ArticlePDF Available

Seasonality in Site Response: An Example from Two Historical Earthquakes in Kazakhstan

Authors:

Abstract

During the past 150 yr, the city of Almaty (formerly Verny) in Kazakhstan has suffered significant damage due to several large earthquakes. The 9 June 1887 Mw 7.3 Verny earthquake occurred at a time when the city mainly consisted of adobe buildings with a population of 30,000, with it being nearly totally destroyed with 300 deaths. The 3 January 1911 Mw 7.8 Kemin earthquake caused 390 deaths, with 44 in Verny itself. Remarkably, this earthquake, which occurred around 40 km from Verny, caused significant soil deformation and ground failure in the city. A crucial step toward preparing for future events, mitigating against earthquake risk, and defining opti- mal engineering designs, involves undertaking site response studies. With regard to this, we investigate the possibility that the extreme ground failure observed after the 1911 Kemin earthquake could have been enhanced by the presence of a shal- low frozen ground layer that may have inhibited the drainage of pore pressure excess through the surface, therefore inducing liquefaction at depth. We make use of information collected regarding the soil conditions around the city at the time of the earthquakes, the results from seismic noise analysis, borehole data, and surface temperature data. From these datasets, we estimated the necessary parameters for evaluating the dynamic properties of the soil in this area. We successively characterize the corresponding sediment layers at the sites of the observed liquefaction. Although the estimated soil parameters are not optimally constrained, the dynamic analysis, carried out using selected strong-motion recordings that are expected to be com- patible with the two considered events, indicated that the extensive ground failure that occurred during the Kemin event could be due to the presence of a superficial frozen soil layer. Our results indicate that for this region, possible seasonal effects should, therefore, be considered when undertaking site effect studies.
Seasonality in Site Response: An Example from
Two Historical Earthquakes in Kazakhstan
by Rami Alshembari, Stefano Parolai, Tobias Boxberger, Denis Sandron,
Marco Pilz, and Natalya Sylacheva
ABSTRACT
During the past 150 yr, the city of Almaty (formerly Verny)
in Kazakhstan has suffered significant damage due to several
large earthquakes. The 9 June 1887 Mw7.3 Verny earthquake
occurred at a time when the city mainly consisted of adobe
buildings with a population of 30,000, with it being nearly
totally destroyed with 300 deaths. The 3 January 1911 Mw7.8
Kemin earthquake caused 390 deaths, with 44 in Verny itself.
Remarkably, this earthquake, which occurred around 40 km
from Verny, caused significant soil deformation and ground
failure in the city. A crucial step toward preparing for future
events, mitigating against earthquake risk, and defining opti-
mal engineering designs, involves undertaking site response
studies. With regard to this, we investigate the possibility that
the extreme ground failure observed after the 1911 Kemin
earthquake could have been enhanced by the presence of a shal-
low frozen ground layer that may have inhibited the drainage
of pore pressure excess through the surface, therefore inducing
liquefaction at depth. We make use of information collected
regarding the soil conditions around the city at the time of the
earthquakes, the results from seismic noise analysis, borehole
data, and surface temperature data. From these datasets, we
estimated the necessary parameters for evaluating the dynamic
properties of the soil in this area. We successively characterize
the corresponding sediment layers at the sites of the observed
liquefaction. Although the estimated soil parameters are not
optimally constrained, the dynamic analysis, carried out using
selected strong-motion recordings that are expected to be com-
patible with the two considered events, indicated that the
extensive ground failure that occurred during the Kemin event
could be due to the presence of a superficial frozen soil layer.
Our results indicate that for this region, possible seasonal
effects should, therefore, be considered when undertaking site
effect studies.
INTRODUCTION
The city of Almaty (Fig. 1), formerly known as Verny, is the
largest city in Kazakhstan, with around 1.85 million inhabi-
tants, and was the countrys capital until 1997. During the past
two centuries, Almaty has suffered significant damage due to
several large earthquakes (Silacheva et al., 2014;Kulikova and
Krüger, 2015;Krüger et al., 2017;Mosca et al., 2019). In par-
ticular, the 1887 Mw7.3 Verny earthquake (Arrowsmith et al.,
2017) struck the newly built town. At that time, the town
mainly consisted of adobe buildings with a population of
around 30,000. As a result of this event, the town was nearly
totally destroyed, with a death toll of nearly 300.
Another damaging event was the 1911 Mw7.8 Kemin
earthquake (Bindi et al., 2014), and, from the reports of that
time, all buildings suffered some degree of damage (Bogdanovic,
1911;Bogdanovich et al., 1914). Because of this earthquake, 390
people died, 44 of them in Verny itself. Remarkably, this earth-
quake generated in Verny (nearly 40 km from the epicenter)
large amounts of soil deformation and ground failure, in particu-
lar in the loam sandy soils. Cracks in some places reached 1 m in
width and 5 m in depth (Fig. 2).
While the Verny earthquake struck the city at the end of
spring (9 June), the Kemin event happened in the middle of
the winter season (3 January). The ground in Almaty, due to
the average air temperature during that period with values
significantly below zero (Razuvaev et al., 2008), is expected to
have been frozen in the uppermost meter, which might have led
to a different ground response to the incoming seismic waves.
It is worth remembering that a key step for seismic hazard
assessment and risk mitigation is the estimation of the ground
motion that earthquakes can generate at a given site. This
estimate must also include possible local effects due to the
propagation of waves in the shallower geological layers. For
small-to-moderate earthquake shaking, the soil can be expected
to respond linearly to ground-motion excitation, and simple
site response analysis, based on empirical data or numerical
simulations, are sufficient to provide an overview of the pos-
sible spatial variation of ground motion. However, when the
level of shaking increases, soft soil material starts behaving non-
linearly and, in particular cases, liquefaction phenomena and
ground failure might occur (Kramer, 1996).
Within several initiatives aiming at seismic risk assessment
and mitigation in Central Asia (e.g., the Earthquake Model
Central Asia [EMCA] project, see Data and Resources), the
site response in Almaty was estimated by means of earthquake
and noise recordings (Pilz et al., 2015). In addition, three array
measurements of ambient seismic noise were carried out to
extract shear-wave velocity profiles, an essential parameter for
doi: 10.1785/0220190114 Seismological Research Letters Volume XX, Number XX 2019 1
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
evaluating the dynamic properties of soil, and to characterize
the corresponding sediment layers at each site. These data to-
gether with those derived by previous investigations (Silacheva
et al., 2014) allow the identification of areas with different site
responses. However, these kinds of analyses are not sufficient
for answering the question of why the Kemin earthquake cause
a different response compared to the Verny event (in particular,
a much greater occurrence of liquefaction and ground failure).
The first reason for explaining the differences might be the
different levels of ground shaking induced by each earthquake.
However, Figure 3shows that when considering the estimated
epicenter positions of the two events and their corresponding
magnitudes, the level of shaking induced in Almaty, at least in
terms of peak ground acceleration (PGA), is similar. The level
of shaking was calculated using the ground-motion prediction
equation (GMPE) proposed by Boore and Atkinson (2008) as
suggested for intraplate areas by the Global Earthquake Model
(GEM) (Ullah, 2016).
Because major differences in the level of shaking can be
ruled out as being responsible for the different ground failure
caused by the two earthquakes, we therefore formulate and
verify the hypothesis that the strong liquefaction and ground
failure during the Kemin earthquake was caused by the pres-
ence of a frozen superficial soil layer. This layer, by preventing
drainage through the surface, did not allow the excess pore
water pressure to dissipate.
The analysis we propose is similar to that carried out by
Finn et al. (1978) to explain the occurrence of liquefaction
during the 1964 Mw9.2 Alaska earthquake. The recent 30
November 2018 Alaska earthquake dramatically provided
new evidence of such effects. In the case considered in this
study, we are not concentrating on the modification of the
ground motion due to the frozen layer as in Vinson (1978),
Wang et al. (2004), and Xu et al. (2011), but rather we mainly
focus on its influence on the pore pressure increase as a result of
the seasonal conditions. Unfortunately, because the amount of
information available about the soil conditions in Almaty is
still limited (in terms of geotechnical data), to verify if the pro-
posed hypothesis could be feasible, we explore several different
possible soil structure models.
The models were derived after analyzing the available
parameters that allow the evaluation of the dynamic properties
Figure 1. (a) Almaty (formerly Verny), southern Kazakhstan. The white stars indicate the epicenters of the two large earthquakes
considered in this study: the 9 June 1887 Mw7.3 Verny and the 3 January 1911 Mw7.8 Kemin events. (b) Present urban area of
Almaty and the position of the epicenters. The ancient urban settlement of Verny is marked by the gray box. The known active faults
are shown as thin black lines. (c) Distribution of macroseismic intensity in the Verny area due to the Kemin event (redrawn after
Nurmagambetov et al., 1999). The black lines indicate the areas where liquefaction and ground failures occurred.
2 Seismological Research Letters Volume XX, Number XX 2019
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
of soil. Then, we selected a strong-motion recording according
to the expected source characteristics (considering the regional
tectonics and information available from the literature) and
epicentral position with respect the target site. The recording
was matched to predefined response spectra using a suitable
GMPE to be compatible with the 1911 Kemin and 1887
Verny earthquakes. The resulting time series were used as input
for the dynamic analysis.
DATA AVAILABILITY AND MODELS
To study the S-wave velocity structure with depth, seismic
noise microarray measurements were carried out in Almaty.
In summer 2014, one array was installed in the northern part
of the city, close to the location of theVerny urban area at the
beginning of the twentieth century (gray box in Fig. 1b).
Array measurements were carried out using 16 broadband
receivers and for durations of not less than 120 min. Seismic
noise data were divided into windows of 120 s, and the
extended spatial autocorrelation method was applied (Ohori
et al., 2002;Parolai et al., 2005). The estimated Rayleigh-wave
dispersion curves were inverted jointly with the horizontal-
to-vertical (H/V) curves following the scheme proposed by
Parolai et al. (2005). The results indicated that the S-wave
velocity in the uppermost 10 m is equal to 167 m=s, while it
increases to 320 m=sover the next 20 m (Fig. 4). Although the
minimum misfit is not changing significantly after 75 itera-
tions, the average one shows large variations since the algorithm
is trying to explore different part of the solution space. The
depth of 30 m was assumed to be the input depth of theground
motion in the numerical simulations, considering that lique-
faction is expected to occur mainly in the uppermost 20 m.
According to the available stratigraphy data for the area
(Silacheva et al., 2014) these two layers can be, to a first order,
assigned to a shallow sand-sandy loam layer and a gravel-sandy-
pebbles layer, respectively.
This information has been used to setup the main starting
parameters for what will be referred to from now as the
summer model. The model with a frozen uppermost one meter
layer will be termed the winter model.
Considering the daily temperature measured in Almaty
since 1915 (Razuvaev et al., 2008), an average daily value
was estimated (Fig. 5). At the time of the 1911 Kemin earth-
quake, consistent with the available photographs taken a few
days after the event that showed snow and ice on the ground
(Fig. 2), the air temperature was likely to be well below zero for
Figure 3. Estimated peak ground acceleration (PGA) for the
1887 Mw7.3 Verny and the 1911 Mw7.8 Kemin earthquakes
(Bindi et al., 2014) using the ground-motion prediction equation
(GMPE) proposed by Boore and Atkinson (2008), as suggested
for intraplate areas by the Global Earthquake Model (GEM). R
is the closest horizontal distance to the earthquakesepicenters.
Figure 2. Ground failure effects documented after the 3 January 1911 Kemin earthquake (modified from Nurmagambetov et al., 1999).
Seismological Research Letters Volume XX, Number XX 2019 3
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
a few weeks prior to the event. Starting from the average air
temperature in Almaty on 3 January, and considering the rela-
tionships relating the ground surface temperature to temper-
ature at different depths (Andersland and Ladanyi, 2004) and
the evidence regarding the frost penetration in the area of
Almaty reported in Teltayev et al. (2016), it is reasonable to
assume that at least the first meter of soil was frozen at the time
of the Kemin earthquake.
Accordingly to Finn et al. (1978), who provided shear
modulus values as a function of soil temperature, the velocity
in the uppermost 1 m layer (considered to be frozen) was fixed
to 867 m=s. This value is valid when considering small strains
and after having interpolated Finns curve according to the
estimated temperature of the ground using an adjustable ten-
sion continuous curvature spline. Finally, we tested different
positions of the water table depth (1, 3, and 5 m).
NUMERICAL SIMULATIONS
To evaluate the role of the frozen surface layer,
we carried out the dynamic analyses in terms of
effective stress. In this study, the analysis was
carried out using the DEEPSOIL software
(Hashash et al., 2017), which has the capability
to simulate the effect of increasing pore water
pressure during earthquake shaking on the dy-
namic properties of the unfrozen soil, and the
drainage and redistribution of pore water pres-
sures under dynamically induced pore pressure
gradients.
A time-domain nonlinear analysis was
carried out based on solving the equation of
motion or dynamic equilibrium equation
EQ-TARGET;temp:intralink-;df1;370;553Mfu
:: gCf_ugKfugMfIgu
::
g;
1
in which [M] is the mass matrix, [C]isthe
viscous damping matrix, [K] is the stiffness
matrix, is the vector of relative acceleration,
f_ugis the vector of relative velocities, fugis the
vector of relative displacements, [u
:: ] is the accel-
eration at the base of the soil column, fIgis the
unit vector, and the [M], [C], and [K] matrices
are assembled using the incremental response of
the soil layers (Hashash et al., 2010). The soil
response is obtained from a constitutive model
that describes the cyclic behavior of the soil.
Equation (1) is solved numerically at each time
step using a time integration method (e.g., the
Newmark, 1959,βmethod).
The soil column is discretized into in-
dividual layers using a multi-degree-of-freedom
lumped parameter model or finite elements
(Kramer, 1996). The calculation process for a
nonlinear model is as follows. First, an input
acceleration time series is used to determine
the motion at the base of the soil profile. Then, the motion
at each layer boundary is calculated, moving from the bottom
of the soil profile to the top.
The stiffness and damping values for each layer were
derived based on the Darendeli (2001) models for sand and
gravel deposits, respectively. For the frozen sandy layer at the
surface (winter model), the Singh and Donovan (1977) rela-
tionships were considered.
The Darendeli (2001) relationships require five input
parameters: plasticity index (PI), over consolidation ratio
(OCR), mean effective confining pressure (in atmospheres
(σ
m), loading frequency (f) in hertz, and the number of load-
ing cycles (N)). In general, a major role is played by the PI and
mean effective confining pressure. To account for the approxi-
mate knowledge of the stratigraphy, the analysis considered
three values of the PI (PI 0, 5, 10). However, considering
Figure 4. Shear-wave velocity profiles for array measurement sites in Almaty
(see Fig. 1b, the grey box). (a) Tested S-wave velocity models (inset refers to Pwave)
from the considered genetic algorithm (gray lines) and the best-fitting model (black
line). (b) Minimum misfit (black dots) and the average misfit (gray line) at each gen-
eration. (c) Observed phase velocities (filled gray circles) and the calculated one for
the best-fitting model (black circles). (d) Observed horizontal-to-vertical (H/V) ratio
(filled gray circles) and the calculated one for the best-fitting model (black circles).
4 Seismological Research Letters Volume XX, Number XX 2019
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
the similarity of the curves for the selected three values, and of
the obtained results, only the results of the calculations for
PI 0(nonplastic soils) will be presented in the following.
Darendeli (2001) found that G=Gmax (in which Gmax is
the shear modulus at zero strain and Gis the secant shear
modulus), increases slightly for cohesive soils, whereas the
ratio does not increase for cohesionless soils, with the OCR.
Therefore, we neglected in this study the effect of the OCR.
Regarding the surficial frozen layer in the winter model, we
adopted the Vinson et al. (1977) parameters for typically fro-
zen soils in interior Alaska.
The adopted procedure, implemented in DEEPSOIL,
allows the change in pore water pressure and soil degradation
due to cyclic loading to be predicted. This includes the strain-
based pore pressure generation model of Matasovićand
Vucetic (1993) for sands, based on the model developed by
Dobry et al. (1985) for saturated sands. Dobry et al. (1985)
showed that the residual excess pore pressure ratio, ru, is pro-
portional to the number of applied shearing cycles Nand the
increase in shear strain γ. This result was formalized by Vucetic
and Dobry (1988) as follows:
EQ-TARGET;temp:intralink-;df2;323;589ruuN
σ
v
p×f×F×N×γγtvs
1f×F×N×γγtvs;2
in which ruis the residual excess pore pressure ratio, uNis the
residual excess pore pressure after Ncycles, σ
vis the initial
vertical effective stress before shearing, γtv is the volumetric thresh-
old shear strain, and f,p,F,andkare curve-fitting parameters.
Some empirical correlation relations for the curve-fitting
parameters Fand sproposed by Carlton (2014) for sands, hav-
ing the following functional form, have been used:
EQ-TARGET;temp:intralink-;df3;323;458F3810V1:55
S;3
EQ-TARGET;temp:intralink-;df4;323;411sFC 10:1252 ;4
in which VSis the shear-wave velocity in m=sand FC is the
percentage of fines content.
Figure 5. Average daily temperature ± standard deviation of
the surface temperature for the city of Almaty. The graph was
obtained by averaging the daily temperatures recorded in Almaty
between 1915 and 2001.
Table 2
Main Characteristics of the Selected and the Matched Recordings
Event MwTPGA (g)R(km) VS30 (m=s)
1999 Chi-Chi earthquake recording (station HWA014) 7.6 Reverse 0.1 52 277
1911 matched Kemin earthquake recording 7.8 Reverse 0.199 40 245
1887 matched Verny earthquake recording 7.3 Reverse 0.202 23 245
Mw, moment magnitude; PGA, peak ground acceleration; R, epicentral distance; T, rupture type.
Table 1
Typical Values of Vertical Permeability and the Consolidation Coefficient (Pestana et al., 1997;Carlton, 2014) Considered in This
Study
Soil Typical Names
Fine Content
Percent (%)
Vertical
Permeability (m=s)
Consolidation
Coefficient
Clayey sands, sand-silt mixtures or silty sands, sands silt mixtures 12 <FC <50 3 ×1050.0612
Poorly graded sand with clay 5<FC <12 8 ×1050.1632
Poorly graded sands or well-graded sand FC <55×1041.02
Gravels FC = 0 5×10210.2
FC, percentage of fine content.
Seismological Research Letters Volume XX, Number XX 2019 5
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
DEEPSOIL models the dissipation and redistribution of
residual excess pore pressures using the Terzaghi 1D consoli-
dation theory (Hashash et al., 2017), expressed as
EQ-TARGET;temp:intralink-;df5;40;213
u
tCV2u
z2;5
in which uis the residual excess pore pressure, tis the time, zis
the depth, and CVis the coefficient of consolidation coeffi-
cient given by the following formula for the cohesionless soils
(sand and gravel) (Carlton 2014):
EQ-TARGET;temp:intralink-;df6;40;117CVk
mv×γm
;6
in which CVis in m2=s,kis the vertical permeability in m=s,
mvis the volumetric compressibility in m2=kN, and γmis the
water unit weight equal to 9:8kN=m3.
Based on the Unified Soil Classification System and
Pestana et al. (1997), both the value of vertical permeability
kand the volumetric compressibility mvwere chosen according
to their fine material content, as listed in Table 1.
Because of the approximations made with regards to the
granulometry of the soil, the numerical simulations have been
carried out for different values of FC in the uppermost 10 m of
the sand layer. The values are nonetheless considered to be rea-
sonable based on the knowledge of the grain size composition
of soils in the study area (Nisii, 2009).
To carry out the numerical simulations, the choice of the
input strong-motion time series is crucial. To be able to carry
Figure 6. Time series of the 1999 Chi-Chi earthquake (top panel) selected to represent the ground motion experienced in Verny during
the Kemin and Verny events. The results of the spectral matching are shown in the central panel for the 1887 Verny earthquake and in the
bottom panel for the 1911 Kemin earthquake.
6 Seismological Research Letters Volume XX, Number XX 2019
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
out these simulations as realistically as possible, because no
strong-motion recordings of large events in the study area are
available (and, obviously, for the studied earthquakes), a selec-
tion has to be done considering the available strong-motion
data bases. The selection mainly considered recordings for
earthquakes with focal mechanisms similar to that estimated
for the Kemin (Kulikova and Krüger, 2015) and Verny events,
with comparable positions with respect to the fault.
After having searched the international data bases, the
input motion was selected from recordings of the 1999
Mw7.6 Chi-Chi earthquake. This event had a reverse focal
mechanism and a hypocentral depth estimated at 21.2 km
(Global Centroid Moment Tensor catalog, see Ekström et al.,
2012). This depth is similar to that calculated by Mosca et al.
(2019) and Kulikova and Krüger (2015) for the Kemin and
Verny earthquakes. The selected recording was that of station
HWA014, which is at a similar location with respect to the
epicenter as the city of Verny was with respect to the ruptures
that generated the Verny and Kemin earthquakes.
VS30, that is, the travel-time averaged S-wave velocity in
the uppermost 30 m, at the selected recording site is equal
to 277 m=s, which is slightly different from that estimated in
Almaty (245 m=s) from the array measurements. The strong-
motion record has been scaled using the SeismoMatch software
(Seismosoft, 2016) to account for the differences in magnitude,
epicentral distance, and VS30. The software, developed by
Abrahamson (1992) and then updated by Hancock et al.
(2006), modifies an acceleration time history in the time
domain to match it with a specified spectrum using the tech-
nique proposed by Lilhanand and Tseng (1987,1988).The
target spectrum was generated using the OpenSHA software
developed by the Southern California Earthquake Center
and the U. S. Geological Survey (Field et al., 2003). A target
spectrum was calculated to simulate the Kemin and Verny
earthquake recordings by considering the GMPE of Boore
and Atkinson (2008), which is recommended by GEM for
active shallow crust conditions, such as those at hand.
Table 2summarizes the main characteristics of the selected
and the matched recordings. A PGA value of nearly 0:2gis
expected, based on the GMPEs of Boore and Atkinson (2008),
to have affect Verny during the considered earthquakes.
It is worth noting that the recording has been matched with
both the Verny and Kemin earthquakes by considering in both
cases summer and winter velocity profiles. However, because the
analysis carried out showed that the resultsare not dependenton
the employed input recordings, but only on the soil velocity pro-
file, only the results obtained for the Kemin event, with winter
soil conditions, and theVerny one, for the summer velocity pro-
file, are shown for sake of brevity. Figure 6shows the original
selected recording and the simulated recordings for the Verny
and Kemin earthquake after spectral matching.
The adequacy of the spectral matching procedure is also
shown in Figure 7, in which the target response spectra deter-
mined by considering the GMPE of Boore and Atkinson
(2008) and the matched one are reported.
RESULTS
The results regarding the pore water pressure generation,
obtained for site response analysis for the Verny and Kemin
earthquakes, are shown in Figure 8, which shows the results as
a function of different water table positions (1, 3, and 5 m) and
vertical permeability k.
The pore water pressure curves are shown for the maxi-
mum values reached during the numerical calculations versus
the effective vertical stress curves. It is remarkable that for the
two highest kvalues, the pore water pressure reaches the value
of the effective stress at the bottom of the frozen layer.
These kvalues are also reasonably consistent, although on
the lower bound, with the thresholds determined by Finn et al.
(1978), who found that when kis larger than 103m=s, the
redistribution of pore water pressure-induced liquefaction in
the tested model. On the contrary, when kwas smaller than
104m=s, no liquefaction occurred.
In all the other cases, drainage through the soil layers up to
the surface is allowed, affecting the internal redistribution of
pore water pressure. The increases of pore water pressure at
around 8 m depth in all of the presented model could be due
to the change in the velocity profile at 10 m depth, where an
impedance contrast and change in permeability exists between
the sand and the gravel layers.
In the following, considering that the results seem to be
independent of the chosen position of the water table, only those
regarding the water table positioned at 1 m depth will be shown.
Figure 9shows the pore water pressure at different times
during the strong shaking as a function of depth. During the
maximum level of shaking, the pore water pressure is increased
mainly at 8 m depth, but it reaches a maximum at nearly 8 m
and 1 m depth after 60 s (Fig. 9). For the Kemin event, as
Figure 7. (a) Target and matched response spectrum for the
1887 Verny earthquake (distance 23 km). (b) Target and matched
response spectrum for the 1911 Kemin earthquake (distance 40 km).
Seismological Research Letters Volume XX, Number XX 2019 7
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
Figure 8. Maximum pore water pressure versus depth while considering different water table (WT) depths (from left to right, 1, 3, and
5 m, respectively). Different line styles (dashed, dotteddashedcontinuous) show the results obtained for three different values of vertical
permeabilities (k). The thick black line indicates the effective vertical stress. (a) Results obtained for the summer profile and the Verny
event. (b) Results obtained for the winter profile and the Kemin event.
8 Seismological Research Letters Volume XX, Number XX 2019
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
shown in Figure 9, the pore water pressure at 1 m depth reaches
the effective pressure, therefore generating liquefaction. At the
end of the strong shaking phase (around 90 s), the pore water
pressure has lowered with depth, although for the Kemin event
it is still as high as the effective one.
Figure 10 shows the development of the pore water pres-
sure ratio (PWPR) versus time under undrained conditions
(Kemin earthquake) and drained conditions (corresponding
to those of the Verny earthquake) at a depth of 1 m for
the three considered vertical permeability values. Clearly, the
PWPR increases with the arrival of the strong ground motion
phase (nearly 40 s). For undrained conditions, when kis equal
to 5×104and 8×105m=s, the PWPR reaches a value of 1
after 55 and 80 s, respectively. This is obviously due to the
larger values of kfacilitating the migration of pore water pres-
sure to the soil in a shorter time. When kis equal to
3×105m=s, the PWPR increases steadily, but remains at val-
ues as low as 20% at the end of the strong ground shaking.
Under drained conditions, the two lower kvalues lead
to a continuous increase of the PWPR, starting after the
strong-motion phase arrival. When the highest value of the
vertical permeability is considered, a decrease of the PWPR
is observed after 60 s. From Figures 810, it is therefore clear
that the frozen layer inhibits the drainage to the surface and
Figure 9. Pore water pressure versus depth at different time for the case of the water table at 1 m depth (see Fig. 6for the reference
time). The dashed line shows the pore water pressure distribution with depth after 40 s, the dasheddotted line shows after 60 s, and the
dotted line shows after 90 s. The thick black line depicts the effective vertical stress. (a) Results for the Verny earthquake. (b) Results for
the Kemin earthquake. WT, water table.
Figure 10. Pore water pressure development at a depth of 1 m, considering the depth of the water table being equal to 1 m for different
values of k.
Seismological Research Letters Volume XX, Number XX 2019 9
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
allows the pore water pressure to reach the value of the effective
stress, therefore increasing the liquefaction potential.
DISCUSSION AND CONCLUSIONS
In this study, we attempted to provide an explanation for the
different well-documented coseismic effects (mainly ground
failure and liquefaction) of two historical earthquakes that
occurred in Central Asia, namely the 1887 Mw7.3 Verny and
the 1911 Mw7.8 Kemin earthquakes. Considering the scarcity
of available geotechnical data, we developed different models of
the subsurface structure and considered different positions of
the water table. After having selected a suitable recording of the
ground motion to match what is believed to have been the
situation during these events, and adequately scaled it by means
of a spectral matching procedure, we carried out dynamic
analysis. The results indicated that, for values of vertical per-
meability compatible with the soil typology of the study area,
the presence of a shallow frozen layer whose presence is likely
in winter considering the available knowledge of soil condi-
tions and the air temperature in the area, could have increased
the liquefaction potential of shallow layers by preventing the
external drainage of pore water pressure. This could have in
turn lead the large amount of ground failure observed during
the 3 January 1911 Kemin earthquake.
In fact, an internal redistribution of pore water pressure
could have occurred because of the seismically induced pore
water pressure gradient. As a result, the pore water pressure
ratio could have increased in the uppermost layers of the
unfrozen soils where the effective stress is low. We showed that
the higher the vertical permeability, the faster the transfer of
seismically induced pressures and the higher the susceptibility
of the unfrozen saturated layer to liquefaction in the upper
levels.
We are aware that our study is based on a still poor data
set. Nevertheless, we think that, because the parameters used in
this work are scientifically sound for the material existing in the
area, and the level of ground motion considered was reasonable
for the size and epicentral of the considered event, the obtained
results shed some light on the importance of secondary effects
that could affect the city during earthquake shaking.
Furthermore, these results highlight the importance, in
areas where due to cold winter the shallow most soil layer
is frozen, of assessing the effect of local soil amplification
on ground motion and considering the seasonal influence
on secondary effects such as liquefaction and ground failure.
We also remark that a careful analysis of the available
description of the effects of an historical earthquake, while
qualitative, could still provide important hints for quantitative
analyses and an improvement of a quantitative assessment of
seismic hazard and risk.
DATA AND RESOURCES
All data used in this article came from published sources listed
in the references. Some plots were made using the Generic
Mapping Tools version 5.4.5 (www.soest.hawaii.edu/gmt;
Wessel and Smith, 1998, last accessed June 2019). The
Earthquake Model Central Asia (EMCA) project is available
at http://www.emca-gem.org (last accessed June 2019).
ACKNOWLEDGMENTS
This research was made possible, thanks to the support of the
International Center forTheoretical Physics (Trieste, Italy). The
authors thank D. Bindi and S. Orunbaev for the help in the field
work. M. Santulin provided suggestions for the spectral match-
ing procedure. K. Fleming kindly improved our English.
REFERENCES
Abrahamson, N. A. (1992). Non-stationary spectral matching , Seismol.
Res. Lett. 63, 30.
Andersland, O. B., and B. Ladanyi (2004). Frozen Ground Engineering,
John Wiley & Sons, Hoboken, New Jersey, 384 pp.
Arrowsmith, J. R., C. J. Crosby, A. M. Korzhenkov, E. Mamyrov, I.
Povolotskaya, B. Guralnik, and A. Landgraf (2017). Surface rupture
of the 1911 Kebin (ChonKemin) earthquake, Northern Tien
Shan, Kyrgyzstan, Geol. Soc. Lond. Spec. Publ. 432, 233253.
Bindi, D., S. Parolai, A. Gómez-Capera, M. Locati, Z. Kalmetyeva, and N.
Mikhailova (2014). Locations and magnitudes of earthquakes in
Central Asia from seismic intensity data, J. Seismol. 18, 121.
Bogdanovic, K. I. (1911). An earthquake of December 22, 1910 (January
4, 1911) in northern chains of the Tien Shan between Verny and
Issyk-Kul, Proc. Geol. Comm. 30, 329419.
Bogdanovich, K. I., I. M. Kark, B. Y. Korolkov, and I. V. Muchketov
(1914). Earthquake of the 4th January 1911 in the northern districts
of the Tien Shan, Trans. Geol. Comm. Ser. 89, 270.
Boore, D., and G. Atkinson (2008). Ground-motion prediction equa-
tions for the average horizontal component of PGA, PGV, and
5%-damped PSA at spectral periods between 0.01 s and 10.0 s,
Earthq. Spectra 24, 99138.
Carlton, B. (2014). An improved description of the seismic response of
sites with high plasticity soils, organic clays, and deep soft soil
deposits, Ph.D. Thesis, University of California, Berkeley.
Darendeli, M. B. (2001). Development of a new family of normalized
modulus reduction and material damping curves, Ph.D. Dissertation,
University of Texas at Austin, Austin, Texas.
Dobry, R., A. Vasquez-Herrera, R. Mohamad, and M. Vucetic (1985).
Liquefaction flow failure of silty sand by torsional cyclic tests, in
Advances in the Art of Testing Soils Under Cyclic Conditions,V.
Khosla (Editor), American Society of Civil Engineers, New York,
New York, 2950.
Ekström, G., M. Nettles, and A. M. Dziewonski (2012). The global CMT
project 2004-2010: Centroid-moment tensors for 13,017 earth-
quakes, Phys. Earth Planet. In. 200/201, 19, doi: 10.1016/
j.pepi.2012.04.002.
Field, E. H., T. H. Jordan, and C. A. Cornell (2003). OpenSHA: A devel-
oping community-modeling environment for seismic hazard analy-
sis, Seismol. Res. Lett. 74, 406419.
Finn, L., R. N. Yong, and K. W. Lee (1978). Liquefaction of thawed
layers in frozen soils, J. Geotech. Eng. Div. ASCE 104, 12431255.
Hancock, J., J. WatsonLamprey, N. A. Abrahamson, J. J. Bommer, A.
Markatis, E. M. M. A. McCoy, and R. Mendis (2006). An improved
method of matching response spectra of recorded earthquake
ground motion using wavelets, J. Earthq. Eng. 10, 6789.
Hashash, Y. M. A., M. I. Musgrove, J. A. Harmon, O. Ilhan, D. R.
Groholski, C. A. Phillips, and D. Park (2017). DEEPSOIL 7.0,
User Manual, Board of Trustees of University of Illinois at
Urbana-Champaign, Urbana, Illinois.
10 Seismological Research Letters Volume XX, Number XX 2019
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
Hashash, Y. M. A., C. Phillips, and D. R. Groholski (2010). Recent
advances in non-linear site response analysis, International Confer-
ences on Recent Advances in Geotechnical Earthquake Engineering
and Soil Dynamics, 8, available at http://scholarsmine.mst.edu/
icrageesd/05icrageesd/session12/8 (last accessed June 2019).
Kramer, S. L. (1996). Geotechnical Earthquake Engineering, Prentice Hall,
Upper Saddle River, New Jersey, 653 pp.
Krüger, F., G. Kulikova, and A. Landgraf (2017). Instrumental magnitude
constraints for the 11 July 1889, Chilik earthquake, Geol. Soc. Lond.
Spec. Publ. 432, 4172.
Kulikova, G., and F. Krüger (2015). Source process of the 1911 M 8.0
Chon-Kemin earthquake: Investigation results by analogue seismic
records, Geophys. J. Int. 201, 18911911.
Lilhanand, K., and W. S. Tseng (1987). Generation of synthetic time
histories compatible with multiple damping design response spectra,
SMiRT-9, Lausanne, Switzerland, K2/10, 105110.
Lilhanand, K., and W. S. Tseng (1988). Development and application of
realistic earthquake time histories compatible with multiple damp-
ing design spectra, Proc. of the 9th WCEE, Vol. II, TokyoKyoto,
Japan, 819824.
Matasović, N., and M. Vucetic (1993). Cyclic characterization of liquefi-
able sands, J. Geotech. Eng. 119, 18051822.
Mosca, I., B. Baptie, S. Sargeant, and R. T. Walker (2019). Integrating
outcomes from probabilistic and deterministic seismic hazard
analysis in the Tien Shan, Bull. Seismol. Soc. Am. 109, 688715.
Newmark, N. M. (1959). A method of computation for structural
dynamics, J. Eng. Mech. Div. 85, 6794.
Nisii, O. (2009). Study on risk management of earthquakes in the city of
Almaty, Republic of Kazakhstan, Final Report, Japan International
Cooperation Agency, OYO International Corp., Nippon Koei Co.,
Ltd., Tokyo, Japan, 134 pp.
Nurmagambetov, A., N. Mikhailova, and W. Iwan (1999). Seismic haz-
ard of the Central Asia region, in Seismic Hazard and Building
Vulnerability in Post-Soviet Central Asian Republics, S. A. King,
V. I. Khalturinand, and B. E. Tucker (Editors), Kluwer Academic
Publishers, Dordrecht, The Netherlands, 143.
Ohori, M., A. Nobata, and K. Wakamatsu (2002). A comparison of
ESAC and FK methods of estimating phase velocity using arbitrarily
shaped microtremor arrays, Bull. Seismol. Soc. Am. 92, 23232332.
Parolai, S., M. Picozzi, S. M. Richwalski, and C. Milkereit (2005). Joint
inversion of phase velocity dispersion and H/V ratio curves from
seismic noise recordings using a genetic algorithm, considering
higher modes, Geophys. Res. Lett. 32, 14.
Pestana, J. M., C. E. Hunt, and R. R. Goughnour (1997). FEQDrain: A
Finite Element Computer Program for the Analysis of the Earthquake
Generation and Dissipation of Pore Water Pressure in Layered Sand
Deposits with Vertical Drains, Report No. UCB/EERC-97-15,
Earthquake Engineering Research Center, University of California,
Berkeley, California.
Pilz, M., T. Abakanov, K. Abdrakhmatov, D. Bindi, T. Boxberger, B.
Moldobekov, S. Orunbaev, N. Silacheva, S. Ullah, S. Usupaev, et al.
(2015). An overview on the seismic microzonation and site effect
studies in central Asia, Ann. Geophys. 58, no. 1, S0104, doi:
10.4401/ag-6662.
Razuvaev, V. N., E. G. Apasova, and R. A. Martuganov (2008). Daily
temperature and precipitation data for 223 former-USSR stations,
ORNL/CDIAC-56, NDP-040, Carbon Dioxide Information
Analysis Center, Oak Ridge National Laboratory, U.S. Department
of Energy, Oak Ridge, Tennessee, doi: 10.3334/CDIAC/cli.ndp040.
Seismosoft (2016). SeismoMatch v2.1A computer program for spec-
trum matching of earthquake records, available at http://www
.seismosoft.com (last accessed June 2019).
Silacheva, N., U. Kulbayeva, and N. Kravchenko (2014). Seismic ground
motion variations resulting from site conditions, Geodes. Geodynam.
5, 915.
Singh, S., and N. C. Donovan (1977). Seismic response of frozen-thawed
soil systems, Proc. of the 6th International Conference on Earthquake
Engineering, New Delhi, India, 1014 January, paper no. 19, session
6, preprints, 611616.
Teltayev, B., A. Baibatyrov, and E. Suppes (2016). Characteristics of high-
way subgrade frost penetration in regions of the Kazakhstan,
Japanese Geotech. Soc. Spec. Publ. 2, 16641668.
Ullah, S. (2016). Seismic hazard assessment in Central Asia: Combining
site effects investigations and probabilistic seismic hazard, Doctoral
Dissertation, Technische Universität Berlin, Berlin, 163 pp.
Vinson, T. S. (1978). Response of frozen ground to dynamic loadings, in
Geotechnical Engineering for Cold Regions, O. B. Andersland and D.
M. Anderson (Editors), McGraw-Hill Book Co., New York, New
York, 405458.
Vinson, T. S., R. Czajkowski, and J. Li (1977). Dynamic properties of
frozen cohesionless soils under cyclic triaxial loading conditions,
Rept. No. MSU-CE-77-1, Division of Engineering Research,
Michigan State University, East Lansing, Michigan.
Vucetic, M., and R. Dobry (1988). Cyclic triaxial strain-controlled testing
of liquefiable sands, in Advanced Triaxial Testing of Soil and Rock,R.
Donaghe, R. Chaney, and M. Silver (Editors), ASTM International,
West Conshohocken, Pennsylvania, 475485, doi: 10.1520/
STP29093S.
Wang, L., X. Ling, X. Xu, and Q. Hu (2004). Study on response spec-
trum characteristics of earthquake acceleration for roadbed on
permafrost site, Chin. J. Rock Mech. Eng. 23, 13301335.
Wessel, P., and W. H. F. Smith (1998). New, improved version of the
Generic Mapping Tools released, EOS Trans. AGU 79, 579.
Xu, G., Z. Yang, U. Dutta, L. Tang , and E. Marx (2011). Seasonally frozen
soil effects on the seismic site response, J. Cold Reg. Eng. 25, 5370.
Rami Alshembari
International Centre for Theoretical Physics
Str. Costiera, 11
34151 Trieste, Italy
ralshemb@ictp.it
Stefano Parolai
Denis Sandron
Istituto Nazionale di Oceanografia e di Geofisica Sperimentale
Borgo Grotta Gigante 42C
34010 Sgonico (TS), Italy
sparolai@inogs.it
dsandron@inogs.it
Tobias Boxberger
Marco Pilz
Helmholtz-Zentrum Potsdam, German Research Center for
Geosciences
Telegrafenberg
14473 Potsdam, Germany
tobias.boxberger@gmx.de
pilz@gfzpotsdam.de
Natalya Sylacheva
Limited Liability Company
Institute of Seismology
Al-Farabi, 75A
050060 Almaty, Kazakhstan
silacheva_nat@mail.ru
Published Online 6 November 2019
Seismological Research Letters Volume XX, Number XX 2019 11
SRL Early Edition
Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/doi/10.1785/0220190114/4862049/srl-2019114.1.pdf
by OGS Inst Naz Oceanografia Geofisica Sperim - Biblioteca, Denis Sandron
on 07 November 2019
... We acknowledge the limitations of using Vs30 derived from slope, especially for local risk assessments (Lemoine et al., 2012;Narciso et al., 2013), but in absence of more precise information we consider this the best option available. The Allen and Wald (2009) study also does not include analysis on whether this is a proxy for dry or water-saturated soil, and as such also does not account for any seasonal changes in site effects as has been suggested is important for Almaty (Alshembari et al., 2019). ...
... All earthquake scenarios are suggested to extend from the surface to 20 km depth, except the Kazakh platform earthquake, which is buried, and extends from 30 to 50 km depth. deformation and ground failure within Almaty itself, and is suggested to be because a shallow frozen soil layer induced liquefaction at depth (Alshembari et al., 2019). ...
... We calculate that with the current exposure of Almaty, a Chon-Kemin event would cause 1,700  E 2,100 fatalities (0.1% of the population) and 6,500  E 3,900 completely damaged residential buildings (12.9%). We note these OpenQuake risk estimations do not include any effects of liquefaction, which was reported to have been induced at depth due to a shallow frozen soil layer (Alshembari et al., 2019). Modeling the sensitivity of the season by adjusting Vs30 to compare results from dry soil (summer) and water-saturated soil (spring and water) would be an interesting area of future study. ...
Article
Full-text available
Major faults of the Tien Shan, Central Asia, have long repeat times, but fail in large (Mw 7+) earthquakes. In addition, there may be smaller, buried faults off the major faults which are not properly characterized or even recognized as active. These all pose hazard to cities along the mountain range front such as Almaty, Kazakhstan. Here, we explore the seismic hazard and risk for Almaty from specific earthquake scenarios. We run three historical‐based earthquake scenarios (1887 Verny Mw 7.3, 1889 Chilik Mw 8.0 and 1911 Chon‐Kemin Mw 8.0) on the current population and four hypothetical scenarios for near‐field faulting. By making high‐resolution Digital Elevation Models (DEMs) from SPOT and Pleiades stereo optical satellite imagery, we identify fault splays near and under Almaty. We assess the feasibility of using DEMs to estimate city building heights, aiming to better constrain future exposure datasets. Both Pleiades and SPOT‐derived DEMs find accurate building heights of the majority of sampled buildings within error; Pleiades tri‐stereo estimates 80% of 15 building heights within one sigma and has the smallest average percentage difference to field‐measured heights (14%). A moderately sized Mw 6.5 earthquake rupture occurring on a blind thrust fault, under folding north of Almaty is the most damaging scenario explored here due to the modeled fault stretching under Almaty, with estimated 12,300±5,000 completely damaged buildings, 4,100 ± 3,500 fatalities and an economic cost of 4,700 ± 2,700 Million US dollars (one sigma uncertainty). This highlights the importance of characterizing location, extent, geometry, and activity of small faults beneath cities.
... The most recent strong earthquake that occurred on June 26, 2016 (M-6.4 H = 28 km) with maximum shock intensity, caused concurrent landslide formations at 27 different sites of the region and large landslides were simultaneously formed on 2−3 sites (Alshembari et al. 2020). Additionally, there are potential sites such as Maylusuu, Baypaza, Altynbel, and Naugarzan prone to landslides, where the rock displacement accelerated and repeated through fixed years (Haberland et al. 2011). ...
... Flooding, lithology and mudflow were only sparsely considered as factor in landslide research, each of them was appearing in seveen articles during 2000-2020 (Strom & Abdrakhmatov 2014;Torgoev et al. 2014). Additionally, aspect and land use/land cover were taken as factor influencing landslide process, and each of those factors was used in five publications, whereas air temperature, distance to roads, mountain glaciers and seismicity were employed in overall 16 publications (Havenith et al. 2002;Zhusupbekov et al. 2013;Kalmet'eva et al. 2019;Alshembari et al. 2020;Zhakulin et al. 2020). Each of 11 factors, such as altitude, bedrock, curvature, terrain morphology, elevation, glacier erosion, ground motions, river distance, soil texture, unstable masses, and vibration, were used in three papers within 2000-2020 (Danneels et al. , 2008Torgoev & Havenith 2013;Torgoev et al. 2014;Torgoev & Havenith 2016;Zhakulin et al. 2020). ...
Article
Full-text available
Landslides are among the major environmental hazards with large-scale socio-economic and environmental impacts that jeopardize socio-economic wellbeing in mountainous regions. Landslides are due to the interaction of several complex factors such as local or regional geology, geomorphology, topography, and seismic motions. The goal of this study is to review published articles on causes and effects of landslides in Central Asia throughout 2000–2020. In line with this goal, we have collected (using Scopus database), reviewed, and analyzed 79 papers published during 2000–2020. Our results revealed an increasing number of landslide studies in Central Asia during the period under investigation, with authors from Belgium dominating in the published outcomes (28% of total), followed by authors from Central-Asian countries. After then, the paper analyses the mostly applied models and frequently identified driving conditions and triggers of landsliding, such as aspect, altitude, soil types, precipitation, earthquakes and human interventions. Geographic information system (GIS) and remote sensing (RS) had not commonly been used in the papers between 2000 and 2010, and they have progressively been applied in landslide studies in Central Asia in the last decade. According to our analysis, geotechnical, geophysical and statistical methods were preferably used for the landslide studies in Central Asia.
... This includes (1) the variability in source azimuth, incidence angle and wave field on the input motion in which, for example, shallow events might generate local surface waves which can contribute to azimuthal-dependent amplification in certain frequency bands (Pilz et al. 2018), (2) the lateral variability of the site (i.e. surface and subsurface topography and (3) temporally variable nature of soil properties due to soil non-linearity (Régnier et al. 2018) and/or climatic factors (Alshembari et al. 2020). With various degrees all these effects will negatively influence the ability for 1-D groundmotion modelling at this site. ...
Article
One-dimensional (1D) site response analysis dominates earthquake engineering practice, while local 2D/3D models are often required at sites where the site response is complex. For such sites, the 1D representation of the soil column can account neither for topographic effects or dipping layers nor for locally generated horizontally propagating surface waves. It then remains a crucial task to identify whether the site response can be modelled sufficiently precisely by 1D analysis. In this study we develop a method to classify sites according to their 1D or 2D/3D nature. This classification scheme is based on the analysis of surface earthquake recordings and the evaluation of the variability and similarity of the horizontal Fourier spectra. The taxonomy is focused on capturing significant directional dependencies and inter-event variabilities indicating a more probable 2D/3D structure around the site causing the ground motion to be more variable. While no significant correlation of the 1D/3D site index with environmental parameters and site proxies seems to exist, a reduction in the within-site (single-station) variability is found. The reduction is largest (up to 20%) for purely 1D sites. Although the taxonomy system is developed using surface stations of the KiK-net network in Japan as considerable additional information is available, it can also be applied to any (non-downhole array) site.
... Also of interest are temporal variations in site response, which can occur for multiple reasons, such as inelastic soil deformations during a significant earthquake. Temporal changes in the groundwater table and the level of frozen soil (including permafrost) can also have significant effects on site response (Alshembari et al., 2019); research on these temporal effects is inextricably linked to ongoing global efforts to better understand the impacts of climate change. There are many potential avenues through which site response estimation may evolve in the decades ahead. ...
Article
Full-text available
In 2010 temperature and moisture sensors were made and installed into the pavement and subgrade of the "Astana-Burabay" highway (North region of the Kazakhstan). In 2013 similar sensors were used for the "Atyrau-Aktobe" (West region) and "Almaty-Bishkek" (South region) highways. In the present paper the study results of subgrade frost penetration of the above stated highways are given. For each highway a subgrade frost penetration starting and finishing times are determined; frost penetration and thawing graphs are constructed; frost penetration and thawing rate values are determined; dependence of the amount of unfrozen water from the value of negative temperature and initial moisture is established.
Article
In this study, we have evaluated the probabilistic and deterministic seismic hazard for the city of Almaty, the largest city in Kazakhstan, which has a population of nearly two million people. Almaty is located in the Tien Shan belt, a low-strain-rate environment within the interior of the Eurasian plate that is characterized by large infrequent earthquakes. A robust assessment of seismic hazard for Almaty is challenging because current knowledge about the occurrence of large earthquakes is limited, due to the short duration of the earthquake catalog and only partial information about the geometry, rupture behavior, slip rate, and the maximum expected earthquake magnitude of the faults in the area. The impact that this incomplete knowledge has on assessing seismic hazard in this area can be overcome using both probabilistic and deterministic approaches and integrating the results. First, we simulate ground-shaking scenarios for three destructive historical earthquakes that occurred in the northern Tien Shan in 1887, 1889, and 1911, using ground-motion prediction equations (GMPEs) and realistic fault-rupture models based on recent geomorphological studies. We show that the large variability in the GMPEs results in large uncertainty in the ground-motion simulations. Then, we estimate the seismic hazard probabilistically using a Monte Carlo-based probabilistic seismic hazard analysis and the earthquake catalog compiled from the databases of the International Seismological Centre and the British Geological Survey. The results show that earthquakes of M w 7.0-7.5 at Joyner-Boore distances of less than 10 km from the city pose a significant hazard to Almaty due to their proximity. These potential future earthquakes are similar to the 1887 Verny earthquake in terms of their magnitude and distance from Almaty. Unfortunately, this is the least well understood of the destructive historical earthquakes that have occurred in the northern Tien Shan.
Book
Preface: Frozen ground Physical and thermal properties Heat flow in soils Thaw behavior of frozen ground Mechanical properties of frozen soil Construction ground freezing Foundations in frozen soil Stability of soil masses in cold regions Earthwork in cold regions Field investigations Appendix A. Symbols Appendix B. SI Units Appendix C. Laboratory and field tests on frozen soils References Index.
Chapter
The territory under review is the north part of Central Asia including five republics of the former Soviet Union — Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, and Turkmenistan. It is a very complicated region in its geological-tectonic aspect, and it is at present one of the most highly seismic geostructural areas in the world. Much research work in various scientific fields has been directed towards studying the nature of earthquakes, the assessment of seismic hazard, and the development of methodologies for forecasting large earthquakes. A large amount of material has been collected on the different aspects of geology, tectonics, and seismic activity of the region, which shows the high level of seismic hazard in most parts of the republics, including the capital cities.
Thesis
Central Asia is one of the world’s most seismically active regions, with the highest level of seismic hazard. Usually seismic hazard is estimated considering ground motion at rock site, but since the ground motion could vary significantly over short distances due to local surficial geology, it is important to consider locally estimated site effects in seismic hazard assessment. Under the GSHAP project carried out in 1992-1999, seismic hazard was calculated at a global scale, including Central Asia. However, an update of this assessment is required in order to consider updated and more recent datasets. Therefore, purpose of this study is to: 1) assess the updated probabilistic seismic hazard at the regional level in Central Asia, and 2) consider the hazard assessment at the local level, including empirically-estimated site effects. As part of the GEM (Global Earthquake Model) initiative, this study is carried out within the EMCA (Earthquake Model Central Asia) project, which aims to calculate an updated cross border harmonized seismic hazard study at the regional level in Central Asia. In this study, the seismic hazard is calculated for Central Asia using an updated earthquake catalogue with respect to the Soviet times and the GSHAP project. The earthquake catalogue has been assembled to cover until 2009 from different sources, containing both instrumental and historical events, and is homogenized to surface wave magnitude MLH from different magnitude scales. Shallow seismicity (< 50 km) is considered for the calculation of seismic hazard assessment in the region. Different seismic source models are used for the calculation of seismic hazard. These include the area source model and smoothed seismicity models. In smoothed seismicity models, the approaches of Frankel (1995) and Woo (1996) approach are used. In particular, along with the Gaussian kernel function with fixed correlation distance (smoothing bandwidth) approach of Frankel (1995), the adaptive kernel function proposed by Stock and Smith (2002) is also implemented inside the Frankel (1995) approach. The seismic hazard is calculated in terms of macroseismic intensity (MSK-64), intended to be used for the seismic risk maps of the region, using the open source software platform OpenQuake. Most of the large cities in Central Asia lie on thick sediments, which influence the level of ground motion. Also, due to the current trend of urbanization, there is an urgent need to address the site effects in an urban level seismic hazard assessment. For this purpose, the empirical site effects are evaluated by considering both earthquake and seismic noise recordings in terms of spectral ratios and from the array analysis in terms of shear wave velocity. In this study, using clustering and correlation analysis, the spatial resolution of ground motion variability is improved upon in terms of standard spectral ratios, using earthquakes recorded at a few selected sites for a relatively short amount of time, and seismic noise data collected over a denser grid. This method is applied to Bishkek, Kyrgyzstan, where a K-means clustering algorithm is used to identify three clusters of site response type based on their similarity of standard spectral ratios. The cluster’s site responses are then adopted for sites where only single station noise measurements are carried out based on the results of correlation analysis. Here a first attempt is made to take into account the influence of the shallow geological structure on the seismic hazard for Bishkek, Kyrgyzstan, by using a proxy of Vs30 that has been estimated from in-situ seismic noise array analyses, and considering response spectral ratios calculated by analysing a series of earthquake recordings of a temporary seismic network. To highlight the spatial variability of the observed ground motion, the obtained results are compared with those estimated assuming a homogeneous Vs30 value over the whole urban area, corresponding to rock site condition. The seismic hazard is evaluated in terms of peak ground acceleration (PGA) and spectral acceleration (SA) at different periods (frequencies). The maximum hazard observed in the regional model reaches an intensity of around 8 in southern Tien Shan for a mean return period of 475 years. The maximum hazard estimated for some of the cities in the region, namely Bishkek, Dushanbe, Tashkent and Almaty, is between 7 and 8 (7-8), 8.0, 7.0 and 8.0 macroseismic intensity, respectively, for 475 years mean return period, using different approaches. Comparing these results, the current study shows that the hazard is generally higher by an order of 2 intensity units compared with that from the GSHAP project. The maximum hazard observed for rock site condition at the urban level for Bishkek is 0.45 g at a period of 0.1 s with a maximum PGA of 0.21 g, for a 475 years mean return period. When site effects are included through the Vs30 proxy in the seismic hazard calculation, the largest spectral acceleration of 0.64 g is obtained for a period of 0.1 s. In terms of PGA, in this case the largest estimated value reaches 0.31 g in the northern part of the city. When the variability of ground motion is accounted for through response spectrum ratios, the largest spectral acceleration reaches a value of 1.13 g at a period of 0.5 s. In general, considering site effects in the seismic hazard assessment of Bishkek leads to an increase in the estimated seismic hazard in the north of the city, which is thus identified as the most hazardous part within the study area and which is in fact further away from the faults and seismic sources. This study represents an update of the seismic hazard at regional and local scale.
Book
PART I: SITE CHARACTERISATION AND SITE AMPLIFICATION 1 Spatially constrained inversion of surface wave data to build shear wave Velocity models S. Foti, L.V. Socco 2 Site Classification and spectral amplification for seismic code provisions A. Anastasiadis, E. Riga PART II: LIQUEFACTION 3 Sand Liquefaction observed during recent earthquake and Basic Laboratory Studies on Aging effect T. Kokusho, Y. Nagao, F. Ito, T. Fukuyama 4 Liquefaction In Tokyo Bay And Kanto Regions In The 2011 Great East Japan Earthquake K. Ishihara, K. Araki 5 Allowable settlement and inclination of houses defined after the 2011 Tohoku - Pacific Ocean Earthquake in Japan S. Yasuda PART III: RIVER LEVEES AND DAMS 6 Seismic Performances of River Levees Experience and Prediction I. Towhata 7 Earthquake Performance Design of Dams using Destructive Potential Factors G. R. Saragoni PART IV: FOUNDATIONS AND SOIL-STRUCTURE INTERACTION 8 Seismic responses of shallow footings: a promising application for the macro- Element approach C. di Prisco, R., M. Maugeri 9 Large-Scale Modeling of Ground and Soil-Structure Earthquake Response K. Kim , A. Elgamal , G. Petropoulos , A. Askan , J. Bielak, G. L. Fenves 10 Seismic displacement based design of structures: relevance of soil structure interaction G. M. Calvi, M. Cecconi , R. Paolucci PART V: UNDERGROUND STRUCTURES 11 Performance and seismic design of underground structures K. Pitilakis, G. Tsinidis PART VI: SPECIAL TOPICS 12 Reinforced Soil Walls during the Recent Great Earthquakes in Japan and Geo- Risk based Design Y. Miyata 13 Performances Based Seismic Design of Geosynthetic Barriers for Waste Containment E. Kavazanjian, M. Arab, P. Fox, N. Matasovic
Article
A dynamic effective stress method is presented for assessing the liquefaction potential of thawed layers of saturated cohesionless soils sealed between frozen surface layers and permafrost. Such layers are common in Arctic regions. Analysis indicates that the liquefaction potential is increased by the presence of a frozen surface layer. The pore-water pressures created by dynamic stress gradients are redistributed upwards to regions of lower effective stresses, and they cannot dissipate because drainage is sealed off, thus leading to increased liquefaction potential. Within limits yet to be established, the coarser the soil, the greater the risk of liquefaction at low densities as upward redistribution of pore-water pressures is facilitated by increased permeability. Field data on liquefaction from the Alaska earthquake of 1964 are examined, and they appear to support the conclusions.
Article
The aseismic problem of roadbed and other relevant earth structure on permafrost site has become a key problem that is not studied sufficiently in the past and needs to be solved urgently. Based on several typical cases of roadbeds of Qinghai-Tibet Railway Project, the response spectrum characteristics of earthquake acceleration for roadbed on permafrost site are studied by using two-dimensional dynamic FEM. The high-frequency El Centro wave and low-frequency Tianjin wave are input as an excitation. The result indicates that whether there is frozen layer, earthquake acceleration response spectrum is mostly the type of high frequency impulse, and the soil failure is bump type. In the mid and low frequency stage, the values of earthquake acceleration response spectrum for roadbed with frozen soil are larger than that without frozen soil when input is El Centro wave. But for Tianjin wave, there is no significant difference. Compared to different topography conditions, the frequence of input earthquake wave has great influence on earthquake acceleration response spectrum of roadbed. Especially, the existence of thawed basin ground enhances the effect of earthquake acceleration response for roadbed.