Improving the Accuracy of Low Resolution Digital Elevation Models (DEM) for Enhanced Performance in Hydrological Modelling

Basin hydrological characteristics including terrain, land use, soil interpretations, etc., are crucial in identifying catchment responses. A digital elevation model (DEM) with reasonable accuracy should be used in the initial stage of hydrologic applications. There are different sources and platforms which offer different resolution DEMs with different accuracy levels. Several parameters decide the DEM quality. Identification of each of these parameters with defined accuracy levels which represent their suitability for distinct studies is crucial for developing applications and mitigation measures. This study focuses on addressing the requirement of a tailormade methodology to assess DEM accuracy levels in tandem with associated applications to ensure their optimum performance. For this study, open source SRTM and ASTER DEMs, TOPO maps of $1:50,000,1:10,000$ resolutions, and LiDAR data were used focusing on the upstream $50\mathrm{~km}^{2}$ subcatchment in Kelani River Basin in Sri Lanka. Accuracy assessments were completed under two different approaches by comparison and correlation of spot elevation points and profiles examined for different geographic directions. As a conclusion for the profile-based observation, it was identified that the accuracy levels are high in the DEMs produced based on the high-resolution LiDAR and satellite images in the study area. An algorithm was developed to ensure improved hydrologic performance based on critical parameters according to the accuracy levels that exist in different DEMs and platforms. Further development of the methodology can ensure the use of even relatively low-resolution DEMs with appropriate corrections according to the relevant critical parameters or applications to achieve improved accuracy levels.