Bathymetry and Benthic Habitat Mapping in Shallow Waters From Sentinel-2A Imagery: A Case Study in Xisha Islands, China

Abstract

Mapping of benthic habitats and bathymetry is crucial for sustainable development and assessment of climate change and human activities. In this study, Hyperspectral Optimization Process Exemplar (HOPE) was modified, renamed as M-HOPE, to simultaneously obtain bathymetry and benthic habitat in shallow waters in Xisha Island, China. A local lookup table (LUT) for benthic reflectance spectra was established. Validation using <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> measurements demonstrated good performance of M-HOPE with a R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.76 for bathymetry using the local LUT. Application of M-HOPE to Sentinel-2A imagery further proved good accuracy of M-HOPE derived bathymetry with a R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.86 against <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> observations and a R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.92 against ICESat-2 measurements. M-HOPE-derived benthic classification also agreed well with field observations with probability of detection (POD) > 0.6 and false alarm ratio (FAR) < 0.5. However, uncertainties exist, mainly caused by mixed benthic types, atmospheric correction, and similarity in spectral reflectance of different benthic types.