Surface water mapping and volume estimation of Lake Victoria using Machine Learning Algorithms

Freshwater mapping is a crucial element for water resource planning and conservation. Recently, the estimation of surface area and its temporal changes have been made easier due to the availability of remote sensing data. However, the quantification of water body volume is limited because the existing remote sensing technologies cannot estimate bathymetry data. In this study, Lake Victoria’s surface water extent and volume are estimated by combining the remote sensing and bathymetry data. The surface water extent is determined by feature extraction and classification using Machine Learning (ML). Gaussian Naïve Bayes (GNB), Decision Tree (DT), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Categorical Boosting (CatBoost) are the ML algorithms considered. Landsat ETM+images have been used for experimentation. Experimental results concluded that LightGBM and DT are the best and least performing ML algorithms for determining surface extent and volume.