Integrating topographic continuity and lake recession dynamics for improved bathymetry mapping from DEMs

Abstract

Accurate lake bathymetry is critical for advancing hydrological and biogeochemical research, yet large-scale and deep-water mapping remains constrained by cost challenges. While remote sensing techniques have been extensively employed for bathymetric mapping, their effectiveness is primarily limited to shallow waters due to the rapid attenuation of optical signals with increasing depth. To overcome this limitation, we propose a novel bathymetric mapping method that leverages topographic continuity to infer underwater terrain by simulated progressive lake recession. This approach relies solely on Digital Elevation Model (DEM) data, using shoreline topographic gradients to estimate depth, providing a robust alternative for regions where conventional surveying is impractical. Validation across 12 lakes on the Tibetan Plateau demonstrated promising accuracy, with an average normalized root-mean-square error of 19.08 % for depth estimation and a mean absolute percentage error of 23.47 % for lake water storage. To evaluate the method's generalizability across diverse hydrological settings, it was applied to Lake Mead in the United States, producing a bathymetric map with a correlation coefficient of 0.66 with in situ measurements. Overall, this study introduces a low-cost solution for bathymetric mapping in data-scarce regions, offering a valuable tool for assessing lake water storage at regional and global scales.