Sensitivity of various topographic data in flood management: Implications on inundation mapping over large data-scarce regions

Topographic data in the form of digital elevation models (DEMs) play a significant role in flood management. Despite the increasing availability of DEMs for large regions, there is a need to evaluate their performance at the inundation/flood level, while considering the overall complexity of flood models. The present study identifies, for the first time, the uncertainties generated in both river channel and overland flooding while considering a set of nine variants from various sources (LiDAR, Cartosat, SRTM, and ASTER) and grid resolutions (resampled versions) in the presence of discharge, rainfall, and tide boundary conditions for a severely flood-prone catchment in the Mahanadi River Basin, India. Extensive geostatistical analyses reveal the existence of significant biases with global DEMs i.e., SRTM and ASTER, whereas interestingly the LiDAR and Carto DEMs exhibit a high degree of isotropy. The global DEMs fail to capture several inundated spots; thus plummeting the flood inundation extents to a sufficient degree of unacceptability. Prominently, the inability in identifying high and very high flood depths (> 1.5 m) over the coastal stretches results in large uncertainties in the majority of the grids. Our analysis reveals the existence of significant noise in global DEMs, which nullifies the hydrodynamic interaction during the coupling of 1-D and 2-D flood models in presence of tidal influence. We recommend that under unavailability of precise LiDAR DEMs, resampled and freely available Carto DEMs, that are as efficient as LiDAR if not more, be given higher preference. We caution against the copious usage of global DEMs for large data-scarce and flood-prone regions, as the DEM uncertainty may be substantially amplified at the inundation level during combined channel and overland flood simulations. Through this study, we would like to recommend the proposed framework as a guided step while selecting appropriate DEM for flood inundation mapping over large data scarce regions.