Satellite derived bathymetry based on ICESat-2 diffuse attenuation signal without prior information

Abstract

Bathymetric information is critical for environmental monitoring and maintenance of marine safety. The signal received by passive optical remote sensors contains the reflectance information from the seabed, which then forms the informational basis of bathymetry; however, passive optical bathymetry models often require partial prior information. Although spaceborne active lidar has been shown to accurately detect bathymetry within narrow range of tracks, for some small islands or coastal sites, these tracks may not accurately traverse the shallow ocean area. To this end, a satellite derived bathymetry method based on ICESat-2 diffuse attenuation signal without prior information was proposed here. The diffuse attenuation coefficients of the water column were estimated according to the ICESat-2 lidar signal of deep areas near the target shallow water area, and the parameters for the bathymetry algorithm without prior depth information were obtained. Combined with the Sentinel-2 multispectral imagery, accurate large-scale bathymetric information of Zhongye Island and Reefs, Guam, and Ngau were obtained. Bathymetric data from ICESat-2 of multiple tracks were used for verification, and the results showed that the new model performed well, as mean absolute error (MAE) for the three study areas ranged 0.35–1.52 m, and root Mean square error (RMSE) fell between 0.74 and 1.89 m. In the absence of any bathymetric information, the new method can achieve higher or similar accuracy to the more commonly employed Stumpf model.