Effect of atmospheric corrections on shallow sea bathymetric mapping using gaofen-2 imagery: a case study in Lingyang Reef, South China Sea

Abstract

Satellite-derived bathymetry (SDB) with high spatial resolution effectively maps detailed information about shallow sea depths. Properly selecting atmospheric correction (AC) methods is crucial to obtaining precise bathymetric information from satellite data. In this work, three different AC methods (FLAASH, 6S, and DOS) were applied to GaoFen-2 imagery in Lingyang Reef, South China Sea. Water depth retrieval models, such as single-band, multiband, and band ratio models, were established by 470 points of in-situ water depth data and used for evaluating the model performances. Additionally, the optimal model was applied to depth inversion. The results show that the multiband model based on four bands performs well in this study area with R2 =0.736. The choice of AC methods significantly affects SDB, although acceptable results can be derived without these methods. Specifically, the DOS method has the highest inversion accuracy, with a mean relative error and root mean square error of 14.05% and 3.31 m, respectively. Furthermore, the inversion accuracy in various depth ranges may be primarily influenced by suspended sediment concentration and bottom-type uniformity. This study provides a valuable reference for selecting AC approaches and inversion models in high-spatial-resolution SDB in shallow seas.