Evaluation of Sediment Pollution in an Abandoned Alkaline Slurry Pond and Its Stratification Using the Machine Learning Techniques

The aim of this study is to characterize the contamination of sediments accumulated in an abandoned sludge storage facility, to develop, and test an automated method for precipitation stratification. The assessment of pol-lution is based on the sludge chemical composition using integral heavy metal indices: the total pollution index (Zc), and the potential environmental risk index (RI). According to the calculations, high level of pond sediments pollution was revealed. An automated approach to sediments stratification has been implemented using machine learning algorithms. This paper describes the Python-based algorithm and presents the results of its testing on sediment samples from a man-made reservoir. During the clustering process, five distinct layers were identified, which differ significantly in their chemical composition. The constructed cross-sections display the heterogeneous structure of the sediment column, and are in good agreement with the sections based on values of integral indices obtained manually. The best results were achieved using the cubic interpolation method.