Remote Sensing and Machine Learning-Based Evaluation of Sediment Plume Impacts on Coastal Water Quality in Eastern England, United Kingdom

Sediment plumes and associated levels of turbidity have immense significance for coastal water quality, health of ecosystems and biogeochemical processes in estuarine and near-shore systems. Remote sensing combined with geographic information systems and machine learning stands as an effective framework for monitoring such parameters on regional to national scales and over long-time spans, particularly applicable where coastal areas are sparsely covered by data. This study presents an RS-GIS-ML amalgamation approach to model sediment plume dynamics, turbidity, and surface water quality in coastal waters of Great Yarmouth, UK, using multi-temporal satellite data between 2020 and 2025. The objective of the study is to design remote sensing-based machine learning models for predicting water quality and sediment plume characteristics, and assess the accuracy and reliability of RS-derived water quality products with statistical performance assessment. Preprocess multi-source satellite data by extracting spectral bands and water quality sensitive indices. Machine learning models were applied to estimate turbidity, TDS, TOC, Chl-a, SSC, WQI, and sediment plume index. GIS techniques were involved in spatial analysis and plume delineation, and spatiotemporal trend assessment. Results of the model performance evaluation indicated an excellent prediction capability with R 2 ranging from 0.81 to 0.89. The RMSE values were 6.82 NTU for turbidity, 78.45 mg/L for TDS, 0.62 mg/L for TOC, 1.15 μg/L for chlorophyll-a, 14.9 mg/L for SSC, 7.25 for WQI, and 0.18 for the sediment plume index based on 36–44 validation points. Overall, the findings offer validation for the robustness of the proposed framework for long-term coastal water quality monitoring.