
Sikkim 0.02g Sonitpur 0.031g
Sikkim 0.1g Ind Burma 0.18g
Ind Burma 0.33g

Fig. 21  Profiles of a PGA, b amplification, c PWP ratio

did not develop any significant pore pressures during the excitation while the high PBRA
motions (SKM 0.10  g, INDBUR 0.18  g and INDBUR 0.33  g) induced liquefaction.
Another interesting observation can be made: the speed of liquefaction increased with
increasing PBRA of input motion. The comparison is made between similar events of vary-
ing intensity (INDBUR 0.18 g and INDBUR 0.33 g). The event 0.18 g took 26 s at 10 m
depth while 0.33 g event required less than 22 s to liquefy.
Liquefaction triggering along the depth can be validated by considering the shear strain
profile. Figure 24 presents the peak shear strain (γpeak) induced along the depth for the con-
sidered ground motions. At 10  m depth from surface, γpeak = 0.01% was induced during
the excitation of low PBRA motions while γpeak of 0.10%, 3.9% and 7.45% were observed
for 0.10 g, 0.18 g and 0.33 g PBRA motions respectively. The critical cyclic shear strain
(γcritical) defined in the CTX tests at different σ′c values, provides an insight into the liq-
uefaction phenomenon. A γcritical-50kPa represents the minimum shear strain required for
the onset of liquefaction till 16  m (based on the effective confining pressure). Similarly,
γcritical-100kPa and γcritical-150kPa obtained as 0.15% showcasing the threshold strain required
for liquefaction till 32 m (150 kPa). It can be observed from Fig. 23 that the low PBRA
motions did not develop γpeak > 0.01% (way less than 0.075% of γcritical-50kPa) and hence
the soil deposit did not liquefy for the low PBRA motions. However, high PBRA motions
induced γpeak > 0.15% (higher than γcritical-100kPa and γcritical-150kPa) leading to the liquefac-
tion of surficial 11 m. It is interesting to note that the induced strains for the four ground
motions (0.02 g, 0.03 g, 0.10 g and 0.18 g) are less than the γcritical-100kPa in the deeper stra-
tum and hence the deeper deposits were not liquefied. However, in case of INDBUR 0.33 g

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2927

0.3
(a) SKM 0.02g Input (b) SONIT 0.03g Tinput=0.20s
Surface Tsurface=0.197s
0.10

0.2
Tinput=0.38s
Tsurface=0.39s

0.05
0.034g 0.1
Peak Spectral Acceleration (g)

0.058g
0.02g
0.03g

0.00 0.0
0.01 0.1 1 10 0.01 0.1 1 10
1.00
(c) SKM 0.10g (d) INDBUR 0.18g
0.4
0.75
Tinput=0.38s Tinput=0.10s
Tsurface=0.39s Tsurface=0.10s

0.50

0.2

0.125g 0.25
0.18g
0.10g 0.16g

0.0 0.00
0.01 0.1 1 10 0.01 0.1 1 10

Period (s)

1.5
(e) INDBUR 0.33g Tinput=0.44s
Peak Spectral Acceleration (g)

Tsurface=0.12s

1.0

0.5
0.33g

0.151g

0.0
0.01 0.1 1 10

Period (s)

Fig. 22  Peak spectral acceleration variation with period for the considered ground motions

Table 10  Amplitude of PSA for

the considered ground motions PBRA (g) 0.02 0.03 0.10 0.18 0.33
Input PSA (g) 0.058 0.134 0.289 0.644 1.32
Surface PSA (g) 0.11 0.249 0.42 0.692 0.606
% Difference + 89 + 86 + 45 + 7.45 − 54

13
 2928 Bulletin of Earthquake Engineering (2019) 17:2899–2933

0.10
Sikkim 2011 (PGA=0.02 g)

0.05
No liquefaction

0.00
0.005 0 20 40 60 80 100 120

0.004 Sonitpur 2009 (PGA=0.031 g)

0.003
Excess pore water pressure ratio, ru

0.002 No liquefaction
0.001
0.000
0 15 30 45 60
1.00

0.75 Sikkim 2011 (scaled PGA=0.1 g)

Liquefied
0.50

0.25

0.00
0 30 60 90 120
1.00

0.75 Indo-Burma 1988 (PGA 0.18 g)

0.50

0.25 Liquefied
26.28 sec
0.00
0 20 40 60 80
1.00

0.75
Indo-Burma 1988 (PGA 0.33 g)
0.50

0.25 Liquefied
22.08 sec
0.00
0 20 40 60 80
Time (s)

Fig. 23  Induced PWP ratio at 10 m depth from surface for different ground motions

motion, peak shear strain of 1.5% (higher than 0.15% of γcritical-150kPa) was noticed at 15 m
leading to liquefaction.

6 Conclusions

Strain dependent dynamic soil properties and pore pressure parameters of a natural river
bed sand (Brahmaputra sand-BS) sampled from a highly active seismic region (north-
east India), were determined using two independent high quality apparatus (Resonant
column-RC and Cyclic Triaxial-CTX). Resonant column (RC) apparatus was adopted
to identify the low to intermediate strain behavior (0.001–0.1%) while CTX was used to
understand the high strain behavior (up to 5%) along with the pore pressure parameters.
Results obtained from both the equipment are combined to provide comprehensive data
of dynamic soil properties (Gmax, G/Gmax and damping variation) over wide range of
shear strains. Analytical formulations are developed to estimate the magnitude of Gmax
of BS at any given confining pressure based on the RC test results at low shear strains.
A modified hyperbolic formulation is proposed to model the G/Gmax behavior of BS at

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2929

(a) (b) (c)

Fig. 24  Peak shear strain profiles compared with the critical shear strains obtained from CTX tests

any given shear strain and effective confining pressure. The G/Gmax and damping ratio at
practical values of confining pressures were established. A pore pressure model is for-
mulated using the CTX results with appropriate curve fitting parameters.
Finally, a nonlinear effective stress ground response study for a typical site (in the
northeast Indian region) is performed utilizing the recorded ground motions of varying
PBRA, to demonstrate the applicability of proposed properties. The loose surficial soil
stratum experienced high amplification for low PBRA motions and vice versa. Also, the
surficial stratum was liquefied beyond a PBRA of 0.10 g and the obtained liquefaction
response in terms of strains induced in the profile are consistent with the CTX results.
As the northeast Indian region frequently witnesses mild to moderate seismic events,
such high amplifications for low PBRA motions should be thoroughly investigated. The
established dynamic soil properties and the liquefaction parameters of BS will be highly
useful for performing seismic resistant design of new structures or seismic requalifica-
tion of existing structures in similar soils.

References
ASTM D2487 (2006) Standard practice for classification of soils for engineering purposes (unified soil
classification system). ASTM Stand Guid D5521–5:1–5. https​://doi.org/10.1520/D2487​-11
ASTM D3999 (2003) Standard test methods for the determination of the modulus and damping properties
of soils using the cyclic triaxial apparatus. Am Soc Test Mater 91:1–16. https​://doi.org/10.1520/D3999​
-11E01​.1.6

13
 2930 Bulletin of Earthquake Engineering (2019) 17:2899–2933

ASTM D4015 (2014) Standard test methods for modulus and damping of soils by resonant-column 1–22.
https​://doi.org/10.1520/d4015​-07.1
ASTM D5311 (1996) Standard test method for load controlled cyclic triaxial strength of soil. ASTM
D5311(92):1–11. https​://doi.org/10.1520/D5311​
Ayothiraman R, Raghu Kanth STG, Sreelatha S (2012) Evaluation of liquefaction potential of Guwa-
hati: Gateway city to Northeastern India. Nat Hazards 63:449–460. https​://doi.org/10.1007/s1106​
9-012-0158-9
Bai L (2011) Preloading effects on dynamic sand behavior by resonant column tests. Doctoral thesis, Tech-
nischen Universität Berlin
Basu D, Dey A, Kumar SS (2017) One-dimensional effective stress non-Masing nonlinear ground response
analysis of IIT Guwahati. Int J Geotech Earthq Eng 8:1–27
Bommer JJ, Acevedo AB (2004) The use of real earthquake accelerograms as input to dynamic analysis. J
Earthq Eng 8(spec01):43–91
Carlton BD (2014) An improved description of the seismic response of sites with high plasticity soils,
organic clays, and deep soft soil deposits. University of California, Berkeley
Chattaraj R, Sengupta A (2016) Liquefaction potential and strain dependent dynamic properties of Kasai
River sand. Soil Dyn Earthq Eng 90:467–475. https​://doi.org/10.1016/j.soild​yn.2016.07.023
Chatterjee K, Choudhury D (2016) Influences of local soil conditions for ground response in Kolkata
city during earthquakes. Proc Natl Acad Sci India Sect A Phys Sci. https​://doi.org/10.1007/s4001​
0-016-0265-1
Chehat A, Hussien MN, Abdelazize M, Chekired M (2018) Stiffness-and damping-strain curves of sen-
sitive Champlain clays through experimental and analytical approaches. Can Geotech J. https​://doi.
org/10.1139/cgj-2017-0732
Chiaradonna A, Tropeano G, Onofrio A, Silvestri F, Park D (2015) Application of a simplified model for the
prediction of pore pressure build-up in sandy soils subjected to seismic loading. In: 6th international
conference on earthquake geotechnical engineering, Christchurch
Chiaradonna A, Tropeano G, d’Onofrio A, Silvestri F (2018) Development of a simplified model for pore
water pressure build-up induced by cyclic loading. Bull Earthq Eng. https​://doi.org/10.1007/s1051​
8-018-0354-4
Chung R, Yokel F, Drnevich V (1984) Evaluation of dynamic properties of sands by resonant column test-
ing. Geotech Test J 7:60. https​://doi.org/10.1520/GTJ10​594J
Dammala PK, Krishna AM, Bhattacharya S (2015) Cyclic threshold shear strains in cohesionless soils
based on stiffness degradation. In: Indian geotechnical conference, Pune, Paper ID: 291
Dammala PK, Krishna AM, Bhattacharya S, Aingaran S (2016) Cyclic repones of cohesionless soil using
cyclic simple shear testing. In: 6th international conference on recent advances in geotechnical earth-
quake engineering. ICRAGEE, pp 1–10
Dammala PK, Bhattacharya S, Krishna AM, Kumar SS, Dasgupta K (2017a) Scenario based seismic re-
qualification of caisson supported major bridges? a case study of Saraighat Bridge. Soil Dyn Earthq
Eng 100:270–275. https​://doi.org/10.1016/j.soild​yn.2017.06.005
Dammala PK, Krishna AM, Bhattacharya S, Rouholamin M, Nikitas G (2017b) Dynamic soil properties for
seismic ground response studies in Northeastern India. Soil Dyn Earthq Eng 100:357–370. https​://doi.
org/10.1016/j.soild​yn.2017.06.003
Darendeli M (2001) Development of a new family of normalized modulus reduction and material damping.
Doctoral thesis, University of Texas
Desai SS, Choudhury D (2012) Site-specific seismic ground response study for nuclear power plants and
ports in Mumbai. Nat Hazards Rev 13:205–220. https​://doi.org/10.1061/(ASCE)NH.1527-6996
Dobry R, Pierce W, Dyvik R, Thomas GE, Ladd R (1985) Pore pressure model for cyclic straining of sand.
Rensselaer Polytechnic Institute, Troy
Drnevich VP, Hall JR, Richart Jr F (1967) Large amplitude vibration effects on the shear modulus of sand.
Michigan University Ann Arbor
El Mohtar CS, Drnevich VP, Santagata M, Bobet A (2013) Combined resonant column and cyclic triaxial
tests for measuring undrained shear modulus reduction of sand with plastic fines. Geotech Test J 36:1–
9. https​://doi.org/10.1520/GTJ20​12012​9
Esteva L (1988) The Mexico earthquake of September 19, 1985-consequences, lessons, and impact on
research and practise. Earthq Spectra 4:413–426
Finn WDL, Bhatia SK (1982) Prediction of seismic pore water pressures. In: 10th international conference
in soil mechanics and foundations, Stockholm, pp 201–206
Govindaraju L (2005) Liquefaction and dynamic properties of sandy soils. Doctoral thesis, IISc Bangalore
Guha S, Bhattacharya U (1984) Studies on prediction of seismicity in northeastern India. In: 8th world
conference on earthquake engineering, San Francisco

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2931

Hardin BO (1978) The nature of stress-strain behavior for soils. In: From Volume I of Earthquake Engi-
neering and Soil Dynamics--Proceedings of the ASCE Geotechnical Engineering Division Spe-
cialty Conference, June 19–21, 1978, Pasadena, California. Sponsored by Geotechnical Engineer-
ing Division of ASCE
Hardin BO, Drnevich VP (1972) Shear modulus and damping in soils: design equations and curves. J
Soil Mech Found Div SM7:667–692
Hardin BO, Richart FE (1963) Elastic wave velocities in Granular soils. J Soil Mech Found Div
89:33–65
Hashash YMA, Musgrouve MI, Harmon JA, Groholski DR, Phillips CA, Park D (2016) DEEPSOIL 6.1,
user manual. Board of Trustees of University of Illinois at Urbana-Champaign, Urbana
Heshmati AA, Shahnazari H, Sarbaz H (2015) The cyclic threshold shear strains in very dense clean
sand. Eur J Environ Civ Eng 19:884–899. https​://doi.org/10.1080/19648​189.2014.98584​8
Hsu C-C, Vucetic M (2004) Volumetric threshold shear strain for cyclic settlement. J Geotech Geoenvi-
ron Eng 130:58–70. https​://doi.org/10.1061/(ASCE)1090-0241(2004)130:1(58)
Idriss IM, Boulanger RW (2006) Semi-empirical procedures for evaluating liquefaction potential during
earthquakes. Soil Dyn Earthq Eng 26:115–130. https​://doi.org/10.1016/j.soild​yn.2004.11.023
Imai T, Tonouchi K (1982) Correlation of N-value with S-wave velocity and shear modulus. In: Proceed-
ings of the 2nd European symposium on penetration testing, Amsterdam, pp 57–72
IS:1893 (2002) Criteria for earthquake resistant design of structures. Indian Stand 1–44
Ishibashi I, Zhang X (1993) Unified dynamic shear moduli and damping ratios of sand and clay. Soils
Found Jpn Soc Soil Mech Found Eng 33:182–191. https​://doi.org/10.1248/cpb.37.3229
Ivsic T (2006) A model for presentation of seismic pore water pressures. Soil Dyn Earthq Eng 26:191–199
Kayal JR, De R (1991) Microseismicity and tectonics in Northeast India. Bull Seismol Soc Am 81:131–138
Khattri KN (1999) Probabilities of occurrence of great earthquakes in the Himalaya. Proc Indian Acad
Sci Earth Planet Sci 108:87–92
Kiku H, Yoshida N (2000) Dynamic deformation property tests at large strains. World Conf Earthq Eng
12:1–7
Kokusho T (1980) Cyclic triaxial test of dynamic soil properties for wide strain range. Soils Found 20:45–59
Kramer SL (1996) Geotechnical earthquake engineering, 1st edn. Prentice-Hall, Upper Saddle River
Kuhlemeyer R, Lysmer J (1973) Finite element method accuracy for wave propagation problems. J Soil
Mech Found Div 99:421–427
Kumar A, Harinarayan NH, Baro O (2015) High amplification factor for low amplitude ground motion:
assessment for Delhi. Disaster Adv 8:1–11
Kumar A, Baro O, Harinarayan NH (2016) Obtaining the surface PGA from site response analyses based
on globally recorded ground motions and matching with the codal values. Nat Hazards 81:543–572.
https​://doi.org/10.1007/s1106​9-015-2095-x
Kumar SS, Krishna AM, Dey A (2017) Evaluation of dynamic properties of sandy soil at high cyclic
strains. Soil Dyn Earthq Eng 99:157–167. https​://doi.org/10.1016/j.soild​yn.2017.05.016
Kumar SS, Krishna AM, Dey A (2018) Importance of site-specific dynamic soil properties for seis-
mic ground response studies Importance of site-specific dynamic soil properties for seismic ground
response studies. Int J Geotech Earthq Eng 9:1–21. https​://doi.org/10.4018/IJGEE​.20180​10105​
Maheshwari BK, Kirar B (2017) Dynamic properties of soils at low strains in Roorkee region using reso-
nant column tests. Int J Geotech Eng 6362:1–12. https​://doi.org/10.1080/19386​362.2017.13654​74
Matasovic N, Vucetic M (1994) Cyclic characterization of liquefiable sands. J Geotech Geoenviron Eng
119:1805–1822
Matasovic N, Vucetic M (1995) Generalized cyclic-degradation–pore-pressure generation model for
clays. J Geotech Eng 121:33–42
Mohanty P, Dutta SC, Bhattacharya S (2017) Proposed mechanism for mid-span failure of pile supported
river bridges during seismic liquefaction. Soil Dyn Earthq Eng 102:41–45
Newmark NM (1959) A Method of Computation for Structural Dynamics. J Eng Mech Div 85:67‒94
Oldham R (1899) Report of the great earthquake of 12 June 1897. Mem Geol Surv India 29:1–379
Pagliaroli A, Aprile V, Chamlagain D, Lanzo G, Poovarodom N (2018) Assessment of site e ff ects in the
Kathmandu valley, Nepal, during the 2015 Mw 7. 8 Gorkha earthquake sequence using 1D and 2D
numerical modelling. Eng Geol 239:50–62. https​://doi.org/10.1016/j.engge​o.2018.03.011
Park D, Hashash YMA (2008) Rate-dependent soil behavior in seismic site response analysis. Can Geo-
tech J 45:454–469. https​://doi.org/10.1139/T07-090
Park T, Park D, Ahn JK (2015) Pore pressure model based on accumulated stress. Bull Earthq Eng
13(7):1913–1926
Phillips C, Hashash YMA (2009) Damping formulation for nonlinear 1D site response analyses. Soil
Dyn Earthq Eng 29:1143–1158. https​://doi.org/10.1016/j.soild​yn.2009.01.004

13
 2932 Bulletin of Earthquake Engineering (2019) 17:2899–2933

PHRI Port and Harbour Research Institute (1997) Handbook on liquefaction remediation of reclaimed
land, Balkema
Poddar M (1950) The Assam earthquake of 15th August 1950. Indian Miner 4:167–176
Raghu Kanth STG, Dash SK (2010) Evaluation of seismic soil-liquefaction at Guwahati city. Environ
Earth Sci 61:355–368. https​://doi.org/10.1007/s1266​5-009-0347-3
Ramirez J, Asce SM, Barrero AR, Chen L, Dashti S, Ghofrani A, Taiebat M, Arduino P (2018) Site
response in a layered liquefiable deposit: evaluation of different numerical tools and methodologies
with centrifuge experimental results. J Geotech Geoenviron Eng 144:1–22. https​://doi.org/10.1061/
(ASCE)GT.1943-5606.00019​47
Romero SM, Rix GJ (2005) Ground motion amplification of soils in the upper Mississippi embayment.
Report no. GIT-CEE/GEO-01-1, National Science Foundation Mid America Earthquake Center
Saxena SK, Reddy KR (1989) Dynamic moduli and damping ratios for Monterey No.0 sand by resonant
column tests. Soils Found 29:37–51. https​://doi.org/10.3208/sandf​1972.29.2_37
Seed HB, Idriss IM (1970) Soil moduli and damping factors for dynamic response analysis. Report EE
70-10, Earthquake Engineering and Research Center, University of Californina Berkeley
Seed HB, Lee KL (1966) Liquefaction of saturated sands during cyclic loading. J Soil Mech Found Div
92:105–134
Seed HB, Martin PP, Lysmer J (1975) The generation and dissipation of pore water pressures during soil
liquefaction. University of California Berkeley, Berkeley
Silvestri F (2001). Looking for objective criteria in the interpretation of laboratory stress–strain tests. Pre-
failure deformation characteristics of geometerials. Jamiolkowski, Lancellotta e Lo Presti (eds). Swets
and Zeitlinger, Lisse, ISBN 9058090752
Singhai A, Kumar SS, Dey A (2016) Site-specific 1-D nonlinear effective stress GRA with pore water pres-
sure dissipation. In: 6th international conference on recent advances in geotechnical earthquake engi-
neering and soil dynamics, New Delhi, pp 1–11
Sitharam T, Govindaraju L, Srinivasa Murthy B (2004) Evaluation of liquefaction potential and dynamic
properties of silty sand using cyclic triaxial testing. Geotech Test J 27:Paper ID GTJ11894
Stokoe KH, Hwang SK, Darendeli M, Lee NJ (1995) Correlation study of nonlinear dynamic soils proper-
ties. Westinghouse Savanah River Company, Aiken
Towhata I (2008) Geotechnical earthquake engineering. Springer, Berlin
Vardanega PJ, Bolton MD (2013) Stiffness of clays and silts: normalizing shear modulus and shear strain. J
Geotech Geoenviron Eng 9:1575–1589. https​://doi.org/10.1061/(ASCE)GT.1943-5606.00008​87
Vucetic M (1995) Cyclic threshold shear strains in soils. J Geotech Geoenviron Eng 120:2208–2228
Vucetic M, Dobry R (1988) Cyclic triaxial strain-controlled testing of liquefiable sands. Adv Triaxial Test
Soil Rock. https​://doi.org/10.1520/stp29​093s
Vucetic M, Dobry R (1991) Effect of soil plasticity on cyclic response. J Geotech Geoenviron Eng
117:89–107
Zhang J, Andrus RD, Juang CH (2005) Normalized shear modulus and material damping ratio relationships.
J Geotech Geoenviron Eng 131:453–464. https​://doi.org/10.1061/(ASCE)1090-0241(2005)131:4(453)

Publisher’s Note  Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Affiliations

Pradeep Kumar Dammala1   · Shiv Shankar Kumar2 · A. Murali Krishna1 ·

Subhamoy Bhattacharya3
Shiv Shankar Kumar
shivshankar.mit@gmail.com
A. Murali Krishna
amurali@iitg.ac.in
Subhamoy Bhattacharya
s.bhattacharya@surrey.ac.uk

13
Bulletin of Earthquake Engineering (2019) 17:2899–2933 2933

1
Department of Civil Engineering, Indian Institute of Technology Guwahati, Guwahati 781039,
Assam, India
2
Department of Civil Engineering, Kalinga Institute of Industrial Technology Bhubaneswar,
Bhubaneswar 751024, Odisha, India
3
Department of Civil and Environmental Engineering, University of Surrey, Guildford GU2 7XH,
UK

13