Retrieval of High-Resolution Nearshore Bathymetry from Sentinel-2 Twin Multispectral Imagers using a Multi-Scene Approach

Determining nearshore bathymetry by traditional surveying methods is a challenging task as it involves huge costs and efforts. Most of the coastal shallowwater zones worldwide either remain unmapped or not updated. Bathymetry estimations from optical satellite imageries have been increasingly implemented as an alternative tool for traditional bathymetry surveys. In this study, we examine the usefulness of freely available, five-day revisit and relatively highresolution Multi Spectral Instruments (MSI) on-board Sentinel-2A and 2B twin satellites. A process workflow has been developed which automatically incorporates a robust atmospheric correction through ACOLITE software and multi-scene compositing of several scenes to improve the reliability and no data gaps. Two study sites in India are explored owing to their variability in submarine morphology. High-resolution bathymetry maps are generated through a log-ratio transform model calibrated with minimal in situ data from the jet ski soundings. The satellite-derived bathymetry obtained has an overall bias of –0.01 and 0.02 m, and ischolar_main mean square error of 1.09 and 0.93 m respectively, at two study sites up to 15 m depth. The consistency in bathymetry retrieval indicates a potential for automated application for the benefit of operational and scientific studies. These high-resolution maps capture small-scale nearshore features like sandbars and rip channels, which are of prime importance for coastal and beach managers.