Analysis of The Utilization of Sentinel-2 Imagery for Satellite-Derived Bathymetry Using Lyzenga Algorithm (Case Study: Bali Province)

Abstract

Bathymetric surveys nowadays are often used by the echosounding method. This method has weakness because the range that can be surveyed is limited due to cannot reach areas that have shallow depths. With advances in technology, there are alternative ways that can be done to map the depth of the sea, with the help of satellite imagery or remote sensing. This method uses a mathematical algorithm based on a combination of spectrum channels called Satellite-Derived Bathymetry (SDB). In this research, Sentinel-2 satellite imagery was used with Lyzenga algorithm. Lyzenga (2006) uses multilinear regression which is a combination of an equation that aims to convert pixel values into in-situ depth values. Within research of six locations in Bali Province, which are Sangsit Harbor, Gunaksa Harbor, Amed Port, Keramas Beach, Serangan Harbor and Sawangan Beach produced an R2 value each of 0.451, 0.747, 0.495, 0.610, 0.451 and 0.676. While the RMSE values were 26,247, 12,377, 31,942, 3,531, 3,000, and 1,992 respectively for water depths of 146 m, 97 m, 183 m, 22 m, 13 m, and 18 m respectively. The parameters that affect the level of accuracy are: water depth, presence of breaking waves, presence of disturbance objects in the waters.