Windowed anisotropic local inverse distance-weighted (WALID) interpolation method for riverbed mapping

Abstract Digital elevation models (DEMs) are pivotal in river engineering and management, making their accuracy paramount. This paper introduces a novel interpolation method that addresses the challenges of data point distributions, anisotropic river topography, and meandering rivers. The windowed anisotropic local inverse distance-weighted (WALID) method is tested and compared with conventional methods through a case study. Elevation data are generated from multi-beam echosounder data to mimic conventional cross-sectional measurements, while ground truth data provides qualitative comparisons. Root-mean-square error and volumetric differences are used as criteria for validation and comparison. Sensitivity analysis of the method’s free parameters demonstrated its robustness. Notably, the proposed algorithm exhibited superior volumetric accuracy compared to other methods when considering literature recommendations for optimal transect spacing.