Geomatics, Natural Hazards and Risk, 2015

Vol. 6, No. 1, 1–20, http://dx.doi.org/10.1080/19475705.2013.815282

An engineering model for seismicity of India

S. JAYALAKSHMI and S.T.G. RAGHUKANTH*

Department of Civil Engineering, Indian Institute of Technology, Madras 600036, India

(Received 5 March 2013; in final form 11 June 2013)

This article explores an engineering approach to model the seismic activity in

India. Finite element analysis is carried out to estimate the stresses and displace-
ments in the Indian plate. Assuming the Himalayan boundary as fixed, the
plate-driving forces are modelled as axial forces applied at the mid-oceanic ridge
between the Indian and African plates. The effect of Aravali, Dharwar and
Bundelkhand cratons on the stress patterns in the plate is also studied. The
obtained results are validated to the extent possible with the available recorded
seismicity data and measurements from Global Positioning System (GPS).

1. Introduction
India broadly consists of three distinct geological units, namely the Himalaya,
the Indo-Gangetic plain and the Indian shield (figure 1). It is well established that
Himalaya is a result of collision between the Indian and the Eurasian plates about
50–60 million years ago (Valdiya 1998). Due to the continuous under-thrusting of
the Indian plate beneath the Eurasian plate, stresses are increasing and accumulating
progressively in the Himalayas. This makes the Himalaya seismically active com-
pared to other geological units. Some of the notable earthquakes that have occurred
in the Himalayan region are the 1897 Shillong earthquake (Mw ¼ 8.7), 1905 Kangra
earthquake (Mw ¼ 7.8), 1934 Bihar–Nepal earthquake (Mw ¼ 8.3), 1950 Assam
earthquake (Mw ¼ 8.7) and the 2005 Muzzaffarabad earthquake (Mw ¼ 7.6). There
are also reports of three major gaps in Himalayan region, namely Kashmir gap,
Garhwal gap and the Assam gap where large events have not occurred in the past
500 years (Khattri & Tyagi 1983; Khattri 1987; Bilham et al. 2001). These gap
regions have generated a lot of interest among seismologists and earthquake engi-
neers to know the seismic potential of these gaps. Although, Himalaya is regarded as
a plate boundary and an active region, the seismicity database indicates that there
are regions in the Indian shield reporting sporadic seismic activity. These events
known as intraplate earthquakes have occurred in the past in Gujarat and peninsular
India (PI). Several investigators attempted to explain the seismic activity in the
Indian plate based on seismicity data and focal mechanism solutions of past earth-
quakes (Ni & Bazrangi 1984; Rao & Rao 1984; Bilham & Gaur 2000; Rajendran
et al. 2004; Lav
e et al. 2005; Kayal 2008). The evolution of precise GPS
geodesy has made it possible to understand the dynamics of Indian plate and also
measure the accumulated strain in India. Several groups have conducted GPS studies

*Corresponding author. Email: [email protected]

Ó 2013 Taylor & Francis

2 S. Jayalakshmi and S.T.G. Raghukanth

Figure 1. Tectonic map of India.

in different regions of the Indian subcontinent (Freymueller et al. 1996; Bilham et al.
1998; Paul et al. 2001). Jade (2004) analysed the GPS data and reported convergence
rate of 10–20 mm/year along the Himalayan arc. Past efforts to study deformation
of Peninsular India (PI) used GPS data from two sites only, namely Hyderabad
(HYDE) and Indian Institute of Science, Bangalore (IISC) (Sella et al. 2002; Betti-
nelli et al. 2006). Hence these studies were confined to estimate the plate motion at
global scale and relative motion between two plates at plate boundaries. Jade et al.
(2007) used GPS data from five sites within the plate interior from 2003 to 2006. The
data from two sites, namely IISC at Bangalore and DELH at Delhi, were used for
the period from 1997/98 to 2006. All the sites used by them are located along a
north–south transect and there is not much coverage in the east–west direction.
Banerjee et al. (2008) used the most comprehensive data set from 12 GPS sites
located in the Indian plate interior. Along with the estimation of the Euler pole for
the Indian plate motion, they reported that the almost east–west trending Narmada
Son failed rift zone accommodates about 2  1 mm/year of shortening in the
 Geomatics, Natural Hazards and Risk 3

Figure 2. Indian plate boundary modified from Bird (2003) and the location of GPS sites in
the Indian reference frame.

north–south direction. Recently, Mahesh et al. (2012) used the data from more than
26 GPS sites (figure 2) in the Indian plate to constrain the motion of the plate and
assessed the stability of the Indian plate and deformation in the boundary regions.
Although these data provided numerous insights into the deformation at some loca-
tions in the Indian plate, a mechanistic approach to model tectonic plates, would be
more effective in identification of seismic gaps and occurrence of future earthquakes
in the Indian plate. There have been efforts in the past three decades to model the tec-
tonic plates by mechanistic approaches on both global and regional scales (Coblentz
et al. 1995; Dyksterhuis et al. 2005; Hieronymus et al. 2008). These studies have pro-
vided valuable information to understand the lithosphere and they have also raised
awareness for many unresolved issues like understanding of the forces driving plate
tectonics. Although, mantle convection is responsible for movement for tectonic
plates, no convection model till date has been able to reproduce the plate tectonics.
The tectonic forces arise due to the density contrasts within the crust and the mantle.
The driving forces on tectonic plates have been categorized into two: forces
acting at the bottom surface of plate and forces acting at plate boundaries
4 S. Jayalakshmi and S.T.G. Raghukanth

(Forsyth & Uyeda 1975). Apart from studies on forces driving tectonic plates,
researchers have also worked on plate models to compute plate velocities (DeMets
et al. 1990). The forces/velocities driving the plate tectonics that should be reflected
in a mechanistic (dynamic/kinematic) model are thus known. Development of a suit-
able dynamic model for Indian plate is likely to provide better understanding the
occurrence of earthquakes. In the present work, a two-dimensional (2D) finite ele-
ment model of the Indian plate is presented.

2. Tectonic setting of India

The seismotectonic map of India is shown in figure 1. Extending about 2500 km from
Kashmir to Arunachal, the Himalayan mountain system comprises four geologically
distinct parallel sections, namely the Outer Himalaya or Siwalik, the Lesser
Himalaya or Himachal, the Greater Himalaya or Himadri and the Tethys Himalaya.
The width of the entire Himalaya varies 250–300 km from south to north (Valdiya
1980). It is now well established that Himalaya is a result of collision between the
Indian and the Eurasian plates about 50–60 million years ago (Valdiya 2010). The
plate collision process gave rise to several active faults in the Himalayan region,
namely the Main Central Thrust (MCT), the Main Boundary Thrust (MBT) and the
Himalayan Frontal Thrust (Valdiya 1976). The focal mechanism solutions over these
faults show thrust-type faulting (Tandon & Srivastava 1975; Gupta 2006; Kayal
2008), which indicates under-thrusting of the Indian plate beneath the Eurasian
plate; therefore, the stresses are increasing and accumulating progressively in the
Himalayas. This makes the Himalaya seismically very active compared to other geo-
logical units. In the north-east, the Himalaya takes a bend and meets the Indo-
Burmese arc. This eastern extremity of the Himalaya is known as the Assam syntaxis.
Similarly, the turning point of the Himalayan mountain system in the north-western
region is known as the Hazara syntaxis, where the north-west end of the Himalaya
meets the Pamir-Hindukush region. The syntaxis zones are regions of high stress con-
centration and seismic activity. The Chaman fault in the west and Sagaing fault in
the east cut the Himalayan mountain chain at its two extremities. These faults are
further linked to much longer and fundamental faults of the Indian ocean, and they
act as links in transmitting the movements of the oceanic faults to the Indian plate.
The Indus Tsangpo suture zone (ITSZ) is considered to be the plate boundary where
the Tethys Ocean was consumed by the subduction process. The Main Karakoram
Thrust (MKT) marks the southern boundary of the Hindu Kush and the
Karakoram. The curvilinear ITSZ and MCT are traced to the south of MKT. The
Main Mantle Thrust (MMT) in the Hazara syntaxis is the western extension of
the MCT. The Chaman fault joins the Herat fault and the two bend eastward and
split into the Karakoram and the Altyn Tagh fault systems in the Pamir region
(figure 1). The Balochistan arc, comprising Sulaiman and Kirthar ranges, is aligned
in a north–south direction.
The Indo-Gangetic plain lies between PI and the Himalayas. This region is charac-
terized by several hidden faults and ridges in the basement of the Ganga basin
(Gansser 1974; Valdiya 1976). These faults have oblique and transverse alignments
across the Himalayan tectonic trend. Gansser (1974) pointed out that Gangetic plain
is not a sediment-filled foredeep, and it represents the depressed part of the peninsu-
lar shield in which several hidden faults exist. The earthquake activity in the Gangetic
 Geomatics, Natural Hazards and Risk 5

plain is broadly associated with strike-slip faulting (Gupta 2006). The Gangetic
plain is moderately seismic when compared to the Himalaya (Quittmeyer & Jacob
1979).
The Indian shield is a land mass that lies far away from the collision zone and is
believed to be relatively stable. It is composed mainly of four rigid cratonic features,
namely Dharwar craton, Bastar craton, Bundelkhand craton and the Singhbum cra-
ton (Valdiya 2010). Among various tectonic structures in PI, the Cambay and Kutch
regions in Gujarat are very active regions. Apart from these two regions, the remain-
ing part of shield region has reliably experienced some 400 earthquakes of magnitude
greater than 3 in a period of 600 years (Rao & Rao 1984). The seismicity of PI, from
a seismological perspective, was discussed in the past, notably by Chandra (1977)
and Rao and Rao (1984). It is generally held that seismic activity is more frequent at
the intersections of the Dharwar, Aravali and Singhbhum protocontinents, which
together constitute the shield region (Iyengar & Raghukanth 2004). These three pro-
tocontinents are separated by rifts. The earthquakes in PI can be classified into rift
and non-rift events. The 1967 Koyna earthquake and 1993 Khillari earthquake are
examples of non-rift events, whereas the 1997 Jabalpur and 2001 Kutch earthquakes
are rift events. The most striking feature in the Indian shield is the Son–Narmada–
Tapti (SONATA) rift zone, which is an East Northeast (ENE)-West Southwest
(WSW) trending zone and runs across the Indian shield from west coast to east
coast.This rift zone, about 1600 km long, separates the northern and southern blocks
of the Indian shield and is a region of moderate seismic activity with infrequent
earthquakes. In southern India, sporadic and low-level seismicity is observed along
the old shear zones. The faults associated with the Godavari graben, namely the
Kaddam fault and the Gundlakamma fault near Ongole on the coast, which trends
North West (NW)-South East (SE), are regarded as the moderately active faults in
south India. These faults separate the Singhbum and Dharwar protocontinents.

3. Finite element model

A brief review of seismotectonic setting of India has been presented above. It remains
to develop a mechanistic model incorporating the geological and tectonic features in
the Indian plate. Although a three-dimensional (3D) model incorporating all the
geological features, lateral variation of material properties, mantle, topography and
tectonics is desirable, difficulties exist in modelling the faults. In the present study,
Indian plate is modelled using 2D plane stress elements. The plate boundary
coordinates of Indian plate are taken from the work of Bird (2003). This is shown in
figure 2. The northern part of the Indian plate boundary passes through the Kirthar,
Sulaiman range and the MBT, and it has been established that MBT marks the visi-
ble divide of India with Eurasia (Valdiya 1998; Bird 2003). However, many evolu-
tionary models proposed for the formation of Himalaya (Molnar et al. 1977; Seeber
& Armbruster 1981; Ni & Barazangi 1984) suggest that the convergence of India
with Eurasia took place along the ITSZ. Although ITSZ has become dormant with
regard to seismicity, it can still be considered as the border of Indian plate. The
boundary has been extended up to the ITSZ to study the stress patterns in the
Himalaya beyond MCT. In the Indian shield region, cratonic features are incorpo-
rated in the model to understand the effect of the cratons on the pattern of stresses
and deformations inside the plate. The cratons included in the analysis are Dharwar
6 S. Jayalakshmi and S.T.G. Raghukanth

craton, Bastar craton, Bundelkhand craton, Singhbhum craton and the Meghalaya
craton (Valdiya 2010). The Son–Narmada rift valley demarcates the boundary of the
Bundelkhand craton against the Bastar and Singhbhum cratons. The Singhbhum
craton is separated from the Bastar craton by the Mahanadi graben. The Godavari
rift valley forms the dividing line between the Bastar and Dharwar cratons. The
southern boundary of the Dharwar craton against the Southern Granulite Terrane is
formed by a system of shear zones, namely the Moyar–Bhavani Shear Zone in the
north and the Palghat–Cauvery Shear Zone in the south. The locations of these
five cratons are modified from those mentioned in Valdiya (2010) based on the geom-
etry of faults given in GSI (2000). These cratons are shown in figure 2.
After modelling the important geological features in the Indian plate, the next
step is to identify the plate-driving forces. The Indian plate boundary includes the
continental collision boundary along the Himalayas between the Indian and the
Eurasian plate, the divergent margins along the mid-oceanic ridge bordering
between Indian and African plates and also the subduction zone in the eastern
edge. The India–Eurasia continental collision has remarkably decreased the rela-
tive motion between the Indian plate and the Eurasian plate. The phenomenal
drop in the continental convergence rate (from 100–180 mm/year to 50 mm/year)
highlights the importance of the plate-boundary forces in driving tectonic plates
(Molnar & Tapponnier 1975). Lithospheric plates created at the ocean ridges are
subjected to a body force that causes the plates on either side of the ridge to move
apart from the ridge crest. This is a component of the gravitational body force
that drives the plate and is known as ridge push force (Forsyth & Uyeda 1975). As
the lithosphere moves away from the ocean ridges, it undergoes thermal contrac-
tion and becomes thicker and denser. This makes the old oceanic lithosphere grav-
itationally unstable compared to the hot underlying mantle rocks. As a result, the
dense lithosphere sinks into the mantle at the ocean trenches. The subduction con-
tinues until the lithosphere remains denser than the adjacent mantle rocks at any
depth. The body force acting on the descending plate is transmitted to the surface
plate, which is pulled towards the ocean trench. This force also contributes to the
plate tectonics and is known as slab pull. However, the estimation of slab pull
requires better understanding of the causative agents (Scholz & Campos, 1995).
The slab pull force is not incorporated in the present study due to the lack of
understanding of the nature of this force at Indo-Burmese and Andaman region.
In the analysis, the ridge push force is used to estimate the stresses and deforma-
tions in the Indian plate. The expression for ridge push is given by the relation
(Turcotte & Schubert 2002),

 
2r av ðTm  T0 Þ
FRP ¼ rm gav ðTm  T0 Þ 1 þ m kt; ð1Þ
pðrm  rw Þ

where rm is the density of the mantle (3300 kg/m3), g is the acceleration due to
gravity (10 m/sec2), av is the thermal expansion coefficient (3  105/K), (Tm  T0)
is the temperature difference between the mantle and the surface (1200 K), rw is
the density of water (1000 kg/m3), thermal diffusivity (k) can be taken as 1 mm2/
sec and t is the age of lithosphere in seconds. It can be observed from equation (1)
that the ridge push force is directly proportional to the age of lithosphere.
 Geomatics, Natural Hazards and Risk 7

3.1. Modelling procedure

The finite element modelling of the Indian plate is carried out using ABAQUS 6.10.1
software. ABAQUS is widely used in automotive, aerospace and industrial applica-
tions. The software is user friendly and has wide material modelling capabilities. The
complete ABAQUS environment (CAE) program was used for the creation and
analysis of the model. ABAQUS is divided into several modules for defining the
geometry, material properties, mesh and also for analysing and monitoring the
results.

3.2. Parts, properties and section

The dimensions of the Indian plate are not subjected to much variation when going
from Earth’s spherical geometry to a flat surface. Therefore, the latitudes and longi-
tudes are converted to equivalent Cartesian coordinates in kilometre. The boundary
of the Indian plate shown in figure 2 is plotted in the part module after converting
the latitudes and longitudes into Cartesian coordinates (kilometres). Since the cra-
tons being relatively stiff compared to other parts of the lithosphere, these geological
units are modelled as regions of large thickness. The cratons are created as
‘partitions’ within the model. The modelling is done assuming plane stress condition
because the thickness of the plate is relatively small compared to the in-plane dimen-
sions. The model has homogeneous solid section of 60 km thickness at the continen-
tal crust (Manglik et al. 2008). The thickness of the cratons is fixed by trial and
error. Homogeneous mechanical properties are assumed throughout the Indian
plate. Young’s modulus and Poisson’s ratio for rock are taken as 75 GPa and 0.25,
respectively (Manglik et al. 2008).
The boundary condition along the continental collision Himalayan boundary can
be hinged or fixed in the model with regard to the fact that the displacements near
the boundary are very small compared to those far away from the boundary. The
assumption of plane stress conditions does not alter the results for both hinged and
fixed boundary conditions. Hence in the analysis, the continental collision boundary
is fixed along the Himalayas.

3.3. Mesh
The Indian plate is meshed predominantly using the four-node bilinear plane stress
quadrilateral elements (CPS4R) and three-node plane stress triangular elements
(CP3R) with reduced hourglass control. The advantage of using these elements is
that in addition to the removal of shear locking by reduced integration, it also con-
trols the element from undergoing spurious modes of deformation. Each node of the
element will have two degrees of freedom (displacements along x and y directions).
ABAQUS allows to fix the size of the ‘seed’ (points on the edges of the plate to spec-
ify the target mesh density), which is taken as 50 km in the present study. To under-
stand the effect of cratons on the stresses inside the plate, two different models are
developed – the first model with uniform thickness throughout the plate; and the sec-
ond with increased thickness at the cratons. The Indian plate is discretized into 6152
elements and the mesh size is suitably fixed after carrying out mesh refinement until
there is not much variation in stresses. The finite element mesh with the specified
loads and boundary conditions on the model is shown in figure 3.
8 S. Jayalakshmi and S.T.G. Raghukanth

Figure 3. Loads, boundary conditions and finite element mesh on the model RP (ridge push).
The boundary along the Himalaya is fixed in the model.

4. Comparison with the GPS measurements

The above developed finite element models with and without cratons are subjected to
ridge push force corresponding to 1 year. The thickness of the cratons is determined
by trial and error and the value that compares reasonably with GPS data is esti-
mated. In the present study, the thickness of cratons is taken as 150 km. The con-
tours of deformation along north and east directions for both the models are shown
in figures 4 and 5. It will be interesting to compare the obtained deformations with
GPS measurements. Table 1 shows the GPS measurements at 26 sites as reported by
Mahesh et al. (2012). A comparison between the magnitudes of deformation pre-
dicted by both the models with the data is shown in figure 6. It can be observed from
 Geomatics, Natural Hazards and Risk 9

Figure 4. Contours of deformation (mm) on Indian plate without cratons: (a) E–W
component, (b) N–S component. The squares indicate the GPS sites.

figure 4 and table 1 that the magnitude of deformation for the sites located in north-
east India (LUMA, IMPH, TZPR, AZWL, SHIL, GWHT) estimated by both the
models is almost 10 times lower than the GPS measurement and there is not much
difference between the predictions of the two models on the deformations on sites
SHIL(5) and GWHT(6). For almost all sites in the Indian shield (site nos. 8–26), the

Figure 5. Contours of deformation (mm) on Indian plate with cratons: (a) E–W component,
(b) N–S component. The squares indicate the GPS sites.
10 S. Jayalakshmi and S.T.G. Raghukanth

Table 1. Comparison of deformations for Indian plate model with GPS measurements.

Site Long Lat Umag Ue Un Umag Ue Un Umag Ue Un

Model without cratons Model with cratons Mahesh et al., 2012

LUMA 94.47 26.22 1.23 0.82 0.74 1.04 0.79 0.68 14.28 3.07 13.95
IMPH 93.92 24.74 2.12 1.16 0.88 1.38 1.12 0.80 20.57 10.81 17.51
TZPR 92.78 26.61 0.44 0.55 0.37 0.77 0.65 0.42 10.46 0.95 10.42
AZWL 92.73 23.72 2.93 1.49 0.84 1.43 1.23 0.72 10.31 6.26 8.2
SHIL 91.88 25.56 0.90 0.83 0.46 0.92 0.80 0.45 6.31 1.13 6.21
GWHT 91.66 26.15 0.58 0.68 0.36 0.80 0.70 0.37 7.59 0.05 7.59
BHUB 85.79 20.26 6.89 2.61 0.25 2.06 2.05 0.23 3.39 3.21 1.11
ALHB 81.80 25.30 1.44 1.17 0.26 1.00 0.97 0.22 0.79 0.04 0.79
PLVM 81.64 17.27 13.27 3.63 0.31 2.96 2.95 0.27 1.11 1.06 0.36
GSIL 80.94 26.89 0.68 0.78 0.28 0.80 0.75 0.26 2.33 1.73 1.57
JBPR 79.87 23.12 3.73 1.86 0.52 1.55 1.47 0.49 1.27 0.68 1.08
TONK 79.60 19.51 9.27 2.98 0.63 2.47 2.40 0.57 0.92 0.83 0.41
HANL 78.97 32.77 – – – – – – 7.72 4.52 6.26
HYDE 78.55 17.41 14.09 3.66 0.82 3.07 2.99 0.68 0.19 0.15 0.12
DEHR 78.05 30.32 0.17 0.27 0.32 0.50 0.28 0.42 0.88 0.06 0.88
RSCL 77.6 34.12 0.04 0.07 0.19 0.30 0.12 0.28 15.98 6.6 14.56
IISC 77.57 13.02 27.75 5.17 1.02 4.37 4.27 0.90 0.43 0.28 0.33
KDKL 77.46 10.23 38.05 6.08 1.07 5.17 5.08 0.95 0.37 0.13 0.35
BHOP 77.44 23.20 4.66 1.97 0.87 1.81 1.61 0.81 0.66 0.43 0.51
DELH 77.12 28.48 0.83 0.68 0.61 0.90 0.59 0.68 3.17 2.72 1.63
ARKW 73.93 17.23 18.36 3.95 1.67 3.58 3.23 1.54 0.83 0.44 0.71
PUNE 73.88 18.55 15.23 3.55 1.62 3.48 3.11 1.55 1.09 1.06 0.27
GOKL 73.72 17.4 17.85 3.88 1.67 3.20 2.82 1.50 0.52 0.42 0.31
UDAI 73.71 24.58 4.97 1.74 1.39 1.91 1.38 1.32 0.32 0.05 0.32
IITB 72.91 19.13 14.51 3.36 1.79 3.23 2.78 1.66 0.47 0.44 0.19
ISRG 72.66 23.21 7.91 2.23 1.71 2.30 1.70 1.56 1.31 0.53 1.2

deformations predicted by the model with cratons are almost of the same order as
that of the GPS measurement. Among these, except for sites DEHR(15), RSCL(16)
and DELH(20), which are near the plate boundary, the deformations predicted by
the model with cratons are almost of the same order as that of the GPS measure-
ment. For most of the sites in the Indian shield, the obtained deformations from the
homogeneous thickness model are much higher than the measured data. The site
HANL(13) lies outside the modelled area. The results predicted by the model with
cratons show fairly good comparison for more than 20 GPS sites.

5. Comparison with the seismicity data

Recently, Raghukanth (2011) compiled the earthquake catalogue from various sour-
ces in the literature for the Indian subcontinent. Indian landmass can be grouped
into 33 seismogenic zones based on historical seismicity, tectonic features and geol-
ogy (NDMA 2011). The 33 source zones are shown in figure 7. It will be interesting
to compare the stresses and strains obtained from the model with the seismicity data
of these seismogenic source zones. The magnitudes of earthquakes are collected from
the earthquake catalogue and the stresses and strains are obtained using the empiri-
cal formulae available in the literature.
 Geomatics, Natural Hazards and Risk 11

Figure 6. Comparison of magnitude of deformation obtained from both the models with
GPS data. The GPS sites are indicated by numbers along x-axis.

6. Comparison with strain rate

The seismic moment M0 is obtained from moment magnitude using the relation of
Hanks and Kanamori (1979) as

log10 M0 ¼ 1:5Mw þ 16:1; ð2Þ

where Mw is the moment magnitude. After estimating seismic moment, the regional
strain rate can be estimated from the relation given by Anderson (1979)

M_ 0
e_ ¼ ; ð3Þ
2:67mAh

where A is the geographical area considered, M0 is the rate of seismic moment, m is

the shear rigidity of the material (3.6  1011 dyne/cm2) and h is the average thickness
of the seismogenic layer. Strain rates are calculated for each seismogenic zone by
assuming uniform rigidity throughout the plate. Although the earthquake catalogue
of India starts from 2474 BC, completeness test shows that the data are complete for
all magnitudes for past 60 years only (Raghukanth 2011). Hence earthquake events
that occurred after 1949 are used to estimate the strain rates. The average thickness
of the seismogenic layer (h) for each seismogenic zone is fixed based on the focal
depth of the events. The strain rates are calculated from equation (3) for all the
events in each of the 32 seismogenic zones and the average strain rate in each zone is
reported in table 2. The contours of maximum principal strain obtained from both
the finite element models corresponding to a ridge push force of 1 year are shown in
12 S. Jayalakshmi and S.T.G. Raghukanth

Figure 7. Seismogenic zone map of India (NDMA).

figure 8. The maximum and minimum principal strains obtained in each seismogenic
zone estimated from figure 8 are also given in table 2. A comparison of strain rates
with data is shown in figure 9. Seismic zones 6, 11, 12, 13, 14, 15 and 30 have been
eliminated from the study since it falls outside the Indian plate. From table 2 and
figure 9, it is clear that for zones near the plate boundary close to Himalayas, both
the models predict almost equal strain rates, which are fairly large compared to the
strain rates from data. This difference arises due to the assumption of fixed boundary
condition along the boundary. However, for most of the zones in the Indian shield,
the strains predicted by the model with cratons compare better with the data.

7. Comparison with stress drop

The above analysis shows that the model with cratons compares better with the data
and is taken further to explain seismic activity in India. The stress drops for
 Geomatics, Natural Hazards and Risk 13

Table 2. Comparison of strain rates for Indian plate model with data.

Seismicity Minimum strain Maximum strain Minimum strain Maximum strain Strain rate
zone year1 year1 year1 year1 from data

Model without cratons Model with cratons

1 6.02E10 3.86E09 6.23E10 3.50E09 1.03E10

2 8.93E10 1.32E09 9.33E10 1.38E09 7.34E11
3 8.06E10 2.11E09 6.98E10 2.12E09 9.56E15
4 7.70E10 2.37E09 6.77E10 2.23E09 7.90E11
5 4.11E11 1.90E09 2.08E12 1.78E09 7.11E10
7 4.96E10 7.84E10 2.30E10 6.36E10 3.78E10
8 5.73E10 9.93E10 2.01E10 8.51E10 1.27E10
9 4.04E10 7.92E10 2.26E10 7.29E10 6.04E11
10 1.56E10 6.71E10 1.02E10 4.95E10 1.66E09
12 9.71E11 4.41E10 9.86E11 4.04E10 2.76E09
14 8.42E12 1.88E10 9.81E12 1.78E10 4.21E09
16 4.95E10 6.70E10 2.40E10 6.41E10 2.14E10
17 2.60E10 4.95E10 1.21E10 4.53E10 4.77E10
18 4.29E10 6.09E10 1.83E10 4.50E10 9.18E10
19 4.28E10 5.39E10 2.36E10 4.70E10 4.40E10
20 2.69E10 4.86E10 1.23E10 5.66E10 1.65E10
21 3.72E10 7.92E10 3.71E10 8.05E10 1.60E10
22 7.81E10 1.22E09 7.53E10 1.25E09 9.85E11
23 5.81E10 1.31E09 5.31E10 1.28E09 8.43E11
24 4.55E10 6.39E10 4.28E10 6.27E10 1.07E10
25 2.85E10 5.35E10 2.81E10 5.39E10 1.23E10
27 3.66E10 4.92E10 3.00E10 5.27E10 2.21E10
28 4.86E10 9.33E10 2.44E10 9.66E10 2.84E10
29 2.77E10 5.13E10 1.09E10 5.54E10 3.74E10
31 5.02E10 8.52E10 2.30E10 9.13E10 1.77E10
32 1.03E10 4.98E10 9.86E11 4.55E10 1.31E11

earthquakes in the past 60 years are obtained from the empirical relation given by
Aki (1972),

M0
Ds ¼ 2:43 ; ð4Þ
A1:5

where Ds is the stress drop in dynes/cm2 and A is the area of rupture in km2.In the
above equation, the constant 2.43 accounts for circular crack model. The seismic
moment of all the past events is obtained from moment magnitude (equation (2)).
The rupture area is estimated from the empirical relation of Wyss and Brune (1968)
as

log10 A ¼ 1:05Mw  2:95: ð5Þ

The cumulative stress released for all the events that occurred in the past 60 years is
calculated for all the zones. These are reported in table 3. It will be interesting to
obtain how many years of ridge push force are required for this stress release. For
this purpose, the ridge push force is varied from a magnitude of 1.5  108 N/km to
14 S. Jayalakshmi and S.T.G. Raghukanth

Figure 8. (a) Contours of maximum principal strain on Indian plate without cratons.
(b) Contours of maximum principal strain on Indian plate with cratons. The seismogenic zones
of India are also marked.

Figure 9. Comparison of maximum principal strain rates with the strain rates from data.
The vertical lines indicate the bound between minimum and maximum strain obtained in the
model.
 Geomatics, Natural Hazards and Risk 15

Table 3. Comparison of stress obtained from the model with stress released due to past
earthquakes.

Minimum stress Maximum stress

Zone Stress drop Age of lithosphere ¼ 5 Ma

1 79.46 29.25 124.23

2 20.94 29.25 109.13
3 13.31 29.25 94.71
4 9.74 29.25 150.54
5 8.55 1.81 160.29
7 2.05 8.51 160.29
8 8.31 8.51 138.16
9 12.09 8.56 76.60
10 39.95 8.56 70.42
12 8.34 8.56 70.42
14 6.11 1.27 70.42
16 3.12 8.56 70.42
17 2.88 8.26 47.89
18 1.47 8.26 50.02
19 1.65 8.26 35.24
20 15.56 8.62 34.03
21 4.14 8.62 34.03
22 12.14 8.62 44.26
23 14.51 8.62 44.92
24 10.70 8.62 34.03
25 7.08 8.62 34.03
27 15.17 8.62 34.03
28 22.78 8.55 55.62
29 4.94 8.32 33.69
31 1.70 8.32 66.66
32 2.07 10.85 40.19

1.9  1015 N/km (corresponding to 60 years – 50 Ma). The ridge push force is esti-
mated by minimizing the error between the average stress obtained from the finite
element model in each seismic zone with that obtained from the data. The ridge push
force corresponding to a spreading age of 5 million years gives a fairly good predic-
tion of stresses in the model. Table 3 compares the stress values obtained from the
model corresponding to a ridge push force equivalent to a spreading age of 5 million
years with the stress drop.
A comparison between the stresses released from seismicity data and the stresses
obtained from the finite element model is shown in figure 10. The stress released does
not exceed the maximum predicted stress. It can be seen that for most of the zones in
the Himalayan belt (zones 2, 3, 4), the stress built up is much higher than the stress
released due to past earthquakes, which agrees with the seismic gap theory. It indi-
cates that although some great events have occurred in the past, more stress is yet to
be released in these zones. For most zones in the Indian shield (zones 16–21, 25, 29,
31, 32), model stresses are higher than the stress released, which indicates that these
zones are prone to more earthquakes in future. Zones 22, 23, 24 and 28 show stress
released within the bounds predicted by the model.
16 S. Jayalakshmi and S.T.G. Raghukanth

Figure 10. Comparison of stress in the model with the stress released in past earthquakes
from data. The vertical lines indicate the bound between minimum and maximum stress
obtained for the corresponding zone in the model.

The stress contours in figure 11 show high stresses near the plate boundary
where large number of earthquakes can be seen. It also gives high stresses at few
regions in the Indian shield especially the Son–Narmada rift zone, which runs
from west coast to east coast of India where many earthquakes have occurred in
the past. Although the model explains the occurrence of few earthquakes in the
Indian shield, there are some events within the cratons that needs to be under-
stood. Modelling of faults could give better insight into the occurrence of these
events.

8. Conclusions
This article has been motivated by the desire to develop a mechanistic model for the
occurrence of earthquakes in India. The Indian plate is modelled using the finite ele-
ment method. The plate boundary at the Himalaya is fixed and ridge push forces are
applied as uniform pressure at the mid-oceanic ridge boundary to obtain the defor-
mation in the Indian plate. ABAQUS 6.10.1 software is used to carry out the analy-
sis. The effect of including the five major cratons in the Indian plate is also studied in
this article. The cratons are modelled as regions with higher thickness compared to
the rest of the Indian plate. The thickness of the cratons is selected based on the com-
parison with the GPS measurements and seismicity data. Table 1 shows the compari-
son between the obtained results with the GPS measurements. The model is able to
explain the deformations at most of the GPS sites. A further comparison with the
strain rates obtained from the seismicity data is also presented in table 2. The order
of the strain rates estimated from the model is similar to that obtained from the data.
 Geomatics, Natural Hazards and Risk 17

Figure 11. (a) Von Mises stress contours on Indian plate model with cratons. (b) Contours on
the Himalayan region. Dots indicate location of epicentres of past earthquakes. (Blue –
Mw < 6; magenta – Mw < 7; green – Mw < 8; red – Mw > 8).

A comparison with the total stress drop of the events that occurred in the past
60 years is shown in table 3. It is observed that ridge push force corresponding to
5 million years age of the lithosphere is required for this stress release. The locations
of high stress concentrations coincide with the occurrence of the past earthquakes
(figure 11).
18 S. Jayalakshmi and S.T.G. Raghukanth

The assumption of linear elasticity with homogeneous material properties is one of

the limitations in the present study. The effect of slab pull forces in the Indo-
Burmese region on the stresses in the Indian plate also has not been attempted in this
study. Modelling the uncertainties associated with rheology of the Indian plate,
faults, slab pull forces and their effect on the seismicity of the Indian plate has to be
investigated further.

References
Aki K. 1972. Scaling law of earthquake source time function. Geophys J Roy Astr Soc. 31:
3–25.
Anderson JG. 1979. Estimating the seismicity from geological structure for seismic risk stud-
ies. Bull Seis Soc Am. 69:135–158.
Banerjee P, B€ urgmann R, Nagarajan B, Apel E. 2008. Intraplate deformation of the Indian
subcontinent. Geophys Res Lett. 35:L18301.
Bettinelli P, Avouac JP, Flouzat M, Jouanne F, Bollinger L, Willis P, Chitrakar GR. 2006.
Plate motion of India and interseismic strain in the Nepal Himalaya from GPS and
DORIS measurements. J Geodesy. 80:567–589.
Bilham R, Blume F, Bendick R, Gaur VK. 1998. Geodetic constraints on the translation and
deformation of India: implications for future great Himalayan earthquakes. Curr Sci.
74(3):213–229.
Bilham R, Gaur VK. 2000. Geodetic contributions to the study of seismotectonics in India.
Curr Sci. 79:1259–1269.
Bilham R, Gaur VK, Molnar P. 2001. Himalayan seismic hazard. Science. 293:1442–1444.
Bird P. 2003. An updated digital model of plate boundaries. Geochem Geophys Geosyst.
4:1027–1052.
Chandra U. 1977. Earthquakes of Peninsular India – a seismotectonic study Bull. Seism Soc
Am. 67(5):1387–1413.
Coblentz DD, Sandiford M, Richardson RM, Zhou S, Hillis R. 1995. The origins of the intra-
plate stress field in continental Australia. Earth Planet Sci Lett. 133:299–309.
DeMets C, Gordon RG, Argus DF, Stein S. 1990. Current plate motions. Geophys J Int.
101:425–478.
Dyksterhuis S, Muller RD, Albert RA. 2005. Paleostress field evolution of the Australian con-
tinent since the Eocene. J Geophys Res. 110:B05102, doi:10.1029/2003JB002728.
Forsyth D, Uyeda S. 1975. On the relative importance of the driving forces of plate motion.
Geophys J Roy Astr Soc. 43:163–200.
Freymueller J, Bilham R, Burgmann R, Larson KM, Paul J, Jade S, Gaur V. 1996. Global
positioning system measurements of Indian plate motion and convergence across the
Lesser Himalaya. Geophys Res Lett. 23(22):3107–3110, doi:10.1029/96GL02518.
Gansser A. 1974. The Ophiolitic Melange, a world-wide problem on Tethyan examples.
Eclogae Geol Helv. 67:479–507.
Geological Survey of India (GSI). 2000. Seismotectonic atlas of India and its environs.
Kolkata: GSI.
Gupta ID. 2006. Delineation of probable seismic sources in India and neighbourhood by a
comprehensive analysis of seismotectonic characteristics of the region. Soil Dyn Earth-
quake Eng. 26:766–790.
Hanks TC, Kanamori H. 1979. A moment magnitude scale. J Geophys Res: Solid Earth.
84:2348–2350.
Hieronymus CF, Goes S, Sargent M, Morra G. 2008. A dynamic model for generating
Eurasian lithospheric stress and strain rate fields: effects of rheology and cratons.
J Geophys Res. 180:148–227, doi:10.1029/2007JB004953
 Geomatics, Natural Hazards and Risk 19

Iyengar RN, Raghukanth STG. 2004. Attenuation of strong ground motion in peninsular
India. Seismol Res Lett. 75(4):530–540.
Jade S. 2004. Estimates of plate velocity and crustal deformation in the Indian subcontinent
using GPS geodesy. Curr Sci. 86:1443–1448.
Jade S, Mukul M, Bhattacharya AK, Vijayan MSM, Jaganathan S, Kumar A, Tiwari RP,
Kumar A, Kalita S, Sahu SC, et al. 2007. Estimates of interseismic deformation in
northeast India from GPS measurements. Earth Planet Sci Lett. 263:221–234.
Kayal JR. 2008. Microearthquake seismology and seismotectonics of South Asia. New Delhi:
Capital Publishing Company.
Khattri KN. 1987. Great earthquakes, seismicity gaps and potential for earthquake disaster
along the Himalaya plate boundary. Tectonophysics. 138:79–92.
Khattri KN, Tyagi AK. 1983. Seismicity patterns in the Himalayan plate boundary and the
identification of areas of high seismic potential. Tectonophysics. 96:19– 29.
Lav
e J, Yule D, Sapkota S, Basant K, Madden C, Attal M, Pandey R. 2005. Evidence for
a great medieval earthquake (1100 A.D.) in the Central Himalayas. Science. 307:
1302–1305.
Mahesh P, Catherine JK, Gahalaut VK, Kundu B, Ambikapathy A, Bansal A, Premkishore
L, Narsaiah M, Ghavri S, Chadha RK, et al. 2012. Rigid Indian plate: constraints
from GPS measurements. Gondwana Res. 22:1068–1072.
Manglik A, Thiagarajan S, Mikhailova AV, Rebetsky Yu. 2008. Finite element modelling of
elastic intraplate stresses due to heterogeneities in crustal density and mechanical prop-
erties for the Jabalpur earthquake region. J Earth Sys Sci. 117(2):1–9.
Molnar P, Tapponier P. 1975. Cenozoic tectonics of Asia: effects of a continental collision.
Science. 189:419–426.
Molnar P, Tapponnier P. 1977. The collision between India and Eurasia. Sci Am. 235(4):
30–41.
NDMA. 2011. Development of probabilistic seismic hazard map of India. New Delhi:
National Disaster Management Authority.
Ni J, Barazangi M. 1984. Seismotectonics of the Himalayan collision zone: geometry of the
underthrusting Indian plate beneath the Himalaya. J Geophys Res. 80:1142–1163.
Paul J, Burgmann R, Gaur VK, Bilham R, Larson KM, Ananda MB, Jade S, Mukal M,
Anupama TS, Satyal G, Kumar D. 2001. The motion and active deformation in India.
Geophys Res Lett. 28:647–650.
Quittmeyer RC, Jacob KH. 1979. Historical and modern seismicity of Pakistan, Afghanistan,
north-western India and south-eastern Iran. Bull Seismol Soc Am. 69:773–823.
Raghukanth STG. 2010, Estimation of seismicity parameters for India. Seismol Res Lett.
81(2):207–217.
Rajendran CP, Rajendran K, Duarah BP, Baruah S, Earnest A. 2004. Interpreting the style of
faulting and paleoseismicity associated with the 1897 Shillong, northeast India, earth-
quake: implications for regional tectonism. Tectonics. 23(4):TC4009.
Rao BR, Rao PS. 1984. Historical seismicity of peninsular India. Bull Seismol Soc Am.
74(6):2519–2533.
Scholz CH, Campos J. 1995. On the mechanism of seismic decoupling and back arc spreading
at subduction zones. J Geophys Res. 100(B11):22103–22115.
Seeber L, Armbruster J. 1981. Great detachment earthquakes along the Himalayan arc and
long term forecasts. In: Simpson D, Richards M, editors. Earthquake prediction: an
international review, Ewing series. Vol. 4. Washington, DC: AGU; p. 259–277.
Sella GF, Dixon TH, Mao A. 2002. REVEL: a model for recent plate velocities from space
geodesy. J Geophys Res. 107(B4):ETG 11-1–ETG 11-30.
Tandon AN, Srivastava HN. 1975. Focal mechanisms of some recent Himalayan earthquakes
and regional plate tectonics. Bull Seismol Soc Am. 65:963–969.
Turcotte D, Schubert G. 2002. Geodynamics. Cambridge: Cambridge University Press
20 S. Jayalakshmi and S.T.G. Raghukanth

Valdiya KS. 1976. Himalayan transverse faults and folds and their parallelism with subsurface
structures of the northern Indian plains. Tectonophysics. 32:353–386.
Valdiya KS. 1980. Geology of the Kumaun Lesser Himalaya. Dehradun: Wadia Institute of
Himalayan Geology; 291.
Valdiya KS. 1998. Dynamic Himalaya. Hyderabad: Universities Press.
Valdiya KS. 2010. The making of India: geodynamic evolution. India: Macmillan Publishers
India Ltd.
Wyss M, Brune JN. 1968. Seismic moment, stress and source dimensions for earthquakes in
the California-Nevada region. J Geophys Res. 73:4681–4694