SIESMIC RISK IN SRI LANKA - A PROBABILISTIC APPROACH

Abstract
The objectives of this study is to develop a uniform hazard response spectra for Colombo City of Sri
Lanka which can be used for design community of Sri Lanka.

In the Current study probabilistic seismic hazard assessment was carried out for the Colombo city to
produce uniform hazard response spectra. A comprehensive earthquake catalogue has been complied
for the region bounded by latitudes 0ºN to 20ºN and longitudes 70ºE to 90ºE based on historical and
instrumental earthquake data from 1063 to 2012. The PGA and SA were calculated within logic tree
frame work incorporating different parameters such as maximum cutoff magnitude and ground
motion predictive equations for shallow crustal intraplate earthquakes. The study produced a PGA of
0.107g for 10% probability of exceedance in 50 years or 475 year return period.

Keywords: Earthquake, PGA, SA, attenuation

1. Introduction
Sri Lanka is located within a tectonic plate known as "Indo-Australia plate". According to the
historical records a very few number of earthquakes were recorded within the country. Historical
records indicate that there was a devastating earthquake (Mw=6.4) in 1615 in Sri Lanka. The
epicenter of this earthquake was situated in Colombo and it caused to damage about 200 houses with a
casualty figure of over 2000. The Indian Ocean Tsunami hit on 26th December 2004 following the
Magnitude of 9.4 due to Sumatra earthquake damaged most of the coastal areas in Sri Lanka. But
there was no observed significant direct damages due to ground shaking. After the Tsunami many
local and international organizations were involved in reconstruction process. Local insurance
companies demanded for paying damaged properties. They highly emphasized that damaged
buildings were not designed to accommodate seismic as well as tsunami loads. So, there was a
demand for seismic resistant designs.

Now Sri Lanka is developing fast after the end of civil war in 2009. Government of Sri Lanka has
taken necessary action to orient the country’s development plans to meet Millennium Development
goals. According to that Colombo city will be developed as commercial hub such as Singapore and
Dubai. Lot of high rise buildings will be constructed in the Colombo city. Then seismic resistant
design will be playing a major role in the design process. There are no specific guidelines for seismic
resistant designs available locally on how to meet the new requirements in order to satisfy the
expectations of various parties. Cost factor is very important in any project. So, new optimizations
tools will be used to carrying out economical designs. The performance based design methodology is
the most common method that engineers can be used to produce economical designs. The objective of
this study was set out to fulfil this requirement by introducing a suitable response spectrum by
carrying out a Probabilistic Seismic Hazard Assessment (PSHA) for Colombo City. Because PSHA is
the most widely used procedure to determine the ground motion parameters such as peak ground
acceleration to which structure has to be designed by an engineer.

The basic procedure of PSHA was initially developed by Cornell [1968] and its computerized form
was implemented by McGuire [1976 and 1978]. Modern PSHA has gradually evolved by
incorporating additional terms and computational tools in order to better represent seismic hazard.
The basic methodology involves computing how often a suite of specified levels of ground motion
will be exceeded at the site. The general procedure for a Cornell-McGuire PSHA comprises four
fundamental steps.

The first step involves the identification and delineation of all potential sources of seismicity that may
affect the site or sites of interest. These sources of seismicity may be represented as area sources, fault
sources, or, point sources, depending upon the geological nature of the sources and available data. In
the second step, the temporal behaviour of earthquakes is assumed to follow a Poissonian process and
it is determined for each source by establishing a magnitude recurrence relationship over the range of
magnitudes that are likely to be generated by each seismic source. The third step involves the use of
GMPEs to establish the conditional probability of exceedance of a pre-specified ground motion value
for each site given the occurrence of an earthquake at a particular magnitude and location and the final
step of the analysis computes the annual number of events that produce a ground motion parameter,
e.g. SA that exceeds a specified level, z. This number of events per year, ν, is also called the “annual
frequency of exceedance”. The inverse of ν is called the “return period”. Several probability
distributions for each seismic source defined in the previous steps are introduced.

2. Earthquake Catalogue

An earthquake catalogue was compiled for an area bounded by latitudes 0ºN to 20ºN and longitudes
70ºE to 90ºE based on historical and instrumental earthquake data from 1063 to 2012. Data was
collected from different sources such as Menon et al.(2010), Chandra (1977), Rao and Rao (1984),
Guha and Basu (1993), Iyengar et al.(1999), Jaiswal and Sinha (2007), Abeykoon (1995), Navin Peris
(2007), Earthquake catalogue prepared by National Disaster Management Authority, New Delhi,
India (2011) and Internationally recognized earthquake databases on the internet, such as the National
Earthquake Information Centre (NEIC), the International Seismological Centre (ISC), the
Incorporated Research Institutions for Seismology (IRIS), and the Geological Survey and Mines
Bureau, Sri Lanka (GSMB). Duplicate events were manually detected and deleted.

To remove foreshocks and aftershocks Declustering method proposed by Gardner and Knopoff
(1974) was used. Completeness analysis of the catalogue was done according to the method
proposed by Stepp (1973). Figure 1 shows plot of earthquake catalogue.

3. Seismic Source Zonation and Magnitude Recurrence Relationship

The seismic zonation is usually defined by associating the seismic to the tectonic and geologic
settings in the area. The major tectonic structures aground Sri Lanka are located on the boundaries of
the Indo-Australian plate. The nearest structure are the Sumatra subduction zone to the east and the
extensional/transform fault structures of the central Indian Ridge to the west.

The geological history of Sri Lanka is such that 90% of the island consists of Precambrian
metamorphic and granitoid rocks. The Precambrian basement is divided into three major units: central
Highland Complex (HC), Wanni Complex (EC) and the Vijayan Complex (VC). The Kadugannawa
Complex (KC) is a minor unit within the Highland Complex. HC is the oldest unit thrusting upward to
form the central highlands with the highest uplift along a SW-NE-ENE belt from Galle-Rakwana-
Horton Plains to Batticaloa. The uplift in the Jurassic (post-Precambrian) period was attributed to the
movement of the Sri Lanka mini-plate in S-SE direction relative to the Indian plate (Vithanage ,1995).
 Figure 1: Plot of earthquake catalogue

Numerous lineaments were therefore identified in the HC with major lineaments along the belt of
highest uplift and the N-S trending Mahaweli lineament where the Mahaweli River takes a straight
course for about 60km. The micro- seismicity within the island recorded by the Kotmale micro-
seismic network could be closely associated with these lineaments in the central highland region
(Vitanage, 1995 and Fernando,Kulasinghe,1986). Seismicity of Sri Lanka that could have an impact
on the island is located outside as well as within the island. Various researchers have studied the
seismicity and tectonics within and outside of Sri Lanka in order to highlight the potential seismic
hazard. Many Indian scientists have carried out studies to define seismic characteristic of peninsular
India. A. Menon, T. Ornthammarath, M. Corigliano, and C. G. Lai had carried out a Probabilistic
Seismic Hazard Macrozonation of Tamil Nadu in Southern India in 2010. They have scientifically
identified eleven seismogenic zones related to the Southern part of the India around Tamil Nadu
including Sri Lanka. Based on above information, Same seismogenic zones classification is used for
this study because Tamil Nadu is geographically very close to Sri Lanka. Figure 2 shows these
seismic zones.

Eight out of these eleven seismic zones are related to the current study. In addition to these seismic
area sources, line source was identified 90 km away from Colombo city along the Cameron ridge in
the western sea. Then, Truncated Gutenberg-Richter relationships were developed to these seismic
area sources. Figure 3 shows the Truncated G-R Recurrence relationship obtained in the current study
for Seismic Zone 9.
 Figure 2: Different Seismic Zones

Variation of Log(N) vs Mw for SZ9

-0.5

-1
Log(N)

-1.5

-2
y = -0.5777x + 1.209

-2.5
3 3.5 4 4.5 5 5.5 6
Magnitude (Mw)

Figure 3: Truncated G-R Relationship for SZ9

4. Attenuation Relationships

Three ground motion attenuation relationships were chosen in this PSHA to derive the PGA and the
spectral acceleration values after comparison with Peninsular India(PI) strong motion data. The
relationship by Abrahamson and Silva(1997) was developed from earthquakes recorded in worldwide
specially west and north America and recommended to use for shallow crustal events worldwide.
Raghu Knath and Iyengar(2007) developed an empirical relationship for Peninsular India based on a
stochastic seismological model and subsequently compared to instrumental data from Koyona (1967)
and Bhuj (2001) earthquakes in India. In addition to these two equations, the attenuation relationship
developed by Campbell and Bozorgnia(2008) which suitable for western United states and similar
tectonically active regions of shallow crustal faulting has been chosen. Attenuation of ground
motion predicted by these three relationship is compared to strong motion records of two earthquakes
from PI. Four strong motion records from the Jabalpur earthquake(1997) and a Bhuj aftershock(2001),
both of Magnitude Mw 5.7 (Singh et al.,2003) have been used in the comparisons(Ref. Figure 4). The
shortest distance to the area source boundary was used as epicentral distance in these attenuation
relationships to calculate the PGAs and spectral accelerations except Zone 9 which Sri Lanka lies.

Figure 4: Comparison of Three GMPEs in terms of PGA for Mw 6.5

5. Deaggregation and Logic Trees

The hazard curve gives the combined effect of all magnitudes and distances on the probability of
exceeding a given ground motion level. Since all of the sources, magnitudes, and distances are
considered together, it is difficult to understand of what is controlling the hazard from the hazard
curve by itself. To provide insight into what events are the most important for the hazard, the hazard
at a given ground motion level is broken down into its contributions from different earthquake
scenarios. This process is called deaggregation [e.g. Bazzurro and Cornell, 1999]. This process was
carrying out for Zone 9 to find the PGA and SA.

Logic tree methodology was used to address the epistemic uncertainty of various parameters and
relationships in the PSHA calculation. The maximum cut off magnitude is based on the maximum
historical earthquake(MHE) in each source zone of the earthquake catalogue. As an alternative, the
MHE increased by 0.3 units has been considered. 60% of weight have been assigned to the maximum
magnitude and balance 40 % of weight was assigned to MHE increased by 0.3 units. For GMPEs, it
was difficult to assign a higher weight to one equation over the other, and hence equal weight have
been assigned. Figure 5 shows this distribution.

Figure5: Parameters and weighting factors adopted in the logic tree

6. Results

Figure 6 shows the calculated Response Spectra(RS) for 50 year, 475 year and 2475 year return
periods at rock sites in Colombo at bed rock level for near field earthquake (Earthquake 30 km away
from Colombo) . The results shows that the PGA at rock site for 10% of probability of exceedance
in 50 years or 475 years return period is 0.107g for Colombo.

Similarly RS was obtained for the PGA at rock site for 2% of probability of exceedance in 50 years
or 2475 years and also for 50 year return periods. The PGA values were 0.207g and 0.025g
respectively. Figure 7 shows the calculated Response Spectra(RS) for 50 year, 475 year and 2475 year
return periods at rock sites in Kandy at bed rock level for near field earthquake.
 Response Spectrum
0.50
0.45
Spectral Acceleration (g)

0.40
0.35 475 yrs
0.30
50 yrs
0.25
0.20
0.15
0.10
0.05
0.00
0 1 2 3 4 5
Period(s)

Figure 6: RS for 50 yr, 475 yr and 2475 yr Return periods in rock site in Colombo

Response Spectrum
0.12
Spectral Acceleration (g)

0.10

0.08 475 yrs

50 yrs
0.06

0.04

0.02

0.00
0 1 2 3 4 5
Period(s)

Figure 7: RS for 50 yr, 475 yr and 2475 yr Return periods in rock site in Kandy

The calculated PGA values for 50 year, 475 year and 2475 year return periods were 0.005g, 0.025g
and 0.057g respectively. PGA values have been reported for main twenty one cities in Sri Lanka for
50 year, 475 year and 2475 year return periods for earthquake occurred 30 km away from Colombo
city are shown in Table 1 and PGA values have been reported for same cities for same return periods
for earthquake occurred 90 km away from Colombo city are shown in Table 2.
 Latitude Longitude PGA/(g)
No City
/(North) /(East) T=50 T=475 T=2475
1 Ampara 7.28 81.67 0.002 0.011 0.027
2 Anuradhapura 8.35 80.38 0.007 0.034 0.074
3 Badulla 6.98 81.06 0.003 0.017 0.040
4 Baticaloa 7.72 81.70 0.002 0.011 0.028
5 Colombo 6.93 79.84 0.025 0.107 0.207
6 Dambulla 7.86 80.65 0.004 0.025 0.056
7 Galle 6.04 80.22 0.009 0.045 0.094
8 Hambantota 6.12 81.12 0.003 0.016 0.038
9 Horana 6.72 80.05 0.013 0.062 0.127
10 Jaffna 9.67 80.00 0.041 0.158 0.325
11 Kandy 7.30 80.64 0.005 0.025 0.057
12 Kurunegala 7.48 80.37 0.007 0.035 0.076
13 Maho 7.82 80.28 0.008 0.040 0.086
14 Mannar 8.98 79.90 0.033 0.134 0.280
15 Matara 5.95 80.53 0.005 0.028 0.061
16 Mullaitivu 9.23 80.53 0.008 0.040 0.101
17 Nuwara eliya 6.97 80.77 0.004 0.022 0.050
18 Polonnaruwa 7.93 81.00 0.003 0.018 0.042
19 Puttalam 8.03 79.84 0.029 0.122 0.233
20 Rathnapura 6.67 80.40 0.006 0.033 0.073
21 Trincomalee 8.57 81.23 0.009 0.026 0.048
Table 1: PGA values 21 cities in Sri Lanka for near field earthquake (30km)

Latitude Longitude PGA/(g)

City
No /(North) /(East) T=50 T=475 T=2475
1 Ampara 7.28 81.67 0.001 0.008 0.021
2 Anuradhapura 8.35 80.38 0.003 0.019 0.044
3 Badulla 6.98 81.06 0.002 0.012 0.029
4 Baticaloa 7.72 81.70 0.001 0.008 0.022
5 Colombo 6.93 79.84 0.006 0.033 0.072
6 Dambulla 7.86 80.65 0.003 0.015 0.037
7 Galle 6.04 80.22 0.004 0.022 0.051
8 Hambantota 6.12 81.12 0.002 0.011 0.028
9 Horana 6.72 80.05 0.005 0.026 0.059
10 Jaffna 9.67 80.00 0.011 0.053 0.125
11 Kandy 7.30 80.64 0.003 0.016 0.037
12 Kurunegala 7.48 80.37 0.003 0.019 0.045
13 Maho 7.82 80.28 0.004 0.021 0.048
14 Mannar 8.98 79.90 0.011 0.053 0.125
15 Matara 5.95 80.53 0.003 0.017 0.039
16 Mullaitivu 9.23 80.53 0.008 0.040 0.101
17 Nuwara Eliya 6.97 80.77 0.002 0.014 0.034
18 Polonnaruwa 7.93 81.00 0.002 0.012 0.030
19 Puttalam 8.03 79.84 0.007 0.035 0.075
20 Rathnapura 6.67 80.40 0.003 0.019 0.044
21 Trincomalee 8.57 81.23 0.009 0.026 0.048
Table 2: PGA values 21 cities in Sri Lanka for far field earthquake (90km)
Sri Lanka Hazard Maps

Seismic hazard maps contour maps have been prepared by performing hazard computations. The
PSHA results obtained by combining six branches of the logic tree. Six number of contour maps have
been produced for mean values of PGA for 50 year, 475 year and 2475 year return periods for near
field and far field earthquakes. The contour maps are illustrated in Figures 8-13.

Figure 8- SL Hazard map for PGA at 50 yr return Figure 9- SL Hazard map for PGA at 475 yr
period for near field earthquake (30km) return period for near field earthquake (30km)
Figure 10- SL Hazard map for PGA at 2475 yr return Figure 11- SL Hazard map for PGA at 50 yr
period for near field earthquake (30km) return period for near field earthquake (90km)

Figure 12- SL Hazard map for PGA at 475 yr return Figure 13- SL Hazard map for PGA at 2475 yr
period for near field earthquake (90km) return period for near field earthquake (90km)

7. Conclusions

The scope of the research work presented here was to carry out probabilistic seismic hazard
assessment and develop response spectra for main Cities of Sri Lanka which can be used for
designing of structures for earthquake loading in Sri Lanka and development of seismic hazard map
for Sri Lanka. The study produced a PGA of 0.107g for 10% probability of exceedance in 50 years or
475 year return period.

8. Acknowledgements

Authors would like to express deepest gratitude to my supervisosr prof. P.B.R.Dissanayake, Senior
Lecturer, Department of Civil Engineering, Faculty of engineering University of Peradeniya and Dr.
K.K. Wijesundara, Senior Lecture, South Asian Institute of Technology and Medicine, for their
untiring guidance and continuous encouragement throughout this research project.Autors would also
like to express their gratitude Dr. D. menon and Dr. T. Ornthammarath and Dr. Navin peris for their
valuable support to the study. Finally I thank members of my family for their kind support provided in
numerous ways to make this study a success
References

Abrahamson, N.A. and Silva, W.A.[1997], Emperical response spectral attenuation relations for
shallow crustal earthquakes, Seism. Res. Lett. 68, no. 1 94-127.

Abrahamson, N.[2006], seismic Hazard assessment: problems with current practice and future
developments, First European conference on earthquake engineering and seismology.

Biswas, S., Majumdar, R.K. [1997] “Seismicity and tectonics of the Bay of Bengal: Evidence for
intraplate deformation of the Northern Indian Plate”, Tectonophysics, Vol. 269, pp. 323-336.

Campbell, K. W., and Y. Bozorgnia[2008] , NGA ground motion model for the geometric mean
horizontal component of PGA, PGV, PGD and 5% damped linear
elastic response spectra for periods ranging from 0.01 to 10 s, Earthq. Spectra 24, no. 1, 139–171.

Chandra, U. [1977] “Earthquakes of Peninsular India - A seismotectonic study”, Bulletin of the

Seismological Society of America, Vol. 87, No. 5, pp. 1387-1413.

Guha, S.K., Basu, P.C. [1993] “Catalogue of earthquakes (M ≥ 3.0) in Peninsular India”, Tech.
Document No. TD/CSE-1, Atomic Energy Regulatory Board, Mumbai, India, pp. 1-70.
Iyengar, R.N., Sharma, D., Siddiqui, J.M. [1999] “Earthquake history of India in medieval times”,
Indian Journal of History of Science, Vol. 34, No. 3., pp. 181-237.

Jaiswal, K., Sinha, R. [2007] “Probabilistic seismic-hazard estimation for peninsular India,” Bull.
Seismol. Soc. Am. Vol. 97, No. 1, pp. 318–330.

Johnston, C., Seismic moment assesment of earthquakes in stable continental regions-I Instrumental
seismicity. Geophysics. J. Int. [1996],124,381-414.
Rao, R.B., Rao, S.P. [1984] “Historical seismicity of Peninsular India”, Bulletin of the Seismological
Society of America, Vo. 74, No. 6, pp. 2519 – 2533.

Rao, R.B. [1992] “Seismicity and geodynamics of the low-to high-grade transition zone of Peninsular
India”, Tectonophysics, Vol. 201, pp. 175-185.

Rao, R.B. [2000] “Historical seismicity and deformation rates in the Indian Peninsular Shield”,
Journal of Seismology, Vol. 4, pp. 247 – 258.

Ramasamy,S.M.[2006]“Remote sensing and

active tectonics of South India”, International Journal of Remote Sensing, Vol. 27, No. 20, pp. 4397 –
4431.

Ramasamy, S.M., Balaji, S. [1995] “Remote sensing and Pleistocene tectonics of Southern Indian
Peninsula”, International Journal of Remote Sensing, Vol. 16, No. 13, pp. 2375 – 2391.

Menon A, Ornthammarath T, Corigliano M, Lai CG [2010] “Probabilistic Seismic Hazard

Macrozonation of Tamil Nadu in Southern India” Bull. Seism. Soc. Am. Vol. 100, no. 3, pp. 1320-
1341.