Bathymetry Extraction from SPOT 7 Satellite Imagery with Semiparametric Regression

Abstract: Bathymetry maps are essential for navigation and resource management, and as an intermediate product for mapping coastal marine habitats. Satellite-derived bathymetry (SDB) uses remote sensing technology to process reflection values from the bottom of the water to produce bathymetric information. This study aims to analyze the effect of bottom types of in-situ depth data on bathymetry estimation using semiparametric regression using spatial coordinates (Kanno-STR). We utilized two modified Kanno-STR methods, e.g., unconsidered and considered benthic habitats (coral reefs, seagrass, macroalgae, and substrates) on SPOT 7 data. The study area focused on the shallow water around Gili Trawangan, Gili Meno, and Gili Air Island, West Nusa Tenggara Province. For the first modification, the result shows that the coefficient of determination of the Kanno-STR method was 0.95, with an accuracy value of 78 % and a kappa coefficient value of 0.71. On the other hand, the second modification coefficient of determination on the Kanno-STR method was 0.84, with an accuracy value of 61% and a kappa coefficient value of 0.49. The interesting finding that SPOT 7 could penetrate up to 29.93 m. Separating benthic habitat cover into coral, seagrass, macroalgae, and substrate does not improve the accuracy of bathymetry results.

Keywords: : Bathymetry, Semiparametric, Kanno-STR, SPOT 7.