Text of Talk Delivered at the International Conference on

“Post Earthquake Reconstruction: Lessons Learnt & Way Forward”

14-16 April, 2011, GSDMA, Gandhinagar, Gujarat

Probabilistic Seismic Hazard Mapping of India and its Importance in

Microzonation

R.N. Iyengar
Centre for Disaster Mitigation
Jain University, Bangalore 562112
(email: aareni@yahoo.com)

Introduction
It is now well recognized that earthquakes are hazardous to human habitats in all parts of
India. Even though seismogenic zones may be restricted to some particular tectonic
provinces, ground motion due to earthquakes can be felt anywhere in the country. The
resultant ground vibration depending on its severity characterized in terms of its
amplitude-frequency distribution can cause damage to buildings and other civil
infrastructure in our cities. This is a well known fact and at present various agencies are
working to develop reliable open source hazard maps which help administrators,
engineers, building owners and other stake holders to evolve better pre-event planning,
emergency preparedness, engineering design, construction and post-event relief
management strategies. The end result expected is safety of life and uninterrupted civic,
industrial and economic activities in our vulnerable cities. For this purpose seismic
microzonation (MZ) exercises are carried out to map hazard, vulnerability and risk to
various levels of details. Microzonation (MZ) as the name indicates is an exercise to map
a quantity of interest on a relatively finer scale. In the context of earthquakes, state level
hazard mapping can be called MZ in comparison with national level or global scale
hazard maps. Similarly one can talk of district or city level hazard mapping as MZ.
Hence the scale of the map decides the end use of MZ exercise. However there has to be
compatibility among maps generated at various levels whether they are said to be macro-
or micro-zone maps. When city vulnerability and consequent risk in terms of rupee value
is discussed it is usual to restrict oneself to the soils, buildings and social aspects specific
to a particular urban habitat which may be spatially limited to an area of about 100-200
sq km. But the risk is sensitively dependent on the prevalent seismic hazard which in turn
will be dictated by the geology and tectonic features of a very large region. Thus, the city
level risk map, although sensitively dependent on the local subsurface soil conditions, has
to be in harmony with the seismic hazard at the basement rock level which will not vary
at short distances. The broad seismogenic zones in and around the subcontinent are
known, the national level seismo-tectonic atlas is available and a fairly representative
catalogue for the country can be assembled. Carrying out deterministic or probabilistic
MZ of individual cities, without homogenizing the results with large scale data would
lead to errors of over/under estimation of the consequent vulnerability and risk results.
Even when scenario based events have been used to arrive at MZ of cities, lack of bench

1
marking soil classes, quality factors and attenuation would lead to results not usable by
policy makers and engineers.

With the above points in background the National Disaster Management Authority
(NDMA) has prepared a probabilistic seismic hazard assessment (PSHA) map for the
country. This work was coordinated by the present author. This article briefly explains
the development of the national level PSHA map and its practical application.

Basis of Hazard Analysis

The basic frame work of seismic hazard analysis, either deterministic or probabilistic, is
well known. However, source definition, site classification, and grid size could lead to
some variations in the final results. To arrive at a consensus on the mode of carrying out
the PSHA a National Workshop was held by NDMA on 16th July 2008. This was
attended by leading R&D specialists in geophysics, seismology and earthquake
engineering representing academia, industry, government and private agencies. The
following broad recommendations emerged from the workshop.

1) Hazard will be computed by considering a circular region of 300 km around the

sample site point, except in the Himalayan region where the above figure would
be 500 km.
2) The catalogue database will consist of all known earthquakes of magnitude 4 and
above recorded in and around the subcontinent in the last ~500 years.
3) All magnitude values will be expressed in terms of moment magnitude Mw.
4) The maximum considered earthquake magnitude will be 0.5 units above the
recorded or estimated past value for the region.
5) (a, b) values will be found by studying for completeness as per established
statistical analysis due to Kijko, Sellvol and others.
6) Only faults identified by official agencies such as GSI, IMD, NGRI will be
included in PSHA.
7) Appropriate attenuation relations will be derived and validated wherever possible
for Himalayan region, Indo-Gangetic plains, NE region, Central India, Gujarat
and southern India separately.
8) PSHA result will be presented as contours for engineering bed rock conditions
corresponding to A- or B-type rock site.
9) Results will be shown for PGA, and Sa for one long period and one short period
for different return periods.
10) All India PSHA maps will be prepared on a grid size of 0.20x0.20.

Seismogenic Zones and Fault Map

Indian subcontinent is known to be inhomogeneous with respect to its seismic characters.
A map of epicenters superposed on fault lines is shown in Figure1. This shows that the
Andaman-Nicobar Islands, the North East, the Himalayas and the Kutch region are more
active than the stable continental region (SCR) of Peninsular India. For PSHA it is
necessary to characterize the various seismogenic zones and the faults within these zones
in terms of recurrence relations, fault lengths and potential maximum magnitude that can
be associated with the sources.

2
The Indian land mass is grouped into thirty-two seismogenic zones based on historical
seismicity, tectonic features and geology. Following Seeber et al (1999) the Indian shield
region is divided into seven tectonic zones. Studies of Bhatia et al (1999) and Gupta
(2006) have been utilized to identify seismogenic zones in the Himalayas, north-east
India and Andaman-Nicobar islands. Source zones in Pakistan which can induce ground
motion across the international borders are selected as per the publication of the Pakistan
Meteorology Department (PMD 2007). In Figure 2 the thirty-two source zones are shown
along with the faults within each zone. The seismic parameters of all the zones are
presented in Table 1. The details of assembling the catalogue, declustering, completeness
study and estimation of maximum potential magnitude are available in the PSHA report
(NDMA 2010).

Quality Factors and Attenuation Relations

With the country organized into source zones of differing potential any future event can
originate within one of these zones only. The actual event will be due to the rupture of a
fault, amounting to a particular Mw, within a given zone. This rupture creates seismic
waves propagating towards target sites through the intervening medium. The ground
motion depends significantly on the source to site distance and the material properties of
the rock medium that forms the path for the seismic waves. How the path affects the
ground motion in different parts of the country is described by deriving empirical strong
motion attenuation relations of the form

S 
( )
ln  a  = C1 + C2 M + C3M 2 + C4 r + C5 ln r + C6eC7 M + C8 log(r )f 0 + ln ( ε )
g (1)
f0 = max(ln(r /100),0)

Here Sa is the spectral acceleration (5% damping), M is the moment magnitude, r is the
hypocentral distance in kilometers. Seven primary geological provinces with differing
quality factors, stress drop and focal depth properties are identified as shown in Figure 3.
The coefficients to be used in the attenuation relation for all the seven provinces are
generated using the stochastic seismological model valid for finite fault sources (Boore
2009). The results are validated by comparing them with recorded results wherever
available. The site condition is uniformly fixed as A-type rock which corresponds to V30
of more than 1.5 km/s. As an example the coefficients in eq.1 valid for the Gujarat region
are shown in Table 2.

PSHA Maps
Standard PSHA using line sources within 300 km (500 km in the Himalayan region)
radius of the 0.20x0.20 grid points is carried out at All India level. The soil type at the site
is fixed as A-type engineering bed rock. A-type rock sites are available in many places of
the country. Typical surface expression of such a site is shown in Figure 4. Hazard results
(PGA and Sa) are made available for return periods of 500, 2500, 5000 and 10,000 years.
Some typical results are shown in Figures 5 (a, b, c, d, e, f). The hazard in terms of PGA
at 48 important cities is shown in Table 3 for four different return periods. This table
shows the relative severity of rock level ground motion to the most populous cities of

3
India. It should be noted that a low PGA value on rock need not necessarily mean low
surface level accelerations. The subsurface soil layers of B, C, D, E and F-type with
progressively decreasing V30 values can modify surface ground motion characters
considerably.

Microzonation Application
Cities rated highly as per Table 3 are the ones facing higher hazard over a length of time.
The results of the NDMA report can be directly used on A-type sites any where in India
to know the PGA values or for constructing the mean and percentile response spectra for
5% damping.

Urban areas with sedimentary deposits exhibiting large variations in V30 values and
ground water table levels are susceptible for soil amplification and some times
liquefaction. Such cities would benefit by local urban microzonation. However, there is
no need to carry out basic seismic hazard exercise afresh. The values presented in the
NDMA report for the particular city grid point can be used as input at A-type rock level.
The surface level hazard can be found by adopting available methods of computing soil
amplification. One of the simplest approaches is to first divide the urban agglomeration
on a fine scale (say 0.5 x 0.5 km) as per the V30 values. Methods are available in the
literature for this type of site classification based on borehole (N-value) data or using
more sophisticated geophysical investigations. Once the site type is fixed, the design
spectrum can be computed using the method described in the International Building Code
(IBC-2009). In the Appendix-II of the NDMA report examples of such computations are
presented for Chennai, Delhi, Kolkata and Mumbai. Such an empirical method is
approximate but, is considered sufficient for routine engineering design of buildings.

In case the local soil conditions are known in greater detail in terms of their layering,
damping and shear wave velocities, options exist to compute the surface level
accelerations using popular software like SHAKE, which handles the local soil response
as one dimensional. More sophisticated analytical methods and tools to include nonlinear
soil properties and 2-D or 3-D basin effects in computing the surface time histories from
rock level inputs are also available in the literature. For such work the results presented in
the report can be used to specify the input response spectrum or an ensemble of
compatible time histories corresponding to A-type rock level above which the soil
formation exists at the city site.

Conclusion
The NDMA report is available for public access at the following links

http://ndma.gov.in/ndma/disaster/earthquake/India-psha-finalreport.pdf
http://ndma.gov.in/ndma/disaster/earthquake/Appendix%20I%20Earthquake%20catalog.
pdf
http://ndma.gov.in/ndma/disaster/earthquake/India-PSHA-AppendixII.pdf

The report shows how surface level hazard at hard rock sites (A-type) can be estimated to
develop charts and tables that can be used by government agencies, architects, engineers
and other interested groups. All known data about past earthquakes and mapped faults

4
are considered to characterize the seismic activity of the thirty-two source zones of the
country structured into seven geological regions with differing quality factors. New
attenuation equations are derived for each of the seven regions incorporating known
quality factors and near source effects due to finite source size. State-of-art probabilistic
hazard analysis has been carried out covering the whole country at 7156 sites on a grid
size of 0.20 x 0.20. The prevalent hazard in terms of peak ground acceleration (PGA),
short period and long period spectral accelerations for different return periods has been
presented in the report.

This article reviews and presents some findings of the recent NDMA work on seismic
hazard estimation at All India level. Results of the report can be directly used by
engineers for developing design response spectra at A-type hard rock sites with shear
wave velocity greater than 1.5 km/s. For other sites IBC-2009 can be used if the site
classification is known. The PGA results can be used as preliminary indicators for
microzonation and planning of cities with high hazard.

Cities near faults with high potential and cities in geologically sensitive regions would
benefit by MZ even when the PGA values are not very high. Nearness to an active fault
makes the sites susceptible for permanent ground displacements which effect is not
explicitly handled in the NDMA report. Sensitive regions are those for which deep
section basin geometry and engineering properties of the sediments pose special
problems. For example, Ahmedabad and Baroda show relatively low 2500-year PGA
values at A-type rock level. But these cities are in the Cambay rift basin (Figure 6) the
triangular type sectional geometry of which will surely influence the free surface motion
significantly. Superposed on the local subsurface soil property variation, the resultant
surface level ground motion can show significant spatial variation which can be captured
only by detailed microzonation. But for such an exercise to be meaningful it should not
ignore the fact that Saurashtra was not contiguous with the main land as recently as 3-
4000 years before present.

Acknowledgement
Contributions of Dr. STG Raghukanth, IIT-M, Chennai and Dr. K.Balaji Rao, CSIR-
SERC, Chennai to the NDMA report are thankfully acknowledged.

References
Bhatia, S.C., Kumar, M.R. and Gupta, H.K. (1999), A probabilistic seismic hazard map
of India and adjoining regions, Annali di Geofisica, 42:1153-1166.
Boore, D.M. (2009). Comparing stochastic point-source and finite-source ground motion
simulations: SMSIM and EXSIM, Bull. Seism. Soc. Am. 99(6):3202-3216.
Choudhary, I.R. (2004) Petroleum geology of the Cambay basin, Gujarat, India. Indian
Allied Publishers.
Gupta, I.D. (2006). Delineation of probable seismic sources in India and neighbourhood
by a comprehensive analysis of seismotectonic characteristics of the region, Soil
Dynamics and Earthquake Engineering, 26,766–790.

5
Seeber, L., Armbruster, J.G. and Jacob, K.H. (1999), Probabilistic assessment of
earthquake hazard for the state of Maharashtra. Report to Government of Maharashtra
Earthquake Rehabilitation Cell Mumbai.
PMD (2007), Seismic Reports, Pakistan Meteorological Department,
www.pakmet.com.pk
NDMA (2010), Development of Probabilistic Seismic Hazard Map of India, Final Report
of the WCE, submitted to NDMA, Govt. of India, New Delhi, 30th September 2010.

Table 1 Seismic Parameters of the thirty-two Zones

Source Zones b-value N(4) Max. Potential No. of
No. Magnitude earthquakes
(Mmax)
1 Western Himalaya 0.88±0.02 5.37 8.8 901
2 Central Himalaya-I 0.73±0.04 3.15 7.8 306
3 Central Himalaya-II 0.78±0.04 2.30 8.8 340
4 Eastern Himalaya 0.71±0.04 3.12 8.0 223
5 Mishmi Block 0.66±0.03 3.72 8.8 219
6 Altya Tegh & Karakoram 0.91±0.03 7.10 7.3 726
7 Naga Thrust 0.67±0.08 0.18 6.8 32
8 Shillong Plateau & Assam 181
valley 0.73±0.04 1.46 8.4
9 Bengal Basin 0.74±0.04 1.99 8.1 289
10 Indo-Burmese Arc 0.80±0.02 11.40 7.8 1055
11 Shan-Sagaing Fault 0.66±0.04 5.28 8.1 260
12 West Andaman-I 0.70±0.03 3.62 8.4 239
13 East Andaman-I 0.63±0.03 5.83 7.5 331
14 West Andaman-II 0.71±0.02 2.55 7.5 158
15 East Andaman-II 0.62±0.01 16.53 7.6 985
16 SONATA 0.64±0.08 0.24 6.8 24
17 Eastern Passive Margin 0.74±0.08 0.27 6.1 40
18 Mahanandi Graben & 0.77±0.09 0.24 5.3 15
Eastern Craton
19 Godavari Graben 0.85±0.09 0.13 6.0 10
20 Western Passive Margin 0.76±0.07 0.37 6.8 70
21 Sindh-Punjab 0.77±0.06 0.60 8.0 89
22 Upper Punjab 1.01±0.05 1.68 7.8 224
23 Koh-e-Sulaiman 0.84±0.04 5.03 7.3 358
24 Quetta-Sibi 0.74±0.04 5.22 7.8 293
25 Southern Baluchistan 0.74±0.05 2.58 7.3 190
26 Eastern Afghanistan 0.89±0.04 5.59 8.3 534
27 Gujarat Region 0.87±0.06 1.31 8.0 93
28 Aravali-Bundelkhand 0.81±0.06 1.16 7.0 114
29 Southern Craton 1.19±0.08 0.47 6.8 45
30 Hindukush and Pamirs 0.93±0.01 83.54 8.0 6790

6
 31 Gangetic region 0.84±0.09 0.17 6.3 25
32 Bay of Bengal 0.60±0.08 0.49 6.7 53

Table 2. Coefficients in the spectral attenuation relation (eq.1) for Gujarat Region,
A-type sites

Period C1 C2 C3 C4 C5 C6 C7 C8 σ(ε)
0.0000 -4.7653 1.3935 -0.0162 -0.0020 -1.4183 0.0187 0.9899 0.1243 0.3596
0.0100 -4.7654 1.3925 -0.0162 -0.0020 -1.4180 0.0186 0.9906 0.1243 0.3597
0.0150 -3.7890 1.2437 -0.0057 -0.0020 -1.4138 0.0170 1.0015 0.1226 0.3661
0.0200 -3.4369 1.2058 -0.0033 -0.0021 -1.3996 0.0150 1.0151 0.1218 0.3650
0.0300 -3.2520 1.2062 -0.0038 -0.0020 -1.3782 0.0146 1.0161 0.1178 0.3631
0.0400 -3.2479 1.2148 -0.0045 -0.0020 -1.3680 0.0140 1.0203 0.1165 0.3638
0.0500 -3.3558 1.2533 -0.0074 -0.0020 -1.3690 0.0143 1.0155 0.1167 0.3643
0.0600 -3.6403 1.3297 -0.0132 -0.0019 -1.3622 0.0165 0.9978 0.1157 0.3664
0.0750 -3.9514 1.3939 -0.0184 -0.0019 -1.3468 0.0158 0.9996 0.1149 0.3681
0.0900 -4.3288 1.4850 -0.0254 -0.0019 -1.3347 0.0165 0.9904 0.1141 0.3688
0.1000 -4.6321 1.5609 -0.0310 -0.0019 -1.3326 0.0177 0.9826 0.1123 0.3696
0.1500 -6.1285 1.9439 -0.0597 -0.0018 -1.3206 0.0263 0.9280 0.1095 0.3758
0.2000 -7.6691 2.3366 -0.0890 -0.0018 -1.2994 0.0327 0.8962 0.1057 0.3798
0.3000 -10.5848 3.1084 -0.1458 -0.0017 -1.2888 0.0486 0.8389 0.1044 0.3886
0.4000 -12.9478 3.7214 -0.1896 -0.0017 -1.2751 0.0612 0.8082 0.1039 0.3950
0.5000 -14.8383 4.1990 -0.2231 -0.0017 -1.2652 0.0657 0.7974 0.1021 0.3964
0.6000 -16.6838 4.6394 -0.2532 -0.0016 -1.2477 0.0644 0.7973 0.0983 0.3964
0.7000 -17.9334 4.9467 -0.2737 -0.0016 -1.2447 0.0703 0.7889 0.0972 0.3949
0.7500 -18.4834 5.0638 -0.2811 -0.0016 -1.2368 0.0650 0.7976 0.0964 0.3959
0.8000 -19.0245 5.1951 -0.2890 -0.0016 -1.2468 0.0702 0.7914 0.0983 0.3954
0.9000 -20.0872 5.4162 -0.3023 -0.0016 -1.2375 0.0622 0.8044 0.0974 0.3949
1.0000 -20.5927 5.5113 -0.3075 -0.0016 -1.2298 0.0575 0.8143 0.0973 0.3944
1.2000 -21.8734 5.7343 -0.3176 -0.0014 -1.2272 0.0525 0.8334 0.0963 0.3904
1.5000 -22.8904 5.8770 -0.3202 -0.0013 -1.2438 0.0522 0.8446 0.0972 0.3882
2.0000 -23.9486 5.9483 -0.3140 -0.0012 -1.2567 0.0516 0.8548 0.1027 0.3902
2.5000 -24.3712 5.8437 -0.2969 -0.0011 -1.2440 0.0334 0.9136 0.1023 0.3970
3.0000 -24.4438 5.6416 -0.2715 -0.0011 -1.2528 0.0237 0.9662 0.1034 0.4019
4.0000 -24.1841 5.2717 -0.2297 -0.0011 -1.2726 0.0229 0.9826 0.1103 0.4096

7
 Table 3 Relative Seismic Hazard of Indian Cities on A-type Sites in terms of PGA

No. Agglomeration Lat. Long. PGA(g) PGA(g) PGA(g) PGA(g)

(TR=500) (TR=2500) (TR=5000) (TR=10000)
1 Guwahati 26.17 91.77 0.23 0.37 0.42 0.51
2 Chandigarh 30.75 76.78 0.10 0.19 0.25 0.32
3 Kolkata 22.55 88.37 0.09 0.19 0.24 0.31
4 Mumbai 19.00 72.80 0.09 0.19 0.25 0.32
5 Jabalpur 23.15 79.93 0.08 0.18 0.24 0.31
6 Asansol 23.68 86.98 0.08 0.17 0.21 0.27
7 Delhi 28.62 77.22 0.08 0.17 0.22 0.29
8 Srinagar 34.08 74.78 0.08 0.16 0.21 0.26
9 Jaipur 26.92 75.82 0.07 0.16 0.17 0.28
10 Meerut 28.99 77.70 0.06 0.14 0.18 0.23
11 Agra 27.18 78.02 0.06 0.14 0.18 0.25
12 Ahmedabad 23.03 72.57 0.07 0.14 0.17 0.24
13 Vijayawada 16.51 80.61 0.07 0.13 0.15 0.21
14 Jamshedpur 22.80 86.18 0.07 0.12 0.14 0.20
15 Dhanbad 23.80 86.45 0.06 0.11 0.13 0.18
16 Pune 18.52 73.85 0.06 0.11 0.13 0.16
17 Kozhikode 11.25 75.77 0.05 0.10 0.09 0.17
18 Kolhapur 16.70 74.23 0.05 0.10 0.11 0.11
19 Rajkot 22.30 70.78 0.05 0.09 0.11 0.13
20 Vadodara 22.30 73.20 0.05 0.09 0.10 0.16
21 T’nanthapuram 8.480 76.95 0.04 0.08 0.10 0.11
22 Kochi 9.970 76.27 0.03 0.08 0.11 0.13
23 Ludhiana 30.91 75.85 0.04 0.08 0.10 0.14
24 Indore 22.42 75.54 0.04 0.08 0.10 0.12
25 Patna 25.60 85.12 0.04 0.08 0.11 0.13
26 Lucknow 26.83 80.92 0.04 0.08 0.11 0.17
27 Kanpur 26.46 80.33 0.03 0.08 0.11 0.17
28 Surat 21.23 72.78 0.04 0.08 0.09 0.14
29 Trichy 10.81 78.69 0.04 0.07 0.09 0.11
30 Ranchi 23.35 85.33 0.03 0.07 0.09 0.11
31 Allahabad 25.45 81.85 0.03 0.07 0.09 0.13
32 Amritsar 31.64 74.86 0.03 0.07 0.09 0.11
33 Coimbatore 11.04 76.96 0.04 0.07 0.09 0.11
34 Nagpur 21.15 79.08 0.04 0.07 0.09 0.10
35 Jodhpur 26.28 73.02 0.03 0.06 0.08 0.11
36 Nashir 20.00 73.78 0.03 0.06 0.08 0.10
37 Gwalior 26.14 78.10 0.02 0.05 0.07 0.10
38 Madurai 9.800 78.10 0.03 0.05 0.07 0.10
39 Varanasi 25.32 82.98 0.02 0.05 0.06 0.07
40 Bhopal 23.25 77.42 0.03 0.05 0.07 0.09
41 Hyderabad 17.37 78.48 0.02 0.05 0.06 0.07
42 Chennai 13.05 80.27 0.03 0.05 0.06 0.07
43 Solapur 17.68 75.92 0.02 0.04 0.05 0.06
44 Bhubaneswar 20.27 85.84 0.02 0.04 0.05 0.06
45 Bangalore 12.97 77.58 0.02 0.04 0.05 0.06
46 Aurangabad 19.78 75.29 0.02 0.03 0.03 0.04
47 Visakhapatnam 17.07 83.25 0.02 0.03 0.04 0.04
48 Raipur 21.23 81.63 0.01 0.02 0.02 0.03

8
Fig.1 Past Epicenters Superposed on known Faults

9
Figure 2 Thirty-two Seismogenic Zones of India for PSHA

10
Fig 3 Seven Geological Provinces and Cities with Population > 1 Million

Fig 4 (a) Typical shear wave velocity profile

at a Granitic outcrop. (A-type site)

11
Fig 4 (b) Typical Surface Expressions at Granitic Outcrop
(Likely A-type Site)

12
Fig 5 (a) PGA Contours with 2% probability of exceedence in 50 years (Return
Period ~2500 years) on A-type Sites

13
Fig 5(b) PGA Contours with Return Period of ~10,000 years on A-type Sites

14
Fig. 5(c) Short Period Spectral Acceleration at T = 0.2 second with Return Period of 2500
years on A-type Sites (5% damping)

15
Fig. 5(d) Short Period Spectral Acceleration at T =0.5sec and 5% damping with 2%
probability of exceedence in 50 years (Return Period ~2500 years) on A-type sites

16
Fig. 5(e) Long Period Spectral Acceleration at T = 1 second with Return Period of 2500
years on A-type Sites (5% damping)

17
Fig. 5(f) Long Period Spectral Acceleration at T =1.25sec and 5% damping with 2%
probability of exceedence in 50 years (Return Period ~2500 years) on A-type Sites

18
 Fig.6. Geological Sketch Map of India showing Cambay Basin separating
Saurashtra from the Mainland. Ahmadabad and Baroda are located on the Cambay
Basin. (Ref: Choudhary 2004)

19