See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228772034

Flood hazard mapping in lower reach of Kelani River, Sri Lanka

Article · January 2008

CITATIONS READS
7 5,035

5 authors, including:

J.s M. Fowze Manzul Kumar Hazarika

Central Engineering Consultancy Bureau Asian Institute of Technology
10 PUBLICATIONS 152 CITATIONS 50 PUBLICATIONS 1,342 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Forest mapping using radar View project

UAV Based Mapping and Applications View project

All content following this page was uploaded by Manzul Kumar Hazarika on 20 February 2014.

The user has requested enhancement of the downloaded file.

 FLOOD HAZARD MAPPING IN LOWER REACH OF KELANI RIVER, SRI LANKA

J. S. M. FOWZE1, I. P. A. GUNASEKARA2, P. P. LIYANAGE3, M. K. HAZARIKA1, L.

SAMARAKOON1
1
Geoinformatics Center, Asian Institute of Technology, Pathumthani, Thailand
emails: [email protected], [email protected], [email protected]
2
Irrigation Department, Colombo, Sri Lanka
email: [email protected]
3
Survey Department, Colombo, Sri Lanka
email: [email protected]

KEYWORDS: Flood Hazard Mapping, Remote Sensing, GIS

ABSTRACT: Although floods are the most frequent natural disasters in Sri Lanka they had
received little attention until the tsunami of December 2004. After the Tsunami, however,
disaster management strategies underwent rapid changes and under this framework flood hazard
mapping has, particularly, been realized as one of the foremost tasks to be accomplished in
support of disaster management activities. As a pilot study, the lower reach of Kelani River
which covers the suburbs of the island’s capital, Colombo, was chosen for flood hazard mapping
as it is frequently affected.
The one dimensional steady flow component of HEC-RAS Software was utilized for the study.
The requisite geo-spatial modeling of the flood plain was carried out by implementing HEC-
GeoRAS utility in ArcView environment and using a TIN model. The land-use map of the study
area was updated using the data remotely sensed by ALOS for attributing the study area with
necessary hydraulic parameters. Peak flows associated with considerate return periods were
obtained from a frequency analysis and the calibration run was executed simulating the 50 year
return period flood event.
Model results and field records corresponding to the above event of 1989 was compared and it
was found that results are very satisfactory. Accordingly, the first ever set of flood hazard maps
associated with 10 year, 20 year, and 50 year return period events to be adopted by the
government of Sri Lanka were prepared successfully. Statistics show that the study area would
be subjected to flood encroachments of 60, 77, and 94 square kilometers for events with return
periods of 10 years, 20 years, and 50 years respectively. Modern day geospatial tools have also
facilitated to identify with ease, flood encroachments on different land use types and the number
of buildings they would affect.

1. INTRODUCTION

Sri Lanka is an island located in the Indian Ocean between the Bay of Begal in the east and the
Gulf of Mannar in the west. Its geographical setting and geomorphologic and climatological
conditions make it prone to several types of natural disasters such as floods, landslides, cyclones,
droughts, wind storms, coastal erosion, and tsunami. Among these, floods have been identified
as the most frequent natural disaster that savages the economic and social activities of the
affected areas and that take away lives. Flood occurrences in Sri Lanka are mainly due to
excessive of rainfall during the monsoon periods. At instances, depressions over the Bay of
Bengal too bring in heavy rains causing floods. Moreover, Sri Lanka has, in the recent past,
experienced several flash flood occurrences which failed to provide ample time for evacuation.
These events meet the expectation that tropical countries would receive less annual rainfall, but
increased rainfall intensities due to global warming.
While this has been the case, until the tsunami of December 2004, disaster management
strategies were at a minimal level (Seneviratne et al, 2005). However, after the tsunami, they
underwent rapid changes and with the formation of the Ministry of Disaster of Management in
2006, flood hazard mapping has been identified as one of the foremost tasks to be accomplished
for sustainable development and was delegated to the Irrigation Department. The existed
practice of carrying out this task was through collection of field data pertaining to previous flood
events which consumes unmanageable labor hours and imposes practical difficulties in
extracting the flood depths which are very important. This paper presents the utilization of a
flood simulation model which could overcome the said limitations through integration of remote
sensing, GIS, and GPS technologies for flood hazard mapping for 10 yr, 20yr, and 50 yr return
periods events in the lower reach of Kelani River prioritized based on economic activities,
administrative functions, population density, etc.

2. STUDY AREA

Figure 1: Location Map of the Study Area

Kelani river is the second longest of Sri Lanka which originates in the central hills and flows
mainly to the West till it meets the sea at the northern boundary of the city of Colombo. The
river basin is entirely located in the wet zone with an extent of 2230 sq. km with average annual
rainfall of about 2400 mm.
With the knowledge of historical flood events occurred in the basin, lower reach of Kelani River
was selected for this study. It lies between northern latitudes of 6°47' to 7°05' and longitudes of
79°52' to 80°13' and occupies an area of 1200 square kilometers It covers the flood plains from
Glencourse gauging station up to Nagalagama street gauging station in Colombo. Total length of
the river in this section is about 60 km. Average river bed level of the reach considered is less
than 1:5000.
Three administrative districts were included in the study area namely, Colombo, Gampaha and
Kegalle. The first two districts are attributed with the highest and second high population
densities of the country.
3. METHODOLOGY

The ‘Steady Flow Water Surface Profiles’ component of the Hydrologic Engineering Center’s
River Analysis System (HEC-RAS) which is designed for applying in flood plain management
and flood insurance studies was utilized in the study by modeling the selected stretch of the
Kelani river as a single river reach. This component of the modeling system is intended for
calculating water surface profiles for steady gradually varied flow, which is capable of modeling
subcritical, supercritical, and mixed flow regimes water surface profiles. The basic
computational procedure is based on the solution of the one-dimensional energy equation.
Energy losses are evaluated by friction (Manning’s equation) and contraction/expansion
(coefficient multiplied by the change in velocity head).
A detailed discussion of HEC-RAS including its theoretical basis could be found at
http://www.hec.usace.army.mil/software/hec-ras

3.1 Model Inputs

Implementation of HEC-RAS requires inputs which come from three basic categories of data;
a. Geometric data
b. Basin Characteristics
c. Flow Data

3.1.1 Geometric Data

The requisite geometric data includes stream centerlines and cross section cut lines and were
prepared using the HEC-GeoRAS graphical user interface. It is a set of procedures, tools, and
utilities for processing geospatial data in ArcView. It also allows the import of the prepared data
into HEC-RAS model. The creation of the import file requires a digital terrain model (DTM) of
the river system in the ArcInfo Triangulated Irregular Network (TIN) format.
A TIN model of the river system was, therefore, created from topographic maps of scale
1:10,000 and spot levels as of year 2000 collected from the Survey Department of Sri Lanka.
The contours were of 10 m interval.
Having imported the data into the HEC-RAS system, the cross section data were adjusted to suit
the surveyed cross sections (as of year 2007) which were obtained from the Irrigation
Department.

3.1.2 Basin Characteristics

Manning’s friction coefficient ‘n’ and, Contraction and expansion coefficients fall under this
category.
a. Manning’s n value is highly variable and depends on a number of factors including surface
roughness, vegetation, channel irregularities, channel alignment, scour and deposition,
obstructions, size and shape of the channel, stage and discharge, seasonal changes,
suspended material and bedload. Reference was made to Chow (1959) together with the
land-use information to attribute the selected part of the basin with initial values of
Manning’s ‘n’.
Land-use map of year 2000 was collected from the Survey Department and was updated
using remotely sensed data of ALOS/AVNIR-2 acquired on February 10, 2007.
b. The contraction and expansion coefficients are used to model the energy loss within a reach
due to the contraction or expansion of flow due to the change in cross section. Typical
values of 0.1 and 0.3 pertaining to locations where the change in river cross sections is
small and the flow is subcritical were adopted based on the variation of selected river reach.
3.1.3 Flow Data

Discharge and water level values make the upstream and downstream boundary conditions. In
the intention of preparing flood hazard maps corresponding to 10 yr, 20 yr, and 50 yr return
period events, frequency analyses were carried out to determine peak discharges employing the
Gumble distribution which has been adopted in Sri Lanka and using annual peak discharge data
recorded at Glencourse gauging station for the last 35 years obtained from the Irrigation
Department.
Water levels corresponding to the above discharges were estimated at the downstream station
using the rating curve collected from the Irrigation Department.

3.2 Flood Simulation

Having completed the set up of the system with the requisite model parameters and variables, a
calibration run was performed using the peak discharge value corresponding to the 50 yr return
period flood event. Results of this simulation were then checked against the 1989 flood event
corresponding to the above event as well as the data collected from a field visit. Comparison
between the observations and model predictions were found to very satisfactory except at some
locations. The initially input Manning’s ‘n’ values at those were then varied to give the best
predictions. Accordingly, simulations of 10 yr and 20 yr return period flood events were
performed. HEC-RAS model simulation results were exported to HEC-GeoRAS for necessary
processing and visualization.

3.3 Flood Hazard Mapping

HEC-GeoRAS utility was used extract flood depths by comparing the TIN models of the water
surface and terrain and thereby to produce the intended flood maps. On the produced maps,
features such as roads, buildings, cultivation, etc. were overlaid to have an insight on the
damage they could cause to different features.

4. RESULTS AND DISCUSSION

Figures 2 and 3 show the produced flood maps corresponding to 50 year and 10 year return
period flood events. As could be seen from the maps flood depths were classed into 4 groups as
up to 1 m, 1 – 3 m, 3 – 5 m, and above 5 m. However, these grouping could easily be modified
to suit different stakeholder requirements and work out hazard levels.
Based on the above classification, flood inundation areas corresponding to the different return
periods were determined and given in Table 1. Results of flood simulations, as could be seen in
Table 1, also indicate that a total of 63, 77, and 94 square kilometers of the study area would be
subjected to flood encroachments due to 10 yr, 20 yr, and 50 return period flood events
respectively.

Return Period Flood depth (m) Total

( yrs ) <1 1-3 3-5 >5
10 19 29 10 5 63
20 20 40 11 6 77
50 24 41 19 10 94
Table 1: Flood Inundation Areas in sq. km
Figure 2: 50 yr Flood Map

Figure 3: 10 yr Flood Map

 Table 2 gives the statistics of the number of buildings that would be affected by flood events in
the districts of Colombo, Gampaha, and Kegalle. This shows the natural tendency of
anthropogenic activities to be put in place at locations of economic centers such as the capital
city of Colombo irrespective of the fact that they are flood prone. It should be acknowledged at
this point that of the factors that are conducive to the creation of flood problems, one of the
principal causes is encroachment by humans on the flood plains of river basins.

Return Period District

Total
( yrs ) Colombo Gampaha Kegalle
10 5421 3589 0 9010
20 8487 5574 3 14064
50 10378 6620 7 17005
Table 2: No. of Building to be Affected; District Wise

5. CONCLUSION

HEC-RAS together with HEC-GeoRAS utility was successfully utilized to prepare the first ever
set of Flood Hazard Maps to be adopted by the Government of Sri Lanka in support disaster
management and sustainable development. The lower reach of Kelani river was chosen for the
study on a priority basis.
Having determined the peak flow discharges corresponding to 10 yr, 20 yr and 50 yr return
period flood events trough a frequency analysis, simulations of the same events were carried out
in the ArcView environment. The produced flood hazard maps are in good agreement with
historical records pertaining to events of similar magnitude. Statistics indicate that the study area
would be subjected to flood encroachments of 60, 77, and 94 square kilometers for events with
10 yr, 20 yr, and 50 yr return period events respectively. Modern day geospatial tools have also
facilitated to identify, with ease, not only the flood encroachments on different land use types
and the number of buildings they would affect but also the appropriate mitigation measures.

ACKNOWLEDGEMENT

The authors wish to thank the Japan Aerospace Exploration Agency (JAXA) for providing
financial support and ALOS data through Mini-Projects, and the Irrigation and Survey
Departments of Sri Lanka for providing with the data used in this study.

REFERENCES

Chow, V. T., (1959), Open Channel Hydraulics. McGraw-Hill Book Co., New York.

Seneviratne, H.N., Ratnaweera, H.G.P.A., Bandara R.M.S.,. (12-13 December 2005).

Geotechnical Aspects of Natural Hazards: Sri Lankan Experience. Proc., First Intl. Conf. on
Geotechnical Engineering for Disaster Mitigation and Rehabilitation, (pp. 185-199). Singapore.

http://www.hec.usace.army.mil/software/hec-ras

View publication stats