Proceedings of the 12th Asian Congress of Fluid Mechanics

18-21 August 2008, Daejeon, Korea

COMPUTATIONAL MODELLING OF THE IMPACT OF 2004 TSUNAMI ON

THE CITY OF HAMBANTOTA IN SRI LANKA

J. J. Wijetunge

Department of Civil Engineering, University of Peradeniya, Sri Lanka, janakaw@pdn.ac.lk

ABSTRACT: On 26th December 2004, coastal belts of Sri Lanka as well as several other countries
bordering the Indian Ocean suffered enormous loss of life and damage to property owing to the tsunami
unleashed by the third largest earthquake ever recorded. In order to mitigate potential loss of lives from a
similar event in the future we need to provide advance warning of an approaching tsunami and then
quickly evacuate vulnerable coastal communities to safer areas. The detailed planning required for such
evacuation is usually carried out based on tsunami hazard zonation maps. Accordingly, the present paper
outlines the numerical modelling carried out to develop a high-resolution tsunami hazard map, as a case
study, for the city of Hambantota on the south coast of Sri Lanka, which was devastated by the 2004
tsunami. The results give the spatial distribution of the dpeth of inundation as well as the maximum flow
velocities due to an event similar to the 2004 tsunami.

1. INTRODUCTION
The mega-tsunami of 26th December 2004 unleashed by the earthquake of moment magnitude 9.1 - 9.3
in the Sumatra-Andaman subduction zone caused enormous loss of lives and damage to property in Sri
Lanka and in several other countries bordering the Indian Ocean. The unprecedented disaster clearly
underscored the need to have a proper system in place for tsunami early warning as well as for quick
evacuation of vulnerable coastal communities to safer areas. One essential pre-requisite in planning such
evacuation is the development of inundation maps which provide a graphical presentation of damage-
prone areas with expected depth of flooding and flow velocities in case of a similar tsunami attack.
Accordingly, the present paper outlines the numerical modelling of tsunami propagation and inundation
carried out to develop a detailed tsunami hazard zonation map, as a case study, for the city of Hambantota
on the south coast of Sri Lanka, which was devastated by the 2004 tsunami.

2. NUMERICAL MODELLING
In the following, we outline the coseismic source model and the hydrodynamic model employed to
simulate the generation and the propagation of the 2004 tsunami, respectively. The nested grid set-up
used in the hydrodynamic model simulations is also described.

2.1 Coseismic Source Model

The present simulations employed the initial surface elevation from the coseismic source model
proposed by Grilli et al. [1] who employed available seismic data as well as hydrodynamic data
pertaining to the generation and propagation of the tsunami to iteratively develop a 5-segment coseismic
tsunami source for the 26th December 2004 event. In a subsequent study [3], the numerical simulations
based on this source model showed good agreement with measured run-up heights along the Andaman
coast of Thailand.
The contours of initial surface elevation from the above co-seismic source model computed using
Okada’s [5] dislocation theory and reported in Grilli et al. [1] is shown in Fig. 1 whilst the corresponding
source parameters of the dislocation segments are given in Table 1. Further, it is assumed here that the
sea surface follows the sea bed deformation instantaneously. Note that, in Fig. 1, the continuous lines
represent uplift and dashed lines represent subsidence, both at 1 m contour intervals in the range -5 m to
+8 m. The background bathymetry is also plotted in grey at 500 m contour intervals.

2.2 Tsunami Propagation and Inundation Model

The mathematical model used in the present work is the Cornell Multi-grid Coupled Tsunami Model
(COMCOT, coded in FORTRAN 90) which solves the non-linear shallow water equations on a
dynamically coupled system of nested grids using a modified leap-frog finite difference numerical
 Proceedings of the 12th Asian Congress of Fluid Mechanics
18-21 August 2008, Daejeon, Korea

scheme. The model has been validated by experimental data [4] and has already been successfully used
to investigate several historical tsunami events [7].
The bathymetry data for the largest grid employed in the simulations, i.e., Region 1 shown in Fig. 2,
was obtained by interpolating ETOPO2 data with a resolution of 2 arc minutes to a grid of 0.6765 arc
minutes (~1250 m) spacing. Region 2, which is embedded in Region 1, is also shown in Fig. 2. The
bathymetry data for Region 2 with a grid resolution of 0.1353 arc minutes (~250 m) shown in Fig. 3 as
well as for Region 3 (grid spacing 50 m, Fig. 4) was at first interpolated from ETOPO2 data and was then
updated with data from the available navigation charts. A fourth level grid, Region 4, with a grid spacing
of 10 m was used to cover the city of Hambantota (see Fig. 4 for location of this grid with respect to
Region 3). The topographic data for this grid was obtained from high resolution LIDAR survey data (The
precise digital earth model of the coastal areas of Sri Lanka, Project Director: Prof. Fabrizio Ferrucci,
Italy) made available to the author by the Ministry of Disaster Management and Human Rights of the
Government of Sri Lanka. These LIDAR data have been acquired at a horizontal resolution of 1 m and a
vertical resolution of not more than 0.3 m and were originally projected on to the UTM WGS84 - Zone
44N coordinate system.

Table 1 Source Parameters Used in Fault

Plane Model of Grilli et al. [1]
Segment/ 1 2 3 4 5
Parameter
Length (km) 220 150 390 150 350
Latitude (in deg.)

Width (km) 130 130 120 95 95

Dip (deg.) 12 12 12 12 12
Rake (deg.) 90 90 90 90 90
Strike (deg.) 323 348 338 356 10
Depth (km) 25 25 25 25 25
Slip (m) 18 23 12 12 12
Longitude of 94.57 93.90 93.21 92.60 92.87
Centroid (oE)
Latitude of 3.83 5.22 7.41 9.70 11.70
Centroid (oN)

Longitude (in deg.)

Fig. 1 Initial surface elevation for tsunami source.
Continuous lines represent uplift and dashed lines
represent subsidence, both at 1 m contour
intervals in the range -5 m to +8 m
(after Ioualalen et al. [3]).

Elevation (m)

Elevation (m)
Latitude (in deg.)
Latitude (in deg.)

Region 3

Region 2

Longitude (in deg.)

Fig. 3 Region 2 of the computational

Longitude (in deg.) domain; location of Region 3 is also shown.
Fig. 2 Region 1 of the computational
domain; location of Region 2 is also shown.
 Proceedings of the 12th Asian Congress of Fluid Mechanics
18-21 August 2008, Daejeon, Korea

The amplitude of the 2004 Indian Ocean Region 4

Latitude (in deg.)

tsunami during its propagation was of the
order of magnitude of 1 m, whilst the
-20 m
wavelength of the leading wave was in the -50 m
order of magnitude of 100 km. Thus, linear
shallow water equations are adequate to
solve tsunami propagation in Regions 1, 2
and 3. On the other hand, in the inundation Longitude (in deg.)
areas in Region 4, the water depth becomes
very small and approaches zero at the tip
of the surging bore. Thus, the non- Elevation (m)
linearity, i.e., the wave amplitude to depth Fig.
Fig. 47 Region 3 of the computational domain; location of
ratio, could become significant. However, Region 4 is also shown.
owing to the shallowness of the water, the
frequency dispersion effects, which are represented by the water depth to wave length ratio, can still be
negligible [7]. Considering these, we use non-linear shallow water equations inclusive of bottom
frictional terms to describe the tsunami overland flow in Region 4. Bottom roughness is represented by
Manning’s formulation with a roughness coefficient of n = 0.013 for the sea bed, n = 0 for the salt lakes
and n = 0.02 for the rest of the city.

3. MODEL RESULTS
The computed arrival time contours for the city of Hambantota are shown in Fig. 5. The numerical
simulations suggest an arrival time of around 130 minutes for the central area of the city whilst Inoue et
al. [2] report that the clock on a tower located near the bus station had stopped at 9:22 AM local time, i.e.,
143 min after the earthquake, probably corresponding to the arrival of the maximum wave which
according to some eyewitnesses was the first wave. However, another eyewitness account reported in [2]
suggests an arrival time of 130 minutes for the first wave.
The model results of the spatial distribution of the
maximum depth of inundation that would be caused by
an event similar to the tsunami in December 2004 in the
city of Hambantota is shown in Fig. 6. The arrows
6.14

indicate the primary flow paths of onshore flooding.

128
Latitude (in deg.)

Following Walsh et al. [6], the computed tsunami

inundation is shown on these maps in three colour-coded 127
depth ranges: less than 0.5 m, 0.5 - 2 m, and greater than
126
6.12

2 m. These depth ranges have been chosen because they

are approximately knee-high or less, knee-high to head- 125
high, and more than head-high. The probable limit of
tsunami inundation is the landward edge of the light grey
zone. 124
6.10

The corresponding map of the magnitude of the

123min.
computed maximum flow velocity is shown in Fig. 7
using three color bands corresponding to velocities less
than 1.5 m/s, between 1.5 and 3.0 m/s, and greater than 81.12 81.14
3.0 m/s. The arrows giving the direction of the maximum Longitude (in deg.)
flow velocities are not shown on these figures for reasons
of clarity. Fig. 5 Computed tsunami arrival time
The numerical simulation indicates that the tsunami contours for the city of Hambantota.
first attacked the southern part of the city followed by the shoreline encompassing the bay with a
remarkably strong surge of water rushing through the neighborhood of location A in Fig. 6; note flow
depths and speeds exceeding 2 m (Fig. 6) and 3 m/s (Fig. 7), respectively. Note that the onshore lands in
the lower southern part of the bay is quite steep and hilly whilst there is a sand dune of height about 5 m -
10 m along the upper part of the bay to the north-east of location A. Most of this strong surge travels
towards Saltern-1 probably picking up further momentum as the land in the vicinity is gently sloping
towards the saltern whilst the remaining part runs to Saltern-2 getting chanelled through the sand dune on
 Proceedings of the 12th Asian Congress of Fluid Mechanics
18-21 August 2008, Daejeon, Korea

the right and the hilly area to the left of the main flow as indicated by the arrows. Meantime, flood water
also rushes through location B, where the sand dune had been breached, towards both Saltern-2 and
Saltern-3. It is clear that, on the whole, the comparatively vast extents of the salterns have also acted as
sinks to absorb and spread water whilst the sand dunes, where present, have helped protect the settlements
in their shadow from direct tsunami attack.

Saltern-3 Saltern-3

Saltern-2

Latitude (in deg.)

Saltern-2
Latitude (in deg.)

6.14 6.14
B B

Saltern-1 Saltern-1
A A
SEA SEA

6.12 6.12

81.12 81.14 81.12 81.14

Longitude (in deg.) Longitude (in deg.)
Depth of inundation: Magnitude of maximum velocity:
< 0.5 m 0.5 – 2 m 2–7m < 1.5 m/s 1.5 - 3 m/s > 3 m/s

Fig. 6 Computed tsunami inundation depths in Fig. 7 Computed magnitude of maximum

the city of Hambantota. velocity in the city of Hambantota.

4. CONCLUSIONS
The onshore flooding caused by the 2004 Indian Ocean tsunami in the city of Hambantota on the south
coast of Sri Lanka has been numerically modelled to develop a tsunami hazard map depicting the spatial
distribution of the maximum depth of flooding as well as the flow velocities. The model simulations
clearly indicate the dominant influence of the geomorphological features of coastal lands on the spatial
distribution of tsunami inundation. The hazard map developed should be useful in planning evacuation
and in public education and awareness activities, but not intended for land-use regulation.

ACKNOWLEDGMENTS
A part of the work described in the present paper was carried out with financial support from USAID/US Indian
Ocean Tsunami Warning System (IOTWS) Project Grant No.04-05-IOTWS-06. The author also wishes to thank
Prof. Philip Liu of Cornell University for making the COMCOT code available to carry out the model simulations.

REFERENCES
[1] Grilli, S. T. et al., Source constraints and model simulation of the December 26, 2004 Indian Ocean tsunami, J.
Waterway, Port , Coastal & Ocean Eng., Vol. 133, No. 6, 414-428, 2007.
[2] Inoue, S. et al., Field survey of tsunami effects in Sri Lanka due to the Sumatra-Andaman earthquake of
December 26, 2004, Pure Appl. Geophys., Vol 164, pp. 395 – 411, 2007.
[3] Ioualalen, M. et al., Modeling the 26 December 2004 Indian Ocean tsunami: Case study of impact in Thailand,
Journal of Geophysical Research, Vol. 112, No. C7, 2007.
[4] Liu, Philip L.-F., Cho, Y.-S., Briggs, M. J., Synolakis, C.E., and Kanoglu, U., Run-up of Solitary Waves on a
Circular Island, J. Fluid Mechanics, Vol. 302, pp. 259-285, 1995.
[5] Okada, S., Surface displacement due to shear and tensile faults in a half-space, Bull. Seismol. Soc. Am., Vol. 75,
pp. 1135–1154, 1985.
[6] Walsh, T. J., Titov, V. V., Venturato, A. J., Mofjeld, H. O., and Gonzalez, F. I., Tsunami hazard map of the
Elliott Bay Area, Seattle, Washington Division of Geology and Earth Resources Report No. 2003-14, 2003.
[7] Wang, X., and Liu, Philip L. -F., An analysis of 2004 Sumatra earthquake fault plane mechanisms and Indian
Ocean tsunami, Journal of Hydraulic Research, Vol. 44, No. 2, pp. 147 – 154, 2006.