Performance and Uncertainty of Bathymetry Derived From Time Series Composite Satellite Imagery

This study aims to evaluate the performance of cloud-based Sentinel-2 time-series composite images for mapping shallow water bathymetry and to assess the accuracy and uncertainty of maps produced from both single-acquisition images and time-series composites, following International Hydrographic Organization (IHO) standards. The research was carried out in the shallow waters around Ontoloe Island, East Nusa Tenggara, Indonesia. A dataset of 409,976 echosounder points and 73 Sentinel-2 images, taken between January 1 and December 31, 2020, was used to test and validate Satellite Derived Bathymetry (SDB) models under different optical and cloud cover conditions. The results show that the water's optical properties vary greatly with seasonal changes. Although there are areas for improvement, this study shows that time-series composite imagery is a useful tool for shallow water bathymetry mapping in regions with limited data. © 2024 IEEE.