Evaluation of bathymetry retrieval using China’s new-generation ocean color satellite HY-1E

Abstract

Satellite-derived bathymetry (SDB) has been widely applied in shallow-water mapping using multispectral sensors such as Sentinel-2, Landsat-8/9, and HY-1D. However, limitations in revisit frequency, spatial resolution, and spectral coverage constrain their performance in dynamic and optically complex coastal environments. HY-1E, the new-generation ocean color satellite following HY-1D, offers improved spatial resolution, an expanded spectral range, and a reduced revisit time. These improvements are expected to advance remote sensing applications in SDB. Prior to its operational application, a systematic assessment of HY-1E’s bathymetric retrieval capability is essential. In this study, we evaluate the performance of HY-1E imagery for SDB using an ICESat-2-assisted approach across multiple study areas, acquisition times, and image types. We also discuss the influence of key spatial, temporal, and spectral parameters on bathymetry retrieval. Across 25 study areas, HY-1E consistently achieved an RMSE below 1 m and a coefficient of determination (R²) exceeding 0.86. On average, HY-1E outperformed HY-1D by 0.13 m, Sentinel-2 10 m by 0.09 m, and Sentinel-2 20 m by 0.01 m in terms of RMSE. In waters shallower than 10 m, most HY-1E-derived bathymetric models met the ZOC C standard. In terms of terrain characterization, HY-1E demonstrated significant improvements in detail resolution over HY-1D and produced DBMs with lower noise levels compared to Sentinel-2. In addition, this study proposes a data source selection strategy tailored to different types of coastal and reef environments, considering factors such as spectral bands, spatial resolution, revisit cycle, and swath coverage, in order to optimize the balance between inversion efficiency and accuracy.