Professional Documents
Culture Documents
net/publication/225583550
CITATIONS READS
80 263
1 author:
Prantik Mandal
National Geophysical Research Institute
123 PUBLICATIONS 2,334 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Prantik Mandal on 30 January 2017.
Abstract—Delineation of the top sedimentary structure and its Qs vs. Qp relationship using the
travel-time difference of direct S and converted Sp phase is key to understanding the seismic hazard of
any sedimentary basin area. We constructed filtered displacement waveforms from local ETNA
Episensor acceleration recordings as well as local velocity recordings of aftershocks of the 2001 Bhuj
earthquake recorded by the Kachchh seismological network of the National Geophysical Research
Institute (NGRI), Hyderabad, India during 2001–2004. Stations are within 15–70km of epicenters, and
the resulting displacement waveforms are generally simple, displaying prominent P, Sp, and S wave
pulses. Particle motion of P and S waves suggest near-vertical raypaths consistent with preliminary
depth estimates. The direct S wave on the horizontal component is characterized by lower frequency
content than the converted Sp phase on the vertical component. This difference in frequency content
between S and Sp phases can be explained in terms of different attenuation effects for P and S waves
in the unconsolidated sediments. The Sp phase is generated by S-to-P phase conversion at the base of
Mesozoic sediments of the Kachchh basin. Travel-time inversion (VELEST) of 2565 P and 2380 S
arrivals from 658 well located aftershocks recorded at 8–14 three-component local seismic stations led
to 1 D velocity models indicated very slow sediments in the upper 0–2 km depth range (Vp: 2.92 km/s
and Vs: 0.90 km/s) and an increasing trend of velocities with depth at 2–40 km depth. The estimated
sediment thicknesses beneath 12 accelerograph and 6 seismograph sites from the estimated velocity
model and the travel-time difference between S and converted Sp phases reaches a maximum of (1.534
± 0.117) km beneath Bandri (near the location of 2001 MS) and attains a minimum sediment
thickness of (0.858 ± 0.104) km beneath Ramvav and Burudia. The spectral ratios between Sp and S
from 159 three-component accelerograms have been used to study seismic wave attenuation in the
Kachchh rift basin. The estimated Qs vs. Qp relations for 12 accelerograph sites vary from Qs = 0.184
Qp (at Chobari) to Qs = 0.505 Qp (at Dudhai). For stations Chobari, Chopdwa, Jahawarnagar, Vondh
and Tapar, the spectral ratio slopes and hence the calculated Qs vs. Qp relations are effectively the
same, and the correlation coefficients are quite high (0.91–0.93). Stations Adhoi, Manfara, New
Dudhai, Dudhai and Sikara have similar Qs vs. Qp relationships to each other and also have high
correlation coefficients (0.78–0.87). The spectral ratios for stations Anjar and Ramvav are small and
poorly constrained, resulting in less reliable Qs vs. Qp relations. This could be due to noisy data, fewer
available waveforms, or scattering due to velocity heterogeneities and/or interface irregularities.
Key words: Converted phases, thickness of sediments, attenuation, quality factors, sedimentary basin.
S B 0 1 0 1 5 8
Journal number Manuscript number B Dispatch: 14.12.2006
Author’s disk received 4
Journal: Pure and applied Geophysics
Used 4 Corrupted Mismatch
No. of pages: 26
Keyed
2 P. Mandal Pure appl. geophys.,
Introduction
Converted phases (S-to-P or P-to-S) from abrupt velocity changes are widely used
in the receiver function analysis of teleseismic P waves for constraining upper mantle
and crustal structure. Travel-time differences between the direct and converted
phases have been used to estimate the thicknesses of the sedimentary layers in the San
Francisco Bay area, USA (HOUGH, 1990). Further, the spectral ratio between Sp-
converted phase and direct S phase has been widely used to study the seismic wave
attenuation in terms of quality factors in sedimentary basins (CLOUSER AND
LANGSTON, 1991; CHEN et al., 1994; CHAPMAN et al., 2003). However, LANGSTON
(2003) drew attention to the fact that the Sp/S spectral ratio method could not
provide any evidence of highly attenuating unconsolidated sediments of the
Mississipi embayment because of wave-propagation effects.
The Kachchh area lies in seismic zone V of India, and has experienced two
earthquakes of Mw ‡ 7.7 within a span of 182 years (BIS, 2003; MANDAL et al.,
2004a). In addition, we know that the region has experienced several moderate-size
earthquakes since 1668 (RAJENDRAN and RAJENDRAN, 1998). The last major
earthquake of Mw 6.1 prior to the 2001 Bhuj earthquake occurred near Anjar in 1956
(CHUNG and GAO, 1995). Such a high level of seismic activity calls for a quantitative
evaluation of the seismic hazard for the region. Understanding attenuation is
necessary to predict ground motion for estimating seismic hazard, however there are
few studies of the attenuation characteristics of the seismically active Kachchh
region. Only three strong ground motion records are available for the 2001 Gujarat,
India earthquake (the peak ground accelerations of 0.1, 0.008, and 0.005 g are
indicated at Ahmedabad, Delhi, and Roorkee, at distances of 240, 850 and 1000 km,
respectively), which indicated an attenuation relation similar to eastern North
America (PETERSEN et al., 2002). Similar results ensued that were further supported
by SRR data of the 2001 Bhuj earthquake (CRAMER and KUMAR, 2003). Recently,
SINGH et al. (2003) reported a large regional Q value on the order of 580 for the
whole Peninsular India including Kachchh, Gujarat, based on their ground motion
estimation using broad-band data of the 2001 Bhuj main shock and two large
aftershocks. The Lg wave Q tomography for Eurasia, using broadband and short-
period seismograms from regional networks, implies a higher Qo (Q at 1 Hz) value
on the order of 700–800 over the Kachchh, Gujarat (MITCHELL et al., 1997). This
higher estimate of Qo can be attributed to the fact that the Lg coda wave averages
over a broad region, and energy does not travel entirely along a great circle path
between source and receiver (MITCHELL et al., 1997). However, MANDAL et al.
(2004), based on their coda Q (Qc) study using 2001 Bhuj aftershocks, suggested an
average low Qc value of 102 at 1 Hz beneath the 2001 Bhuj epicentral zone.
Therefore, it is important to study the attenuation of seismic waves in the sediments
of the Kachchh basin, to aid in the prediction of ground motion for this seismically
important region in India.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 3
The kachchh region in western India consists of 2000-to 3000-m thick Mesozoic
and 1000 m of Tertiary sediments deposited in a rift basin. The main structural
features in the region are a series of E-W trending uplifts along master faults (thrusts)
(Fig. 1). The uplifted portions are areas occupied by Mesozoic and Tertiary rocks,
whereas the residual depressions and low-lying regions between the uplifts consist of
Quaternary sediments in the alluvial mud flats and salt pans of the Great and Little
Ranns and Banni plains (BISWAS, 1987). All the major uplifts are bounded at least on
one side by a fault or monoclinal flexures, and on the other side by gently dipping
planes. These uplifts comprise blocks, which are tilted towards south (MERH, 1995).
The major faults in the region are: i) Nagar Parkar fault (NPF), ii) Allah Bund fault
(ABF), iii) Island Belt fault (IBF), iv) Wagad fault, v) Kachchh Mainland fault
(KMF) and vi) Katrol Hill fault (KHF).
Figure 1
Map showing major E-W trending faults in the Kachchh region, Gujarat India. NPF = Nagar Parkar
Fault; IBF = Island Belt Fault; NWF = North Wagad Fault; KMF = Kachchh Mainland Fault; KMU
= Kachchh mainland Uplift; KHF = Katrol Hill Fault. Accelerographs and seismographs are marked by
open diamonds and solid triangles, respectively. A star shows the epicenter of the main earthquake of
January 26, 2001 of Mw 7.7. The focal mechanism solution of USGS showing reverse faulting on a
southwest dipping fault is also shown. The inset shows the key map for the area with reference to Indian
plate boundaries.
consisted of a 24-bit recorder with an external hard disk (2 GB) and a GPS timing
system. Six of these were equipped with short-period seismometers (frequency range
1–40 Hz) and two with broadband sensors (frequency range 0.01–40 Hz). The
distance between stations and epicenters varies from 14 km to 90 km. Recording was
done in a continuous mode at 100 samples per second. Some station locations were
shifted during the study period making a total of 12 station sites. Further, we used
data from temporary seismic networks installed by CERI, University of Memphis,
USA (from 13–27 February, 2001) and Hirosaki University, Japan (from 28
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 5
February to 7 March, 2001), which provided a total of 28 stations with phase data
from a maximum at any one time of 18 stations to locate earthquakes delineating the
immediate aftershock zone (MANDAL et al., 2004 a,b).
In August 2002, this network was reduced to five seismic stations: three equipped
with three-component broadband CMG-40T sensors and the other two with three-
component short-period L4-3D sensors. Additionally, NGRI installed a digital
network of ten strong motion accelerographs (within about 30 km radius of the
aftershock zone covering an area of 30 km · 40 km), to obtain a better idea of
seismotectonics and seismic hazard of the region. The study area lies in the latitude
range 23.27–23.55oN and longitude range 70.12–70.50oE (Fig. 1). The accelerograph
stations were equipped with Kinematrics three-component Episensor and ETNA
recorders. The triggering and de-triggering levels for accelerographs vary from 0.5 to
0.1 % of the maximum of 2.5 volts. All accelerographs are set for triggered mode for
recording at trigger levels of 0.2 to 0.5% of maximum with a sample rate of 200
samples per second and continuous GPS time synchronization.
Aftershock Data
The aftershock activity has been very intensive and is continuing with occasional
occurrences of M 3 and M 4 aftershocks. The most recent M 5.6 aftershock occurred
on 6th April, 2006, more than five years after the main shock. Until April 2006, 14
earthquakes of magnitude exceeding 5, over 170 earthquakes of 5 ‡ M > 4.0 and
several thousand earthquakes of M £ 4 have occurred. We used 389 selected
aftershocks of magnitude 2.5 to 4.6 for the site response study, recorded by the
NGRI’s local seismic network from August, 2002 to August, 2004. Figure 2 shows
three-component accelerograms of one M4 aftershock.
Initial aftershock locations were obtained using the program HYPO71Pc (LEE
and VALDES, 1985) and the one-dimensional velocity model that was obtained from
the Geophysical integrated survey and controlled source seismics (GUPTA et al., 2001;
KAILA et al., 1981). The average P-wave velocity of the Mesozoic sediments was
estimated to be 2.4 km/sec from the controlled seismics results (GUPTA et al., 2001;
KAILA et al., 1981; MANDAL et al., 2004a). Recently, the inversion of teleseismic
receiver function revealed that the Moho depth for the Kachchh region varies from
39–49 km and an average Moho depth of 40 km at Bhachau (MANDAL, 2006). We
performed our velocity inversion using VELEST (KISSLING, 1995) at Bhachau
because a crustal model beneath Bhachau was available from the integrated
Geophysical survey (GUPTA et al., 2001). The velocity model consists of nine layers,
6 P. Mandal Pure appl. geophys.,
2 P Sp VERT
0
Acceleration (cm/s*s)
-2
6
s N-S
4
2
-2
-4
6 E-W
2
-2
-6
0 1 2 3 4 5
Figure 2
A plot showing the three-component accelerograms for an Mw 4.0 aftershock recorded at Adhoi
accelerograph station of the NGRI local seismological network. The direct P and S, as well as the
converted Sp wave have been marked clearly.
the tops of which are at 0.0, 2.0, 10.0, 15.0, 20.0, 25.0, 30.0, 35.0, and 40.0 km, with
P-wave velocities of 2.4, 6.00, 6.10, 6.20, 6.40, 6.68, 6.68, 6.68 and 8.20 km/s,
respectively (Fig. 3). This velocity model was constructed based on seismic
refraction, gravity, magnetic, electrical and teleseismic receiver function studies of
the region (GUPTA et al, 2001; KAILA et al, 1981; MANDAL, 2006). Since the NGRI
stations are closely spaced in the epicentral zone, at least one station was always
within a distance of the hypocentral depth, which enabled good control on the focal
depth estimation. 70% weight was assigned to impulsive clear P picks, while 10% to
weak emergent P phases. The P-to-S converted phases are weak S arrivals followed
by a strong S. By avoiding the weak beginning of the S-wave train and picking up S
at the strong arrival, P-to-S converted phase is excluded. A maximum of 50% weight
was assigned to clear S picks and 5% to weak S phases. The Vp/Vs ratio is fixed at
1.73. The root-mean-square errors in P residual are on the order of 0.3 s, and, root-
mean-square errors in Sresidual are on the order of 0.35 s. The mean horizontal and
vertical single 90% confidence estimates are 1.0 and 2.0 km, respectively, for the
aftershocks.
In this study, an inversion technique (KISSLING, 1995) was applied to estimate a
more reliable 1-D velocity model with station corrections. At first we selected high-
quality 2565 P arrivals, and 2380 S arrivals from 658 well-located aftershocks (during
February to May, 2001) recorded at 8 to 18 three-component digital stations (error in
epicentral location (ERH) <1.0 km, error in focal depth (ERZ) <2.0 km; azimuthal
gap between recording stations (GAP) <180o and root-mean-squares error in P
residual (RMS) <0.2 s), and using the initial crustal velocity structure around the
study area discussed above (e.g., KAILA et al., 1981; GUPTA et al., 2001; MANDAL et
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 7
Vp (km/sec)
(a) 2 4 6 8 10
0
Vp Model
Depth to Initial Final
the top of (km/s) (km/s)
10 layer (km)
0.0 2.40 2.92
2.0 6.00 5.99
Depth (km)
5.0 6.10 5.90
20 10.0 6.20 6.18
16.0 6.40 6.07
22.0 6.68 6.59
29.0 6.68 7.20
30 34.0 6.68 6.78
40.0 8.20 8.20
40
Initial
Final
50
(b) Vs (km/sec)
0 1 2 3 4 5
0
Vs Model
Depth to Initial Final
10 the top of (km/s) (km/s)
layer (km)
0.0 1.00 0.90
2.0 3.49 3.49
Depth (km)
40
Initial
50 Final
Figure 3
Initial (marked by dashed line) and final velocity (marked by solid line) models. (a) for P wave and (b) for S
wave. The velocities are also shown in tabular form.
al., 2004a, b; MANDAL and PUJOL, 2006; MANDAL, 2006). Then, we relocated events,
as shown in Figure 4 assuming the above-mentioned initial velocity structure model
(Fig. 3). Finally, we estimated a more reliable P and S wave velocity model and
station corrections, using the method of local earthquake data inversion.
(a) (b)
(c)
Figure 4
(a) Relocated epicenters of selected 658 aftershocks using VELEST, which have occurred during 12–28
February, 2001. The inferred causative fault (NWF) is shown by a dotted line. (b) Hypocentral depth plots
of selected earthquakes in the N-S cross section. A solid grey line shows the inferred causative fault. (c)
Hypocentral depth plots of selected earthquakes in the E-W cross section. Depth and horizontal errors of
the hypocenters are relatively small inside the network, while they are relatively large outside the network.
A solid star shows the epicenter for the 2001 Bhuj main shock.
search for the minimal RMS misfit solution by varying velocity model parameters
within a certain reasonable range. After several runs using different initial velocity
models (varying layer thicknesses and velocities), hypocentral parameters and
control parameters, the final 1-D P-wave velocity model (Fig. 3) was obtained on
the basis of the minimum RMS misfit. The final models for P as well as S waves
obtained from inversion are shown in Figures 3(a, b). On an average both Vp and
Vs show an increase from surface to Moho depth (i.e., 40 km) except for a low
velocity in the 34–40 km depth range. The first upper crustal layer (0–2 km)
consisting of Tertiary and Jurassic sediments suggests an increase in Vp from
2.4 km/s to 2.92 km/s and a decrease in Vs from 1.0 km/s to 0.9 km/s in
comparison to the above-discussed initial model. The second crustal layer (2–5 km)
shows in significant change in Vp or Vs in comparison to the initial model. In the
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 9
Observed Waveforms
are generally simple, displaying prominent P, Sp, and S wave pulses. Particle motion
of P and S waves show near-vertical ray paths. The direct P and S waves, and the
converted Sp phases are easily identifiable on the three-component accelerogram
(Fig. 2). In general, the converted Sp waves are impulsive and strong on the vertical
seismogram therefore, they are easier to pick than the weak Ps phases on the radial
and transverse components. Hence, the travel-time differences between the S and the
Sp phases are used in this study to estimate the sedimentary layer thickness beneath
each NGRI station.
t (s-sp)(insec)
1.8 t (S -Sp) = 0.76 0.07 sec 1.8 t (S -Sp) = 0.91 0.14 sec
1.5 1.5
1.2 1.2
0.9 0.9
0.6 0.6
0.3 0.3
0.0 0.0
0 20 40 60 0 20 40 60
Epic. Dist. (km) Epic. Dist. (km)
t (s-sp)(insec)
1.8 t (S -Sp) = 0.77 0.08 sec 1.8 t (S -Sp) = 0.85 0.10 sec
1.5 1.5
1.2 1.2
0.9 0.9
0.6 0.6
0.3 0.3
0.0 0.0
0 20 40 60 0 20 40 60
Epic. Dist. (km) Epic. Dist. (km)
t (s-sp)(insec)
1.8 t (S -Sp) = 0.8477 0.1019 sec 1.8 t (S -Sp) = 0.77 0.08 sec
1.5 1.5
1.2 1.2
0.9 0.9
0.6 0.6
0.3 0.3
0.0 0.0
0 20 40 60 80 100 0 20 40 60 80
Epic. Dist. (km) Epic. Dist. (km)
Figure 5
Plots of travel-time differences between direct S and the converted Sp waves for different earthquakes as a
function of epicentral distances recorded at the same station. (i) Adhoi, (ii) Dudhai, (iii) Chobari, (iv)
Chopadwa, (v) Anjar and (vi) Bhachau.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 11
differences between the direct S and Sp arrivals are independent of epicentral distance
and azimuth, which suggest that the large amplitude, impulsive Sp phase on the
vertical seismograms must be generated at a near constant depth beneath the station.
The travel-time differences between the direct S waves and the converted Sp waves
vary at different stations, indicating the variable conversion depths beneath different
stations (Figs. 5(i–vi)). The travel-time differences between the direct S waves and the
converted Sp phases beneath each NGRI station are listed in Table 1. The maximum
and minimum values of travel-time differences between the direct S waves and the
converted Sp phases are 1.14 sec. at Bandri, and 0.66 sec. at both Ramvav and
Burudia, respectively.
Table 1
t(s-sp) for NGRI Strong Motion/Seismograph seismological network
where, dt (S-Sp) is the travel-time difference between the direct S wave and the
converted Sp phase, h* is the thickness of the sediments beneath the station, and Vp
and Vs are the averaged P- and S- wave velocities in the sediments, respectively. A
Shear-wave velocity of 0.90 km/s and a P-wave velocity of 2.92 km/s have been used
for the computation of the travel times. The estimated travel-time differences between
the direct S-wave and the converted Sp phase, as well as their standard deviation and
sediment thickness for 20 NGRI accelerograph and six seismograph stations, are
listed in Table 1. The maximum and minimum values of estimated sediment thickness
are (1.534 ± 0.117) km beneath Bandri (near the location of the 2001 Bhuj main
shock) and (0.858 ± 0.104) km beneath Ramvav and Burudia, respectively.
where f is frequency (in Hz), As (f) is the spectrum of the S-wave pulse incident at the
basin/basement interface, TSSU is the upgoing S-wave transmission coefficient, and
TSPU is the upgoing S-to-P conversion coefficient. We note that scattering may play a
role in frequency-dependent attenuation, and the values of Q estimated on the basis
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 13
of equations (2) and (3) necessarily include the combined effects of anelastic loss and
any unmodeled loss by scattering. In the exponentials, ts and tsp are the S-wave and
Sp-wave travel times through the sedimentary basin, and QP and QS are the
frequency-independent P-wave and S-wave quality factors.
Taking the logarithm (Log10) of the ratio between the Sp and S amplitude
spectra, the source terms cancel, and we obtain
log½Sp(f)/S(f) ¼ log½TU U
Sp =TSS þ p f½tS =QS tSp =QP : ð4Þ
Thus, a plot of the log spectral ratio in log-amplitude frequency space should be a
straight line with slope
m ¼ p½tS =QS tSp =QP : ð5Þ
The zero frequency intercept of this line is the logarithm of the conversion/
transmission coefficient ratio. Rearranging terms in the equation (5) one can obtain
the relation between QS vs. QP:
(i)
(b) -2
(a)10-2 10
-3 -3
10 10
-4 -4
10 10
-5 -5
10 10
-6 -6
10 10
-7 -7
10 10
0 10 20 30 40 50 0 10 20 30 40 50
Frequency (Hz) Frequency (Hz)
-3 -4
10 10
-4
10 -5
10
-5
10
-6
10
-6
10
-7 -7
10 10
0 30 0 10 20 30 40 50
10 20 40 50
Frequency (Hz) Frequency (Hz)
(ii)
(a) (b) -2
-2
10 10
-3 -3
10 10
-4 -4
10 10
-5 -5
10 10
-6 -6
10 10
-7 -7
10 10
0 10 20 30 40 50 0 10 20 30 40 50
Frequency (Hz) Frequency (Hz)
(c) -3 (d) 10 -2
10
-3
10
-4
10 -4
10
-5 -5
10 10
-6
10
-6
10 -7
10
-7 -8
10 10
0 10 20 30 40 50 0 10 20 30 40 50
Frequency (Hz) Frequency (Hz)
Figure 6
(i): Spectra of S, Sp and noise (from vertical) for ADHOI SMA stations (a) event1, (b) event2, (c) event3,
and (d) event4. The solid line represents the spectra for the S waves. Longer dashes represent the noise
spectra, while smaller dashes mark the Sp spectra. (ii): Same as Figure 6 (i) for CHOPDWA SMA station
(a) event1, (b) event2, (c) event3, and (d) event4. (iii): Same as Figure 6 (i) for JAWAHARNAGAR SMA
station (a) event1, (b) event2, (c) event3, and (d) event4.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 15
Figure 6
(Contd.)
Data from more than one event were used to determine the Qs vs. Qp relationship
independently for each station of the strong motion accelerograph network of the
National Geophysical Research Institute (NGRI), Hyderabad, India. The slope of
the spectral ratio, the standard error, cross-correlation coefficient and the number of
data used for each station are listed in Table 3. First the average spectral ratio (Sp/S)
of all the events considered for each station is estimated. After obtaining the average
spectral ratio over the range in frequencies, a least-squares linear fit is used to
estimate the slope, standard error, and correlation coefficient at each station. Table 3
shows that the correlation coefficients are small (0.69) for Anjar and Ramvav
(0.74), whereas for other stations they suggest larger values more than 0.75. Thus,
our estimates for Anjar and Ramvav have more errors than other stations. All plots
are drawn for the frequency range 1–25 Hz. Table 4 shows Qs vs. Qp relations in the
sedimentary basin of the Kachchh, Gujarat, India. Included curves are of Qs vs. Qp
for each of twelve SMA stations.
The local digital NGRI seismological network in the Kachchh, Gujarat, India
has provided sufficient high-resolution three-component accelerograms and
16 P. Mandal Pure appl. geophys.,
0.5 0.5
0.0 0.0
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
0.0 0.0
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
0.0 0.0
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
Figure 7
Spectral ratio functions for twelve stations, and the best fit lines with slopes. (a) Adhoi (ADH), (b) Anjar
(ANJ), (c) Chobari (CHB), (d) Chopadwa (CHP), (e) Dudhai (DUD), (f) Jahawarnagar (JAH), (g)
Manfara (MNF) (h) New-Dudhai (NDUD), (i) Ramvav (RAM), (j) Sikara (SIK), (k) Tapar (TAP), and (l)
Vondh (VND). Here, logarithm base 10 is used instead of natural logarithm. Therefore, the slopes shown
are off the values for ‘m’ by a factor of 0.43.
seismograms over the past four years (February 2001 to February 2005), to enable
us to carry out a comprehensive study to estimate the sediment thickness of the
basin and Qs vs. Qp relationships using travel-time differences between converted
Sp and direct S phases.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 17
log10(Sp/S)
0.5 0.5
log10(Sp/S)
0.0 0.0
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
1.5 1.5
(i) (j) SIKARA
RAMVAV
1.0 1.0
log10(Sp/S)
0.5 0.5
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
1.5 1.5
(k) (l)
TAPAR VONDH
1.0 1.0
log10(Sp/S)
0.5 0.5
log10(Sp/S)
0.0 0.0
-0.5 -0.5
-1.0 -1.0
-1.5 -1.5
0 5 10 15 20 25 0 5 10 15 20 25
Frequency(Hz) Frequency(Hz)
Figure 7
(Contd.)
(a) (b)
24.0 24.0
23.6 23.6 a
i
23.2 23.2
22.8
69.6 70.0 70.4 70.8 69.6 70.0 70.4 70.8
Correction for P wave(s) Correction for S wave(s)
(c)
24.0
23.6
23.2
Figure 8
Bubble plots for station corrections, (a) P wave and (b) S wave. Symbol size scaled with the size of the
residual. (c) Bubble plot of the estimated sediment thickness. Symbol size scaled with the thickness of the
sediment.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 19
larger values of station corrections (both +ve and –ve), while negative station
corrections (i.e., relatively higher velocities) characterize the SW part of the study
area. The contours of station corrections are aligned in an almost E-W orientation,
which suggests an east-west lateral variation closely related to the trends of major
tectonic units in the region.
C. Qs vs. Qp
Based ons Table 2, within the error bounds given by the standard deviations, the
slopes and the correlation coefficients for Adhoi, Dudhai, Manfara, and Sikara are of
the same order. Stations having larger correlation coefficients (i.e., Chobari,
Chopdwa, Jawaharnagar, Vondh, and Tapar) provide larger slopes (0.040–0.074)
except at New Dudhai, whose spectral ratio is 0.033 with a very high correlation
coefficient of 0.93.
Figure 9
Contour map of estimated sediment thickness in km. The gray star marks the epicenter of the 2001 Mw 7.7
Bhuj earthquake. The contour interval is 0.05 km.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 21
For twelve accelerograph sites, the Qs vs. Qp relationships have been estimated
using equation 6 (Table 3). The stations (CHB, CHP, JAH, VND and TAP) whose
spectral slopes have very high correlation coefficients (0.091–0.093) have very similar
Qs vs. Qp relations, which suggest relatively less attenuative sediments beneath these
stations (Tables 3 and 4). Except at New Dudhai station, whose spectral slope is 0.33
with a very high correlation coefficient (0.93), the Qs vs. Qp relation suggests a
relatively more attenuative sediment beneath this station. At ADH, DUD, MNF,
and SIK, stations with reasonably high correlation coefficients (0.074–0.078), the
spectral slopes lead to relatively more attenuative sediments beneath these stations.
The spectral ratios for stations Anjar and Ramvav are small and are poorly
constrained, resulting in less reliable Qs vs. Qp relations. This could be due to noisy
data, fewer available waveforms, or scattering due to velocity heterogeneities and/or
interface irregularities.
Since these stations fit the linear model well, we feel that we are accurately
measuring an anelastic Q whose effect is described by equations (2) and (3). The
estimated Qs vs. Qp relations for 12 accelerograph sites vary from Qs = 0.184 Qp (at
Chobari) to Qs = 0.505 Qp (at Dudhai). Note that the S and Sp phases have similar
frequency content and relatively large Sp amplitudes compared to S. Scattering due
to velocity heterogeneities and/or interface irregularities would probably significantly
distort the linear relation between Qs and Qp. Perhaps the material and interfaces
encountered by the ANJ and RAM, S and Sp waves are more heterogeneous than
along paths to other stations. Further, having a more heterogeneous media is not
consistent with the plane layered and laterally homogeneous medium, which is
assumed to calculate the travel times for S and Sp phases. Therefore, Qs vs. Qp
obtained at these stations are probably not representative of the true picture.
Table 2
Spectral Ratio Slopes estimated at 12 Strong Motion Accelerograph Sites
Table 3
Q s Vs. Qp relationships obtained at 12 Strong Motion Accelerograph Sites
Adhoi Qs=0.37Qp 15
Chobari Qs=0.18Qp 09
Chopdwa Qs=0.22Qp 14
Dudhai Qs=0.51Qp 21
Manfara Qs=0.40Qp 17
Ramvav Qs=0.29Qp 13
Sikara Qs=0.43Qp 06
Anjar Qs=0.44Qp 08
Jawaharnagar Qs=0.22Qp 25
New Dudhai Qs=0.41Qp 15
Tapar Qs=0.25Qp 09
Vondh Qs=0.29Qp 16
130 down to 100 m near El Centro, California. It is clear from the above-mentioned
discussion that near-surface Q values vary widely from region to region. In 1958,
MCDONAL et al. reported Qp and Qs for Pierre shale as 32 and 10, respectively. From
a borehole using vertically propagating waves from a nearby deep event, HAUKSSON
et al. (1987) measured a Qp and Qs of 44 and 25, respectively, between 420 and
1500 m. However, they attributed a high Qs of 108 in the upper 420 m to frequency
dependent near-surface amplification or a real effect (high Qs).
Based on their study of the theoretical elastic Qs vs. Qp relations, VASSILIOU et al.
(1982) suggested that for the case of zero bulk loss Qp and Qs can be related by the
equation,
Based on the receiver function analysis of the Kachchh, Gujarat, the average Vp/
Vs is estimated to be 1.8 in the upper sedimentary layer (MANDAL, 2006). This leads
to a linear relation Qs = 0.41 Qp, which might be the lower bound for Qs vs. Qp
relationships for the Kachchh basin.
We can work out another Qs vs. Qp relationship, assuming there are no apparent
differential attenuation effects and the slope of log spectral ratio is zero (i.e., m = 0),
using equation (5). For normal incidence, this linear relation is,
Qs ¼ ðVp =Vs ÞQp : ð8Þ
We know from the RF study for Kachchh basin region that the Vp/Vs for the upper
sedimentary layer range from 1.73 to 2.99 (MANDAL, 2006). Using Vp/Vs = 2.99, we
obtained Qs = 2.99 Qp, which might be the upper bound for Qs vs. Qp relationships
for the Kachchh basin.
Conclusions
In this study, the thickness of sediments beneath the numerous NGRI stations
has been estimated using the travel-time difference between direct waves and
converted waves and the available subsurface velocity structure obtained from the
inversion of receiver functions from teleseismic P waves. The estimated travel-time
differences between the direct S wave and the converted Sp phase for 20
accelerograph and 6 seismograph stations suggest a variation from (0.66 ± 0.08)
to (1.14 ± 0.09) (Table 1). These differences suggest a maximum sediment thickness
of (1.534 ± 0.117) km beneath Bandri (near the location of 2001 MS) and a
minimum sediment thickness of (0.858 ± 0.104) km beneath Ramvav and Burudia.
A large thickness of sediments at the Jawaharnagar station could explain the damage
in this region during the 2001 Bhuj as well as the 1956 Anjar earthquakes. Based on
the estimated travel-time differences (tS – tSp) and P- and S-station corrections, it is
inferred that in general, the large +ve and –ve station corrections for both P and S
24 P. Mandal Pure appl. geophys.,
waves characterize an area of large sediment thickness (>1 km) covering Daneti,
Newdudhai, Sikara, Mai, Meghpar, Bandri and Jahwarnagar except a few places like
Adhoi and Ramvav. A good agreement between the estimated small sediment
thickness and small station corrections (P as well as S) has also been observed at
SMA sites such as Burudia, Vondh, Bhachau and Manfara.
The estimated sedimentary layer thickness and spectral ratios of the S wave and
Sp conversion have been used to estimate the Qs vs. Qp relations for sedimentary
layers below twelve strong motion stations of the local network of NGRI. The
estimated Qs vs. Qp relations for 12 accelerograph sites vary from Qs = 0.184 Qp
(at Chobari) to Qs = 0.505 Qp (at Dudhai). Similar Qs vs. Qp relationships
obtained for the five strong motion accelerograph stations, namely Chobari,
Chopdwa, Jawaharnagar, Vondh and Tapar, whose spectral ratio slopes have very
high correlation coefficients (0.91–0.93), imply a relatively less attenuative near-
surface structure. These Qs vs. Qp relationships are velocity-model dependent,
which agrees quite well with the relations obtained for other studies in the world
(CHEN et al., 1994). However, the Qs vs. Qp relationship for the Adhoi, Dudhai,
Manfara and Sikara stations, whose spectral ratio slopes also show high correlation
coefficients (0.78–0.87), suggests a relatively more attenuative near surface. The
stations having larger correlation coefficients (i.e., Chobari, Chopdwa, Jawaharna-
gar, Vondh, and Tapar) provide larger slopes (0.040–0.074) except at New Dudhai,
whose spectral ratio is 0.033 with a very high correlation coefficient of 0.93. The Qs
vs. Qp relations imply relatively more attenuative sediments beneath the New
Dudhai station. The small correlation coefficients of the spectral ratio slopes (0.69–
0.74) estimated for stations Anjar and Ramvav, could be due to noisy data, fewer
available waveforms, or scattering due to velocity heterogeneities and/or interface
irregularities.
The seismic response of sedimentary basins to earthquake ground motion is an
important problem for both seismologists and earthquake engineers because ground
motion over sedimentary basins is usually larger that on nearby hard rocks. Strong
ground motion and the concomitant damage to buildings during an earthquake are
critically determined by the geologic setting of a site. The effect of variation in
sediment thickness, shear-wave velocity of sediments and Qs vs. Qp relations of
sediments should be taken into account in the numerical simulation of ground
motion for estimating seismic hazard for any region.
Acknowledgements
The author is thankful to Dr. V. P. Dimri, Director, NGRI for his encourage-
ment and kind permission to publish this work. This study was supported by the
Department of Science and Technology, New Delhi.
Vol. 164, 2006 Qs vs. Qp Relations in the Kachchh Rift Basin 25
REFERENCES
BAKUN, W.H. and BUFE, C.G. (1975), Shear-wave attenuation along the San Andreas Fault Zone in central
California, Bull. Seismol. Soc. Am. 65, 439–459.
BARKER, T.G. and STEVENS, J.L. (1983), Shallow shear-wave velocity and q structures at the El Centro
strong motion accelerograph array, Geophys. Res. Lett. 10, 853–856.
BIS (2003) jjj
BISWAS, S.K. (1987), Regional framework, structure and evolution of the western marginal basins of India,
Tectonophysics 135, 302–327.
CHAPMAN, M.C., TALWANI, P., and CANNON, R.C. (2003), Ground motion attenuation in the Atlantic
coastal plain near Charleston, South Carolina, Bull. Seismol. Soc. Am. 93(3), 998–1011.
CHEN, KOU-CHENG, JER-MING CHIU, and YANG, Y.T. (1994), Qp-Qs relations in the sedimentary basin of
the upper Mississippi embayment using converted phases, Bull. Seismol. Soc. Am. 84(6), 1861–1868.
CHUNG and GAO (1995) jjj
CLOUSER, R.H. and LANGSTON, C.A. (1991), Qp-Qs relations in a sedimentary basin using converted phases,
Bull. Seismol. Soc. Am. 81(3), 733–750.
CRAMER, C.H. and KUMAR, A. (2003), 2001 Bhuj, India, Earthquake Engineering Seismoscope Recordings
and Eastern North America Ground-Motion Attenuation Relations, BSSA 93(3), 102–104.
GUPTA, H.K., HARINARAYANA, T., KOUSALYA, M., MISHRA, D.C., MOHAN, I., PURNACHANDRA RAO, N.,
RAJU, P.S., RASTOGI, B.K., REDDY, P.R., and SARKAR, DIPANKAR (2001), Bhuj earthquake of 26
January 2001, J. Geolog. Soc. In. 57, 275–278.
HAUGE, P.S. (1981), Measurements of attenuation from vertical seismic profiles, Geophysics 46, 1548–1558.
HAUKSSON, E.T., TENG, T.L., and HENYEY, T.L. (1987), Results from a 1500 m deep, 3-level downhole
seismometer array: site response, low Q values, and fmax, Bull. Seismol. Soc. Am. 77, 1883–1904.
HOUGH, S.E. (1990), Constraining sediment thickness in the San Francisco Bay area using observed
resonances and Ps conversions, Geophys. Res. Lett. 17, 1469–1472.
JOHNSTON, D.H., and TOKSÖZ, M.N. (1980), Ultrasonic P and S wave attenuation in dry and saturated rocks
under pressure, J. Geophys. Res. 85, 925–936.
KAILA, K.L., TEWARI, H.C., and SARMA, P.L.N. (1981), Crustal Structure from deep seismic sounding
studies along Navibandar - Amreli profile in Saurashtra, India. In Deccan Volcanism and Related Basalt
Provinces in Other Parts of the World. Memoir 3, Geol. Soc. India., pp. 218–232.
KISSLING, E. (1995), Velest Users Guide, Internal report, Institute of Geophysics, ETH Zürich, 26 pp.
KURITA, T. (1975), Attenuation of shear waves along the San Andreas fault zone in Central California, Bull.
Seismol. Soc. Am. 65, 277–292.
LANGSTON, C.A. (2003), Local earthquake wave propagation through Mississipi embayment sediments, Part
II: Influence of local site velocity structure on Qp-Qs determinations, Bull. Seismol. Soc. Am. 93(6), 2685–
2702.
LEE, W.H.K. and VALDES, C. M. (1985), HYPO71Pc: A Personal Computer Version of the HYPO71
Earthquake Location Program, U.S. Geol. Surv. Open File Report 85–749, 43 pp.
MANDAL, P., RASTOGI, B.K., SATYANARAYANA, H.V.S., KOUSALYA, M, VIJAYRAGHAVAN, R, SATYAMUR-
THY, C., RAJU, I.P., SARMA, A.N.S., and KUMAR, N. (2004a), Characterization of the causative fault
system for the 2001 Bhuj earthquake of Mw 7.7, Tectonophysics 378, 105–121.
MANDAL, PRANTIK, RASTOGI, B.K., SATYANARAYANA, H.V.S., and KOUSALYA, M. (2004b), Results from
Local Earthquake Velocity Tomography: Implications toward the Source Process Involved in Generating
the 2001 Bhuj Earthquake in the Lower Crust beneath Kachchh (India), Bull. Seismol. Soc. Am. 94(2),
633–649.
MANDAL, P., JAINENDRA, JOSHI, S., KUMAR, S., BHUNIA, R., and RASTOGI, B.K. (2004c), Low coda-Qc in
the epicentral region of the 2001 Bhuj Earthquake of Mw 7.7, Pure. Appl. Geophys. 161, 1635–1654.
MANDAL, P. (2006), Receiver Function analysis of teleseismic P-waves: Implication toward sedimentary basin
and Moho depth variation beneath Kachchh and Saurashtra regions, Gujarat (India), Phys. Earth and
Planet. Interiors 155, 286–299.
MANDAL, P. and PUJOL, J. (2006), Seismic imaging of the aftershock zone of the 2001 Mw7.7 Bhuj
earthquake, India, Geophys. Res. Lett. 33, L05309, 1–4.
26 P. Mandal Pure appl. geophys.,
MCDONAL, F.J., ANGONA, F.A., MILLS, R.L., SENGBUSH, R.L., VAN NOSTRAND, R.G., and WHITE J.E.
(1958), Attenuation of shear and compressional waves in Pierre Shale, Geophysics 23, 421–439.
MCCALPIN, J.P., and THAKKAR, M.G. (2003), 2001 Bhuj-Kachchh earthquake: Surface faulting and its
relation with neotectonics and regional structures, Gujarat, Western India, In (Pantoshi, D., Berrymann,
K., Yeats, R. and Kinugasa, Y. eds.), Ten Years of Paleoseismology in the ILP, Progress and Prospects,
Ann. Geophys. 46(5), pp. 937–956.
MERH, S.S. (1995), Geology of Gujara, Geological Society of India.
MITCHELL, B.J., PAN, Y., XIE, J., and CONG, L. (1997), Lg coda Q variation across Eurasia and its relation
to crustal evolution, J.G.R. 102 (B10), 22767–22779.
PETERSEN, M.D., RASTOGI, B.K., SCHWEIG, E.S., and GOMBERG, J.S. (2002), Sensitivity of seismic hazard
for the northwestern portion of the state of Gujarat, India, Tectonophysics (in press).
PUJOL, J. (1988), Comments on the joint determination of hypocenters and station corrections, Bull. Seismol.
Soc. Am. 78, 1179–1189.
RAJENDRAN et al. (2002) jjj
RAJENDRAN and RAJENDRAN et al. (1998) jjj
REDDY, P.R., SARKAR, D., SAIN, K., and MOONEY, W.D. (2001), A Report on a collaborative scientific
study at USGS, Menlo Park, USA (30 October – 31 December, 2001), 19 pp.
SINGH, S.K., APSEL, R.J., FRIED, J., and BRUNE, J.N. (1982), Spectral attenuation of SH waves along the
Imperial Fault, Bull. Seismol. Soc. Am. 72, 2003–2016.
SINGH, S.K., BANSAL, B.K., BHATTACHARYA, S.N., PACHECO, J.F., DATTATRAYAM, R.S., ORDAZ, M.,
SURESH, G., KAMAL, and HOUGH, S.E. (2003), Estimation of ground motion for Bhuj (26 January, 2001;
Mw 7.6) and future earthquakes in India, B.S.S.A. 93, 353–370.
TULLOS, F.N. and REID, A.C. (1969), Seismic attenuation of Gulf Coast sediments, Geophysics 34, 516–528.
WINKLER, K., and NUR, A. (1982), Seismic attenuation: Effects of pore fluids and frictional sliding,
Geophysics 47, 1–15.
VASSILIOU, M., SALVADO, C.A., and TITTMANN, B.R. Seismic attenuation. In CRC Handbook of Physical
Properties of Rocks, vol. III, R.S. Carmichael (ed.), (CRC Press, Boca Raton, Florida 1982).