Geoscience Frontiers 6 (2015) 847e859

HOSTED BY Contents lists available at ScienceDirect

China University of Geosciences (Beijing)

Geoscience Frontiers

journal homepage: www.elsevier .com/locate/gsf

Research paper

Probabilistic assessment of surface level seismic hazard in Indiausing topographic gradient as a proxy for site condition

T.G. Sitharam a, Sreevalsa Kolathayar b,*, Naveen James a

aDepartment of Civil Engineering, Indian Institute of Science, Bangalore 560012, IndiabDepartment of Civil Engineering, Indian School of Mines, Dhanbad 826004, India

a r t i c l e i n f o

Article history:Received 9 September 2013Received in revised form12 June 2014Accepted 23 June 2014Available online 5 July 2014

Keywords:PSHAPGATopographySite amplificationIndia

* Corresponding author. Tel.: þ91 9482088377.E-mail address: sreevalsakolathayar@gmail.com (S. KoPeer-review under responsibility of China University

http://dx.doi.org/10.1016/j.gsf.2014.06.0021674-9871/� 2015, China University of Geosciences (

a b s t r a c t

This paper presents spatial variation of seismic hazard at the surface level for India, covering 6e38� Nand 68e98� E. The most recent knowledge on seismic activity in the region has been used to evaluate thehazard incorporating uncertainties associated with the seismicity parameters using different modelingmethodologies. Three types of seismic source models, viz. linear sources, gridded seismicity model andareal sources, were considered to model the seismic sources and different sets of ground motion pre-diction equations were used for different tectonic provinces to characterize the attenuation properties.The hazard estimation at bedrock level has been carried out using probabilistic approach and the resultsobtained from various methodologies were combined in a logic tree framework. The seismic site char-acterization of India was done using topographic slope map derived from Digital Elevation Model data.This paper presents estimation of the hazard at surface level, using appropriate site amplification factorscorresponding to various site classes based on VS30 values derived from the topographic gradient. Spatialvariation of surface level peak horizontal acceleration (PHA) for return periods of 475 years and 2475years are presented as contour maps.

� 2015, China University of Geosciences (Beijing) and Peking University. Production and hosting byElsevier B.V. All rights reserved.

1. Introduction

Earthquakes are known to have occurred in the region of Indiansubcontinent from ancient times. Large magnitude earthquakeshave occurred in the northern subcontinent and in the Andamanand Nicobar regions. In the southern peninsula, damaging earth-quakes have occurred but less frequently and with lower magni-tudes than at the plate boundaries. The rapid drifting of Indian platetowards Himalayas in the north east direction with a high velocityalong with its low plate thickness might be the cause for highseismicity of the Indian region. Bureau of Indian Standard haspublished seismic zonation map in 1962 and revised it during 1966,1970, 1984 and 2002. The map was entirely based on past earth-quake history, seismotectonics and geophysical data, not based on adetailed hazard analysis.

lathayar).of Geosciences (Beijing).

Beijing) and Peking University. Pro

Various researchers have attempted to evaluate the expectedground motion due to future earthquakes in and around India.Khatri et al. (1984) developed Peak Horizontal Acceleration (PHA)hazard map with 10% annual probability of exceedance in 50 yearswith the use of the ground motion prediction equation developedby Algermissen and Perkins (1976). PHA hazard map for entire In-dia, with 10% annual probability of exceedance in 50 years waspresented by Bhatia et al. (1999) using the ground motion predic-tion equation of Joyner and Boore (1981). However the resultsobtained in the studies of Khatri et al. (1984) and Bhatia et al. (1999)were debatable because of the use of single ground motion pre-diction equation for the entire country. Iyengar et al. (2010)developed probabilistic seismic hazard map for Indian landmassusing linear seismic sources and attenuations relations developedby them using ground motion simulations for various parts of thecountry. There were several other efforts by various researchers toestimate the seismic hazard for various isolated regions in thecountry using different methodologies. Most of the past works onestimating the seismic hazard for various parts of India focused on asmall region without extending it to the adjoining areas that

duction and hosting by Elsevier B.V. All rights reserved.

Figure 1. Tectonic provinces in and around India.

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859848

exhibits different seismotectonic characteristics. The few worksthat focused on entire country either use only one specific sourcemodel or a single ground motion prediction equation withouttaking into account varied attenuation properties of different

Figure 2. Distribution of Earthquake events (mainshocks onl

tectonic provinces in the country. This necessitates the need for acomprehensive seismic hazard analysis for entire country consid-ering various source models and different ground motion pre-dictions to suit various tectonic provinces with variedseismotectonic characteristics.

Estimation of PHA is done using different sets of ground motionprediction equations specific to various tectonic provinces in thecountry. In context of recent development in seismic hazardassessment and its application to earthquake engineering, it is clearthat we cannot bestow completely with either probabilistic ordeterministic elements, but both are equally important (Bommer,2003). Deterministic Seismic Hazard Analysis (DSHA) adopts atransparent process, whereas Probabilistic Seismic Hazard Analysis(PSHA) deals with various uncertainties. PSHA gives a richerdescription of the expected ground motions since it provides theprobability of exceeding a specific groundmotion level whereas theground motions given by DSHA are associated with an unknownprobability of exceedance. PHA estimated from probabilisticmethod handles various uncertainties, where as deterministicapproach generates distinct value of groundmotion parameter for aspecific scenario.

When seismic wave travels from bedrock to ground surface, itundergoes modification due to impedance contrast and thismodification is termed as the local site effects. Thus local geologyand soil conditions have significant effect on the intensity of groundshaking due to earthquake. In most of the buildings where foun-dation is not extending to bedrock, local site effect becomes crucialin dictating the damage intensity. The estimation of intensity of

y) in and around India (Kolathayar and Sitharam, 2012).

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 849

ground shaking at surface level requires site characterization of soiloverburden mass either based on geology or dynamic soil proper-ties. Generally geotechnical and geophysical tests are carried out tocharacterize the soil overburden. However, the use of these tests forseismic site characterization is limited for regions having smallerarea and not viable for characterizing a larger area such as a state, ora country. For the seismic site characterization at macro-scale level,terrain slope data can be very useful. Studies already proved thattopographic slope data derived fromDigital Elevation Model (DEM)is an index for the seismic site conditions.

Allen and Wald (2009) described a technique to derive first-order site-condition maps directly from topographic data. Theyhave used global 30 arc sec topographic data and VS30 measure-ments collected from several studies in the United States, as well asin Taiwan, Italy, and Australia. They correlated VS30 values with thetopographic slope to develop two sets of parameters for obtainingVS30: one for active tectonic regions where topographic relief ishigh, and one for stable shield regions where topography is gentler.By taking the gradient of the topography and choosing ranges ofslope that maximize the correlation with shallow shear-velocityobservations; Allen and Wald (2009) have recovered, to first or-der, many of the spatially varying features of the site-conditionmaps developed in California. They found that maps derived fromthe slope of the topography correlate well with other indepen-dently derived, site-condition maps. In the present study, anattempt was made to develop the VS30 map for India using topo-graphic slope map derived from DEM data.

Based on the VS30 values, the site can be characterised as perNational Earthquake Hazard Reduction Program (NEHRP) norms.Using appropriate site amplification factors, ground motionparameter at surface can be estimated from that estimated at thebedrock level. Various hazard assessment studies (Deterministic aswell as probabilistic methodologies) have been carried out for Indiain the past to estimate the groundmotion parameter at the bedrocklevel. This paper presents estimation of the hazard at surface level,using appropriate site amplification factors corresponding tovarious site classes based on VS30 values derived from the topo-graphic gradient.

Figure 3. Linear seismic sources identified in India.

2. Seismicity of the study area

2.1. Tectonic provinces

Geographical statistics of India show that almost more than halfof the land is vulnerable to earthquakes. The seismicity of Indiansubcontinent can be broadly characterized by three general seis-motectonic regions (Fig. 1): tectonically active shallow crustal re-gion, subduction zones and stable continental region. Thesubduction zone earthquakes can be further classified as regionswith intraslab and interface earthquakes.

The tectonically active shallow crustal regions include theHimalayas and southern Tibetan Plateau, northwest frontier prov-ince of Indian plate (Kayal, 2008; Nath and Thingbaijam, 2011). Themovement of Indian plate in the north eastern direction and itscollisionwith the Eurasian plate created the giganticmountain rangee the Himalayas with an average height of 4600 m and the largestplateau region in theworlde the Tibetan Plateau. The Indian plate isconsidered as one of the fastest moving plates; current movement isestimated to be around 5 cm/yr (Kumar et al., 2007). The collisionand the subsequent formation of the Himalayas and the TibetanPlateau are associated with very high seismicity. Seismically activenortheeast Indian region falls at the junction of NeS trending Bur-mese arc and EeW trending Himalayan Arc resulting in numerousgeological structures (Sharma and Malik, 2006).

The subduction zones include that of Hindukush-Pamir to thenorthwest frontier province, Indo-Myanmararc to thenorth east, andAndaman-Sumatra seismic belt to the south east. The Indo-Burmesearc is an important tectonic feature, the seismicity of which isrelated to the subduction of the Indian plate underneath the South-east Asian plate due to northeastward motion of India (Deshikachar,1974). The northeastern corner of India, sandwiched between theHimalayan and Burmese arcs, is characterized by a complex seismo-tectonic setup and very high level of seismicity (Evans, 1964).

Peninsular India is delineated as Stable Continental Region (SCR)with low to moderate seismic activity (Chandra, 1977). The seis-micity of this region is of intraplate nature and appears to beassociated with some local faults and weak zones (Rao and Murty,1970). The ENEeWSW trending Son-Narmada-Tapti zone is aprominent tectonic province in the northern margin of the penin-sular shield. Themajor tectonic elements in the southern part of thepeninsula can be listed as the massive Deccan Volcanic Province,the Southern Indian Granulite Terrain, the Dharwar Craton, theCuddapah Basin, the Godavari and the Mahanadi Grabens, and theEastern and Western Ghats on the east and west coasts, respec-tively (Gupta, 2006).

2.2. Earthquake catalog

Recently Kolathayar et al. (2012) prepared a comprehensiveearthquake catalog for India and adjoining areas by compiling thedata from various national and international agencies as well asfrom the literature. They developed correlations connecting variousmagnitude scales and homogenized the catalog in a unifiedmoment magnitude scale. This declustered earthquake catalog isused to characterize seismic sources in the present study. Thedistribution of earthquake events in India and adjoining areas isgiven in Fig. 2.

2.3. Linear sources

The Seismotectonic Atlas (SEISAT, 2000), published by Geolog-ical Survey of India (GSI), has been taken as an authentic referencemanual for identifying the seismic sources by various researcherslike, Iyengar and Ghosh (2004) for Delhi, Nath (2006) for

Figure 4. Identified regional seismic source zones in and around India (Kolathayar and Sitharam, 2012).

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859850

microzonation of Sikkim Himalayas, Raghukanth and Iyengar(2006) for Mumbai, Boominathan et al. (2008) for Chennai,Anbazhagan et al. (2009) for Bangalore and Vipin et al. (2009) forSouth India. The details of the faults, lineaments, shear zones andthe geological features in India and adjoining areas are presented inSEISAT (2000). In the present study, all the sheets of SEISAT (2000)were scanned and georeferenced and the individual images weremerged together to form the complete map India. Georeferencingand digitization of seismotectonic maps were done using MapInfoProfessional version 2006. The declustered earthquake data weresuperimposed on the fault map andmaximum reportedmagnitudealong each of these sources was noted. The linear sources thusselected are shown in Fig. 3.

2.4. Gridded seismicity model

The gridded seismicity model is based on the seismic activityrate obtained from the earthquake catalogue and it is one of themost widely adopted methods to model seismic sources for theregions in the absence of clearly identified seismic sources (Frankel,1995; Woo, 1996; Martin et al., 2002). In this method the study areais divided into grids of size 0.1� � 0.1�. The number of earthquakes,which are having magnitude higher than a cutoff magnitude, ineach cell, is counted. This will give the activity rate for thatparticular grid cell. In the present study, the cutoff magnitude(Mcut) was selected asMw ¼ 4.0. Based on this value, the recurrencerates for different magnitude intervals were calculated and thesevalues were smoothed using a Gaussian function to get the final

activity rate for each grid cell (Eq. 1). The uncertainty involved inestimating the location of the earthquake event and the size of theseismic source can be accounted by this smoothing.

bni ¼

Pjnje

��D2

ij

�c2�

Pje�

�D2

ij

�c2� (1)

where njenumber of earthquakes in the jth grid cell, bniesmoothednumber of earthquakes in ith cell, cecorrelation distance to accountfor the location uncertainties, and Dijedistance between the ith andjth cells.

2.5. Areal sources

For the hazard estimation using areal sources, the territory isdivided into seismic source zones according to geotectonic con-siderations where it can be assumed that, within each zone, anindependent earthquake-occurrence process is taking place. Foreach seismic source zone, magnitude exceedance rates are esti-mated bymeans of statistical analysis of earthquake catalogs. Theserates are the number of earthquakes, per unit time, in whichmagnitude M is exceeded, and they characterize the seismicity ofthe source (Ordaz et al., 2007).

Considering India’s complex sesismotectonic setup, Kolathayarand Sitharam (2012) divided entire Indian region into 104 sourcezones based on the trend of tectonic features, predominant source

Table 1Site classification as per NEHRP scheme (BSSC, 2003).

NHERP site class Description VS30 (m/s)

A Hard rock >1500B Firm and hard rock 760e1500C Dense soil, soft rock 360e760D Stiff soil 180e360E Soft clays, special study

soils, eg. liquefiable soil<180

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 851

mechanism solutions and the epicentral distribution of pastearthquakes. CRISIS software (Ordaz et al., 2007) was used tomodelareal sources and to estimate the seismic hazard with polygon-dipping areas. These zones allow for local variability in seismicitycharacteristics for other two types of source models; linear sourcesand zoneless approach (changes in b-value, lm, Mmax etc.). Theregional seismic source zones identified by Kolathayar andSitharam (2012) are shown in Fig. 4. Kolathayar and Sitharam(2012) have performed seismicity analysis using ZMAP software(Wiemer, 2001) and characterized 104 regional seismic sourcezones in the country. The Mmax was estimated for all the zonesbased on the methodology suggested by Kijko (2004).

In this method, a spatial integration process is carried out toaccount for all possible focal locations with an assumption that,within a seismic source, all points are equally likely to be anearthquake focus. CRISIS assumes that, within a source, seismicityis evenly distributed by unit area and to correctly account for thismodeling assumption, CRISIS performs a spatial integration bysubdividing the original sources. Once subdivided into sub-sources,CRISIS assigns to a single point all the seismicity associated to a sub-source, and then the spatial integration adopts a summation form.

3. Ground motion models

The seismicity of Indian subcontinent is spatially varied andcomplex as it encompasses various tectonic zones with differentattenuation characteristics. Appropriate Ground Motion PredictionEquations (GMPEs) should be used for different tectonic regions.Hence for the selection of the GMPEs, study area was divided intofour categories e active tectonic shallow crustal region, Stablecontinental region, subduction intraslab region and subductioninterface region.

In India, there is scarcity of strong motion data and this in turnhas resulted in the development of only very few region specificGMPEs. Since only a few ground motion prediction equations wereavailable for the study area, in the present studywe have used someof the well accepted GMPEs which were developed for other re-gions of the world having similar seismic attenuation characteris-tics. In the present analysis, different GMPEs were used to modelthe attenuation properties of the plate boundary region, stablecontinental region and subduction zones. Several ground motionprediction equations capable of predicting ground motion in theseseismic provinces were identified and studied. The best suitedthree ground motion prediction equations were selected for eachtectonic province, based on strong motion records from recentsmall number of recorded events reported in these areas.

The relations used for Shield region are Campbell and Bozorgnia(2003), Atkinson and Boore (2006), and Raghukanth and Iyengar(2007). Out of these, the relation by Raghukanth and Iyengar(2007) was developed for the Peninsular Indian Shield regions.Raghukanth and Iyengar (2007) observed that their model havepredictions similar to those of the available models for otherintraplate regions. Ground motion prediction equations given byCampbell and Bozorgnia (2003) and Atkinson and Boore (2006)were developed for the Eastern North America (ENA). Based onthe study of aftershocks of Bhuj earthquake, Cramer and Kumar(2003) came to the conclusion that the ground motion attenua-tion in ENA and Peninsular Indian shield are comparable. Similarityof the regional tectonics of ENA and peninsular India has beennoted by Bodin et al. (2004) also.

The GMPE used for active tectonic regions are Boore andAtkinson (2008), Sharma et al. (2009) and Akkar and Bommer(2010). Of these, the relation suggested by Sharma et al. (2009)was developed for Himalayan regions of India. Sharma et al.(2009) used data from the Himalayas and Zagros regions on the

premises that seismotectonics of the two regions have considerablesimilarity (Ni and Barazangi, 1986). On the other hand, the relationby Boore and Atkinson (2008) was developed for active tectonicregions across the world and Akkar and Bommer (2010) wasdeveloped for the active tectonic regions of Europe andMiddle East.

For Indo-Myanmar Subduction zone, we used the ground mo-tion prediction equations suggested by Zhao et al. (2006), Lin andLee (2008) and Gupta (2010) as all three are capable of predictingground motion from intraslab subduction earthquakes. For thesubduction zone with interface earthquakes, the GMPEs used wereof Gregor et al. (2002), Atkinson and Boore (2003) and Lin and Lee(2008). The GMPE by Gupta (2010) is developed specifically forIndo-Myanmar subduction zone whereas Zhao et al. (2006) and Linand Lee (2008) were developed for the subduction regions (bothintraslab and interface) of Japan and Taiwan respectively. Groundmotion prediction equations given by Gregor et al. (2002) andAtkinson and Boore (2003) were developed for Cascadia subduc-tion zone.

4. Logic tree structure

Uncertainties in the models used for seismic hazard assessmentmake the selection of a seismic hazard model difficult. The use ofthe logic tree approach allows characterization of epistemic un-certainties in various models by including alternative models in theanalysis (Budnitz et al., 1997; Stepp et al., 2001; Bommer et al.,2005). Logic tree consists of a series of nodes and branches andthese branches denote different models (hypothesis). A subjectiveweightage, based on engineering judgment, can be given to each ofthese branches depending on the likelihood of being correct.

Theweightage for all the branches at a particular node should beequal to unity. The weightage of the terminal branch of the logictree can be obtained by multiplying the weightage of all thebranches leading to it. The present study considers three types ofsource models and three different ground motion predictionequations each for various tectonic provinces in the study area. Thesource models employed are linear sources, gridded seismicitymodel and areal sources. Two different maximummagnitudes havebeen considered for linear sources; the maximum historicalmagnitude and the previous increased by 0.5 units. This is thecriterion followed and defined by Gupta (2002) for assigningmaximummagnitude to a linear source.Mmax associated with eachfault was identified separately. TheMmax andMmax þ 0.5 have beenconsidered in the analysis in a logic tree framework. All thedifferent models were combined using the logic tree.

5. Estimation of hazard

The seismic hazard evaluation of the Indian landmass based on astate-of-the art PSHA study has been performed using the classicalCornelleMcGuire approach. Regional seismic source zones werecharacterized to allow for local variability in seismicity character-istics (b-value, lm, Mmax). The seismicity parameters estimated forregional seismic source zones by Kolathayar and Sitharam (2012)

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859852

have been used as input into the present hazard analysis. Differentground motion prediction equations were used to suit various tec-tonic provinces of the country. The hazard curves obtained fromPSHAwill show the variation of peak horizontal acceleration (PHA)or spectral acceleration (Sa) againstmeanannual rateof exceedance.The occurrence of an earthquake in a seismic source is assumed tofollow a Poisson’s distribution. The probability of ground motionparameter, Z, at a given site, will exceed a specified level, z, during aspecified time, T is represented by the expression:

P�Z>z

�¼ 1� e�nðzÞT � n

�z�T (2)

where nðzÞ is the mean annual rate of exceedance of ground motionparameter, Z, with respect to z. The function nðzÞ incorporates thealeatory variability in time, size and location of future earthquakesand uncertainty in the ground motion they produce at the site. It isgiven by:

n�z� ¼

XNn¼1

Nn�m0

� Zmmax

m¼m0

fn�m�264 ZN

r¼0

fnðrjmÞPðZ>zjm; rÞdr

375dm(3)

where Nn (m0) is the frequency of earthquakes on a seismic sourcen, having a magnitude equal to or greater than a minimummagnitude m0 (in this study it is taken as 4.0); fn(m) is the proba-bility density function for a minimum magnitude of m0 and amaximum magnitude of mmax; fn(rjm) is the conditional proba-bility density function (probability of occurrence of an earthquakeof magnitudem at a distance r from the site for a seismic source n);P(Z > zjm, r) is the probability at which the ground motionparameter Z exceeds a predefined value of z, when an earthquake ofmagnitude m occurring at a distance of r from the site. The integralin Eq. (3) can be replaced by summation and the density functionsfn(m) and fn(rjm) can be replaced by discrete mass functions. Theresulting expression for nðzÞ is given by:

n�z� ¼

XNn¼1

Xmi ¼mmax

mi ¼m0

ln�mi

�24 Xrj ¼ rmax

rj ¼ rmin

Pn�R ¼ rj

��mi�P�Z>z

��mi; rj�35(4)

where lnðmiÞ is the frequency of occurrence of magnitudemi at thesource n obtained by discretizing the earthquake recurrence rela-tionship for the source n.

For doing the seismic hazard analysis using linear sources andgridded seismicity source model, a set of new programs weredeveloped in MATLAB and the entire analysis was done using theseset of programs. The hazard using areal sources was estimated byusing the software CRISIS (Ordaz et al., 2007).

Figure 5. Digital elevation model (SRTM DEM) for entire India (resampled to0.1� � 0.1� grid size).

6. Site characterization

Characteristic of local soil condition can amplify the earthquakeground shaking to a large extent, thus aggravating damage potentialof earthquake.Hence to evaluate the ground response and site effects,accurate knowledge about the geological, geomorphological andseismotectonic settings are necessary. Thus the site characterizationhas its significance in seismicmicrozonation studies. The thickness ofthe overburden soil, its density, soil type, geometry of the subsoilstructure, and the surface topography influence the amplification ofearthquake ground motion. Site characterization is generally carriedout either based on local geology or on geotechnical dynamic soilproperties. Site characterization based on local geology can be effec-tive forcharacterizing largerarea suchasstateora country.Medvedev

(1962), Borcherdt and Gibbs (1976), Shima (1978), Evernden andThomson (1985) and Midorikawa (1987) have proposed site charac-terization based on local geology and evaluated hazard intensity andamplification factors foreachgeological category.However, thismodeof site characterization is considered as primitive and not appreciatedin modern seismic microzonation practices. Hence shear wave ve-locity (Vs), which is a dynamic geotechnical property, has beenaccepted for seismic site characterization. Most of the codes such asEurocode-8 (2003), NEHRP (BSSC, 2003), International Building Code(IBC, 2009) etc. specify the site classification based on the averageshearwavevelocity values in the top30msoil column(VS30) asperEq.(1) (Table 1), where 30 m is the depth corresponding to one quarterwave length of the period of interest.

VS30 ¼ 30Pni¼1

�divi

(5)

where di is depth of each soil layer, vi is the corresponding shearwave velocity for each layer and n is the number of soil layers.

For small sites, average shear wave velocity for top 30 m can beevaluated using geotechnical and geophysical field tests.Geophysical tests include seismic cross-hole, suspensions loggingtest, seismic uphole and downhole, Spectral Analysis of SurfaceWave and Multichannel Analysis of Surface Waves while geotech-nical tests include standard penetration test (SPT) in boreholes andcone penetration test (CPT). However for characterizing large area,estimation VS30 based on these tests are not economically as well asphysically viable. Hence in the present study, site characterization

Figure 6. Slope map of India (gradient values).

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 853

for India has been carried out using topographic slope map whichwere generated from Digital Elevation Model (DEM) data.

Topographic variations are an effective index of near-surfacegeomorphology and lithology, with steep mountains represent-ing rocky terrain and flat basins indicating soil, and intermediateslopes representing a transition between rock and soil (Allen andWald, 2009). Matsuoka et al. (2005) has found good correlationsbetween VS30 and topographic slope in Japan, similar studies havebeen done by Chiou and Youngs (2006) for Taiwan and by Allenand Wald (2009) for California. In present study site character-ization for entire India has been done based on the slope map. Theslope maps were derived from DEM where the data wasdownloaded from various sources such as Cartosat DEM fromNational Remote Sensing Centre Bhuvan, Shuttle RadarTopography Mission (SRTM) from United States Geological SurveyEarth Resources Observation and Science, interpolated seamlessSRTM data v4.1 from Consultative Group on InternationalAgricultural Research-Consortium for Spatial Information(CGIAR-CSI) and void filled data in mountainous terrain from

Table 2Slope ranges for NEHRP site classes (Allen and Wald, 2009).

Class VS30 range (m/s) Slope range (m/m)

Active tectonic Stable continent

E <180 <1.0 Ee4 <2.0 Ee5D 180e240 1.0Ee4 to 2.2Ee3 2.0Ee5 to 2.0Ee3

240e300 2.2Ee3 to 6.3Ee3 2.0Ee3 to 4.0Ee3300e360 6.3Ee3 to 0.018 4.0Ee3 to 7.2Ee3

C 360e490 0.018e0.050 7.2Ee3 to 0.013490e620 0.050e0.10 0.013e0.018620e760 0.10e0.138 0.018e0.025

B >760 >0.138 >0.025

viewfinder panoramas. The absolute vertical accuracy of theelevation data will be 16 m (at 90% confidence).

6.1. Generation of slope map

Slope map for the study area was developed from DEM usingArcGIS 10 software. The downloaded DEM were found to be inWGS-1984 geographic co-ordinate system. Before proceeding toany raster analysis, the DEM was projected to UTM co-ordinatesystem which is recommended for any terrain analysis andresampled to grid size of 10 km � 10 km (Fig. 5). Using the spatialanalyst tool provided by ArcGIS, slope at each grid point wasevaluated. Slope map for the study area is presented in Fig. 6. Slopemap thus obtained was converted to vector data (points) and thenexported into a table form. The geographic co-ordinates and slopevalues were computed for each grid point.

6.2. Site characterization based on topographic slope

After preparing the slope map, site characterization was carriedout using methodology suggested by Allen and Wald (2009). Basedon the correlation studies conducted for active tectonic and stablecontinental regions, Allen and Wald (2009) proposed slope rangescorresponding to each NEHRP site class as given in Table 2.

Based on the slope range given in Table 2, site class for each gridpoint was determined. Maps showing the spatial variation in siteclasses throughout Indiawas developed and presented in Fig. 7. Thesite classes obtained for various locations in study areas were thencompared with the existing site characterization map for the samelocation generated for microzonation purpose by various re-searchers. Table 3 presents the comparison between site classes for

Figure 7. VS30 (km/s) contour map derived from the slope values.

Table 3Comparison of site classification of various cities.

Location Longitude Latitude Site class

Present study Previous studies

Bengaluru 77.6 13 Site class D Site class D (Anbazhaganand Sitharam, 2008)

Chennai 80.25 13.07 Site class D Site class D (Boominathanet al., 2008)

New Delhi 77.20 28.58 Site class D Site class D (HanumanthoRao and Ramana, 2008)

Imphal 93.95 24.80 Site class D Site class E (Kumar, 2011)

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859854

major locations obtained from present methodology and previousmicrozonation works.

7. Evaluation of PHA at ground surface

For a small area of site, ground response analysis techniques areavailable for evaluating PHA at ground surface, however for a largerlandmass such as a state or a country, this technique is not viable.Hence in this paper the seismic hazard at ground surface is esti-mated using non-linear site amplification technique proposed byRaghukanth and Iyengar (2007). The amplification factors fordifferent NEHRP site classes can be evaluated using the followingequation (Raghukanth and Iyengar, 2007):

lnðFsÞ ¼ a1ybr þ a2 þ ln ds (6)

where a1 and a2 are regression coefficients, ybr is the spectral ac-celeration at rock level and ds is the error term. The values of theregression coefficients a1 and a2 will vary for different site classesand for different time periods. These values were derived based onthe statistical simulation of ground motions (Raghukanth andIyengar, 2007) and they take into account the nonlinear siteresponse of soils. For NEHRP site classes A to D, ten random samplesof soil profiles were considered in evaluating the amplificationfactors. The amplification factor for soft and medium dense soilvaries with the rock level PHA values. The value of the dampingratio, x, and the modulus reduction, Gred, will vary with shear strain(Idriss, 1990). The method adopted for evaluation of the amplifi-cation factor (Fs) values considers this effect and the value of Fsvaries with the rock level PHA values. The value of spectral accel-eration for different site classes can be obtained from:

ys ¼ ybr Fs (7)

where Fs is the amplification factor and is the spectral accelerationat the ground surface for a given site class. The value of theamplification factor Fs varies with PHA values and it accounts forthe non-linear behaviour of soil in amplifying the seismic signals.

8. Results and discussions

The seismic hazard analysis of India was done by dividing theentire country into grids of size 0.1� � 0.1� (about 11 km � 11 km).For each grid point the Peak Horizontal Acceleration (PHA) valueswere evaluated at bedrock level corresponding to probability ofexceedance (PE) of 10% and 2% in 50 years. These exceedance valuescorrespond to return periods of 475 and 2475 years respectively.

While evaluating the seismic hazard using PSHA method, thehypocentral and magnitude uncertainties were considered. Thiswas done by deaggregating the hypocentral distance into smallintervals of 2 km and the magnitude range (betweenminimum andmaximummagnitude) into small incremental values of 0.2 (a lowervalue of the hypocentral incremental distance and magnitude in-terval was not improving the accuracy of the results significantly

Figure 8. Parameters and weighting factors adopted in the logic tree.

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 855

and it was increasing the computation time). For each grid point, allthe sources within a radius of 300 km were considered for evalu-ation of PHA and Sa values. For the plate boundary regions, sincethe earthquake magnitudes were higher, this distance wasincreased to 500 km. The PHA value at each grid point was esti-mated using 24 different methods (various combinations of groundmotion prediction equations; source models and different ap-proaches of Mmax estimation) and later combined to get the finalvalue using the logic tree framework as shown in Fig. 8.

The spatial variation of PHAvalues obtained for 10% and 2% PE in50 years are shown in Figs. 9 and 10. PHA values at the rock level for10 most populous cities of India are given in Table 4. It can be seenthat the seismic hazard is high along the plate boundary regionsviz. North and Northeast India. Along the shield region, the highesthazard is observed at Bhuj region and in Koyna region. For 475years return period, the PHA values for plate boundary regions vary

from 0.2 to 0.4 g where as for shield region the values were lessthan 0.2 g except for the Kutch region in Gujarat. The Indian seismiccode BIS-1893 (2002) divides the country into four zones, viz. ZonesII, II, IV and V. The maximum expected accelerations in each ofthese zones are 0.1, 0.16, 0.24 and 0.36 g, respectively. The PHAvalues obtained in this study for Northeast India and most parts ofJammu and Kashmir are higher.

Using appropriate amplification factors corresponding to siteclasses, we have generated a single map for entire India showingthe spatial variation of surface level PHA. Fig. 11 presents the PHAmaps for 475 years and 2475 years respectively. It can be seen thatPHA values got significantly amplified in regions those belong toSite Class C. Most part of India belongs to Site class D and regions oflow PHAvalues at bedrock level showed significant amplification atsurface level. Regions of high PHA values at bedrock level (e.g.,Indo-Myanmar Subduction region), there is a de-amplification of

Figure 9. PHA values (g) corresponding to a return period of 475 years (10% probability of exceedance in 50 years).

Figure 10. PHA values (g) corresponding to a return period of 2475 years (2% probability of exceedance in 50 years).

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859856

Table 4PHA values at rock level for 10 most populous cities of India.

Major cities Location PHA value (g)

Longitude (�E) Latitude (�N) For returnperiod of475 years

For return periodof 2475 years

Mumbai 72.82 18.90 0.10 0.19Delhi 77.20 28.58 0.27 0.51Bangalore 77.59 12.98 0.13 0.22Kolkata 88.33 22.53 0.13 0.23Chennai 80.25 13.07 0.13 0.21Hyderabad 78.48 17.38 0.01 0.19Ahmedabad 72.62 23.00 0.10 0.18Pune 73.87 18.53 0.07 0.13Kanpur 80.40 26.47 0.11 0.17Jaipur 75.87 26.92 0.12 0.2

Table 5PHA values at rock level for 10 most populous cities of India.

Location PHA (g)

Present study Previous studies

Mumbai 0.10 0.1Seeber et al. (1999)

0.13Raghukanth andIyengar (2006)

Delhi 0.24 0.23Iyengar and Ghosh (2004)

e

Bengaluru 0.11 0.06 to 0.1Jaiswal and Sinha (2007)

0.13Sitharam andVipin (2011)

Kolkata 0.13 0.08 to 0.13Giardini et al. (1999)

0.22 to 0.24Mohanty andWalling (2008)

Chennai 0.10 0.09Menon et al. (2010)

0.08Jaiswal andSinha (2007)

Ahmedabad 0.09 0.07Singh (2009)

e

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 857

the PHA at the surface level. The present surface level PHA valuesmatch well with those obtained for various cities whose micro-zonation studies have been carried out earlier (Table 5). The PHAobtained at ground level using this approach is a simple methodand first approximation which may be used for preliminary design.It is to be noted that several factors affect the surface level accel-eration depending on the soil conditions such as effect of soillayering which need to be explored for a site specific study(Raghunandan, 2012) (Table 6).

9. Conclusions

The seismic hazard evaluation of the Indian landmass based on astate-of-the art PSHA study has been performed using the classicalCornelleMcGuire approach with different source models andgroundmotion prediction equations. Themost recent knowledge of

Figure 11. Spatial variation in PHA value at surface level

seismic activity in the region has been used to evaluate the hazardincorporating uncertainty associated with different modeling pa-rameters as well as spatial and temporal uncertainties. The PSHAhas been performed with currently available data and their bestpossible scientific interpretation using an appropriate instrumentsuch as the logic tree to explicitly account for epistemic uncertaintyby considering alternative models (source models, maximummagnitude in hazard computations, and ground motion predictionequation ships).

The hazard maps have been produced for horizontal groundmotion at bedrock level (Shear wave velocity � 3.6 km/s). Theevaluation of surface level PHA value is of very high importance inthe engineering design. In this paper, the surface level PHA values

throughout India (corresponding to 475 year, PSHA).

Table 6Comparison of Surface level PHA values with earlier studies.

Major cities Location PHA (g)

Longitude(�E)

Latitude(�N)

Present PSHA(475 year returnperiod)

Previous PSHA studies

Gangtok 88.4 27.2 0.702 0.6 (Nath et al., 2008)Delhi 77.21 28.61 0.42 >0.24 (Iyengar and

Ghosh, 2004)Bangalore 77.6 13.0 0.3 0.3 to 0.45 (Anbazhagan

et al., 2009)Guwahati 91.6 26.2 0.4 0.45 (Nath et al., 2008)Chennai 80.2 13.1 0.28 0.141 (Boominathan

et al., 2008)Imphal 93.95 24.8 0.63 0.5 (Kumar, 2011)

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859858

were evaluated for the entire study area for four NEHRP site classesby considering the site conditions. VS30 characterization of entirecountry based on the topographic gradient is presented here, usingexisting correlations connecting slope and VS30. The PHA value wasbrought to the surface level using appropriate amplification factors.Thus this method provides a simplified methodology for evaluatingthe surface level PHA values. The present study shows that theseismic hazard is moderate in peninsular shield, but the hazard inthe North and Northeast India are extremely high. The PHA valuesobtained in this study for Northeast India and most parts of Jammuand Kashmir are higher than what is specified by BIS-1893 (2002).The ground motion predicted from the present study will not onlyaid in safe design of structures, but also will help in deciding thelocations of important structures such as nuclear power plants.

Reliable information on the active faults and geodetic mea-surements of fault movement rates would go a long way in rede-fining seismogenic sources in India. Strong ground-motion recordsof earthquakes in the region would be essential to develop regionspecific attenuation equations for the various parts of the country.The updated and reliable information on these factors would go along way in redefining the seismic hazard of the region in a betterway. The limitation of this work is that the amplification effects dueto surface topography and basin geometry were not considered.The present work characterizes the country into various site classesbased on VS30 estimated from correlations with topographicalgradient. Site characterization of the region by performinggeophysical investigations will give a better picture of site classi-fication which is a huge task that needs to be undertaken at na-tional level.

References

Akkar, S., Bommer, J.J., 2010. Empirical equations for the prediction of PHA, PGV, andspectral accelerations in Europe, the Mediterranean region, and the MiddleEast. Seismological Research Letters 81, 195e206.

Algermissen, S.T., Perkins, D.M., 1976. A probabilistic estimate of maximum accel-eration in rock in the contiguous United States. U.S. Geol. Surv. Open File Rep.,pp. 76e416, 45 pp.

Allen, T.I., Wald, D.J., 2009. On the use of high-resolution topographic data as aproxy for seismic site conditions (VS30). Bulletin of the Seismological Society ofAmerica 99, 935e943.

Anbazhagan, P., Sitharam, T.G., 2008. Mapping of average shear wave velocity forBangalore region: a case study. Journal of Environmental & EngineeringGeophysics 13 (2), 69e84.

Anbazhagan, P., Vinod, J.S., Sitharam, T.G., 2009. Probabilistic seismic hazard anal-ysis for Bangalore. Journal of Natural Hazards 48, 145e166.

Atkinson, G.M., Boore, D.M., 2003. Empirical ground-motion relations forsubduction-zone earthquakes and their application to Cascadia and other re-gions. Bulletin of Seismological Society of America 93, 1703e1729.

Atkinson, G.M., Boore, D.M., 2006. Earthquake ground-motion prediction equationsfor Eastern North America. Bulletin of the Seismological Society of America 96(6), 2181e2205.

Bhatia, S.C., Kumar, M.R., Gupta, H.K., 1999. A probabilistic seismic hazard map ofIndia and adjoining regions. Annali di Geofisica 42, 1153e1166.

BIS-1893, 2002. Indian Standard criteria for earthquake resistant design of struc-tures, Part 1-general provisions and buildings. Bureau of Indian Standards NewDelhi.

Bodin, P., Malagnini, L., Akinci, A., 2004. Ground-motion scaling in the KachchhBasin, India, deduced from aftershocks of the 2001 MW 7.6 Bhuj earthquake.Bulletin of the Seismological Society of America 94, 1658e1669.

Bommer, J.J., Scherbaum, F., Bungum, H., Cotton, F., Sabetta, F., Abrahamson, N.A.,2005. On the use of logic trees for ground-motion prediction equations inseismic hazard analysis. Bulletin of Seismological Society of America 95,377e389.

Bommer, J.J., 2003. Uncertainty about the uncertainty in seismic hazard analysis.Engineering Geology 70 (1/2), 165e168.

Boominathan, A., Dodagoudar, G.R., Suganthi, A., Uma Maheswari, R., 2008. Seismichazard assessment of Chennai city considering local site effects. Journal of EarthSystem Sciences 117 (S2), 853e863.

Boore, D.M., Atkinson, G.M., 2008. Ground-motion prediction equations for theaverage horizontal component of PHA, PGV, and 5%-damped PSA at spectralperiods between 0.01 s and 10.0 s. Earthquake Spectra 24, 99e138.

Borcherdt, R.D., Gibbs, J.F., 1976. Effects of local geological conditions in the SanFrancisco Bay region on ground motions and the intensities of the 1906earthquake. Bulletin of the Seismological Society of America 66, 467e500.

BSSC, 2003. The 2003 NEHRP Recommended Provisions for New Buildings andOther Structures, Part 2: Commentary (FEMA 450). Building Seismic SafetyCouncil, National Institute of Building Sciences, Washington, DC, U.S.A.

Budnitz, R.J., Apostolakis, G., Boore, D.M., Cluff, L.S., Coppersmith, K.J., Cornell, C.A.,Morris, P.A., 1997. Recommendations for Probabilistic Seismic Hazard Analysis:Guidance on Uncertainty and Use of Experts. U.S. Nuclear Regulatory Com-mission Report NUREG/CR- 6372. Census of India.

Campbell, K.W., Bozorgnia, Y., 2003. Updated near-source ground motion (attenu-ation) relations for the horizontal and vertical components of peak groundacceleration and acceleration response spectra. Bulletin of the SeismologicalSociety of America 93, 314e331.

Chandra, U., 1977. Earthquakes of Peninsula India-a seismotectonic study. Bulletin ofthe Seismological Society of America 67, 1387e1413.

Chiou, B., Youngs, R.R., 2006. ChioueYoungs PEER-NGA Empirical Ground MotionModel for the Average Horizontal Component of Peak Acceleration and Pseudo-spectral Acceleration for Spectral Periods of 0.01 to 10 Seconds. PEER ReportDraft. Pacific Earthquake Engineering Research Center, Berkeley, CA, 219 pp.

Cramer, C.H., Kumar, A., 2003. 2001 Bhuj, India, earthquake engineering seismo-scope recordings and Eastern North America ground motion attenuation re-lations. Bulletin of the Seismological Society of America 93, 1390e1394.

Deshikachar, S.V., 1974. A review of the tectonic and geological history of easternIndia in terms of ‘‘plate tectonics’’ theory. Journal of Geological Society of India15, 137e149.

Eurocode-8, 2003. BS-EN 1998-1, Design of structures for earthquakeresistance epart 1: General rules, seismic actions and rules for buildings. European Com-mittee for Standardization, Brussels.

Evans, P., 1964. The tectonic framework of Assam. Journal of Geological Society ofIndia 5, 80e96.

Evernden, J., Thomson, J.M., 1985. Predicting Seismic Intesnsities. U S Geol. Survey.Paper 1360.

Frankel, A., 1995. Mapping seismic hazard in the Central eastern United States.Seismological Research Letters 66 (4), 8e21.

Gregor, N.J., Silva, W.J., Wong, I.G., Youngs, R., 2002. Ground motion attenuationrelationships for Cascadia subduction zone mega-thrust earthquakes based on astochastic finite-fault modeling. Bulletin of the Seismological Society of Amer-ica 92, 1923e1932.

Gupta, I.D., 2006. Delineation of probable seismic sources in India and neighbor-hood by a comprehensive analysis of seismotectonic characteristics of the re-gion. Soil Dynamics and Earthquake Engineering 26, 766e790.

Gupta, I.D., 2010. Response spectral attenuation relations for intraslab earthquakesin Indo-Burmese subduction zone. Soil Dynamics and Earthquake Engineering30, 368e377.

Gupta, I.D., 2002. The state of the art in seismic hazard analysis. ISET Journal ofEarthquake Technology 39 (4), 311e346.

Hanumantho Rao, C., Ramana, G.V., 2008. Dynamic soil properties for micro-zonation of Delhi, India. Journal of Earth System Science 117 (S2), 719e730.

IBC, 2009. International Building Code. International Code Council, Washington.Idriss, I.M., 1990. Response of soft soil sites during earthquakes. In: Duncan, J.M.

(Ed.), Proc. H. Bolton Seed Memorial Symposium, vol. 2, pp. 273e290.Iyengar, R.N., Ghosh, S., 2004. Microzonation of earthquake hazard in greater Delhi

area. Current Science 87, 1193e1202.Iyengar, R.N., Chadha, R.K., Balaji Rao, K., Raghu Kanth, S.T.G., 2010. Development of

Probabilistic Seismic Hazard Map of India. Technical Report of the WorkingCommittee of Experts (WCE) Constituted by the National Disaster ManagementAuthority. Govt. of India, New Delhi.

Jaiswal, K., Sinha, R., 2007. Probabilistic seismic-hazard estimation for peninsularIndia. Bulletin of the Seismological Society of America 97 (1B), 318e330.

Joyner, W.B., Boore, D.M., 1981. Peak horizontal acceleration and velocity fromstrong-motion records including records from the 1979 Imperial Valley,California, earthquake. Bulletin of the Seismological Society of America 71,2011e2038.

Kayal, J.R., 2008. Microearthquake Seismology and Seismotectonics of South Asia,first ed. Springer Verlag and Capital Publishing Company, India, ISBN 978-1-4020-8179-8.

T.G. Sitharam et al. / Geoscience Frontiers 6 (2015) 847e859 859

Khatri, K.N., M Rogers, A., Algermissen, S.T., 1984. A seismic hazard map of India andadjacent areas. Tectonophys 108, 93e134.

Kijko, A., 2004. Estimation of the maximum earthquake magnitude, mmax. Pureand Applied Geophysics 161, 1655e1681.

Kolathayar, S., Sitharam, T.G., Vipin, K.S., 2012. Spatial variation of seismicityparameters across India and adjoining areas. Natural Hazards 60 (3),1365e1379.

Kolathayar, S., Sitharam, T.G., 2012. Characterization of regional seismic sourcezones in and around India. Seismological Research Letters 83 (1), 77e85.

Kumar, P., 2011. Seismic Microzonation of Imphal City and Probalisitc SeismicHazard Assessment of Manipur State. PhD Thesis. Indian Institute of Technol-ogy, Guwahati.

Kumar, P., Yuan, X., Ravi Kumar, M., Kind, R., Li, X., Chadha, R.K., 2007. The rapid driftof Indian tectonic plate. Nature 449, 894e897.

Lin, P.S., Lee, C.T., 2008. Ground-motion attenuation relationships for subduction-zone earthquakes in Northeastern Taiwan. Bulletin of the Seismological Soci-ety of America 98, 220e240.

Martin, C., Secanell, R., Combes, Ph, Lignon, G., 2002. In: 12th European Conferenceon Earthquake Engineering, Paper Ref. 870. Elsevier Science Ltd.

Matsuoka, M., Wakamatsu, K., Fujimoto, K., Midorikawa, S., 2005. Nationwide siteamplification zoning using GIS-based Japan engineering geomorphologic clas-sification map. In: International Conference on Structural Safety and Reliability,9, Rome, Italy, June 19e23, 2005, Proceedings. Millpress, Rotterdam,Netherlands, pp. 239e246.

Medvedev, S.V., 1962. Engineering Seismology. Institute Fiziki Zemli Akad. NaukSSSR, Moscow, 260 pp. [In Russian, English translation, 1965]. (Jerusalem: IsraelProgramme for ScientificTranslations).

Midorikawa, S., 1987. Prediction of iso-seismal map in the Kanto Plain due to hy-pothetical earthquake. Journal of Structural Engineering 33B, 43e48.

Mohanty, W.K., Walling, M.Y., 2008. First order seismic microzonation of Haldia,Bengal basin (India) using a GIS platform. Pure and Applied Geophysics 165,1325e1350.

Nath, S.K., Thingbaijam, K.K.S., 2011. Peak ground motion predictions in India: anappraisal for rock sites. Journal of Seismology 15 (2), 295e315. http://dx.doi.org/10.1007/s10950-010-9224-5.

Nath, S.K., Thingbaijam, K.K.S., Raj, Abhishek, 2008. Earthquake hazard in NortheastIndia e a seismic microzonation approach with typical case studies from Sik-kim Himalaya and Guwahati city. Journal of Earth System Science 117 (S2),809e831.

Nath, S.K., 2006. Seismic Hazard and Microzonation Atlas of the Sikkim Himalaya,Published by Department of Science and Technology. Government of India, NewDelhi.

Ni, J., Barazangi, M., 1986. Seismotectonics of the Zagros continental collision zoneand a comparison with the Himalayas. Journal of Geophysical Research 91,8205e8218.

Ordaz, M., Aguilar, A., Arboleda, J., 2007. CRISIS2007 e Ver. 1.1: Program forComputing Seismic Hazard. Instituto de Ingenieria, UNAM, Mexico.

Raghukanth, S.T.G., Iyengar, R.N., 2006. Seismic hazard estimation for Mumbai city.Current Science 91 (11), 1486e1494.

Raghukanth, S.T.G., Iyengar, R.N., 2007. Estimation of seismic spectral accelerationin Peninsular India. Journal of Earth System Sciences 116 (3), 199e214.

Raghunandan, M.E., 2012. Effect of soil layering on the ground response parame-ters: a parametric study (2012). Natural Hazards 63, 1115e1128.

Rao, B.V., Murty, B.V.S., 1970. Earthquakes and tectonics in peninsular India. Journalof Indian Geophysical Union 7, 1e8.

SEISAT, 2000. Seismotectonic Atlas of India, Geological Survey of India. New Delhi.Sharma, M.L., Malik, S., 2006. Probabilistic seismic hazard analysis and estimation of

spectral strong ground motion on Bed rock in North East India. In: 4th Int. Conf.Earthquake Engineering, Taipei, Taiwan, Oct. 12e13, 2006.

Sharma, M.L., Douglas, J., Bungum, H., Kotadia, J., 2009. Ground-motion predictionequations based on data from the Himalayan and Zagros regions. Journal ofEarthquake Engineering 13, 1191e1210.

Shima, E., 1978. Seismic Microzoning Map of Tokyo. Proc. Second Inter. Conf. onMicrozonation (1), pp. 433e443.

Sitharam, T.G., Vipin, K.S., 2011. Evaluation of spatial variation of peak horizontalacceleration and spectral acceleration for South India: a probabilistic approach.Natural Hazards 59 (2), 639e653.

Stepp, J.C., Wong, I., Whitney, J., Quittemeyer, R., Abrahamson, N., Toro, G.,Youngs, R., Coppersmith, K., Savy, J., Sullivan, T., 2001. Yucca Mountain PSHAproject members, probabilistic seismic hazard analyses for ground motions andfault displacements at Yucca Mountain, Nevada. Earthquake Spectra 17,113e151.

Vipin, K.S., Anbazhagan, P., Sitharam, T.G., 2009. Estimation of peak ground accel-eration and spectral acceleration for South India with local site effects: prob-abilistic approach. Natural Hazards Earth System Science 9, 865e878.

Wiemer, S., 2001. A software package to analyze seismicity: zmap. SeismologicalResearch Letters 72 (2), 374e383.

Woo, G., 1996. Kernel estimation methods for seismic hazard area source model.Bulletin of Seismological Society of America 86 (2), 253e362.

Zhao, J.X., Zhang, J., Asano, A., Ohno, Y., Oouchi, T., Takahashi, T., Ogawa, H.,Irikura, K., Thio, H.K., Somerville, P.G., Fukushima, Y., 2006. Attenuation re-lations of strong ground motion in Japan using site classification based onpredominant period. Bulletin of the Seismological Society of America 96,898e913.