Nat Hazards

DOI 10.1007/s11069-015-1764-0

ORIGINAL PAPER

Seismic hazard maps and spectrum for Patna considering

region-specific seismotectonic parameters

P. Anbazhagan1 • Ketan Bajaj1 • Satyajit Patel2

Received: 28 April 2014 / Accepted: 13 April 2015

Ó Springer Science+Business Media Dordrecht 2015

Abstract The objective of this paper was to develop the seismic hazard maps of Patna
district considering the region-specific maximum magnitude and ground motion prediction
equation (GMPEs) by worst-case deterministic and classical probabilistic approaches.
Patna, located near Himalayan active seismic region has been subjected to destructive
earthquakes such as 1803 and 1934 Bihar–Nepal earthquakes. Based on the past seismicity
and earthquake damage distribution, linear sources and seismic events have been con-
sidered at radius of about 500 km around Patna district center. Maximum magnitude
(Mmax) has been estimated based on the conventional approaches such as maximum ob-
served magnitude (Mobs max) and/or increment of 0.5, Kijko method and regional rupture
characteristics. Maximum of these three is taken as maximum probable magnitude for each
source. Twenty-seven ground motion prediction equations (GMPEs) are found applicable
for Patna region. Of these, suitable region-specific GMPEs are selected by performing the
‘efficacy test,’ which makes use of log-likelihood. Maximum magnitude and selected
GMPEs are used to estimate PGA and spectral acceleration at 0.2 and 1 s and mapped for
worst-case deterministic approach and 2 and 10 % period of exceedance in 50 years.
Furthermore, seismic hazard results are used to develop the deaggregation plot to quantify
the contribution of seismic sources in terms of magnitude and distance. In this study,
normalized site-specific design spectrum has been developed by dividing the hazard map
into four zones based on the peak ground acceleration values. This site-specific response
spectrum has been compared with recent Sikkim 2011 earthquake and Indian seismic code
IS1893.

Electronic supplementary material The online version of this article (doi:10.1007/s11069-015-1764-0)

contains supplementary material, which is available to authorized users.

& P. Anbazhagan
anbazhagan@civil.iisc.ernet.in
1
Department of Civil Engineering, Indian Institute of Science, Bangalore, India
2
Applied Mechanics Department, S.V. National Institute of Technology, Surat, India

123
 Nat Hazards

Keywords Earthquake  Hazards  Maximum magnitude  GMPE  PGA  SA

1 Introduction

Earthquake has played a devastating role in terms of human casualties and infrastructural
damages. Its unpredictable nature can cause multiple hazards such as ground motion,
ground shaking, site effects, ground displacement, fire, liquefaction, landslide and
Tsunami. In India, urban centers and cities are more susceptible to earthquake hazards due
to high population density, improper planning, poor land use and substandard construction
practices. Many great events including 1905 Kangra, 1934 Bihar–Nepal, 1960 Chilean,
1985 Mexico, 1989 Loma Prieta, 2004 Sumatra and 2011 Sendai earthquakes have
originated from subduction zone. Subduction zones having a seismic gap make the sce-
nario worse, and high-frequency and large-magnitude earthquakes originate from these
zones. Various researchers have highlighted that the Central Seismic Gap in Himalayan
region is one of the most seismically active regions in the world (Khattri 1987), which may
be the source of great events in the future. Even though many enhancements have been
made in earthquake resistant design of structures, still it requires more systematic under-
standing and construction practice to minimize loss of lives and damages in developing
countries. Due to rapid development, population growth and increase in seismicity in India,
there is a demand for the estimation of region-specific seismic hazard parameters such as
site classification, amplification, site effects, liquefaction and landslide for the seismic
microzonation of cities near active Himalaya. During the last few decades, India has
experienced several destructive earthquakes. About 60 % of the country is susceptible to
damaging levels of seismic hazard (NDMA 2010). The first and foremost step toward
reducing earthquake hazard is forecasting the same precisely on a regional scale. Deter-
mining the seismic hazard parameters such as peak ground acceleration and response
spectra is important for infrastructure and building design and also for disaster planning
and management. Such study is mandatory and needs to be carried out for cities close to or
in the highly active seismic region of the world, where the occurrence of a large-magnitude
earthquake is comparatively frequent such as an active region of India.
Seismic hazard estimation is a prime step in the seismic microzonation, where micro-
level variations of seismic hazard and its effects are quantified and mapped. In this study,
the seismic hazard of Patna district has been estimated considering region-specific pa-
rameters such as seismic study area, maximum magnitude and suitable attenuation rela-
tions. The seismic study area has been selected considering past damage distribution, i.e.,
isoseismal map. Maximum magnitude from each linear source has been estimated using
conventional methods of increment of 0.5 in maximum observed magnitude (Mmax obs ) based
on ‘b’ values, Kijko method (Kijko and Sellevoll 1989) and regional rupture characteristics
(Anbazhagan et al. 2014, 2015). The maximum of all these methods has been assigned to
each source. Further best-suited ground motion prediction equation (GMPE) for a region
has been selected from the twenty-seven applicable GMPEs for Patna seismic study area
(SSA). Segmented GMPE ranking has been followed to select GMPEs by carrying out
‘efficacy test,’ which makes use of log-likelihood (LLH) given by Scherbaum et al. (2009)
and Delavaud et al. (2009). The seismic hazard maps have been generated by considering
the worst-case deterministic scenario and classical probabilistic approaches. Probabilistic
seismic hazard analysis (PSHA) values are estimated for 2 and 10 % probability of

123
Nat Hazards

exceedance for 50 years, i.e., for the return period of 2475 and 475 years. Furthermore, the
computed hazard in terms of spectral acceleration at 0.2 and 1 s is mapped, and deag-
gregation plot has been developed to understand the hazard contribution from various
combinations of magnitude and hypocentral distance. In addition to that, site-specific
normalized design spectrum has been developed by deterministic approach and PSHA for 2
and 10 % probability of exceedance in 50 years and compared with Sikkim 2011 earth-
quake and Indian Standard IS-1893 (2002).

2 Study area

Patna, the capital of Bihar, is one of the oldest continuously inhabited places in the world.
Ancient Patna, known as Pataliputra, was the capital of the Magadha Empire under Har-
yanka, Nanda, Mauryan, Sunga, Gupta and Pala. The Patna district center point having
latitude 25.611°N and longitude 85.144°E is situated on the southern bank of the Ganges.
The city also straddles the rivers Sone, Gandak and Punpun. The study area of Patna
belongs to the Seismic zone IV in current Seismic Zonation map of India (IS: 1893 2002),
with zone factor of 0.24.
Various researchers have taken different radii of consideration around the city center to
generate a seismotectonic map of the study area. Anbazhagan et al. (2013b) recommended
that a seismic study area (SSA) radius should be decided by taking into account seismicity
of the region and past damage distribution. In this study, isoseismal map, i.e., damage
distribution map and location of Main Boundary Thrust, Main Central Thrust and Hi-
malayan Frontal Thrust (HFT) has been considered to select SSA. Structural damage of
European Macroseismic (EMS) intensity V and above has been reported beyond 350 km
during Bihar–Nepal earthquake (1934) of moment magnitude (Mw) 8.0 with epicenter at
26.6°N and 86.80°E (Nath et al. 2009). Also, 1833 Nepal earthquake of 7.6 Mw which
damaged the Indo-Gangetic basin (IGB) for more than 450 km was reported (Ambraseys
and Douglas 2004). These two intensity maps are shown in Fig. 1a, b, respectively. Based
on this information, it can be inferred that any earthquake occurring up to 500 km radius
around Patna may cause damage to Patna city. So, the radius of SSA has been selected as
500 km, and seismotectonic parameters were collected and discussed in the next section.
Figure 2 shows the study area of Patna with Himalayan belt and IGB.

3 Seismotectonics of Patna

In order to perform seismic hazard analysis, information about seismic feature such as
faults, shear zones and lineaments with all earthquake events occurred in the SSA is
mandatory (Anbazhagan et al. 2013b). In this study, seismic features observed around
500 km radius of Patna have been collected. SSA covers most part of the IGB, which
extends between 24°300 N latitude to 77°880 E longitude and covers an area of ap-
proximately 250,000 km2. It is identified as one of the most densely populated regions of
India, and around 200 million people reside in the basin (Kumar et al. 2013). The for-
mation of IGB is a consequence of collision between Eurasian and Indian plate, which has
caused the rise of Himalayas since Cenozoic era. The Ganga is the main river of the basin,
which is formed owing to the upliftment of Himalaya after the collision of Indian and
Asian Plates (Dewey and Bird 1970). The Ganga River flows from the Himalayas in the

123
 Nat Hazards

Fig. 1 a Damage distribution map of 1934 Bihar–Nepal earthquake (modified after Kayal 2008), b damage
distribution map of 1833 Bihar–Nepal earthquake (modified after Ambraseys and Douglas 2004)

north to the Bay of Bengal in the north-west. The weathering by river Ganga during its
course of flow results in deposition of sediments in the lower course. The significant
amount of deposition of these sediments in the Indo-Gangetic basin over a long period
resulted in thick fluvial deposit. This deposit consists of different layers of sediments with
an overall thickness of up to several kilometers in many parts of IGB (Sinha et al. 2005;
Anbazhagan et al. 2012; Kumar et al. 2013). Many important cities such as Patna, Meerut,
Lucknow, Kanpur, Aligarh, Gorakhpur, Agra and Jhansi located in different parts of the

123
Nat Hazards

Fig. 2 Study area of Patna along with river system and structural feature based on Dasgupta et al. (1987),
Dasgupta et al. (1993) and GSI (2000). MCT Main Central Thrust, MBT Main Boundary Thrust, HFT
Himalayan Frontal Thrust (modified after Gansser 1964)

IGB are susceptible to earthquake damages due to the proximity to seismically active
Himalayan belt and situated on thick soil deposits. Also, IGB consists of many active
tectonic features such as Munger–Saharsa Ridge Fault, Monghyr–Saharsa Ridge Fault,
East Patna Fault, West Patna Fault, Delhi–Haridwar Ridge, Delhi–Muzaffarabad Ridge
and Faridabad Ridge. The major earthquakes such as 1833 Bihar, 1934 Bihar–Nepal, 1988
Bihar–Nepal and 2011 Delhi Earthquakes have occurred in IGB. The study area is sur-
rounded by several active faults and covered with thick soil deposits as shown in Fig. 2.
Regional seismic records designate that deep regions of North Bihar Plains (area between
24.33°E–27.52°E latitude and 82.33°N–88.29°N longitude) are tectonically active. This
part has documented more than 100 seismic events with 46 events of magnitude larger than
4.5 through the period of 1934–1993 (GSI 2000). In the northeast Patna region, the key
faults are West and East Patna Faults in the East Ganga basin (See Fig. 2). These faults are
acknowledged as transverse faults, and the occurrence of seismic events is due to stimulus
of fluvial dynamics in the North Patna plains transverse faults (Valdiya 1976; Dasgupta
et al. 1987). The East Patna Fault (EPF) is considered to be the most active fault, and its
interaction with Himalayan Frontal Thrust is characterized by a cluster of earthquakes
(Banghar 1991; GSI 2000). This fault is located in the part of the city area. The Gandak
River in the western basin of the Baghmati river basin is flowing along the Gandak Fault
(Mohindra et al. 1992), which is also located near Patna district center (see Fig. 2).
Dasgupta et al. (1993) ventured that all other faults between Motihari and Kishanganj have
the same possibility of seismic hazard as they form a part of related fault system. The study

123
 Nat Hazards

area Patna is near the above-mentioned fault. Till date, the state of Bihar has faced a
number of earthquakes, which include devastating earthquakes like 1934 Bihar–Nepal
Earthquake, having a magnitude of 8.0, which killed nearly 10,700 people. Many earth-
quakes have also occurred during 1833, at Bihar–Nepal border. Various earthquakes like
1927 Madhya Pradesh earthquake, 1985 Rajauli area earthquake and 1988 Udaypur Gary
earthquake have affected Patna in terms of the financial loss and loss of lives. Apart from
the local seismic activity around Patna, the area also located within a radial distance of
approximately 250 km from Main Boundary Thrust (MBT) and the Main Central Thrust
(MCT), where many major earthquakes have been reported and are also considered.
Considering the above seismic aspects of areas in and around Patna, Patna district center
can be considered under a high seismic risk.

4 Seismicity around Patna

Seismicity of the study area is a basic and indispensable concern to be scrutinized in

seismic hazard analysis for assessing seismic risk. The tectonic feature of the study area
has been collected from the Seismotectonic Atlas (SEISAT) published by the Geological
survey of India (GSI) (SEISAT 2000). The well-defined and acknowledged seismic source
has been compiled in published seismotectonic maps as hard copy in 2000 and updated soft
copy in 2010. It contains forty-three maps in 42 sheets of 3° 9 4° size with a scale of
1:10,00,000. As part of the present study for the preparation of seismotectonic map, linear
sources are taken from SEISAT and verified with seismotectonic map prepared by Ko-
lathayar et al. (2012) for whole India and Kumar et al. (2013) for Lucknow region. These
maps were scanned using a high-resolution scanner and digitized for identification of the
linear sources in 500 km radius from Patna city center. No aerial source has been con-
sidered in this study as proper distribution of linear source is available. Demarcation of
MBT and MCT has been done in the seismotectonic map, and all the faults have also been
named.
The earthquake event details within 500 km radius around the Patna district compiled
from several resources like National Earthquake Information Centre (NEIC), International
Seismological Centre, Indian Meteorological Department (IMD), United State Geological
Survey (USGS), Northern California Earthquake Data Centre (NCEDC) and Geological
Survey of India have been used. A total of 2326 events have been collected, which consist
of epicenter coordinates, year, month, date, focal depth and magnitude in different mag-
nitude scale. The collected data were reported in different magnitude scale as body
magnitude (mb), Richter or local magnitude (ML), surface wave magnitude (MS) and
moment magnitude (Mw). In order to attain consistency and homogeneity, all the reported
events were converted into Mw, which is more reliable. The numerous empirical rela-
tionships were developed between these magnitudes by various researchers (Stromeyer
et al. 2004; Castellaro et al. 2006; Scordilis 2006; Bormann et al. 2007; Thingbaijam et al.
2008; Sreevalsa et al. 2011). Scordilis (2006) relation has been used in the present study,
which includes worldwide data.
In order to filter the main event from the dependent event (i.e., clusters, foreshocks and
aftershocks), declustering of earthquake catalogue is required. For achieving the best result
of hazard analysis (Wiemer and Wyss 1994; 1997) and for seismicity rate study (Frankel
1995), declustering of earthquake catalogue is necessary. As far as the phenomenon of the
Poisson model of earthquake occurrence is concerned, i.e., the earthquake took place

123
Nat Hazards

Fig. 3 Seismotectonic map of Patna

randomly, a seismicity model needs to be declustered. For the removal of dependent events
such as aftershocks and beforeshocks, several methods were suggested (Savage 1972;
Gardner and Knopoff, 1974; Reasenberg 1985; Davis and Frohlich 1991; Molchan and
Dmitrieva 1992). In the present study, algorithms developed by Gardner and Knopoff
(1974) and modified by Uhrhammer (1986) are used. As per Stiphout et al. (2010), seis-
micity derived by a static window method (Reasenberg 1985) does not go behind Poisson
distribution. Out of 2325 events, 54 % were found to be depended events; a total 1262
events have been acknowledged as main shock. For further analysis, Mw C 4 (i.e., 818
events) were considered, as smaller magnitude would not generate considerable ground
motions for building damage. The complete catalogue contains 444 events with Mw less
than 4. In order to develop seismotectonic map, declustered earthquake events are su-
perimposed with the source map as shown in Fig. 3. It has been seen from Fig. 3 that
events are more densely located near MBT and MCT as compared to other areas. The study
area is divided into two regions, i.e., Region I (belonging to MBT and MCT) and Region II,
depending upon the events allocation. List of numbers of earthquake events with Mw equal
to or greater than 4 is given in Table 1 for both Regions I and II. These regions are divided
using trapezoid as shown in Fig. 3, Region I belongs to events inside the trapezoid, and
Region II belongs to events outside the trapezoid. Both the regions are evaluated separately
for the seismic hazard estimation.

5 Data completeness and G–R recurrence relationship

Assessments of the seismic parameters are the basic requirement in the determination of
the seismic hazard map of a region. In order to predict the ground motion due to forth-
coming earthquakes, it is obligatory to estimate these parameters. These parameters in-
clude the ‘a’ and ‘b’ parameters of Gutenberg–Richter (G–R) recurrence relationship
(Gutenberg and Richter 1956). The earthquake catalogue presents the feature of seismicity

123
 Nat Hazards

Table 1 Summary of earth-

S. No. Earthquake magnitude Number of events
quake events having Mw greater
(Mw) range
than or equal to 4 Region I Region II

1. 4 B Mw \ 4.5 66 39
2. 4.5 B Mw \ 5 214 129
3. 5 B Mw \ 5.5 140 88
4. 5.5 B Mw \ 6 57 33
5. 6 B Mw \ 6.5 18 16
6. 6.5 B Mw \ 7 8 5
7. Mw C 7 4 1

(a) 1000 M<4 4<M<4.99

5<M<5.99 6<M<6.99

M>7 Total
Number of Earthquakes

100

10

0.1
1829 1849 1869 1889 1909 1929 1949 1969 1989 2010
1800 1830 1850 1870 1890 1910 1930 1950 1970 1990
Year

(b)1000 M<4 4<M<4.99

5<M<5.99 6<M<6.99
M>7 Total
Number of Earthquakes

100

10

0.1
1829 1849 1869 1889 1909 1929 1949 1969 1989 2013
1800 1830 1850 1870 1890 1910 1930 1950 1970 1990
Years

Fig. 4 a Histogram of earthquake data for Region I, b histogram of earthquake data for Region II

of a region, which is the backbone for seismic hazard analysis and involves in enlarging
seismogenic zoning circumstances in combination with seismotectonic and geological
information (Lai et al. 2009). To determine the seismicity characteristic of a region,

123
Nat Hazards

complete catalogue needs to be analyzed. Figure 4a, b shows the histogram of the Regions
I and II, respectively, in SSA. Based on the observation from Fig. 4, one can predict that
instrumented data for study area Patna might be recorded after Agra observatory laboratory
shifted to Delhi and the number of observation increased to 15 in 1960 for Region II. As far
as Region I is concerned, it has high instrumentally recorded seismic data after the seis-
mology observatory started in Shimla with Omori Ewing seismograph in 1905. It can also
be noted from these figures that no historic data are present with Mw B 3.
The seismicity parameters around the Patna site for Regions I and II can be quantified
by the standard Gutenberg–Richter (G–R) recurrence relationship (Gutenberg and Richter
1956). It hypothesizes the existence of an exponential correlation between the mean
annual rate of exceedance of an earthquake of specified magnitude and the magnitude for
the period of completeness. The seismic recurrence rate can be assessed correctly if the
collected data of the earthquake events are complete. Therefore, the composed data of
Patna for the regions have to be scrutinized for its completeness. Stepp (1972) proposed
a method to evaluate the duration of completeness of homogenized earthquake data by
distributing it into small bins, seeing the variance of each bin as the same. The existence
of earthquakes can be demonstrated as Poisson’s distribution for the evaluation of ef-
fectual variance. Anbazhagan et al. (2010) has described the detailed procedure for
completeness analysis as per Stepp (1972). The total compiled earthquake data in Region
I cover a time period from 1816 to 2010 (or 194 years), while the data for Region II
cover a time period from 1823 to 2013 (or 190 years). Both earthquake catalogues were
examined independently to check the data completeness for each region. It has been
pffiffiffiffi
comprehended that standard variation is found to be approximately parallel to 1= T for
the last 80 years for the earthquakes having moment magnitude less than 5.0 and for
110 years for higher magnitude. For Region II, the earthquake having moment magni-
tude less than 5.0 is complete for the last 70 years, whereas the higher magnitudes are
complete for 110 years.
Maximum magnitude and its recurrence in the region depend on recurrence relation of
regional seismicity data. This relation can also help to quantify uncertainty in the earth-
quake size of the region and/or every seismic source (Gutenberg and Richter 1956). The
relation assumes exponential distribution of magnitude on every source and is also useful
to estimate the minimum and maximum earthquake for any region. The recurrence law is
defined by Gutenberg and Richter (1956) as given by the following equation
logðN Þ ¼ a  bM ð1Þ
where N resembles the number of earthquakes of magnitude M, ‘a’ and ‘b’ are positive real
constants in which ‘a’ denotes the seismic activity (log number of events with M = 0) and
‘b’ describes the relative abundance of large to small shocks (Gutenberg and Richter 1956).
After checking the completeness analysis of catalogue for both the regions, ‘a’ and ‘b’
parameters have been calculated. After determining the frequency of exceedance versus
magnitude value, Gutenberg–Richter recurrence law for the zone can also be estimated.
Figure 5 shows the G–R recurrence law for the Regions I and II with a correlation coef-
ficient of 0.99 and 0.98, respectively. The ‘b’ value for the Region I is 0.91 and for Region
II is 1.01. These values are compared with that of the other researchers and shown in
Table 2. This ‘b’ value will be further used for the determination of maximum magnitude
in the next section. The ‘a’ value for Regions I and II for the present study is given as 5.32
and 4.98, respectively. These values are comparable with NDMA (2010) and Kumar et al.
(2013).

123
 Nat Hazards

Region I
1.5

Log (Cumulative events per year)

Region II

1 Linear (Region I)

Linear (Region II)

0.5

Log (N) = 5.32 -0.91Mw

-0.5 R² = 0.9979

-1 Log (N) = 4.98-1.01 Mw

R² = 0.9834
-1.5

-2
4 4.5 5 5.5 6 6.5 7 7.5
Moment Magnitude (Mw)

Fig. 5 Gutenberg–Richter relation for Regions I and II

Table 2 Comparison of ‘b’ pa-

Region I Region II
rameter of the present study with
previous data
0.91 (present work) 1.01 (present work)
0.86 (Kumar et al. 2013) 0.80 (Kumar et al. 2013)
0.73 (NDMA 2010) 0.81 (NDMA 2010)
1.0 (Sreevalsa et al. 2011) 0.85 (Sreevalsa et al. 2011)
0.80 (Mahajan et al. 2010)
0.65 (Kumar 2012)

6 Maximum probable earthquake magnitude (Mmax) and focal depth

The maximum probable earthquake magnitude is defined as the upper limit of earthquake
magnitude for a given region and is synonymous with the magnitude of the largest possible
earthquake (EERI Committee on Seismic Risk 1984; Working Group on California
Earthquake Probabilities 1995). It assumes a sharp cutoff magnitude at a maximum
magnitude, Mmax, so that, by definition, no earthquakes are to be expected with magnitude
exceeding Mmax (Joshi and Sharma 2008). The maximum magnitude regarding each fault
has been calculated using three methods mentioned below:
1. Kijko and Sellevoll (1989) have proposed a method to estimate maximum magnitude
considering doubly truncated Gutenberg–Richter relation. This method is only valid
when b for the region is known (CASE I; Kijko and Sellevoll 1989).
E1 ðn2 Þ  E1 ðn1 Þ
Mmax ¼ mobs
max þ þ mmin expðnÞ ð2Þ
bexpðn2 Þ
where Mmax is the largest possible earthquake magnitude, mobs
max is the maximum observed
magnitude on each fault, n is the total earthquakes above magnitude of completeness

123
Nat Hazards

(mmin), n1 ¼ n=f1  exp½bðmmax  mmin Þg, n2 = n1{ exp [ -b(mmax - mmin)]}, E1-
()denotes an exponential integration function, which can be estimated as
2
E1 ðzÞ ¼ zðzz2þa1 zþa2
þb1 zþb2 Þ expðzÞ, where a1 = 2.334733, a2 = 0.250621, b1 = -3.330657
and b2 = -1.681534 (Abramowitz and Stegun 1970), mobs max for each fault and mmin, the
value of Mmax has been estimated using Eq. 2. For the estimation of Mmax, mmin is
calculated as per the methods described by Woessner and Stefan (2005). Detailed de-
scription of calculating mmin and ‘a’ and ‘b’ value for Patna region has been discussed in
Anbazhagan et al. (2015). It has been observed that magnitude of completeness varies
from 1.7 to 5 for Region I, but for Region II, it varies from 1.7 to 4.9. So taking that into
account, 4.5 Mw has been taken as the minimum magnitude for further analysis for
Regions I and II. This method for calculating Mmax has been widely used by various
researchers worldwide as well as in India. In the present study area, regional values are
used as per Kijko and Sellevoll (1989).
2. Mmax has been also estimated by adding a constant value of 0.3 if the mobs max is less than
5 (Mw) and add 0.5 for mobs obs
max greater than 5 to the mmax value of each fault similar to
NDMA (2010).
3. Mmax is also estimated using regional rupture characteristics by considering the
maximum magnitude observed and possible seismic source of SSA. The whole
procedure to find region-specific rupture characteristic was presented in Anbazhagan
et al. (2013b, 2014). The same procedure is followed for determining Mmax for each
seismic source (see Fig. 3). Subsurface rupture length (RLD) of each seismic source
has been estimated by using well-accepted correlation between RLD and Mw by Wells
and Coppersmith (1994) from the maximum observed magnitude of each source.
Percentage fault rupture (PFR) which is the ratio of subsurface rupture length (RLD) to
total fault length (TFL) is expressed in percentage. As per Fig. 6, the plotting of PFR
against TFL shows that PFR follows a unique trend for interplate region. Possible
worst scenario PFR is established by considering minimum, maximum and average
PFR in four length bins as shown in Table 3. For each length bin, PFR for worst
scenario earthquake has been taken as five times the average PFR, which is also more

25
For 500 km radius
Power (For 500 km radius)
Percentage Fault Rupture (PFR)

20

15

10

0
0 100 200 300 400 500 600 700 800 900
Total Fault Length, TFL (km)

Fig. 6 Regional rupture characters for Patna

123
 Nat Hazards

Table 3 Regional rupture character for various distance bins

Length bins PFR (% TFL) PFR (% TFL) Ratio of PFR
for worst for WS to
Maximum Minimum Average scenario (WS) maximum PFR

\100 25 4.71 15.06 33 1.32

100–300 22.2 2.32 8.74 32 1.44
300–600 3 1.13 1.56 5.5 1.83
[600 22.55 7.67 15.11 25 1.11

than the maximum reported PFR. PFR for the worst scenario (see Table 3) is taken as
the regional rupture character of the seismic study area. The subsurface rupture length
is calculated based on the length of each source, which is further used to estimate the
Mmax of each source using well-established Wells and Coppersmith (1994) relation-
ship. The whole procedure is explained in Anbazhagan et al. (2015).

The absolute Mmax for each source has been calculated from the above-mentioned three
approaches. Table 4 gives the Mmax value from the seismic source having an estimated
maximum magnitude greater than 6.5. So the final Mmax for each source was taken as
maximum from these three approaches and given in the last column of Table 4.

6.1 Focal depth of earthquake

Determination of focal depth of future earthquake is an indispensable and complicated job

for seismic hazard analysis. Comprehension about the depth division of the diffuse seis-
micity (i.e., derived from the seismology database) should be included in the seismic
hazard analysis (Anbazhagan et al. 2013b). Most of the studies regarding seismic hazard
analysis incorporate the lowest focal depth or depth considering the minor earthquakes
(Anbazhagan et al. 2013b). In the present study, whole catalogue has been analyzed for the
determination of the appropriate focal depth of future earthquake. Hence, focal depth is
analyzed with magnitude and epicentral distance for Patna SSA. Figure 7 shows the plot of
depth versus epicenter distance of the events having a moment magnitude between 5 and 6
and greater than 6 for the whole Patna region. Depth of earthquakes having a moment
magnitude less than 5 is not considered because Mmax of all sources are 5 and above. It has
been seen from Fig. 7 that for a moment magnitude between 5 and 6, focal depth varies
from 5 to 75 km, whereas for moment magnitude greater than 6, it varies from 10 to 75 km
for the 500 km as epicentral distance. Considering the worst-case scenario, focal depth of
10 km is adopted for epicenter distance up to 200 km (marked in Fig. 7) for all the
magnitude and focal depth of 5 km is taken for Mw \ 6 and 10 km for Mw [ 6 beyond an
epicenter distance of 200 km as per Fig. 7.

7 Ground motion prediction equation (GMPE)

Ground shaking through an earthquake is accountable for the structural damage and ground
failures either within the epicentral region or at far distances. The region-specific GMPE is
an important component in the seismic hazard analysis for both seismic macro- and

123
Nat Hazards

Table 4 Mmax values from three approaches and assigned Mmax value for each source
Seismic Observed Regional rupture characteristics By incremental Kijko and Mmax taken
source magnitude a
value Sellevoll for hazard
(Mw) TFL (km) RLD (% TFL) Mmax (1989) analysis

MBT 8 844.92 211.23 8.1 8.5 8 8.5

MCT 7 638.78 159.7 7.9 7.5 7 7.9
S01 5.4 135.37 43.32 6.9 5.9 5.4 6.9
S02 5.4 206.58 66.1 7.2 5.9 5.4 7.2
S03 5.2 183.13 58.6 7.1 5.7 5.2 7.1
S05 5.2 142.64 45.65 6.9 5.7 5.2 6.9
S06 5.8 128.34 41.07 6.9 6.3 5.9 6.9
S07 5.8 58.19 19.2 6.3 6.3 6.6 6.6
S08 4.5 202.28 64.73 7.2 4.8 4.5 7.2
S09 5.6 128.31 41.06 6.9 6.1 5.6 6.9
S10 6 118.41 37.89 6.8 6.5 6 6.8
S104 5.4 101.82 32.58 6.7 5.9 5.5 6.7
S105 6.8 166.32 53.22 7.1 7.3 6.8 7.3
S106 6.8 213.24 68.24 7.2 7.3 6.8 7.3
S109 6.7 163.44 52.3 7 7.2 6.7 7.2
S11 5.6 181.92 58.22 7.1 6.1 5.6 7.1
S111 6 53.94 17.8 6.3 6.5 7 7
S112 6.2 78.44 25.89 6.5 6.7 6.3 6.7
S12 6.4 125.31 40.1 6.9 6.9 6.4 6.9
S125 5.8 77.77 25.66 6.5 6.3 6.6 6.6
S128 6.4 244.38 78.2 7.3 6.9 6.4 7.3
S13 5.2 126.41 40.45 6.9 5.7 5.2 6.9
S130 6.8 203.22 65.03 7.2 7.3 6.8 7.3
S131 5.3 90.58 29.89 6.6 5.8 5.4 6.6
S139 5.6 134.34 42.99 6.9 6.1 5.6 6.9
S14 5.1 124.58 39.87 6.8 5.6 5.2 6.8
S141 6.2 108.66 34.77 6.7 6.7 6.3 6.7
S146 5.3 127.89 40.92 6.9 5.8 5.3 6.9
S15 5.5 161.13 51.56 7 6 5.6 7
S157 5.8 117.45 37.58 6.8 6.3 5.9 6.8
S158 5.5 84.81 27.99 6.6 6 5.7 6.6
S16 5.5 110.65 35.41 6.8 6 5.6 6.8
S161 6.2 208.22 66.63 7.2 6.7 6.2 7.2
S17 5.1 121.48 38.87 6.8 5.6 5.2 6.8
S172 5.2 113.76 36.4 6.8 5.7 5.2 6.8
S18 5.5 121.87 39 6.8 6 6.1 6.8
S19 5.5 186.75 59.76 7.1 6 5.5 7.1
S20 5.5 139.32 44.58 6.9 6 5.5 6.9
S25 4.5 280.49 89.76 7.4 4.8 4.5 7.4
S27 5.4 82.36 27.18 6.6 5.9 5.6 6.6
S28 5.2 111.79 35.77 6.8 5.7 5.2 6.8

123
 Nat Hazards

Table 4 continued

Seismic Observed Regional rupture characteristics By incremental Kijko and Mmax taken
source magnitude a
value Sellevoll for hazard
(Mw) TFL (km) RLD (% TFL) Mmax (1989) analysis

S31 5.2 90.25 29.78 6.6 5.7 5.3 6.6

S32 5.2 87.61 28.91 6.6 5.7 5.7 6.6
S33 5.9 86.13 28.42 6.6 6.4 6.2 6.6
S34 5.4 197.04 63.05 7.2 5.9 5.4 7.2
S35 4.5 115.3 36.89 6.8 4.8 4.5 6.8
S36 5.5 135.63 43.4 6.9 6 5.5 6.9
S39 5.2 81.66 26.95 6.6 5.7 5.3 6.6
S43 4.6 86.35 28.49 6.6 4.9 4.7 6.6
S44 5.3 170.39 54.52 7.1 5.8 5.3 7.1
S45 6.8 217.61 69.63 7.3 7.3 6.8 7.3
S47 4.8 27.72 9.15 5.8 5.1 8 8
S48 5.3 84.09 27.75 6.6 5.8 5.4 6.6
S49 5.7 108.2 34.62 6.7 6.2 5.8 6.7
S50 6.1 115.38 36.92 6.8 6.6 6.1 6.8
S52 5.9 82.28 27.15 6.6 6.4 6 6.6
S53 5.9 104.4 33.41 6.7 6.4 6 6.7
S54 6.2 132.04 42.25 6.9 6.7 6.2 6.9
S57 6.7 221.6 70.91 7.3 7.2 6.7 7.3
S58 5.4 204.19 65.34 7.2 5.9 5.4 7.2
S59 6.8 174.9 55.97 7.1 7.3 6.8 7.3
S60 5.4 210.41 67.33 7.2 5.9 5.4 7.2
S62 7 220.63 70.6 7.3 7.5 7 7.5
S63 6.3 80.58 26.59 6.6 6.8 6.5 6.8
S64 4.9 82.04 27.07 6.6 5.2 5 6.6
S69 6.2 121.58 38.9 6.8 6.7 6.3 6.8
S73 6.3 101.99 32.64 6.7 6.8 6.5 6.8
S76 6.1 74.98 24.74 6.5 6.6 6.1 6.6
S77 4.9 83.29 27.49 6.6 5.2 5.1 6.6
S78 6.1 58.28 19.23 6.3 6.6 7.2 7.2
S82 4.9 84.24 27.8 6.6 5.2 5 6.6
S83 5.8 171.1 54.75 7.1 6.3 5.8 7.1
S90 5.8 61.81 20.4 6.4 6.3 6.6 6.6
a
Calculated as per Table 3

microzonation. Developed countries are working on the next generation of ground motion
attenuation (NGA) for the better prediction of ground shaking due to any future earthquake
events (Campbell and Bozorgnia 2006; Kaklamanos and Baise 2011). But the limited
number of GMPEs is available for seismic hazard estimation, both in bedrock as well as at
surface by accounting the local site effects in India and other parts of the world (Atkinson
and Boore 2006; NDMA 2010). An indispensable step in hazard analysis for any region is
the selection of appropriate GMPE for forecasting the ground shaking.

123
Nat Hazards

Fig. 7 Distribution of focal Epicentral Distance (km)

depth of earthquake with 0 100 200 300 400 500
epicentral distance 0

10

20

Depth (km)
30

40

50
5<Mw<6
60
Mw>6
70

80

90

Various researchers have analyzed the attenuation characteristics of the Himalayan

region based on the available data. Region-specific GMPEs developed by Singh et al.
(1996), Sharma (1998), Nath et al. (2005, 2009), Das et al. (2006), Sharma and Bungum
(2006), Baruah et al. (2009), Sharma et al. (2009), Gupta (2010), NDMA (2010) and
Anbazhagan et al. (2013a) are based on recorded as well as simulated earthquake data.
Table EM1 (submitted as electronic material) summarizes various GMPEs developed in
different parts of the Himalayan belt along with the range of magnitude and distance to
which each GMPE is valid for. In addition to these GMPEs, there are several GMPEs
developed for similar tectonic conditions, which can also be applicable to the Himalayan
region. GMPEs developed elsewhere and applicable to Himalayan regions are Abrahamson
and Litehiser (1989), Youngs et al. (1997), Campbell (1997), Spudich et al. (1999),
Atkinson and Boore (2003), Takahashi et al. (2004), Ambraseys et al. (2005), Kanno et al.
(2006), Zhao et al. (2006), Campbell and Bozorgnia (2008), Idriss (2008), Boore and
Atkinson (2008), Abrahamson and Silva (2007), Aghabarati and Tehranizadeh (2009), Lin
and Lee (2008) and Akkar and Bommer (2010). Table EM2 (submitted as electronic
material) shows summary of 16 applicable GMPEs for Himalayan region. The list of
GMPE applicable for the study region along with abbreviations is given in Table 5. Fig-
ure 8 shows the plot of region-specific available GMPEs and applicable GMPEs for Mw 6.8
and hypocenter distance of 10–500 km. From Fig. 8, it may be difficult to identify the
appropriate region-specific GMPEs for the hazard analysis by comparison.
The seismic hazard assessment requires appropriate selection of GMPEs, which will
help to forecast representative level of ground shaking (Bommer et al. 2010). The selected
GMPE should be capable of apprehending the essence of ground motion, i.e., earthquake
source, path and site attributes at the same time. GMPE developed from the past four
decades has shown rather consistency in the related inconsistency and epistemic uncer-
tainty, notwithstanding the increasing complexities (Strasser et al. 2009; Douglas and
Mohais 2009; Douglas 2010; Nath and Thingbaijam 2011; Anbazhagan et al. 2013b). This
gives rise to the obligation of the selection and ranking of GMPEs (Bommer et al. 2005;
Cotton et al. 2006; Sabetta et al. 2005; Scherbaum et al. 2004, 2005; Hintersberger et al.
2007; Nath and Thingbaijam 2011) and which results in usage of multiple GMPEs in a
logic tree framework for hazard analysis. Many researchers in India generally use two or
three GMPEs for predicting the hazard values without any physical and mathematical

123
 Nat Hazards

Table 5 Available GMPEs with their abbreviations considered for the seismic study area
S. No. Ground motion prediction equation (GMPE) Abbreviation of the equations

1. Singh et al. (1996) SI-96

2. Sharma (1998) SH-98
3. Nath et al. (2005) NATH-05
4. Das et al. (2006) DAS-06
5. Sharma and Bungum (2006) SHBU-06
6. Baruah et al. (2009) BA-09
7. Nath et al. (2009) NATH-09
8. Sharma et al. (2009) SH-09
9. Gupta (2010) GT-10
10. National Disaster Management Authority (2010) NDMA-10
11. Anbazhagan et al. (2013a, b) ANBU-13
12. Abrahamson and Litehiser (1989) ABLI-89
13. Youngs et al. (1997) YONG-97
14. Campbell (1997) CAMP-97
15. Spudich et al. (1999) SPUD-99
16. Atkinson and Boore (2003) ATKB-03
17. Takahashi et al. (2004) TAKA-04
18. Ambraseys et al. (2005) AMB-05
19. Kanno et al. (2006) KANO-06
20. Zhao et al. (2006) ZHAO-06
21. Campbell and Bozorgnia (2008) CABO-08
22. Idriss (2008) IDRS-08
23. Boore and Atkinson (2008) BOAT-08
24. Abrahamson and Silva (2007) ABSI-08
25. Aghabarati and Tehranizadeh (2009) AGTE-08-09
26. Lin and Lee (2008) LILE-08
27. Akkar and Bommer (2010) AKBO-10

reasoning; however, considering 2–3 GMPE randomly and comparing with the observed
value may give inconsistent results because of the absence of comprehensive procedure
(Delavaud et al. 2012). Hence, for the present study area, the best-suited GMPE is selected
for hazard analysis considering past earthquake data.
The best-suited GMPE has been selected considering the criteria proposed by Bommer
et al. (2010) and by performing the efficacy test recommended by Scherbaum et al. (2009)
and Delavaud et al. (2009). The determination of order of ranking of GMPEs is based on
the observed earthquakes in a particular region. In the present study, the information-
theoretic approach recommended by Scherbaum et al. (2009) has been used. The efficacy
test makes use of average sample log-likelihood (LLH) for the ranking purpose of the
available GMPE of a particular SSA. The efficacy test using average LLH has been
performed successfully by Delavaud et al. (2009) and applied to India by Nath and
Thingbaijam (2011). Hence, for the present study, efficacy test has been carried out by
considering macroseismic intensity map of 1833 and 1934 Bihar–Nepal earthquake (shown
in Fig. 1a, b) and PGA-European Macroseismic Scale (EMS, Grünthal 1998) relation

123
Nat Hazards

Idriss(2008) Youngs et al(1997)

Lin and Lee(2008) Spudich et al (1999)
Campbell (1997) Atkinson and Boore (2003)
Ambraseys et al (2005) Zhao et al (2006)
Campbell and Bozorgnia (2008) Kanno et al (2006)
Boore and Atkinson (2008) Abrahamson and Litehiser (1989)
Abrahamson and Silva (2008) Takahashi et. al.
Singh et al (1996) Akkar and Bommer (2010)
1 Anbazhagan et al (2013) NDMA (2010)
Peak Ground Acceleration (g)

Sharma et al (2009) Das et al. (2006)

Nath et al. (2009) Nath et al. (2005)
Sharma and Bahuguna (2006) Gupta (2010)
Sharma (1998) Abrahamson and Litehiser (1989)

0.1

0.01

0.001
0 50 100 150 200 250 300 350 400 450 500
Hypocentral Distance (km)

Fig. 8 Comparison of ground motion prediction equations applicable to Himalayan region for earthquake
moment magnitude of 6.8

proposed by Nath and Thingbaijam (2011) for Indian crustal earthquakes. The LLH is
calculated using the equation given by Delavaud et al. (2009) and given as Eq. 3.

1X N
LLHðg; xÞ ¼  log2 ðgðxi ÞÞ ð3Þ
n i¼1

where x ¼ fxi g; i ¼ 1; . . .:N are the empirical data and g(xi) is the likelihood that model
g has produced for the observation xi. In this case of GMPE selection, g is the probability
density function given by a GMPE to predict the observation produced by an earthquake
with magnitude M at a site i that is located at a distance R from the source (Delavaud et al.
2012).
The PGA variation with distance of all the applicable GMPEs has been discussed (see
Fig. 8). Based on the trend of variation, the hypocentral distance is divided into three
length bins as 0–100, 100–300 and 300–500 km. For the hazard analysis, the ranking of
GMPE has been considered for all the three bins. As macroseismic intensity map for 1934
earthquake is available up to 300 km, so for more than 300 km, 1833 earthquake has been
considered for ranking of GMPEs. Some of the GMPEs such as Singh et al. (1996), Sharma
(1998), Das et al. (2006), Baruah et al. (2009), Sharma et al. (2009) and Gupta (2010) are
not used for efficacy test as the isoseismal map used in the present study area has a
magnitude greater than 7.6 Mw. The LLH values along with the ranking of GMPEs are
given in Table 6. The EMS values are used to estimate the LLH values and data support
index (DSI), which are further used to rank the GMPEs. LLH values are not a measure of
closeness, but a measure of the distance between a model and the data-generating process
(Delavaud et al. 2012). Delavaud et al. (2012) have given a data support index (DSI) to
know the percentage by which the weight on a model is increased or decreased through

123
 Table 6 Segmented ranking of GMPEs for Patna region
Sl. GMPEs 0–100 100–300 300–500
No.

123
1833 1934 1833 1934 1833

LLH DSI R W LLH DSI R W LLH DSI R W LLH DSI R W LLH DSI R W

1. ABLI- 16.68 -99.95 15 – 28.33 -100 11 – NA NA NA

89_Hort
2. ABLI- 17.77 -99.98 16 – 26.38 -100 10 – NA NA NA
89_Vert
3. YONG-97 26.16 -100 22 – 53.45 -100 15 – 8.06 -64.67 5 – 17.87 -99.75 6 – 6.50 -76.14 2 –
4. CAMP-97 8.47 -85.22 12 – 13.28 -97.79 5 – NA NA NA
5. SPUD-99 19.11 -99.99 17 – NA NA NA NA
6. TAKA-04 9.95 -94.71 10 – 15.91 -99.64 7 – 11.16 -95.89 8 – 15.87 -99.01 6 – NA
7. AMB-05 22.96 -100.00 20 – 70.48 -100 16 – NA NA NA
8. NATH-05 20.71 -100.00 18 – 39.57 -100 12 – NA NA NA
9. KANO-06 5.45 19.29 3 0.12 10.91 -88.56 4 – 4.58 295.22 1 0.32 6.98 370.44 2 0.45 NA
10. ZHAO-06 9.95 -94.71 9 – 15.91 -99.64 6 – 11.16 -95.89 7 – 15.87 -99.01 5 – NA
11. SHBU-06 5.75 -2.93 4 – NA 26.39 -100 10 – NA NA
12. IDRS-08 14.84 -99.82 14 – 25.58 -100 9 – 29.29 -100 11 – 45.51 -100 9 – NA
13. BOAT-08 11.13 -97.66 13 – 19.34 -99.97 8 – 5.85 63.24 4 0.16 10.00 -41.68 4 – 10.32 -98.31 4 –
14. ABSI-08 9.31 -91.78 8 – 44.92 -100 14 – 16.12 -99.87 9 – NA NA
15. CABO-08 24.56 -100.00 21 – 72.44 -100 18 – 51.43 -100 12 – 149.01 -100 10 – NA
16. LILE-08 22.72 -100.00 19 – 42.91 -100 13 – 10.62 -94.01 6 – 22.04 -99.99 8 – 8.55 -94.23 3 –
17. AGTE-08- 7.69 -74.76 7 – NA NA NA NA
09_Vert
18. AGTE-08- 6.95 -57.77 6 – NA NA NA NA
09_Hort
19. NATH-09 5.79 -5.24 5 – 9.87 –76.50 3 – NA NA NA
Nat Hazards
 Table 6 continued

Sl. GMPEs 0–100 100–300 300–500

No.
1833 1934 1833 1934 1833
Nat Hazards

LLH DSI R W LLH DSI R W LLH DSI R W LLH DSI R W LLH DSI R W

20. AKBO-10 10.15 -95.41 11 – NA NA NA NA

21. NDMA-10 2.81 645.32 2 0.35 6.09 222.97 2 0.26 4.84 229.98 3 0.22 9.74 -30.27 3 – 2.55 268.69 1 1
22. ANBU-13 2.23 1011.90 1 0.53 4.04 1241.71 1 0.74 4.70 261.84 2 0.30 6.90 399.06 1 0.55 NA

R ranking, W weight
Positive DSI values for different GMPES are marked as bold

123
 Nat Hazards

data. DSI of an equation (given as Eq. 5) shows the percentage increase or decrease of
weight of a model with respect to its state of non-informativeness (Delavaud et al. 2012).
2LLHðgi ;xÞ
wi ¼ Pn LLHðgi ;xÞ
ð4Þ
k¼1 2

wi  wunif
DSIi ¼ 100 ð5Þ
wunif
where wunif ¼ 1=M and M are the number of models used for the calculation of LLH value.
Segmented-based ranking of GMPEs has been attempted in order to avoid over/un-
derestimation of the PGA of shorter and longer distances. To select the best-suited GMPEs
for each region past earthquake location, a DSI criterion has been used. For each distance
segment, positive DSI values are identified and ranked based on maximum to minimum
values. A maximum positive DSI value is considered first rank, and minimum is considered
as lowest rank. Positive DSI values for different GMPEs for Patna SSA are marked as bold
in Table 6. This study shows that GMPE developed by ANBU-13, NDMA-10 and KANO-
06 is best suitable up to 100 km, and ANBU-13, NDMA-10, KANO-06 and BOAT-10 are
best suitable for 100–300 km distances. GMPE given by NDMA-10 is only suitable GMPE
for distance above 300 km and up to 500 km. These LLH values are further used to
evaluate the LLH-based weight factor as per Delavaud et al. (2012) as it infers to what
extent the data increase or decrease the weight of model with respect to the non-infor-
mativeness (see Table 6). In the present study, DSI is directly calculated using LLH and
weight is calculated later from only those GMPE having positive DSI. The weight factor
corresponds to particular GMPE for different segments are further used in evaluating the
hazard of Patna SSA. Seismic hazard values in terms of PGA and SA can be calculated
considering these equations for each seismic source.

8 Hazard maps for Patna region

To derive the hazard value, deterministic seismic hazard analysis (DSHA) and probabilistic
hazard seismic hazard (PSHA) analysis have been widely practiced. In this study, both
PSHA and DSHA have been used to estimate the peak ground acceleration (PGA) and
spectral acceleration (SA) based on past seismicity and future maximum magnitude. The
detailed procedures for both the methods are given in Anbazhagan et al. (2009) and Kumar
et al.(2013). These hazard maps are most widely used for significant structures and seismic
disaster planning and mitigation. GMPEs have been selected, and weights are calculated
based on regional data. The weight factor was 0.53, 0.35 and 0.12 for ANBU-13, NDMA-
10 and KANO-06, respectively, up to 100 km, and 0.32, 0.30, 0.22 and 0.16 for GMPEs of
KANO-06, ANBU-13, NDMA-10 and BOAT-08 for segmented hypocentral distance of
100–300 km and a factor of 1 for NDMA-10 for more than 300 km hypocentral distance.
Separate MATLAB code has been generated to determine PGA deterministically and
probabilistically by considering magnitude, source-to-site distance and site condition.
These codes have been validated with the results of EM-1110 (1999). The whole Patna
district’s SSA has been divided into 2500 grids of size 0.0228 9 0.0148 along the longitude
and latitude, respectively. Following procedure has been used to determine the PGA value
for each of 2500 grids. Kriging interpolation technique has been used for the estimation of
intermediate values of PGA for the development of the seismic hazard map.

123
Nat Hazards

8.1 Deterministic seismic hazard analysis (DSHA)

Usually in DSHA, one or more earthquakes are identified by magnitude and location with
respect to site. In this approach, the earthquake is assumed to occur in the portion of the
site closest to the site. In the present study, the hazard map of Patna district center has been
prepared considering the entire seismic source with corresponding maximum magnitudes
and systematically selected GMPEs having different weight factors. In total, 178 seismic
sources have been found, which have experienced an earthquake magnitude of 4 and within
500 km radial distance around Patna (shown in Fig. 3). A MATLAB code has been de-
veloped for the DSHA, which has also been verified with the manual calculation of Patna
district center. The minimum hypocentral distance has been estimated from the center of
each grid to each fault using the code. The Patna SSA has been divided into 2500 grids of
size 0.0228 9 0.0148 along the longitude and latitude, respectively. The peak ground
acceleration at each grid has been estimated considering the maximum magnitude and
GMPE. The maximum PGA from the entire 178 seismic source will be assigned as the
PGA for that grid. The similar technique has been adopted for all the 2500 grids for the
development of new seismic hazard map of Patna. Kriging interpolation technique has
been used for the estimation of intermediate values of PGA for the development of the
seismic hazard map. Figure 9 shows the worst-case map of the Patna city center. The PGA
variation has been found to be 0.14 g in the central part of Patna district, whereas it
increased to 0.5 g in northwestern periphery. The high hazard values are resulted due to
location of East Patna Fault and West Patna Faults within the city. In addition to that,
spectral acceleration (SA) maps at 0.2 and 1 s have been developed and shown in Figs. 10
and 11. The weight factor used for SA calculation is similar to PGA calculations. This is
because of limited data available in determining the weight factor for the SA hazard map at
0.2 and 1 s, but in future, this may be taken into consideration as these weight factors have
different impact at different periods. Areas like Nehru Nagar, Patna High Court, Kothia,
Pataupura Colony, Makhdhumpur and surrounding areas are less prone to earthquake-

Fig. 9 Peak ground acceleration (PGA) map of Patna urban center

123
 Nat Hazards

Fig. 10 Spectral acceleration at 0.2 s map of Patna urban center

Fig. 11 Spectral acceleration at 1 s map of Patna urban center

induced ground shaking; however, places near and surrounding Hanuman Nagar, Sad-
hanpuri, Vigrahpur, Loknayak Jaiprakash Airport, Bihar Chak, Patrakar Nagar are more
susceptible to earthquake shaking. Since, these levels of ground shaking are evaluated at
bedrock level, and no changes in PGA contours along the alignment of river Ganga can be
seen here. The large variation in PGA value within the city may be due to Monghyr–
Saharsa Ridge (see Fig. 2), Gandak Fault (S161), East Patna Fault (S59) and West Patna

123
Nat Hazards

Fault (S60) as they located within the southeastern part of the city. The maximum PGA
from DSHA for the present study is found to be 0.74 g. Parvez et al. (2003) developed the
DSHA Map for the entire Indian subcontinent and found PGA range between 0.3 and 0.6 g
for the Patna city. However, the PGA value found from the present study is higher than that
of deterministic seismic hazard macrozonation carried out by Kolathayar et al. (2012),
which is in the range of 0.15–0.25 g.

8.2 Probabilistic seismic hazard analysis (PSHA)

The probability of exceedance of a given ground motion in a particular time period can be
estimated once the probability of its size, locations and level of ground shaking is known
cumulatively. The seismic hazard map for Patna has been generated using PSHA using
probabilistic method proposed by Cornell (1968), which was later improved by Alger-
missen et al. (1982). For evaluating the seismic hazard using the classical approach
(Cornell 1968), the entire SSA of Patna has been divided into 2500 grids of size
0.022° 9 0.014°. The uncertainties associated with magnitude, hypocentral distance and
probability of exceedance for GMPEs for 178 seismic sources have been computed using a
program developed in MATLAB. The program computed the frequency of exceedance of a
particular magnitude ‘mi’ occurring at a particular hypocentral distance ‘R’ with a known
probability of exceedance with respect to ‘z’, and the combined frequency of exceedance
of a particular ground motion can be estimated by merging all types of uncertainties for
each seismic source. The detailed methodology for determining the PGA using probability
seismic hazard analysis is explained in Anbazhagan et al. (2009) in this journal.
A hazard curve, which is defined as the frequency of exceedance of various levels of
ground motion for 10 most vulnerable sources at Patna center, is shown in Fig. 12. It can
be seen from Fig. 12 that S60 (West Patna Fault) is the most vulnerable source located at a
hypocentral distance of 55.11 km with a maximum magnitude of 7.5 (Mw). Other sources,
which have been found vulnerable for Patna, are also shown in Fig. 12 as S139, S58, S61,
S57, S138, S62, S39, S54 and S20. The hazard curve for any SSA can be obtained by the
summation of all the hazard curves obtained from all the active sources. Thus, merging all
the hazard curves from 178 sources at the Patna center will give the hazard curve for Patna
district center. Figure 13 shows the cumulative hazard curve obtained at the Patna district

Fig. 12 Hazard curve for ten S138 S139 S20 S39 S54
1.0E+01 S57 S58 S60 S61 S62
most contributing seismic source
1.0E+00
at Patna
1.0E-01
Probability of exceedance

1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1.0E-07
1.0E-08
1.0E-09
1.0E-10
1.0E-11
1.0E-12
1.0E-13
1.0E-14
0.02 0.2
Peak Ground Acceleration (g)

123
 Nat Hazards

Fig. 13 Hazard curve at Patna 1.0E+01

for different periods using 1.0E+00
classical approach

Frequency of exceedance
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1.0E-07 0 sec 0.05 sec
0.1 sec 0.2 sec
0.3 sec 0.4 sec
1.0E-08 0.5 sec 0.6 sec
0.8 se 1 sec
1.0E-09 1.6 sec 2 sec
2% Prob in 50 years 10% Prob in 50 years
1.0E-10
0.002 0.02 0.2
Spectral acceleration (g)

center for 0, 0.05, 0.1, 0.2, 0.3, 0.4, 0.6, 0.8, 1.0, 1.6 and 2 s. Hazard curve subsequent to
different periods presents the spectral acceleration values for an identified probability of
exceedance in a particular time period. In can be observed from Fig. 13 that the frequency
of exceedance for 0.075 g at zero second is 0.0074311, which will give the return period
135 years (return period is the inverse of the frequency of exceedance). This indicates that
PGA of 0.075 g has a 31.03 % probability of exceedance in 50 years at the Patna center.
Similarly, for 0.5 g, the frequency of exceedance at zero seconds is 2.55E-05, which will
give a return period of 39.2 thousand years or a probability of exceedance of
1.28 9 10-1 % in 50 years at Patna city center. As the period on interest increases from
zero second to 0.8 s, a huge change in return period has been observed from Fig. 13.
Initially, the frequency of exceedance decreases from 135 years at zero periods to 22 years
in 0.1 s, which further increases to 46 years in 0.2 s and again till 2.53E ? 05 years for
2 s. In order to understand the hazard contribution from various combinations of magni-
tude and hypocentral distance, degradation plot is generated, which is a function of
magnitude and hypocentral distance for all the levels of spectral period associated with the
GMPEs. The mean degradation plot for Patna SSA for 2 and 10 % probability of ex-
ceedance at 50 years have been made in order to understand the hazard contribution for
various magnitudes at a different hypocentral distance and are shown in Fig. 14a, b. It has
been observed from Fig. 14 that the motion for 6.0 Mw at 40 km hypocentral distance is
predominant for 2 % probability of exceedance at 50 years. Similarly, for 10 % probability
of exceedance at 50 years, the motion for 5.5 Mw at 50 km hypocentral distance is pre-
dominant. Hazard curve has been generated at each grid for Patna SSA; the level of ground
motion for frequency of exceedance ‘v(z)’ can be determined from it. The level of ground
motions has been estimated from the zero period hazard curves (PGA value) of each grid
for 2 and 10 % probability of exceedance in 50 years. Figure 15a, b is the PSHA maps for
Patna center for 2 and 10 % probability of exceedance in 50 years, respectively. It can be
observed from Fig. 15a that PGA varies from 0.44 g in the northwestern and 0.4 g in the
north eastern periphery to 0.08 g toward the central part. PGA value in the southeastern
part of Patna is 3–4 times as compared to southwestern part. Similarly, for 10 % prob-
ability of exceedance in 50 years, PGA value is low at the central part of the city and
increases about fivefold toward the northeastern part of the city. In addition to that, spectral
acceleration at, respectively, 0.2 and 1 s has been given as Fig. 16a–d for the return period
of 2475 and 475 years, respectively. The increment is due to East Patna Fault (S59) and

123
Nat Hazards

(a) 5.0E-04
4.0E-04
Frequency
3.0E-04

2.0E-04

1.0E-04

0.0
15.2
25.3
0.0E+00

35.4
45.5
4.5
4.75

55.6
5
5.25
5.5

70.7
5.75
6

90.9
6.25
6.5
6.75

111.1
7
(b) 3.0E-04
2.5E-04

2.0E-04
Frequency

1.5E-04

1.0E-04

5.0E-05

0.0
15.2
25.3

0.0E+00
35.4
45.5
4.5
4.75

55.6
5
5.25
5.5

70.7
5.75
6

90.9
6.25
6.5
6.75

111.1
7

Fig. 14 a Deaggregation of hazard value at Patna at bed rock at PGA for 2 % probability of exceedance in
50 years, b deaggregation of hazard value at Patna at bed rock at PGA for 10 % probability of exceedance in
50 years

West Patna Fault (S60), which lie within the city and are source of devastating earth-
quakes. Southwestern part and central part include areas like Nehru Nagar, Patna High
Court, Kothia, Pataupura Colony, Makhdhumpur and surrounding areas, which are less
prone to earthquake-induced ground shaking. However, areas which fall in eastern and
northern part of the city such as Hanuman Nagar, Sadhanpuri, Vigrahpur, Loknayak Jai-
prakash Airport, Bihar Chak, Patrakar Nagar and their nearby areas are more susceptible to
earthquake shaking. Since these levels of ground shaking are evaluated at bedrock level,
however, no changes in PGA contours along the alignment of river Ganga can be seen.
Recently, National Disaster Management Authority (NDMA 2010) and Nath and Thing-
baijam (2012) have developed a PSHA map for entire India. Nath and Thingbaijam (2012)
predicted the PGA value at Patna considering 10 % probability of exceedance in 50 years
as 0.13 g, whereas as per NDMA (2010), PGA value at 2 and 10 % probability of ex-
ceedance for 50 years was 0.08 and 0.04 g, respectively. Bhatia et al. (1999) presented a
PSHA of India under the Global Seismic Hazard Assessment Program (GSHAP) frame-
work. As per Bhatia et al. (1999), PGA value is in between 0.1 and 0.15 g of Patna

123
 Nat Hazards

Fig. 15 a PSHA map for Patna district for 2 % probability of exceedance in 50 years, b PSHA map for
Patna district for 10 % probability of exceedance in 50 years

considering 10 % probability of exceedance in 50 years. The predicted PGA value in this

study is comparable and slightly higher than the previous studies, and it may be due to
updated seismicity and considering regional-specific maximum magnitude and GMPE.

9 Site-specific spectrum

The site-specific design spectrum is necessary to design and to understand amplification

character of the region. In the present study, design spectrum is derived from the GMPE as
mentioned above with respect to different segments. For deriving the design spectrum, the

123
Nat Hazards

Fig. 16 a SA map at 1 s for 2 % probability of exceedance in 50 years, b SA map at 0.2 s for 2 %

probability of exceedance in 50 years, c SA map at 1 s for 10 % probability of exceedance in 50 years and
d SA map at 1 s for 10 % probability of exceedance in 50 years

whole district is divided into four zones based on PGA values calculated using DHSA
results (see Fig. 9). These zones are Zone 1 (0.14 B PGA \ 0.26), Zone 2
(0.26 B PGA \ 0.44), Zone 3 (0.44 B PGA \ 0.56) and Zone 4 (PGA C 0.56). Similarly,
design spectrum (5 % damping) for 2 and 10 % probability of exceedance in 50 years has
been developed by dividing the hazard map into four different zones in Fig. 15a, b. These
zones are Zone 1 (0.08 B PGA \ 0.16), Zone 2 (0.16 B PGA \ 0.32), Zone 3
(0.32 B PGA \ 0.4) and Zone 4 (PGA C 0.4) for a 2 % probability of exceedance in
50 years, and Zone 1 (0.03 B PGA \ 0.075), Zone 2 (0.075 B PGA \ 0.12), Zone 3
(0.12 B PGA \ 0.135) and Zone 4 (PGA C 0.135) for a 10 % probability of exceedance
in 50 years. For each of the zones, a spectral acceleration has been estimated at the center
of the zone (which is similar for both PSHA and DSHA) from the valid GMPEs. The site-
specific GMPEs ANBU-13, NDMA-10 and KANO-06 are used up to 100 km and ANBU-
13, NDMA-10, BOAT-08 and KANO-06 from 100 to 300 km for deriving design spectra.
Spectral acceleration which is a function of period for the respective GMPEs at 5 %
damping has been taken from respected research papers. So, Mmax and the average shortest
distance from each grid have been identified from the vulnerable seismic sources for
developing design spectra at 5 % damping level. Averaged smoothed design spectrum has
been developed as per Malhotra (2006). Design spectrum comprises of a peak, valley and
shape variation in response spectrum from each GMPE. The design spectra are normalized
with respect to spectral acceleration at zero periods (PGA) and shown as spectral ratio
versus time period in Fig. 17. Figure 17a, b also shows the normalized design spectrum for
5 % damping for all the four zones considering DSHA values and PSHA with IS-code and
Sikkim Earthquake spectrum comparison. This can be considered as the site-specific
normalized design spectrum curve for 5 % damping at the rock level for all the four zones,

123
 Nat Hazards

(a) 3 Zone 1 PSHA

Zone 2 PSHA
Zone 3 PSHA
2.5 Zone 4 PSHA
Zone 1 DSHA

Spectral Ratio (Sa/g)

Zone 2 DSHA
2 Zone 3 DSHA
Zone 4 DSHA
IS 1893 (2002)
1.5 Sikkim, 2011 EQ

0.5

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Time Period (sec)

(b) 3 Zone 1 PSHA

Zone 2 PSHA
Zone 3 PSHA
2.5 Zone 4 PSHA
Zone 1 DSHA
Spectral Ratio (Sa/g)

Zone 2 DSHA
2 Zone 3 DSHA
Zone 4 DSHA
IS 1893 (2002)
1.5 Sikkim, 2011 EQ

0.5

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Time Period (sec)
Fig. 17 a Normalized design spectrum for Patna for 5 % damping from four zones and spectrum from 2011
Sikkim earthquake and IS 1893 (2002) for 2 % probability of exceedance in 50 years comparing with worst-
case scenario DSHA result, b normalized design spectrum for Patna for 5 % damping from four zones and
spectrum from 2011 Sikkim earthquake and IS 1893 (2002) for 10 % probability of exceedance in 50 years
comparing with worst-case scenario DSHA result

considering venerable seismic sources of Patna. These rock-level normalized design

spectrums have been compared with the 2011 Sikkim Earthquake recorded at Gangtok
seismic station, located at rock site. It has been seen that present design spectrum hazard
analysis is predicting slightly higher value than recorded values. It can be noted that the
Sikkim earthquake magnitude is less than the maximum magnitude in the region. Site-
specific spectrum developed for Patna is also compared with the design spectrum of IS-
1893 (2002) of soil category of type I (Rock or Hard soil). The value of spectral accel-
eration developed from the present is low for zone 1 and higher for zone 2, 3 and 4 when
compared with IS-1893 (2002) for larger period in case of DSHA, which may be because
IS 1893 (2002) deals on a macrolevel (see Fig. 17a), and the present study is in microlevel
and region specific. However, in case of PSHA considering 10 % probability, spectral
acceleration is low for all the four zones, whereas in case of 2 % probability, high for zone
3 and zone 4 and low for zone 1 and zone 2. Additionally, the design calculated using
DSHA is closely comparable with 2 % probability of exceedance in 50 years (see

123
Nat Hazards

Fig. 17a). It has been observed that normalized design spectra developed in this study both
deterministically and probabilistically are either higher or lower as compared to IS-code.
This might be due to incorporation of regional-specific parameters and using site-specific
GMPE and maximum magnitude.

10 Conclusion

This paper presented seismic hazard map and site-specific design spectrum for Patna
district both deterministically and probabilistically also considering region-specific data. A
seismic study area of 500 km was arrived based on past earthquake damage distribution,
and seismotectonic map has been generated. Seismotectonic map consists of declustered
and homogenized past earthquake data and all the linear sources. The maximum magnitude
has been estimated by considering three methods, i.e., incremental method, Kijko method
and regional rupture-based characteristic. The maximum magnitude at each source was
selected by considering the maximum of three methods. About 27 GMPEs are applicable to
the study region, and suitable GMPES are identified for performing the efficacy test. The
segmented-based efficacy test has been carried out, and GMPEs are selected. It was found
that three GMPEs of ANBU-13, NDMA-10 and KANO-06 performed better up to 300 km
epicentral distance and NDMA-10 for more than 300 km. Hazard curve for 0, 0.05, 0.1,
0.2, 0.3, 0.4, 0.6, 0.8, 1.0, 1.6 and 2 s has also generated. The hazard map for both 2 and
10 % probability of exceedance in 50 years has been developed. In addition to that spectral
acceleration, hazard map has been developed at a period of 0.2 and 1 s for DSHA and
PSHA. PGA varies from 0.14 to 0.74 g for DSHA, from 0.08 to 0.44 g in case for 2 %
probability, and 0.03 to 0.165 g for 10 % probability. Furthermore, site-specific design
spectrum developed using DSHA is comparable with 2 % probability of exceedance in
50 years; however, it is either low or high as compared to IS-code for DSHA as well as
PSHA. The present result is slightly more advanced than previous studies and can be
further used for estimating microzonation parameter of Patna district. Seismic hazard
values given in this paper are at rock condition with V30 s [ 1500 m/s. These values may
alter when site effects based on site-specific soil properties are considered.

References
Abrahamson NA, Litehiser JJ (1989) Attenuation of vertical peak accelerations. Bull Seismol Soc Am
79:549–580
Abrahamson NA, Silva WJ (2007) Abrahamson & Silva NGA ground motion relation for the geometric
mean horizontal component of peak and spectral ground motion parameters. Report, Pacific Earth-
quake Research Center, Berkeley
Abramowitz M, Stegun IA (1970) Handbook of mathematical functions, 9th edn. Dover Publication, New
York
Aghabarati H, Tehranizadeh (2009) Near-source ground motion attenuation relationship for PGA and PSA
of vertical and horizontal components. Bull Earthq Eng 7:609–635. doi:10.1007/s10518-009-9114-9
Akkar S, Bommer JJ (2010) Empirical equations for the prediction of PGA, PGV and spectral acceleration
in Europe, the Mediterranean region and the Middle East. Seismol Res Lett 81:195–206
Algermissen ST, Perkins DM, Thenhaus PC, Hanson SL, Bender BL (1982) Probabilistic estimates of
maximum acceleration and velocity in rock in the contiguous United States. Open-File Report
82–1033. U.S. Geological Survey, Washington, DC, p 99
Ambraseys NN, Douglas J (2004) Magnitude calibration of north Indian earthquakes. Geophys J Int
158:1–42

123
 Nat Hazards

Ambraseys N, Douglas JS, Sarma K, Smit PM (2005) Equation for the estimation of strong ground motions
from shallow crustal earthquakes using data from Europe and the Middle East: horizontal peak ground
acceleration and the spectral acceleration. Bull Earthq Eng 3:1–53
Anbazhagan P, Vinod JS, Sitharam TG (2009) Probabilistic seismic hazard analysis for Bangalore. Nat
Hazards 48:145–166
Anbazhagan P, Vinod JS, Sitharam TG (2010) Evaluation of seismic hazard parameters for Bangalore
region in South India. Disaster Adv 3(3):5–13
Anbazhagan P, Kumar A, Sitharam TG (2012) Seismic site classification and correlation between standard
penetration test N value and shear wave velocity for Lucknow city in Indo-Gangetic basin. Pure Appl
Geophys 170(2013):299–318. doi:10.1007/s00024-012-0525-1
Anbazhagan P, Kumar A, Sitharam TG (2013a) Ground motion prediction equation considering combined
data set of recorded and simulated ground motions. Soil Dyn Earthq Eng 53:92–108
Anbazhagan P, Smitha CV, Kumar A, Chandran D (2013b) Seismic hazard assessment of NPP site at
Kalpakkam, Tamil Nadu, India. Nucl Eng Des 259:41–64
Anbazhagan P, Smitha CV, Abhishek Kumar (2014) Representative seismic hazard map of Coimbatore,
India. Eng Geol 171:81–95
Anbazhagan P, Bajaj K, Moustafa SSR, Al-Arifi NSN (2015) Maximum magnitude estimation considering
the regional rupture character. J Seismol. http://link.springer.com/article/10.1007%2Fs10950-015-
9488-x (published online)
Atkinson GM, Boore DM (2003) Empirical ground-motion relations for subduction-zone earthquakes and
their applications to Cascadian and other regions. Bull Seismol Soc Am 93:1703–1717
Atkinson GM, Boore DM (2006) Earthquake ground-motion prediction equations for eastern North
America. Bull Seismol Soc Am 96:2181–2205
Banghar AR (1991) Mechanism solution of Nepal–Bihar earthquake of August 20, 1988. J Geol Soc India
37:25–30
Baruah S, Gogoi NK, Erteleva Q, Aptikaev F, Kayal JR (2009) Ground motion parameters of Shillong
plateau: one of the most seismically active zones of Northeastern India. Earthq Sci 22:283–291
Bhatia SC, Ravi MK, Gupta HK (1999) A probabilistic seismic hazard map of India and adjoining regions.
Ann Geofis 42:1153–1164
Bommer JJ, Scherbaum F, Bungum H, Cotton F, Sabetta F, Abrahamson NA (2005) On the use of logic trees
for ground-motion prediction equations in seismic-hazard analysis. Bull Seismol Soc Am 95:377–389
Bommer JJ, Douglas J, Scherbaum F, Cotton F, Bungum H, Fäh D (2010) On the selection of ground-motion
prediction equations for seismic hazard analysis. Seismol Res Lett 81(5):783–793
Boore DM, Atkinson GM (2008) Ground-Motion Prediction Equations for the average horizontal compo-
nent of PGA, PGV and 5% damped PSA at spectral periods between 0.01 and 10.0 s. Earthq Spectra
24(1):99–138
Bormann P, Liu R, Ren X, Gutdeutch R, Kaiser D, Castellaro S (2007) Chinese national network magni-
tudes, their relation to NEIC magnitudes and recommendations for new IASPEI magnitude standard.
Bull Seismol Soc Am 95:58–74
Campbell KW (1997) Empirical near-source attenuation relationships for horizontal and vertical compo-
nents of peak ground acceleration, peak ground velocity and pseudo-absolute acceleration response
spectra. Seismol Res Lett 68(1):154–179
Campbell KW, Bozorgnia Y (2006) Next generation attenuation relation (NGA) Empirical ground motion
models: can they be used for Europe. In: Proceedings of first European conference on earthquake
engineering and seismology, Geneva, Switzerland, paper no. 458
Campbell KW, Bozorgnia Y (2008) NGA ground motion model for the geometric mean horizontal com-
ponent of PGA, PGV, PGD and 5 % damped linear elastic response spectra for period ranging from
0.01 to 10 s. Earthq Spectra 24:139–171
Castellaro S, Mulargia F, Kagan YY (2006) Regression problems for magnitudes. Geophys J Int
165:913–930
Cornell CA (1968) Engineering seismic rick analysis. Bull Seismol Soc Am 58:1583–1606
Cotton F, Scherbaum F, Bommer JJ, Bungum H (2006) Criteria for selecting and adjusting ground-motion
models for specific target regions: application to Central Europe and rock sites. J Seismol 10:137–156
Das S, Gupta ID, Gupta VK (2006) A probabilistic seismic hazard analysis of Northeast India. Earthq
Spectra 22:1–27
Dasgupta S, Mukhopadyay M, Nandy DR (1987) Active transverse features in the central portion of the
Himalaya. Tectonophysics 136(1987):255–264
Dasgupta S, Sengupta P, Mondal A, Fukuoka M (1993) Mineral chemistry and reaction textures in meta-
basites from the Eastern Ghats belt, India and their implications. Miner Mag 57:113–120

123
Nat Hazards

Davis SD, Frohlich C (1991) Single-link cluster analysis, synthetic earthquake catalogs and aftershock
identification. Geophys J Int 104:289–306
Delavaud E, Scherbaum F, Kuehn N, Allen T (2012) Testing the global applicability of ground-motion
prediction equations for active shallow crustal regions. Bull Seism Soc Am 102(2):702–721
Delavaud E, Scherbaum F, Kuehn N, Riggelsen C (2009) Information-theoretic selection of ground-motion
prediction equations for seismic hazard analysis: an applicability study using Californian data. Bull
Seismol Soc Am 99:3248–3263
Dewey JF, Bird JM (1970) Mountain belts and the new global tectonics. J Geophys Res 75(14):2625–2647
Douglas J (2010) Consistency of ground-motion predictions from the past four decades. Bull Earthq Eng
8:1515–1526
Douglas J, Mohais R (2009) Comparing predicted and observed ground motions from subduction earth-
quakes in the Lesser Antilles. J Seism 13(4):577–587
EERI committee on seismic risk (1984) Glossary of terms for probabilistic seismic risk and hazard analysis.
Earthq Spectra 1:33–36
Frankel A (1995) Mapping seismic hazard in the central eastern United States. Seismol Res Lett 66(4):8–21
Gansser A (1964) Geological history of Himalayan. Geology of the Himalayas, Wiley, Switzerland,
pp 235–245
Gardner JK, Knopoff L (1974) Is the sequence of earthquakes in southern California, with aftershocks
removed, poissonian? Bull Seismol Soc Am 64(5):1363–1367
Grünthal G (1998) European macroseismic scale 1998. Cahiers du Centre EuropéendeGéodynamiqueet de
Séismologie, vol 15, Luxembourg
GSI (2000) Eastern Nepal Himalaya and Indo-Gangetic Plains of Bihar. In: Narula PL, Acharyya SK,
Banerjee J (eds) Seismotectonics Atlas of India and its environs. Geological Survey of India, India,
pp 26–27
Gupta ID (2010) Response spectral attenuation relations for inslab earthquakes in Indo-Burmese subduction
zone. Soil Dyn Earthq Eng 30:368–377
Gutenberg B, Richter CF (1956) Earthquake magnitude, intensity, energy and acceleration. Bull Seismol
Soc Am 46:105–145
Hintersberger E, Scherbaum F, Hainzl S (2007) Update of likelihood-based ground motion model selection
for seismic hazard analysis in western central Europe. Bull EarthqEng 5:1–16
Idriss IM (2008) An NGA empirical model for estimating the horizontal spectral values generated by
shallow crustal earthquakes. Earthq Spectra 16:363–372
IS 1893 (2002) Indian standard criteria for earthquake resistant design of structures, part 1-general provi-
sions and buildings. Bureau of Indian Standards, New Delhi
Joshi GC, Sharma ML (2008) Uncertainties in the estimation of Mmax. J Earth Syst Sci 117(S2):671–682
Kaklamanos J, Baise LG (2011) Model validations and comparisons of the next generation attenuation of
ground motions (NGA–West) project. Bull Seism Soc Am 101(1):160–175
Kanno T, Narita A, Morikawa N, Fujiwara H, Fukushima Y (2006) A new attenuation relation for strong
ground motion in Japan based on recorded data. Bull Seismol Soc Am 96:879–897
Khattri KN (1987) Great earthquakes, seismicity gaps and potential for earthquakes along the Himalayan
plate boundary. Tectonophysics 38:79–92
Kijko A, Sellevoll MA (1989) Estimation of earthquake hazard parameters from incomplete data files. Part I,
utilization of extreme and complete catalogues with different threshold magnitudes. Bull Seismol Soc
Am 79:645–654
Kolathayar S, Sitharam TG, Vipin KS (2012) Deterministic seismic hazard microzonation of India. J Earth
Syst Sci 121:1351–1364
Kumar S (2012) Seismicity in the NW Himalaya India: fractal dimension, b-value mapping and temporal
variation for hazard evaluation. Geosci Res 3(1):83–87
Kumar A, Anbazhagan P, Sitharam TG (2013) Seismic hazard analysis of Lucknow considering local and
active seismic gaps. Nat Hazards 69:327–350. doi:10.1007/s11069-0.13-0712-0
Lai CG, Menon A, Corigliano M, Ornthamarrath, Sanchez HL, Dodagoudar GR (2009) Probabilistic seismic
hazard assessment and stochastic site response analysis at the archaeological site of Kancheepuram in
southern India. Research Report EUCENTRE 2009/01, IUSS Press, Pavia, ISBN 978-88-6198-037-
2:250
Lin PS, Lee CH (2008) Ground-Motion Attenuation relationship for subduction-zone earthquakes in
Northeastern Taiwan. Bull Seismol Soc Am 98(1):220–240
Mahajan AK, Thakur VC, Sharma ML, Chauhan M (2010) Probabilistic seismic hazard map of NW
Himalaya and its adjoining area, India. Nat Hazards 53:443–457
Malhotra PK (2006) Smooth spectra of horizontal and vertical ground motions. Bull Seismol Soc Am
96(2):506–518

123
 Nat Hazards

Mohindra R, Parkash B, Prasad J (1992) Historical geomorphology and pedology of the Gandak Megafan,
Middle Gangetic Plains, India. Earth Surf Process Land 17:643–662
Molchan G, Dmitrieva O (1992) Aftershock identification: methods and new approaches. Geophys J Int
109:501–516
Nath SK, Thingbaijam KKS (2011) Peak ground motion predictions in India: an appraisal for rock sites.
J Seismol 15:295–315
Nath SK, Thingbaijam KKS (2012) Probabilistic seismic hazard assessment of India. Seismol Res Lett
83:135–149
Nath SK, Vyas M, Pal I, Sengupta P (2005) A hazard scenario in the Sikkim Himalaya from seismotectonics
spectral amplification source parameterization and spectral attenuation laws using strong motion
seismometry. J Geophys Res 110:1–24
Nath SK, Raj A, Thingbaijam KKS (2009) Kumar A (2009) Ground motion synthesis and seismic scenario
in Guwahati city, a stochastic approach. Seismol Res Lett 80(2):233–242
NDMA (2010) Development of probabilistic seismic hazard map of India. Technical report by National
Disaster Management Authority, Government of India, New Delhi
Parvez IA, Vaccari F, Panza GF (2003) A deterministic seismic hazard map of India and adjacent areas.
Geophys J Int 155:489–508
Reasenberg P (1985) Second-order moment of central California seismicity 1969–1982. J Geophys Res
90:5479–5495
Sabetta F, Lucantoni A, Bungum H, Bommer JJ (2005) Sensitivity of PSHA results to ground motion
prediction relations and logic-tree weights. Soil Dyn Earthq Eng 25:317–329
Savage WU (1972) Microearthquake clustering near Fairview Peak, Nevada, and in the Nevada seismic
zone. J Geophys Res 77(35):7049–7056
Scherbaum F, Cotton F, Smit P (2004) On the use of response spectral reference data for the selection and
ranking of ground-motion models for seismic hazard analysis in regions of moderate seismicity: the
case of rock motion. Bull Seismol Soc Am 94:1–22
Scherbaum F, Bommer JJ, Bungum H, Cotton F, Abrahamson NA (2005) Composite ground-motion models
and logic trees: methodology, sensitivities, and uncertainties. Bull Seismol Soc Am 95:1575–1593
Scherbaum F, Delavaud E, Riggelsen C (2009) Model selection in seismic hazard analysis: an information
theoretic perspective. Bull Seismol Soc Am 99:3234–3247
Scordilis EM (2006) Empirical global relations converting MS and mb to moment magnitude. J Seismol
10:225–236
SEISAT (2000) Seismotectonic Atlas of India and its environs. Geological Survey of India, India
Sharma ML (1998) Attenuation relationship for estimation of peak ground horizontal acceleration using data
from strong motions arrays in India. Bull Seismol Soc Am 88:1063–1069
Sharma ML, Bungum H (2006) New strong ground motion spectral acceleration relation for the Himalayan
region. In First European conference on earthquake engineering and seismology, p 1459
Sharma ML, Douglas J, Bungum H, Kotadia J (2009) Ground-motion prediction equations based on data
from Himalayan and Zagros regions. J Earthq Eng 13:1191–1210
Singh RP, Aman A, Prasad YJJ (1996) Attenuation relations for strong ground motion in the Himalayan
region. Pure appl Geophys 147:161–180
Sinha R, Tandon SK, Gibling MR, Bhattarcharjee PS, Dasgupta AS (2005) Late quaternary geology and
alluvial stratigraphy of the Ganga basin. Himal Geol 26(1):223–340
Spudich P, Joyner WB, Lindh AG, Boore DM, Margaris BM, Fletcher JB (1999) SEA99: a revised ground
motion prediction relation for use in Extensional tectonic regions. Bull Seism Soc Am
89(5):1156–1170
Sreevalsa K, Sitharam TG, Vipin KS (2011) Spatial variation of seismicity parameters across India and
adjoining area. Nat Hazards. doi:10.1007/s11069-011-9898-1
Stepp JC (1972) Analysis of completeness of the earthquake sample in the Puget Sound area and its effect on
statistical estimates of earthquake hazard. Proceeding of the International conference on microzona-
tion, vol 2. Seattle, USA, pp 897–910
Stiphout VT, Zhuang J, Marsan D (2010) Seismicity declustering, community online resource for statistical
seismicity analysis. http://www.corssa.org. Accessed 14 Jan 2014
Strasser FO, Abrahamson NA, Bommer JJ (2009) Sigma: issues, insights, and challenges. Seismol Res Lett
80(1):40–56
Stromeyer D, Grunthal G, Wahlstrom R (2004) Chi square regression for seismic strength parameter
relations, and their uncertainty with application to an Mw based earthquake catalogue for central,
northern and north-western Europe. J Seismol 8:143–153
Takahashi T, Saiki T, Okada H, Irikura K, Zhao JX, Zhang J, Thoi HK, Somerville PG, Fukushima Y,
Fukushima Y (2004) Attenuation models for response spectra derived from Japanese strong-motion

123
Nat Hazards

records accounting for tectonic source types. 13th world conference of earthquake engineering,
Vancouver, B.C., Canada, paper 1271
Thingbaijam KKS, Nath SK, Yadav A, Raj A, Walling MY, Mohanty WK (2008) Recent seismicity in
Northeast India and its adjoining region. J Seismol 12:107–123
Uhrhammer RA (1986) Characteristics of northern and central California seismicity. Earthq Notes 1:21
Valdiya KS (1976) Himalayan transverse faults and their parallelism with subsurface structures of north
Indian plains. Tectonophysics 32:352–386
Wallace K, Bilham R, Blume F, Gaur VK, Gahalaut V (2006) Geodetic constraints on the Bhuj 2001
earthquake and surface deformation in the Kachchh Rift basin. Geophys Res Lett 33:L10301. doi:10.
1029/2006GL025775
Wells DL, Coppersmith KJ (1994) New empirical relationships among magnitude, rupture length, rupture
width, rupture area, and surface displacement. Bull Seismol Soc Am 4(84):975–1002
Wiemer S, Wyss M (1994) Seismic quiescence before the landers (M = 7.5) and big bear (M = 6.5) 1992
earthquakes. Bull Seismol Soc Am 84:900–916
Wiemer S, Wyss M (1997) Mapping the frequency-magnitude distribution in asperities: an improved
technique to calculate recurrence times? J Geophys Res 102:15115–15128
Woessner J, Stefan W (2005) Assessing the quality of earthquake catalogues: estimating the magnitude of
completeness and its uncertainty. Bull Seism Soc Am 95(2):684–698
Youngs RR, Chiou SJ, Silva WJ, Humphrey JR (1997) Strong ground motion relationship for subduction
earthquakes. Seismol Res Lett 68:58–73
Zhao JX, Zhang J, Asano A, Ohno Y, Oouchi T, Takahashi T, Ogawa H, Irikura K, Thio HK, Somerville
PG, Fukushima Y, Fukushima Y (2006) Attenuation relations of strong ground motion in Japan using
site classification based on predominant period. Bull Seismol Soc Am 96:898–913

123