Indian Geotech J

https://doi.org/10.1007/s40098-024-01011-7

ORIGINAL PAPER

Seismic Source Parameters and High‑Frequency Spectral Decay

Model for the Region of North India
N. H. Harinarayan1 · Shiv Shankar Kumar2 · M. M. Krishna1 ·
Chithira Krishnakumar1 · Avinash Kumar3

Received: 31 July 2023 / Accepted: 17 June 2024

© The Author(s), under exclusive licence to Indian Geotechnical Society 2024

Abstract This study examines the seismic source param- and is found to range between 0.015 and 0.071 for the hori-
eters and high-frequency spectral attenuation factor (κ) of zontal component, and 0.01 and 0.036 for the vertical com-
36 minor to moderate earthquake (EQ) events that have been ponent. The epicentre distance for various site classes is
occurred in the north-west Himalayas and its foothills of used to construct a regional model for κ. Finally, stochastic
north India from 2007 to 2016. A non-reference generalised ground motion simulations of historical EQ records utilising
inversion technique has been used to compute the source estimated source parameters and constructed kappa models
spectra, and the point source model is used to estimate the are performed in order to validate the findings of the pre-
seismic moment (M0) and corner frequency (fc). Further, sent work. The comparison of the original and simulated
a scaling relation between M0 and fc has been developed EQ records is in very good agreement, demonstrating the
for the study area. Additionally, based on the determined fc efficacy of the obtained parameters in capturing the local
and M0 values, source characteristics such as seismic energy ground motion features.
(Es), stress drop (Δσ) and apparent stress drop (σa) are esti-
mated for each of the 36 events. In the present work, σa and Keywords GNIV · Kappa parameter · Source
∆σ were found to range from 0.1 MPa to 12.36 MPa and characteristics · Stress drop · Apparent stress
0.17 MPa to 14.61 MPa, respectively. Low values of σa and
∆σ can be associated with the low effective stress model
EQs. Furthermore, a majority of EQs appear to have a partial Introduction
stress drop mechanism according to the Zúñiga parameter (ε)
determined for the region. Also, the parameter 𝜿 is estimated Himalayan region is one of the seismically active regions
worldwide, which includes the entire north-west Himalayan
range [1]. According to the Global Seismic Hazard Assess-
* Shiv Shankar Kumar ment Program, the region around the north-west Himalayan
[email protected]
arc is classified as a high seismic risk zone with an expected
N. H. Harinarayan peak acceleration of 0.45–0.50 g for 10% possibility of
[email protected]
exceedance in 50 years [2, 3]. The region has also experi-
M. M. Krishna enced numerous destructive earthquakes (EQs) such as the
[email protected]
1885 Kashmir EQ (Mw = 6.3); the 1905 Kangra—Himachal
Chithira Krishnakumar Pradesh EQ (Ms = 7.8); the 1991 Uttarakashi EQ (Mw = 6.6);
[email protected]
the 1999 Chamoli EQ (Mw = 6.4) and the 2005 Muzaffar-
Avinash Kumar abad—Kashmir (Mw = 7.6) [4–8]. This region is one of most
[email protected]
densely inhabited areas with the population of 96 million
1
Universal Engineering College, Thrissur, India as per 2011 census. Therefore, it can be concluded that the
2
National Institute of Technology (NIT) Patna, Patna, Bihar, accurate assessment of regional seismic hazards scenarios,
India the regional seismic parameters as well as the associated
3
Shankar College of Engineering, Ghaziabad, India source parameters considerably reduce the regional damage

13
Vol.:(0123456789)
 Indian Geotech J

potential of resulting from such similar EQs. The EQs source accelerogram data from the recording stations situated on
parameters, that provide information regarding the phys- rock sites (called reference sites) to eliminate the trade-off
ics of EQs, are the indication of nature of rupture and the between the source and the site parameters [22]. Therefore,
amount of energy released after rupture from a fault. The EQ events with at least one record, on the rock site, can
parameters such as M0, rupture dimensions, fc, Es and Δσ only be taken for the analysis. For the present study, total 37
are the typical EQ source parameters; where, Es is defined EQ events of strong-motion database have been considered.
as the wave energy released if an EQ occurred in an infinite However, among all these earthquakes, 21 EQ events have at
medium with no energy loss, M0 quantify the size of the least one record on the rock site. Therefore, both the source
seismic disturbance, Δσ defines the variations in the average and site spectra for 21 EQ events have been estimated using
stress value prior to rupture process as well as during the conventional generalised inversion method, also mentioned
rupture process. Since Δσ influences high-frequency energy in Harinarayan and Kumar [23], whereas only source spec-
radiation, it is an essential input parameters for generating tra are computed for all 37 events irrespective of whether
synthetic ground motion [8]. These parameters become more the records are on rock or soil sites. Brune’s point source
important while assessing the seismic potential of tectoni- model [24] is then used to interpret the evaluated source
cally active locations. Furthermore, Parolai et al. [9] and spectra in order to estimate source parameters like M0 and fc.
Saito and Masuda [10] have also reported that the seismic Further, the relationship between the aforementioned source
risk of a region can be estimated by using the temporal var- parameters has been developed for estimating the parameters
iation of Δσ. The parameter fc represents the intersection such as Es, Δσ and σA, which provide more insight into the
point, in log–log displacement/acceleration vs frequency tectonic environment of the study region. Furthermore, the
plot, of the flat low-frequency region of the spectrum with value of κ is also estimated and a regional κ-model has been
a power-law fall-off at higher frequencies [11]. Using afore- developed, for different site classes for the study region, con-
mentioned parameters, a set of correlations have been estab- sidering hypocentral distance as a model parameters. Finally,
lished, i.e. called as a scaling relation which depends on this study also discusses the stochastic ground motion simu-
the region-specific/regional tectonic settings as well as on lations for few selected events have been performed using
the energy associated with EQs [12–14]. The information the above-mentioned source parameters and κ-model.
about the scaling relationships and regional-specific source
parameters are significantly important characteristics for the
stochastic method to simulate the seismic ground motion Study Area
[15]. Allmann and Shearer [16] emphasised the dependency
of Δσ on the source mechanism and it was reported that the The current study area includes the Indian states such as
strike-slip EQ reflects greater Δσ in comparison with normal Himachal Pradesh, Punjab, Uttarakhand, Haryana, Rajasthan
and reverse type EQs. Therefore, the dependency of Δσ on and Delhi, which falls between latitude of 28.1° N to 34.1° N
M0 considering the regional scaling relations is needed to be and longitude from 75.8° E to 80.5° E. It comprises various
investigated, since the lack of dependency between Δσ and localised thrusts, faults and folds as well as several north-
M0, suggests self-similar nature of EQs for the region [17, dipping thrust systems, such as Main Central Thrust (MCT),
18]. Furthermore, κ is another important input parameter, Main Boundary Thrust (MBT) and Himalayan Frontal
used in stochastic ground motion simulation, which repre- Thrust (HFT), which is extended down to a depth of approxi-
sents the behaviour of Fourier amplitude spectrum at a high mately 10 km [25, 26]. The MCT and MBT run parallel to
frequency that governs the rate of declination of the ampli- one another and are formed during Cenozoic shortening in
tude with frequency beyond the cut-off frequency. Moreo- the north-western Himalayan region [27, 28]. Several EQs
ver, in the field of geotechnical earthquake engineering, the have been set off in this area by the aforementioned major
dependency of κ either on the source parameter [19] or on thrust systems, out of which the 1905 Kangra—Himachal
the interaction of site and path parameter [20] is still a matter Pradesh EQ and the 2005 Muzaffarabad—Kashmir EQ
of discussion. However, Tsai and Chan [21] have reported are two of the most destructive EQ events [29, 30]. Nearly
high level ground motion, when κ becomes lower [21]. 20,000 people died in the 1905 Kangra EQ, mostly in the
Therefore, the main objective of this work is to obtain north Punjab and the Dehradun region. This EQ also resulted
the source parameters and κ of the study area utilising the in an upliftment of 15 cm, 250 kms from the epicentre [31].
recorded data from accelerograms. Initially, the source spec- The Muzaffarabad—Kashmir EQ in 2005 resulted the loss
tra for EQ events have been developed from the available EQ of 80,000 people causing significant infrastructure damages
dataset using a modified generalised inversion approach pro- [8, 32–34]. The MCT zone has also experienced few moder-
posed by Andrews [22]. For this analysis, the source and site ate EQs such as 1991 Uttarkashi and 1999 Chamoli caus-
spectra are generated simultaneously by inverting the path of ing tremors which were felt even in the far-flung regions of
corrected S-wave spectra. This method, however, needs the Chandigarh and Delhi [35]. 1991 Uttarkashi EQ killed 769

13
Indian Geotech J

people and extensively damaged poorly built buildings in the at 72 strong-motion recording sites located in north India
Uttarkashi district [4]. This EQ inflicted multiple landslides (Fig. 1). The locations of each recording station as well as
and also reflected that an intensity of IX on MSK scale was site class information based on NEHRP classification [38]
reported [4, 36], whereas 1999 Chamoli EQ recorded 64 are provided in Table 1. From Table 1, it can be seen that
fatalities and the extensive structural damage in Chamoli, out of 72 recording stations, 2 are assigned to site class A, 4
Rudraprayag, and Tehri regions [5]. Therefore, based on are assigned to site class B, 26 are assigned to site class C,
the prevalent study, it is clear that the selected study area 31 are assigned to site class D, and 10 are assigned to site
has been subjected to moderate to major EQs, and these class E. The range of EQ magnitude as well as the distance
EQs have caused building damages and fatalities outside of varies from 2.5 to 5.7 and 9 to 265 km, respectively (as
the epicentral region. Thus, it can be stated that this study shown in Table 2).
will be beneficial in assessing the ground motion features, Initially, the recorded ground motions, used in this study,
that can be expected during futuristic regional EQs, for the are processed. All EQ records are first given equal win-
improved preparedness. dow lengths to compute the signal to pre-event noise ratio
(SNR < 5), and only records with SNR > 5 are used for the
further analyses. In addition, baseline correction is made to
Strong‑Motion Dataset EQ records between the frequencies of 0.25 Hz and 50 Hz
using a 5% cosine taper and a band-pass Butterwort filter.
The strong-motion database collected from PESMOS The range of S-wave window length, used in this study, is
(http://​www.​pesmos.​in/), which contains accelerograms 4–15 s, which is estimated as the periods beginning at 0.5
for more than 300 strong-motion recording stations spread- s before S-wave arrival and ending at 90 per cent of the
ing throughout several seismically active regions of India, EQ’s overall energy (similar to Bindi et al. [39]). Further, the
is used in the present study. To capture the strong-motion acquired Fourier amplitude spectrum (FAS) of EQ record,
data, an inbuilt AC-63 GeoSIG triaxial force balanced accel- for S-wave section, is then smoothed using Konno and Ohm-
erometer and an external GPS have been included in the achi’s [40] approach with 20 as a b-parameter value. Finally,
recording equipment [37]. The dataset used in this study the processed ground motions are used to develop source
consists of 253 strong-motion records from 36 EQs, acquired spectra and κ as discussed in the next sections.

Fig. 1  Map of the region under

study with location of EQs
(stars) and recording stations
(triangles). Note: A-Himachal
Pradesh, B-Punjab, C-Haryana,
D-Rajasthan, E-Uttarakhand,
F-Delhi

13
 Table 1  Detail of strong-motion recording stations and estimated values of κ
Si. no Station Code Lat. (°) (N) Lon. (°)(E) SC* 𝜿 EW 𝜿 NS 𝜿H 𝜿V Si. no Station Code Lat. (°) (N) Lon. (°)(E) SC* 𝜿 EW 𝜿 NS 𝜿H 𝜿V

13
(1) (2) (3) (4) (5) (6) (7) (8) (9) (1) (2) (3) (4) (5) (6) (7) (8) (9)

1 BHA 31.6 77.9 C 0.0540 0.0573 0.0556 0.0272 37 ALM 29.6 79.7 C 0.0596 0.0490 0.0543 0.0300
2 CHM 30.4 79.3 E 0.0306 0.0568 0.0437 0.0273 38 BAG 29.8 79.8 E 0.0452 0.0546 0.0499 0.0235
3 DEH 31.9 76.2 B 0.0407 0.0421 0.0414 0.0232 39 BAR 30.8 78.2 D 0.0560 0.0533 0.0546 0.0280
4 DHH 32.2 76.3 D 0.0587 0.0526 0.0556 0.0282 40 CHM 32.6 76.1 C 0.0560 0.0559 0.0560 0.0282
5 HAM 31.7 76.5 D 0.0583 0.0451 0.0517 0.0297 41 CHP 29.3 80.1 C 0.0566 0.0600 0.0583 0.0300
6 JUB 31.1 77.7 C 0.0694 0.0511 0.0603 0.0299 42 CKR 30.7 77.9 D 0.0583 0.0575 0.0579 0.0284
7 KLG 32.6 77 B 0.0407 0.0421 0.0414 0.0232 43 DHA 29.8 80.5 C 0.0583 0.0575 0.0579 0.0284
8 KUL 32 77.1 C 0.0416 0.0600 0.0508 0.0115 44 DNL 30.4 78.2 D 0.0596 0.0600 0.0598 0.0274
9 MAN 31.7 76.9 D 0.0577 0.0383 0.0480 0.0299 45 DUN 30.3 78 D 0.0372 0.0590 0.0481 0.0177
10 RAM 31.4 77.6 D 0.0514 0.0586 0.0550 0.0295 46 GAR​ 30.1 79.3 D 0.0319 0.0442 0.0381 0.0254
11 SAL 32.7 76.1 D 0.0592 0.0530 0.0561 0.0254 47 GHA 30.4 78.7 C 0.0583 0.0600 0.0592 0.0298
12 SND 31.5 76.9 C 0.0582 0.0570 0.0576 0.0296 48 GLTR 30.3 79.1 D 0.0306 0.0568 0.0437 0.0273
13 PLM 32.1 76.5 D 0.0361 0.0325 0.0343 0.0286 49 KAP 29.9 79.9 C 0.0482 0.0589 0.0535 0.0271
14 ANS 31.2 76.5 E 0.0319 0.0442 0.0381 0.0254 50 MUN 30.1 80.2 C 0.0454 0.0462 0.0458 0.0291
15 ASR 31.6 74.9 E 0.0508 0.0423 0.0465 0.0172 51 PAU 30.2 78.8 C 0.0361 0.0325 0.0343 0.0286
16 GSK 31.2 76.1 E 0.0592 0.0592 0.0592 0.0183 52 PTI 29.4 79.9 C 0.0285 0.0301 0.0293 0.0271
17 JAL 31.3 75.6 D 0.0600 0.0600 0.0600 0.0300 53 RIS 30.1 78.3 C 0.0271 0.0177 0.0224 0.0187
18 KAT 31.4 75.4 D 0.0456 0.0360 0.0408 0.0300 54 ROO 29.9 77.9 E 0.0599 0.0600 0.0599 0.0300
19 MOG 30.8 75.2 D 0.0594 0.0596 0.0595 0.0292 55 RUD 30.3 79 E 0.0596 0.0600 0.0598 0.0300
20 MUK 31.9 75.6 E 0.0319 0.0442 0.0381 0.0254 56 TAN 29.1 80.1 C 0.0526 0.0507 0.0517 0.0274
21 NAW 31.1 76.1 E 0.0315 0.0437 0.0376 0.0251 57 THE 30.4 78.4 D 0.0574 0.0430 0.0502 0.0213
22 NKD 31.1 75.5 E 0.0302 0.0560 0.0431 0.0269 58 UDH 29 79.4 D 0.0596 0.0490 0.0543 0.0300
23 PHG 31.2 75.8 D 0.0600 0.0600 0.0600 0.0300 59 UTK 30.7 78.4 D 0.0600 0.0585 0.0592 0.0300
24 TAR​ 31.4 74.9 D 0.0456 0.0360 0.0408 0.0300 60 VIK 30.5 77.8 D 0.0566 0.0600 0.0583 0.0300
25 ARI 26.1 77.5 D 0.0514 0.0511 0.0512 0.0297 61 PAL 28.1 77.3 D 0.0414 0.0489 0.0451 0.0286
26 IGN 28.5 77.2 C 0.0376 0.0357 0.0367 0.0252 62 JAFR 28.6 76.9 C 0.0508 0.0506 0.0507 0.0296
27 JNU 28.5 77.2 B 0.0534 0.0612 0.0573 0.0295 63 REW 28.2 76.6 D 0.0487 0.0520 0.0504 0.0300
28 DJB 28.7 77.2 D 0.0553 0.0546 0.0550 0.0244 64 CRRI 29 77.1 C 0.0443 0.0343 0.0393 0.0295
29 NDI 28.7 77.2 B 0.0407 0.0421 0.0414 0.0232 65 RGD 28.7 77.1 D 0.0379 0.0605 0.0492 0.0280
30 IMD 28.7 77.2 C 0.0542 0.0546 0.0544 0.0244 66 GGI 28.7 77.2 A 0.0529 0.0558 0.0544 0.0277
31 NTPC 28.5 77.3 D 0.0486 0.0426 0.0456 0.0280 67 DLU 28.7 77.2 D 0.0545 0.0517 0.0531 0.0265
32 ANC 28.5 77.3 C 0.0542 0.0546 0.0544 0.0244 68 DCE 28.8 77.1 C 0.0487 0.0441 0.0464 0.0276
33 JAMI 28.6 77.3 C 0.0561 0.0516 0.0539 0.0199 69 IGI 28.6 77.1 D 0.0559 0.0544 0.0551 0.0296
34 VCD 28.6 77.2 C 0.0460 0.0456 0.0458 0.0270 70 ZAKI 28.6 77.2 C 0.0526 0.0507 0.0517 0.0274
35 IIT 28.6 77.3 C 0.0434 0.0304 0.0369 0.0255 71 ALIP 28.8 77.1 D 0.0585 0.0577 0.0581 0.0274
72 NSIT 28.6 77 D 0.0514 0.0511 0.0512 0.0297

*Site Class
Indian Geotech J
Indian Geotech J

Source Spectra F(f )ij

F P (f )ij = = S(f )j A(f )i (3)
P(f )ij
The following subsections contain a thorough discussion
regarding the non-reference generalised inversion approach
that is used to develop the source spectra in this work. ln F P (f )ij = ln S(f )j + ln A(f )i (4)

Further, Eq. 4 can be rearranged considering:

Methodology
ln S(f )j = sj (f ), ln A(f )i = ai (f ) and ln F P (f )ij = zij(f), and it
can be expressed in the form of Eq. 5:
To define the ground motion spectrum, the following expres-
sion Eq. 1, in the frequency domain, proposed by Andrews ai (f ) + sj (f ) = zij (f ) (5)
[23] and Oth and Kaiser [9] is used;
Further, for station 1, Eq. 5 can be rewritten like Eq. 6,
F(f )ij = S(f )j P(f )ij A(f )ij (1) which corresponds to “n” sample frequencies, can be recast
for a specific EQ event that has been recorded at mth number
where F(f )ij is FAS of the ith EQ event recorded, at the
of stations as follows:
jth recording station; S(f )j is source term; P(f )ij is the path
attenuation; and A(f )j is the site term. The path attenuation ( ) ( ) ( )
a1 f1 + s f1 = z1 f1
characteristics can be estimated using Eq. 2, which has been ( ) ( ) ( )
a1 f2 + s f2 = z1 f2
proposed by Castro et al. [41].
[ −𝜋⋅f ⋅R ] ∶ ∶ ∶ (6)
( ) ij
P f , Rij = G(f , Rij ) e Q𝛽 (2) ∶ ∶ ∶
( ) ( ) ( )
a1 fn + s fn = z1 fn
In Eq. 2, G(f , Rij ) denotes the geometric spreading (taken
as 1∕Rij according to Parvez et al. [42]), Rij denotes the Similarly, there will be n set of equations (as shown in
hypocentre distance, Q denotes the quality factor (taken as Eq. 7) as listed below for the mth recording station (for the
Q = (105)f (0.94) as per Harinarayan and Kumar [24]), and 𝛽 same EQ event considered in Eq. 6);
denotes the average shear wave velocity in the crustal medium ( ) ( )
am f1 + s f1 = z1 f1
( )
(taken as 3.5 km/s following Mukhopadhyay and Kayal [43] ( ) ( ) ( )
for the study region). Further, the path attenuation term is am f2 + s f2 = z1 f2
removed from the spectral content from Eq. 1, and the expres- ∶ ∶ ∶ (7)
sion is reported in Eq. 3. Further, to derive a linear expression ∶ ∶ ∶
of Eq. 3, a natural logarithm is applied on both sides of Eq. 3, ( ) ( ) ( )
am fn + s fn = z1 fn
and it can be rewritten like Eq. 4.
Finally, Eqs. 6 and 7 can be written in form of matrix,
shown in Eq. 8, using Menke’s [40] notations that shows the
source term for the same EQ event:

| ← 1st station → ← mth station → ← Source → |||| s1 (f1 ) || || z1 (f1 ) ||

|
| 1 2 … N 1 2 … n 1 2 … n |||| ∶ || || ∶ ||
|
| 1 0 … 0 … 0 0 0 1 0 … 0 |||| ∶ || || ∶ ||
|
| || | | |
| 0 1 0 … 0 0 0 0 1 … 0 || ∶ | | ∶ |
| || | | |
| ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ || s1 (fm ) | | z1 (fn ) |
| || | | |
|
| ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ |||| sm (f1 ) || || |
|
|
| 0 0 0 1 0 0 … 0 0 0 … 1 |||| ∶ || = || |
| (8)
| For mth station || s (f ) | | z (f ) |
| || m( m) | | m n |
| 0 0 … 0 … 1 … 0 1 0 … 0 |||| a f1 || || zm (fn ) ||
|
| || | | |
| 0 0 … 0 … 0 1 … 0 0 1 … 0 || a(f2 ) | | ∶ |
| || | | |
|
| ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ |||| ∶ || || ∶ ||
|
| ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ ∶ |||| ∶ || || ∶ ||
|
| 0 0 … 0 … 0 0 … 1 0 0 … 1 |||| a(fn ) || || zm (fn ) ||

Since there are � (m + 1) × n� unknown parameters

and � n × m� data, it is evident that Eq. 8 depicts a purely

13
 Table 2  List of source parameters determined in this study
�
Event No Date (dd-mm-yyyy) (hh:mm) Lat. (°) (N) Lon. (°) (E) Depth (Km) Magnitude M0(N-m) fc(Hz) Υ r(m) Δ𝜎(MPa) Es(J) 𝜎A(MP) ε

13
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)

1 25-11-2007 28.6 77.0 20.3 4.3 9.00E+14 2.06 2.10 634.45 0.96 6.07E+08 0.17 0.85
2 19-08-2008 30.1 80.1 15.0 4.3 3.47E+14 2.78 2.60 469.49 0.17 1.69E+08 0.12 0.85
3 31-01-2009 32.5 75.9 10.0 3.7 1.17E+14 3.44 2.50 378.62 1.58 7.54E+09 1.64 0.65
4 04-09-2008 30.1 80.4 10.0 5.1 8.40E+14 2.78 2.50 469.49 4.89 2.06E+09 0.52 0.92
5 17-07-2009 32.3 76.1 39.3 3.7 4.50E+14 4.22 2.60 308.88 1.67 4.74E+10 1.10 0.87
6 11-01-2010 29.7 80.0 15.0 3.9 4.26E+13 4.78 3.10 272.96 2.29 2.07E+09 0.83 1.16
7 24-02-2010 28.6 76.9 17.0 2.5 3.33E+13 6.11 2.00 213.40 1.37 3.73E+07 0.19 0.99
8 31-05-2010 30.0 79.8 10.0 3.6 6.78E+13 5.00 3.00 260.83 4.18 3.84E+10 9.66 0.36
9 04-04-2011 29.6 80.8 10.0 5.7 4.78E+14 1.67 3.00 782.48 14.53 4.60E+12 12.35 0.74
10 12-03-2012 28.9 77.3 5.0 3.5 8.45E+12 3.44 2.00 378.62 0.17 5.61E+07 0.11 0.86
12 05-03-2012 28.7 76.6 14.0 4.9 4.55E+14 2.67 2.30 489.05 1.70 6.14E+10 5.76 0.26
13 11-11-2012 29.3 80.1 5.0 5.0 7.00E+14 2.67 2.40 489.05 4.29 1.35E+11 5.03 0.60
14 02-01-2013 29.4 81.1 10.0 4.8 4.50E+14 2.89 2.80 451.43 3.35 5.71E+10 2.17 0.87
15 09-01-2013 29.8 81.7 5.0 5.0 5.50E+14 2.78 2.60 469.49 3.81 8.34E+10 2.29 0.91
16 11-11-2013 (19:11:18) 28.5 77.2 10.0 3.1 8.55E+12 3.89 1.60 335.35 0.25 1.17E+08 0.23 0.69
17 11-11-2013 (22,10,42) 28.4 77.2 11.0 2.8 1.20E+13 5.11 2.00 255.16 0.79 3.17E+08 0.45 0.94
18 11-11-13 ( 20:11:30) 28.4 77.2 13.0 3.1 2.76E+13 3.67 1.60 355.67 0.67 1.07E+09 0.66 0.67
19 29-11-2015 30.6 79.6 15.0 4.0 2.25E+14 4.00 3.00 326.04 5.70 3.53E+10 3.34 0.92
20 25-09-2016 30.0 79.5 11.0 3.7 6.99E+13 4.00 3.00 326.04 2.21 3.39E+09 0.83 1.14
21 01-12-2016 30.6 79.6 19.0 4.0 6.78E+14 3.33 3.00 391.24 4.95 1.16E+12 2.50 1.00
22 21-10-2008 31.5 77.3 10.0 4.5 7.82E+13 3.33 2.00 391.24 3.57 2.75E+10 2.40 0.85
23 01-05-2009 29.9 80.1 10.0 4.6 9.82E+13 3.11 2.50 419.19 3.64 2.45E+11 2.00 0.95
24 27-08-2009 30.0 80.0 14.0 3.9 8.62E+13 4.00 3.00 326.04 6.80 4.39E+10 3.48 0.99
25 21-09-2009 30.9 79.1 13.0 4.7 7.42E+14 2.61 3.00 499.46 6.51 6.68E+11 7.00 0.63
26 03-10-2009 30.0 79.9 15.0 4.3 2.42E+14 3.06 3.00 426.81 5.67 1.14E+11 4.82 0.74
27 22-02-2010 30.0 80.1 2.0 4.7 2.42E+14 3.06 3.00 426.81 5.67 1.14E+11 4.82 0.74
28 03-05-2010 30.4 78.4 8.0 3.5 1.21E+14 4.58 3.00 284.54 3.07 2.27E+09 1.84 0.91
29 28-05-2010 31.2 77.9 43.0 4.8 5.54E+14 2.89 2.75 451.43 6.59 8.88E+10 3.73 0.94
30 06-07-2010 29.8 80.4 10.0 5.1 6.04E+14 2.89 2.60 451.43 4.79 1.13E+11 4.80 0.67
31 10-07-2010 29.9 79.6 10.0 4.1 3.54E+14 3.33 2.80 391.24 6.47 5.42E+10 3.61 0.95
32 18-02-2011 28.6 77.3 5.0 2.3 1.30E+13 5.00 1.60 260.83 0.80 4.68E+08 0.61 0.79
33 09-02-2011 30.9 78.2 10.0 5.0 1.05E+15 2.22 3.00 586.86 3.68 1.32E+11 2.50 0.85
34 15-06-2011 30.6 80.1 10.0 3.6 2.29E+14 3.33 2.00 391.24 14.61 4.60E+11 9.83 0.85
35 20-06-2011 30.5 79.4 12.0 4.6 3.00E+14 3.11 2.60 419.19 6.56 4.24E+11 3.39 0.98
36 27-03-2012 26.1 87.8 12.0 3.5 1.29E+14 3.56 3.00 366.79 3.99 9.88E+10 9.37 0.35
Indian Geotech J
Indian Geotech J

under-determinate system (where n represents the number a lower range of magnitude (Mw = 3.3–5.4) compared to
of sample frequency and m represents the number of record- the other studies used for comparison in Table 3. The
ing stations for an EQ event). To compute S(f ) for each EQ value of f c obtained in this study (tabulated in Table 2,
event, Eq. 8 is solved in this study utilising the Moore–Pen- column 8) are in the range of 1.67–6.11 Hz, with an aver-
rose matrix inversion approach (minimum norm inversion) age value of 3.34 Hz; whereas the value of � Υ (tabulated
described by Penrose [44]. in Table 2, column 9) are in the range of 1.3–3.1, with an
average value of 2.5. A good fit between the theoretical
M̈ 0 (estimated using the values of M0, fc and � Υ in Eq. 11)
Source Parameters and the computed M ̈ 0 as seen in Fig. 2 shows the robust-
ness of M0, fc, and � Υ obtained in this study. Estimation
The source spectra for 36 EQ events are computed using of Δσ, Es and σA for each of the 36 EQ events are done as
the inversion method outlined above. The resulting source discussed in further subsections.
spectrum for each event is fitted to the Brune’s point source
model [25] in order to establish the source parameters. The Stress Drop
Brune’s EQ source model [25] requires three independent
parameters namely; M0, fc and � Υ as mentioned in Eq. 9, and Assuming a circular dislocation, Brune [25] has developed
further Eq. 9 can be rearranged like Eq. 10. an empirical equations to define the dependancy of Δσ with
[ ( ) ] M0, and the radius of circular rupture (r) as shown in Eqs. 12
4𝜋 2 f 2 R� VF and 13.
̈ 0j
S(f )j = M (9)
4𝜋𝜌𝛽 3 R0 ( )
7M0
Δ𝜎 = × 10−6 (12)
16r3
S(f )j
̈ 0j = [
M 2 2
( ) /
4𝜋 f R� VF ( )
] (10) 2.34𝛽
4𝜋𝜌𝛽 3 R0 r= (13)
2𝜋fc
R∅ is the average shear wave radiation pattern (assumed
to be 0.55 following Mandal and Dutta [45]; F stands for where r , 𝛽 , M0 and Δσ are in the unit of m, m/s, Nm and
free surface amplification (taken as 2 in accordance with MPa, respectively. The value of r is in the range of 213–782
Zafarani et al. [46]; V accounts for the division of�
S-wave m, with an average value of 406.6 m (tabulated in Table 2
√ column 10). Further, the value of Δσ is in the range of
energy into two horizontal components (taken as 1 2 in 0.17–14.61 MPa for 36 events, with an average value of
accordance with [46]; 𝜌 is the mass density (taken as 2.8g/ 3.96 MPa (tabulated in Table 2 column 11). It can also be
cc, in accordance with [45]; R0 = 1 km is the reference dis- seen that the range of Δσ computed in this study is lower
tance, and M ̈ 0j is the moment rate spectrum.
in comparison with the other studies conducted worldwide
Further, Eq. 10 can be used to calculate M ̈ 0j , of 36 EQ
based on small to moderate EQs. Meirova and Hofstetter
events based on S(f )j , obtained from the inversion. Further- [47] reported Δσ in the range of 0.005–20 MPa for small to
more, M ̈ 0j for each of the 36 EQ events is compared and
moderate EQs in Israel. Dutta et al. [48] reported Δσ in the
fitted to the theoretical model presented in Eq. 11 (after range of 0.2–24 MPa for small to moderate EQs in Alaska.
Brune [25]) using a nonlinear least squares method (similar Sharma and Wason [49] also reported a lower value of Δσ
to Bindi et al. [39]), and the values of M0 , fc and � Υ are (0.004 MPa < Δσ < 5.3 MPa) for the Garhwal Himalaya (a
estimated. part of north-west Himalaya). According to Sharma and
M0 Wason [49], low Δσ value for EQs in this region is an indi-
̈ 0j =
M cation that the crustal medium has low strength to withstand
( )� Υ (11)
1 + f ∕fcj considerable value of accumulated strain. A similar observa-
tion of low value of Δσ for small to moderate EQ was also
The value of M0, estimated in this study (tabulated in
reported for other parts of Himalayas like Arunachal Pradesh
Table 2, column 7), are ranging from 3.33 × 1012 Nm to
[50] and north-east Himalaya [51]. The phenomenon of low
1.05 × 1015 Nm, with an average value of 2.05 × 1014 Nm.
∆for small to moderate EQ was also reported namely: partial
It can be seen that the range of M0, obtained in the present
stress drop model [25] and low effective stress model [52].
study, is smaller in comparison with other studies for the
In the case of low effective stress model (i.e. low ∆σ), EQ
Himalayan region (see Table 3). The lower range of M0
occurs when the available stress required for accelerating
estimated in this study can be attributed to the data set
the fault is low. In the case of low effective stress model,
used for the analysis. These consist of EQ events having
the event propagation is characterised as “creep events” or

13
 Indian Geotech J

Table 3  Compilation of source Region Magnitude M0(Nm) r(km) fc(Hz) Δσ(MPa)

parameters from the region
of Himalaya including those Garhwal Himalaya [49] 1.4–4.2 11
7 × ­10 –6.23 × ­1014
0.347–0.545 – 0.004–3.89
estimated in this study
Himachal Pradesh [65] 5.4 2.1 × ­1017 2.8 – 3.6
Sikkim [66] 4.0–5.1 7.9 × ­1014–6.31 × ­1016 0.225–0.781 1.8–6.5 4.7–38.91
North east Himalaya [51] 3.7–5.8 1.10 × ­1015–2.63 × ­1025 0.5–2 0.99–3.9 2.3–30.9
Arunachal Pradesh [50] 1.0–3.6 6.13 × ­1011–3.15 × ­1014 0.174–0.357 3.2–6.4 0.02–5.6
This study 2.3–5.7 3.33 × ­1012–1.05 × ­1015 0.213–0.782 1.67–6.11 0.17–14.61

“slow EQs” due to minimal particle velocity resulting from (

A = 15𝜋𝜌𝛼 5
)−1 (
+ 10𝜋𝜌𝛽 5
)−1
(15)
very low effective stress.
where α represents P-wave velocity of the source region’s,
Seismic Energy and Apparent Stress which is 6.4 km/s as per [45]. For 36 EQ occurrences, Es
values range from 3.73 × ­107 to 4.68 × ­1012 J, with an aver-
Seismic waves are the results of an EQ releasing its accu- age value of 2.43 × ­1011 J (Table 2). Furthermore, σA is cal-
mulated strain energy and, the total energy emitted during culated for each event using Eq. 16 [25] depending on the
an EQ can be utilised to determine its state of stress [53]. value of Es obtained.
According to Vassiliou and Kanamori [54], Es for each event
𝜎A = 𝜇Es ∕M0 (16)
in the current work is calculated in the frequency range of
0.25–15 Hz utilising Eq. 14. The rigidity modulus of the medium is denoted by the
+∞ 2
symbol 𝜇 in Eq. 16 which is further calculated as 𝜇 = 𝛽 2 𝛼 .
| ̈ 0j || df
Es = A ∫ |2𝜋f M (14) The values of σA calculated for this work (refer to Table 2)
| |
has an average value of 3.18 MPa and ranges from 0.1 MPa
−∞

0.25 2.5 25 0.25 2.5 25 0.25 2.5 25

0.25 2.5 25 0.25 2.5 25 0.25 2.5 25

0.25 2.5 25 0.25 2.5 25

Fig. 2  Computed (thick line) and best-fit (dotted line) moment rate spectra

13
Indian Geotech J

to 12.36 MPa. Zúñiga [55] parameter (ε) is estimated in M0 and fc restricted to − 3, which results in the following
accordance with Eq. 17 to further analyse the mechanism expression:
of stress release for the research area.
log M0 = −3 log fc + 16.06 (19)
Δ𝜎
ε= Δσ (17)
σA + 2 Further, Eq. (19) can be rewritten as follows (see Eq. 20),
As stated by Zuniga [55]: 𝜀 < 1 indicates partial stress which has been further used to estimate the size of sub-faults
drop mechanism, whereas ε > 1 indicates the frictional [57].
overshoot mechanism (when the final stress is lower than M0 fc3 = 1.15 × 1016 Nm/s3 (20)
the frictional stress). The value of ε obtained in this study
is given in Table 2, Column 14. The range of ε is found Several such relationships also exist for the Himala-
to be 0.06–1.38. It can be observed from Fig. 3 that about yan region. However, for the Bilaspur region of Himachal
90% of events belong to the partial stress drop mechanism Pradesh, Vandana et al. [58] proposed M0 fc3 = 2 × 1015 Nm∕s3
while the remaining 10% events suggest frictional overshoot based on 41 EQ incidents (0.5 < M < 2.9). In another
mechanism, indicating that the partial stress drop mecha- study for north-east India, Kumar et al. [51] developed a
nism is prominent in this region. The partial stress drop relation M0 fc3 = 1.35 × 1017 Nm∕s3 using 50 EQ events
model accounts for EQ events occurring when the fault locks (3.7 < M < 5.8). Furthermore, for the Kameng region of
itself immediately after the rupture of the main event. As a the Himalayas, a correlation M0 fc3.34 = 2 × 1015 Nm∕s3 has
result, the average slip release cannot reach to the optimum been developed based on 16 EQ events of seismic magnitude
dynamic stress drop over the entire fault or where there is ranging from 2.2 < M < 3.7. It indicates that the each seismic
non-uniform and coherent stress release over the whole fault regions have distinct scaling relations, thus it can be stated
plane. The self-locking of faults is either due to barriers or that there is a need to develop a regional scaling relation-
due to complexities on the faults. ship for an appropriate seismic hazard assessment. Figure 5
highlight the variations of Es with M0 for each of the 36 EQ
Scaling Relations events considered for this analysis. Further, based on the
linear regression analysis a correlation between Es and M0
A plot of M0 versus fc in logarithmic units is created (similar has been established, mentioned in Eq. 21. From Fig. 5, it
to Aki [56]) to comprehend the relationship between M0 and can be observed that Es increases with M0.
fc for the study region. The following expression (Eq. 18) is
the result of regression analysis between M0 and fc: log Es = 1.7 log M0 − 14.4 (21)

log M0 = −3.1 log fc + 16.10 (18)

Further, the variation of Δσ vs M0 for each of the 36 EQ
According to Aki [56], the scaling relation between M0 events considered in the current investigation as shown in
and fc follow a correlation, which is M0 fc3 = constant. Fig. 6. It is clear from Fig. 6 that M0 does not exhibit any
Regression analysis was further performed (see Fig. 4) change with Δσ . Also, Δσ is found to have no significant
with the slope component of the relationship between variation with M (see Fig. 7). The self-similar behaviour

Fig. 3  Zuniga parameter for the 2

36 EQ events
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37

13
 Indian Geotech J

of EQs [56] in the research area is implied by the relation be vary in between 0.01 and 0.071 with an average value of
c and the non-dependence of
M0 ∝ f−3 Δσ on M0. 0.051. Similarly, 𝜿 V was found to be vary in between 0.01
and 0.036, with an average value of 0.027.
A major concern with regards to the values of fe and fmax
Estimation of Kappa Factor pointed out by Paralai and Bindi [60] is the influence of site
amplification, which could bias κ measurement. Paralai and
In the present study, κ has been estimated using Anderson Bindi [60] recommended that κ would be unbiased if the pre-
and Hough approach [59]. Initially, a plot of Fourier ampli- dominant frequency of the site is less than the value of fe. In
tude spectrum (FAS) in logarithmic scale versus frequency the present study, the predominant frequency of soil beneath
is generated, which reflects a linear decay of FAS beyond a the recording stations is in the range 1.1–5 Hz [24], which is
particular frequency. The frequencies corresponding to the less than the range of fe . Thus, as per the above discussion
starting and ending of the decay line of FAS are identified as it can be concluded that 𝜿 values estimated in the present
fe and fmax , respectively. For the present dataset, the values study are not affected by site amplification. Further, varia-
of fe and fmax are in the range of 7–15 Hz and 30–45 Hz, tions in the value of 𝜿 estimated for the vertical, east–west
respectively. Further, the linear spectral decay portion of the and north–south components are discussed in detail in the
plot in the range of fmax > f > fe has been modelled, using next section.
Eq. 22, in accordance with Anderson and Hough [59] and
further, the value of 𝜿 has been estimated using standard
least-square regression procedure.

Af = A0 e−𝜋𝜿f ; fmax > f > fe (22)

where A0 is a parameter that depends on the source and

path parameters, and Af represents FAS within the limits 14
fmax > f > fe. 14.7
13
Figure 8 illustrates an example of FAS (in log scale) ver-
sus frequency plot for Champawat recording station (located 12
in the Uttarakhand region). A clear linear downward trend in 11
FAS, within the range of frequency 8 ( fe ) and 33 ( fmax ) can 10
be observed in Fig. 8 along with model fit (indicated by the
9
straight line). Further, the values of 𝜿 for east–west (denoted
by 𝜿 EW), north–south (denoted by 𝜿 NS) and vertical (denoted 8
by 𝜿 V) components of EQ records used in the present dataset 7
are estimated. The average value of 𝜿 EW and 𝜿 NS for each of 14 14.5 15 15.5 16 16.5
the recording stations is further estimated, which is listed in
Table 1. 𝜿 EW was found to be vary between 0.02 and 0.076
with an average value of 0.0501, whereas 𝜿 NS was found to Fig. 5  Energy released versus corner frequency in logarithmic unit.
The regression relation is indicated by solid line

16.5 6
= + 16.06
16
15.5 5
15
4
14.5
14 3
13.5
2
13
12.5 1
12
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0

Fig. 4  Scalar moment versus corner frequency in logarithmic unit.

The regression relation is indicated by solid line Fig. 6  Stress drop versus seismic energy

13
Indian Geotech J

0
2 3 4 5 6

0.1 10.1 20.1 30.1

Fig. 7  Stress drop versus magnitude

Fig. 8  An example of manual picking of frequencies fe and fmax

used for estimating kappa factor from shear wave spectrum. (The
Variations in Kappa Estimated for the Vertical, East– sample accelerogram is from an event that occurred on 27 August
West and North–South Components 2009 recorded at Champawat recording station)

Initially, the variations in the values of 𝜿 EW and 𝜿 NS have

been studied. Figure 9 shows the ratio 𝜿 EW to 𝜿 NS versus 𝜿 H = 0.047 + 0.00006R (24a)
epicentre distance. It can be seen that despite few discrepan-
cies at some epicentre distance, for the majority of EQs the
ratio of 𝜿 EW to 𝜿 NS is approximately close to unity. How-
𝜿 V = 0.026 + 0.00002R (24b)
ever, in this study, a single value of κ represented by 𝜿 H , is
used for horizontal component, which has been calculated For the zero value
based on the arithmetic mean of 𝜿 EW and 𝜿 NS. 𝜿 H is found ( of epicentral
) distance, the horizontal
(𝜿 (0,H)) and vertical 𝜿 (0,V) component of 𝜿 was found to be
to vary between 0.015 and 0.07 with an average value of 0.047 and 0.026, respectively. Further, for the consideration
0.051. Further, Fig. 10 presents a comparison between 𝜿 H of site effect, the value of 𝜿 0 can be utilised along with hori-
and 𝜿 V along with the epicentral distance. It can be seen that zontal to vertical spectral ratio method [63].
for the majority of EQ records the value of 𝜿 H is twice than
the value of 𝜿 V . Similar observations have been reported by Effect of Site Class on Kappa
Askan et al. [61] for the north-west Turkey region, Douglas
et al. [62] for the France region and by Motazedian [63] for The 𝜿 model is separately estimated for NEHRP site classes
the North Iran region. C and D (where a sufficient number of EQ records are avail-
able) [32]. Figures 12 and(13 present) the variations of hori-
zontal (𝜿 (0,H)) and vertical 𝜿 (0,V) component of 𝜿 along with
Regional Kappa Model epicentral distance for site classes C and D, respectively. The

For the present study region, the dependency of κ on epicen-

tral distance dependence has been performed using Ander-
son and Hough model [59], as shown in Eq. 23. 3
𝜿 = 𝜿 0 + CR R (23) 2.5

where 𝜿 0 is the value of 𝜿 at zero epicentral distance ( R = 0). 2

However, Douglas et al. [62] reported that the value of 𝜿 0
is be station dependent, which is related to the near-surface 1.5
attenuation in the top few kilometres under the site and CR 1
depends on the regional attenuation characteristics.
Figure 11a, b presents the variations of 𝜿 along with epi- 0.5
central distance for horizontal (𝜿 H ) and vertical (𝜿 V ) com-
0
ponents, respectively. Based on the regression analysis for 0 50 100 150 200 250
the data presented in Fig. 11a, b, the following Eqs. 24a and
24b have been developed.
Fig. 9  Variation of the ratio of 𝜿 EW to 𝜿 NS with epicentre distance

13
 Indian Geotech J

0.08

0.06

0.04

0.02

0
0 50 100 150 200 250

Fig. 10  Variation of the ratio of 𝜿 H to 𝜿 V with epicentre distance

0.04

𝜿 model for site class C has been computed for 𝜿 (0,H) and
0.03
𝜿 (0,V), and presented in Eqs. (25a) and (25b), respectively.
𝜿 H = 0.0452 + 0.00009R (25a) 0.02

𝜿 V = 0.0258 + 0.00003R (25b)

0.01

Further, the 𝜿 model for site class D has also been devel-
0
oped for 𝜿 (0,H) and 𝜿 (0,V) as shown in Eqs. 26a and 26b, 0 50 100 150 200 250
respectively.
𝜿 H = 0.0482 + 0.00008R (26a) Fig. 11  a Dependency of 𝜿 H on epicentre distance; b Dependency of
𝜿 V on epicentre distance
𝜿 V = 0.0285 + 0.000007R (26b)
From Eqs. (25a, 25b) and (26a, 26b), the value of 𝜿 (0,H) of all sub-fault is added after incorporating pertinent time
for site classes C and D was found to be 0.0452 and 0.0482, delays to get the contribution of the entire fault length.
respectively, whereas the value of 𝜿 (0,V) for site classes C and Details of stochastic finite fault method and the EXSIM code
D were found to be 0.0258 and 0.0285, respectively. Higher can be found in [64].
values of 𝜿 (0,H) and 𝜿 (0,V) for site classes D in comparison Further, for ground motion simulations, four different
with site classes C indicate higher attenuation of seismic EQs have been considered, which include EQs of 06th July
wave in softer soil sites. Similar observations have been 2010 (M = 5.1) recorded at PAL recording station, 15th July
reported by Askan et al. [61] for Turkey region. 2011 (M = 3.6) recorded at IMD recording station, 10th July
2010 (M = 4.1) recorded at TEH recording station, and 21st
October-10–08 event (M = 4.5) recorded at DHH recording
Validation of Estimated Source Parameters station. The above records have been considered in such a
and Kappa Model in Stochastic Strong‑Motion way that the far-field and near field simulations can be car-
Simulation ried out for robust validation. The input parameters for sto-
chastic strong-motion simulations are presented in Table 4.
In order to validate the source parameters and 𝜿 model The path parameters have been considered from Harinarayan
obtained in the present study, ground motion simulations and Kumar [24], whereas site amplification curves for the
for selected EQ events have been carried out. The computer recording stations have been chosen from Harinarayan and
program EXSIM, developed by Motazedian and Atkinson Kumar [38]. Comparison between observed (represented
[64], has been used for ground motion simulations. EXSIM by a straight line) and simulated (represented by a dashed
is based on stochastic finite fault method that considers the line) response spectra given in Fig. 14 indicates that EQs are
entire fault as a number of sub-faults; further each sub-faults effectively simulated confirming the robustness of the source
is modelled as a stochastic point source. The contribution parameters and 𝜿 model predicted in this work.

13
Indian Geotech J

0.08 Conclusions

0.06 Strong-motion records of 36 small to moderate size EQs

(2.5 < M < 5.7), recorded at 72 recording stations, are ana-
0.04
lysed and interpreted to study the source parameters and κ
for the region around the north-western Himalayas within
north India. The source spectrum is developed separately
0.02
for each event using a modified generalised inversion tech-
nique that does not necessarily require any groung motion
0 records from rock site. Thus, this method enables to esti-
0 50 100 150 200 250
mate source spectra for larger number of EQ events in
comparison with the conventional generalised inversion
0.04
method, where only those events with at least one groung
motion record from rock site available can be analysed.
0.03 Afterwards, source parameters like M0 , fc , Δσ , Es and σA
for each event are estimated adopting point source model.
0.02 Finally, the value of κ for each record is estimated. Based
on the present work, following important conclusions are
0.01 made for the study region:

• The scaling relation for the region was found to be

0
0 50 100 150 200 250 M0 fc3 = 1.15 × 1016 Nm/s3 . Furthermore, estimates
of Δσ in the present study indicate that the EQs in
the region are characterised by low Δσ, which can be
Fig. 12  a Dependency of 𝜿 H on epicentre distance for site class C; b understood by partial stress drop model or low effective
Dependency of 𝜿 V on epicentre distance for site class C stress model. Moreover, the computation of Zúñiga,
(1993) parameter, based on the estimated Δσ and σA
0.08 values for all the 36 Eqs i ndicates that 90% of events
follow the partial stress drop model.
0.06
• Δ𝜎 is found to shows no significant variation with mag-
nitude and M0 indicating the/; small to moderate size
EQs in the study area follow a self-similar nature.
0.04
• The value of Es, computed for the study area, was found
to be in the range of 3.73 × ­107 J and 4.68 × ­1012 J.
0.02 • The value of κH was found to be in the range of 0.015–
0.07 with an average value of 0.051. However, κV was
0 found to be in the order of 0.01–0.036, with an average
0 50 100 150 200 250 value of 0.027. Further, a regional 𝜿 model, as a function
of epicentre distance, has been developed for the entire
0.04 study area and separately for different site classes.

Moreover, the outcome of the present study provides an

important information that are needed for the generation of
0.02
synthetic ground motions and seismic hazard assessment
for northern region India.

0
0 50 100 150 200 250

Fig. 13  a Dependency of 𝜿 H on epicentre distance for site class D; b

Dependency of 𝜿 V on epicentre distance for site class D

13
 Indian Geotech J

Table 4  Parameters used for Ground-Motion Simulations with EXSIM

Parameter 06-07-10 event 15-06-11 event 10-07-10 event 21-10-08 event References
recorded at PAL recorded at IMD recorded at TEH recorded at DHH
recording station, recording station recording station recording station

Epicentre location 29.8° N, 80.4° E 30.6° N, 80.1° E 29.9° N, 79.6° E 31.5° N, 77.3° E
Magnitude 5.1 3.6 4.1 4.5
Strike 108° 122° 342° 98° Prasanth et al. [67]
Dip 85° 53° 58° 78° Prasanth et al. [67]
Stress drop 4.79 MPa 14.61 MPa 6.47 MPa 3.57 MPa Present study
Crustal density 2.8 g/cm3 Mandal and Dutta [45]
Crustal shear wave 3.5 km/s Mukhopadhyay and
velocity Kayal [43]
Quality factor (Qs ) Qs = (105)f (0.94) Harinarayan and
Kumar [24]
Geometric spreading 1∕Rij Parvez et al. [42]
NEHRP based site D C D C Harinarayan and
class Kumar [57]
Site amplification Obtained from HVSR estimates reported by Harinarayan and Kumar [57]
Kappa 0.053* 0.047* 0.058* 0.049* Present study

*Kappa values obtained according to the kappa model for the corresponding site classes (using Eqs. 25a and 26a)

100

10

0.1

0.01
0.1 1 10 100 0.1 1 10 100

0.1 1 10 100 0.1 1 10 100

Fig. 14  Comparison of simulated and observed Fourier amplitude recorded at TEH recording station; d 21-10-08 event recorded at
spectra for: a 06-07-10 event recorded at PAL recording station; b DHH recording station
15-06-11 event recorded at IMD recording station; c 10-07-10 event

13
Indian Geotech J

Author Contributions HNH was involved in conceptualization, for- 12. Kaneko Y, Shearer PM (2014) Seismic source spectra & estimated
mal analysis, SSK helped in revision and editing of drafted paper with stress drop derived from cohesive-zone models of circular sub-
some formal analysis; writing—original preparation; KMM, CK & shear rupture. Geophys J Inter 197(2):1002–1015
AK contributed to formal analysis, writing—original draft preparation. 13. Frankel A (1981) Source parameters and scaling relationships of
small earthquakes in the northeastern Caribbean. Bull Seismol
Soc America 71(4):1173–1190
14. Archuleta RJ, Cranswick E, Mueller C, Spudich P (1982) Source
parameters of the 1980 Mammoth Lakes, California, earthquake
Funding Not applicable.
sequence. J Geophys Resear: Solid Earth 87(B6):4595–4607
15. Dysart PS, Snoke JA, Sacks IS (1988) Source parameters and
Data Availability The data of this study are available from the cor- scaling relations for small earthquakes in the Matsushiro region,
responding author upon reasonable request. southwest Honshu. Jpn Bull Seismol Soc Am 78(2):571–589
16. Allmann BP, Shearer PM (2009) Global variations of stress drop
Declarations for moderate to large earthquakes. J Geophys Res: Solid Earth
114:B1
Conflict of interest The authors declare that they have no known 17. Boore DM (2003) Simulation of ground motion using the stochas-
competing financial interests or personal relationships that could have tic method. Pure Appl Geophys 160(3–4):635–676
appeared to influence the work reported in this paper. 18. Kumar S, Kumar D, Rastogi BK (2014) Source parameters &
scaling relations for small earthquakes in the Kachchh region of
Ethical Approval Not applicable. Gujarat. India Nat Haz 73(3):1269–1289
19. Oth A, Bindi D, Parolai S, Di Giacomo D (2010) Earthquake
scaling characteristics and the scale-(in) dependence of seismic
energy-to-moment ratio: Insights from KiK-net data in Japan.
Geophys Res Let 37:19
References 20. Papageorgiou AS, Aki K (1983) A specific barrier model for the
quantitative description of inhomogeneous faulting and the predic-
1. Kumar P, Kumar SS, Harinarayan HN (2023) Development of tion of strong ground motion. I. Description of the model. Bull
synthetic ground motion-based attenuation relationship for Bihar Seismol Soc Am 73(3):693–722
region for seismic ground response analysis considering cen- 21. Tsai CCP, Chen KC (2000) A model for the high-cut process
tral seismic gap. Ann Geophys 63:3–4. https://​doi.​org/​10.​4401/​ of strong-motion accelerations in terms of distance, magnitude,
ag-​8889 & site condition: an example from the SMART 1 array, Lotung.
2. Bhatia SC, Kumar MR, Gupta HK (1999) A probabilistic Taiwan Bull Seismol Soc Am 90(6):1535–1542
seismic hazard map of India & adjoining regions. Ann Geofis 22. Mena B, Mai PM, Olsen KB, Purvance MD, Brune JN (2010)
42:1153–1164 Hybrid broad band and ground-motion simulation using scatter-
3. Ansari A, Rao KS, Jain AK, Ansari A (2023) Formulation of ing Green’s functions: application to large-magnitude events. Bull
multi-hazard damage prediction (MhDP) model for tunnel- Seismol Soc Am 100(5A):2143–2162
ling projects in earthquake and landslide prone regions: a novel 23. Andrews DJ (1986) Objective determination of source parameters
approach with artificial neural networking (ANN). J Earth Syst & similarity of earthquakes of different size. Earthq Sour Mech
Sci 132(164):1–11. https://​doi.​org/​10.​1007/​s12040-​023-​02178-y 37:259–267
4. Ambraseys NN, Douglas J (2004) Magnitude calibration of north 24. Harinarayan NH, Kumar A (2019) Estimation of path attenuation
Indian earthquakes. Geophys J Intern 159(1):165–206 and site characteristics in the north-west Himalaya & its adjoining
5. Ambraseys N, Bilham R (2000) A note on the Kangra earthquake area using generalized inversion method. Annl Geophys 62:8
of 4 April 1905. Current Sci 79:46–50 25. Brune JN (1970) Tectonic stress & the spectra of seismic shear
6. Kayal JR (1996) Precursor seismicity, foreshocks and aftershocks waves from earthquakes. J Geophys Res 75(26):4997–5009
of the Uttarkashi earthquake of October 20, 1991 at Garhwal 26. Valdiya KS (1984) Evolution of the Himalaya. Tectonophy
Himalaya. Tectonophys 263(1–4):339–345 105(1–4):229–248
7. Mandal PRANTIK, Padhy S, Rastogi BK, Satyanarayana HVS, 27. Srivastava P, Mitra G (1994) Thrust geometries & deep struc-
Kousalya M, Vijayraghavan R, Srinivasan A (2001) Aftershock ture of the outer & lesser Himalaya, Kumaon & Garhwal (India):
activity & frequency-dependent low coda Qc in the epicentral implications for evolution of the Himalayan fold-&-thrust belt.
region of the 1999 Chamoli earthquake of Mw 6.4. Pure Appl Tectonics 13(1):89–109
Geophys 158:1719–1735 28. Malik JN, Nakata T (2003) Active faults & related late quater-
8. Mandal P, Chadha RK, Kumar N, Raju IP, Satyamurty C (2007) nary deformation along the northwestern Himalayan Frontal Zone,
Source parameters of the deadly M w 7.6 Kashmir earthquake of India. Annl Geophy 46:5
8 October, 2005. Pure Appl Geophys 164:1963–1983 29. Ansari A, Rao KS, Jain AK (2022) Seismic vulnerability of
9. Oth A, Kaiser AE (2014) Stress release & source scaling of the tunnels in Jammu and Kashmir during post-seismic function-
2010–2011 Canterbury New Zeal and earthquake sequence from ality. Geotech Geol Eng 40(11):1–26. https://​doi.​org/​10.​1007/​
spectral inversion of ground motion data. Pure Appl Geophys s10706-​022-​02341-0
171(10):2767–2782 30. Ansari A, Rao KS, Jain AK, Ansari A (2022) Deep learning model
10. Parolai S, Bindi D, Trojani L (2001) Site response for the RSM for predicting tunnel damages and track serviceability under seis-
seismic network & source parameters in the central apennines mic environment. Model Earth Syst Environ 8(4):1–20. https://​
(Italy). Pure Appl Geophys 158:695–715 doi.​org/​10.​1007/​s40808-​022-​01556-7
11. Saito K, Masuda T (1981) Precursory change of spectral charac- 31. Seeber L, Armbruster JG (1981) Great detachment earthquakes
teristics before the 1978 Miyagi-ken-oki Earthquake. The science along the Himalayan arc & long-term forecasting. Earthq Predic:
reports of the Tohoku University, Fifth series. Tohoku Geophys J Inter Rev 4:259–277
27(3):95–110 32. Ansari A, Rao KS, Jain AK (2023) An integrated approach to
model seismic loss for the Himalayan infrastructure projects:

13
 Indian Geotech J

decision making and functionality concept for disaster miti- 51. Kumar V, Kumar D, Chopra S (2019) Source parameters scal-
gation. Bull Eng Geol Env 82:1–12. https://​doi.​org/​10.​1007/​ ing relations for moderate size earthquakes in North-East India
s10064-​023-​03422-x Region. Pure Appl Geophy 176(1):45–64
33. Ansari A, Rao KS, Jain AK (2023) Seismic response and fragility 52. Brune JN, Fletcher J, Vernon F, Haar L, Hanks T, Berger J (1986)
evaluation of circular tunnels in the Himalayan region: implica- Low stress-drop earthquakes in the light of new data from the
tions for post-seismic performance of transportation infrastructure Anza, California telemetered digital array. Earthq Sou Mech
projects in Jammu and Kashmir. Tunn Undergr Space Technol 37:237–245
137:1–13. https://​doi.​org/​10.​1016/j.​tust.​2023.​105118 53. Kanamori H (2001) Energy budget of earthquakes and seismic
34. Ansari A, Zahoor F, Rao KS, Jain AK (2023) Seismic response efficiency. Int J Geophys 76:293–305
and vulnerability evaluation of Jammu Region (Jammu and Kash- 54. Vassiliou MS, Kanamori H (1982) The energy release in earth-
mir). Indian Geotech J 53(3):509–522. https://​doi.​org/​10.​1007/​ quakes. Bull Seism Soc Am 72(2):371–387
s40098-​022-​00694-0 55. Zúñiga FR (1993) Frictional overshoot & partial stress drop.
35. Mundepi AK, Galiana-Merino JJ, Lindholm C (2010) Soil charac- Which one? Bull Seism Soc Am 83(3):939–944
teristics & site effect assessment in the city of Delhi (India) using 56. Aki K (1967) Scaling law of seismic spectrum. J Geophy Res
H/V & f–k methods. Soil Dyn Earthq Eng 81:767–773 72(4):1217–1231
36. GSI (1992) Macrosesimic investigation of Uttarkashi earthquake 57. Kumar A, Harinarayan NH, Verma V, Anand S, Borah U, Bania
of 20th October, 1991. In: Uttarkashi earthquake. Geolog Survey M (2018) Seismic site classification and empirical correlation
India Spec Pub, vol. 30, pp 1–188 between standard penetration test n-value and Shear wave veloc-
37. Kumar A, Mittal H, Sachdeva R, Kumar A (2012) Indian strong ity for Guwahati based on thorough subsoil investigation data.
motion instrumentation network. Seism Res Let 83(1):59–66 Pure Appl Geophy 175:2721–2738
38. Harinarayan NH, Kumar A (2018) Determination of NEHRP Site 58. Vandana KA, Gupta SC, Mishra OP, Kumar A (2017) Source
class of seismic recording stations in the Northwest Himalayas parameters & high frequency characteristics of local events (0.5≤
& its adjoining area using HVSR method. Pure Appl Geophy M L≤ 2.9) around Bilaspur region of the Himachal Himalaya.
175(1):89–107 Pure Appl Geophy 174(4):1643–1658
39. Bindi D, Pacor F, Luzi L, Massa M, Ameri G (2009) The Mw 6.3, 59. Anderson JG, Hough SE (1984) A model for the shape of the
2009 L’Aquila earthquake: source, path & site effects from spec- Fourier amplitude spectrum of acceleration at high frequencies.
tral analysis of strong motion. Geophy J Intern 179(3):1573–1579 Bull Seism Soc Am 74(5):1969–1993
40. Konno K, Ohmachi T (1998) Ground-motion characteristics esti- 60. Parolai S, Bindi D, Baumbach M, Grosser H, Milkereit C, Karak-
mated from spectral ratio between horizontal and vertical compo- isa S, Zünbül S (2004) Comparison of different site response esti-
nents of microtremor. Bull Seism Soc Am 88(1):228–241 mation techniques using aftershocks of the 1999 Izmit earthquake.
41. Castro RR, Pacor F, Sala A, Petrungaro C (1996) S wave attenua- Bull Seism Soc Am 94(3):1096–1108
tion and site effects in the region of Friuli. J Geophy Res B: Solid 61. Askan A, Sisman FN, Pekcan O (2014) A regional near-surface
Earth 101(B10):22355–22369 high frequency spectral attenuation (kappa) model for northwest-
42. Parvez IA, Yadav P, Nagaraj K (2012) Attenuation of P, S & coda ern Turkey. Soil Dyn Earth Eng 65:113–125
waves in the NW-Himalayas, India. Inter J Geosc 3(1):179 62. Douglas J, Gehl P, Bonilla LF, Gelis C (2010) A κ-model for
43. Mukhopadhyay S, Kayal JR (2003) Seismic tomography struc- Mainl & France. Pure Appl Geophy 167:1303–1315
ture of the 1999 Chamoli earthquake source area in the Garhwal 63. Motazedian D (2006) Region-Specific key seismic parameters for
Himalaya. Bull Seism Soc Am 93(4):1854–1861 earthquakes in northern Iran. Bull Seism Soc Am 96:1383–1395
44. Penrose R (1955) A generalized inverse for matrices. Math Proc 64. Motazedian D, Atkinson GM (2005) Stochastic finite-fault mod-
Camb Philos Soc 51:406–413 eling based on a dynamic corner frequency. Bull Seism Soc Am
45. Mandal P, Dutta U (2011) Estimation of earthquake source param- 95(3):995–1010
eters in the Kachchh seismic zone, Gujarat, India, from strong- 65. Ram VS, Kumar D, Khattri KN (2005) The 1986 Dharamsala
motion network data using a generalized inversion technique. Bull earthquake of Himachal Himalaya: estimates of source parame-
Seism Soc Am 101(4):1719–1731 ters, average intrinsic attenuation and site amplification functions.
46. Zafarani H, Hassani B, Ansari A (2012) Estimation of earthquake J Seismol 9:473–485
parameters in the Alborz seismic zone, Iran using generalized 66. Hazarika P, Ravi Kumar M (2012) Seismicity and source param-
inversion method. Soil Dyn Earthq Eng 42:197–218 eters of moderate earthquakes in Sikkim Himalaya. Nat Hazs
47. Meirova T, Hofstetter A (2017) Source parameters of regional 62:937–952
earthquakes recorded by Israel Seismic Network: implications for 67. Prasath RA, Singh S (2017) Upper crustal stress and seismotec-
earthquake scaling. Bull Earthq Eng 15(9):3417–3436 tonics of the Garhwal Himalaya using small-to-moderate earth-
48. Dutta U, Biswas N, Martirosyan A, Papageorgiou A, Kinoshita quakes: implications to the local structures and free fluids. J Asian
S (2003) Estimation of earthquake source parameters & site Earth Sci 135:198–211
response in Anchorage, Alaska from strong-motion network
data using generalized inversion method. Phy Earth Plan Inter Publisher’s Note Springer Nature remains neutral with regard to
137(1–4):13–29 jurisdictional claims in published maps and institutional affiliations.
49. Sharma ML, Wason HR (1994) Occurrence of low stress drop
earthquakes in the Garhwal Himalaya region. Phy Earth Plan Inter Springer Nature or its licensor (e.g. a society or other partner) holds
85(3–4):265–272 exclusive rights to this article under a publishing agreement with the
50. Kumar R, Gupta SC, Kumar A, Mittal H (2015) Source param- author(s) or other rightsholder(s); author self-archiving of the accepted
eters & f­ max in lower Siang region of Arunachal lesser Himalaya. manuscript version of this article is solely governed by the terms of
Arab J Geosc 8(1):255–265 such publishing agreement and applicable law.

13