Determining Atikhisar Reservoir’s Bathymetry from Landsat-5 TM Satellite Images using the Stumpf Algorithm

Abstract

Determining the bathymetry of shallow waters is important for managing coastal areas, river basins, and water resources. However, economic and practical difficulties in collecting bathymetric data cause disruptions in bathymetric studies. To overcome these challenges, a recent focus has involved the use of remote sensing technology as an alternative approach to the bathymetric mapping of shallow waters. This study uses Landsat-5 TM satellite imagery, which is free and open data, to determine the digital bathymetric model (DBM) of Atikhisar Reservoir. The study also uses the Normalized Difference Water Index (NDWI) and Modified Normalized Difference Water Index (MNDWI) to determine the reservoir's surface area and the Stumpf algorithm to perform the bathymetric mapping. Satellite image-based DBMs were obtained using the linear regression equations created from the blue/green log-ratio values from the Landsat-5 TM satellite image and the values obtained from a 1/5000 scale digital bathymetric map for five different training reference point sets. The root mean square error (RMSE) values were calculated by comparing the DBMs with the test data. The model with the best results showed the regression determination coefficient (R2) to be 0.701 and the RMSE to be 2.1 m. These results reveal the potential of low-cost bathymetric map production for preliminary investigation and general evaluation in reservoirs with easy data processing from Landsat images.