Comparative assessment of SRTM and UAV-Derived DEM in flood modelling

Abstract

Flooding is one of the most devastating natural disasters occurring annually in many parts of the world. This study investigates the performance of DEM with different spatial resolution in flood modelling along the floodplain of Shiroro dam, Nigeria. Three different DEMs of 30 m, 15 m and 3 m spatial resolution covering the study area were investigated. The performance of the DEMs for flood modelling was evaluated with the aid of flood inundation modelling (3D analyst) tools in the ArcScene environment of ArcGIS 10.4 using results obtained from the shallow water St. Venant’s equation. The shallow water St. Venant’s equation was implemented in MATLAB software using the given river bathymetric information obtained from the dam authorities. The study determined the discharge quantity and flow velocity at some identified nodes along the river and its tributaries, thus, providing a basis for determining possible flood levels within the study area. Three flood levels were identified for each of the three used DEMs. For the original and re-sampled SRTM DEMs, the identified flood levels were 200 m, 250 m and 280 m, while for the UAV-derived DEM, 150 m, 205 m and 250 m were identified as the flood levels. Flood events covering the three identified flood levels were then simulated in ArcScene in order to estimate inundating settlements within the study area. The original 30 m and 15 m resampled resolution SRTM DEM had maximum and minimum height values of 612 m, 125 m and 585 m, 136 m respectively. In contrast, the UAV-derived DEM shows different characteristics, with maximum and minimum values of 497.22 m and 113.53 m, respectively. Further analysis showed that while the UAV derived DEM reliably predicted the flood risk situation due to its high resolution, the other two DEMs over predicted the flood risk situation of the area.