Comparative analysis of three key atmospheric correction techniques for bathymetric mapping in nearshore areas with Sentinel-2 Data - case study: Kish Island, Persian Gulf

Abstract

This study addresses the critical need for accurate bathymetric data in coastal and nearshore zones, which are essential for ecological balance and resource management. Traditional depth measurement methods are costly, labor-intensive, and spatially limited, further complicated by environmental factors. Remote sensing technologies, particularly the Sentinel-2 satellite imagery, offer a promising solution for efficient and extensive data acquisition. This research evaluates the impact of three atmospheric correction (AC) methods—FLAASH, Sen2Cor, and ACOLITE—on depth estimation accuracy using Sentinel-2 imagery over Kish Island, a biodiverse coral reef habitat in the Persian Gulf. Field measurements at 932 points around the island were used to train and test the performance of the AC methods. An integrated linear and ratio transformation model, utilizing green and blue bands of Sentinel-2, was applied to derive depth values. Statistical analyses, including the coefficient of determination (R²), root-mean-square error (RMSE), and mean absolute percentage error (MAPE), indicate that ACOLITE consistently outperforms the other methods, achieving R² values often exceeding 0.8, the lowest RMSE values of ~ 1.41 m, and a MAPE of ~ 41.56%. In contrast, Sen2Cor exhibits greater variability, with an R² of up to 0.78 and an RMSE of up to 1.75 m and MAPE of 47.10%, while FLAASH offers stable but less precise performance, with R² values ~ 0.74, RMSE ranging from 1.70 m to 1.91 m, and MAPE up to 50.26%. Thus, ACOLITE emerges as the most accurate and reliable method for atmospheric correction, enhancing the accuracy of bathymetric data and aiding the conservation and management of coastal environments.

Keywords

• Satellite derived bathymetry (SDB)
• Coastal mapping
• Remote sensing
• Depth estimation