Evaluation of topographic data sources for 2D flood modelling: case study of Kelani basin, Sri Lanka

Abstract

Frequent flooding in Sri Lanka underscores the necessity of flood modelling as inundation extent and flood depth can easily be identified for implementing flood control measures. Accuracy of flood modelling is primarily influenced by topographic data sources and their resolution. Due to the lack of Light Detection and Ranging (LiDAR) data source in most regions of Sri Lanka, alternative topographic data sources need to be assessed. This paper investigates the accuracy of 2D flood model results in terms of flood depth and inundation extent developed based on open source topographic data sources, namely Shuttle Radar Topography Mission (SRTM) and Advanced Spaceborne Thermal Emission (ASTER) with 30 and 90 m resolutions. This study was carried out at the downstream of Kelani river basin as it is prone to frequent flooding. The 1 m resolution LiDAR data were used as the reference data to assess the accuracy of aforesaid data sources, and were resampled to 30 and 90 m to investigate the effect of resolution with SRTM and ASTER data sources. The results show that reduction in the resolution of LiDAR data source does not significantly affect the model accuracy as even 90 m resolution LiDAR data source produced higher accurate results (flood depth, root mean square error of 0.95 m; inundation extent, F-statistic of 70.21%) than the 30 m resolution SRTM and ASTER data sources.