EXPLAINABLE MACHINE LEARNING FOR WATER QUALITY ASSESSMENT AND POLLUTION RISK PREDICTION: AN INTEGRATED FRAMEWORK FOR SUSTAINABLE WATER RESOURCE MANAGEMENT
Keywords:
Water Quality Index, Explainable Machine Learning, Pollution Risk Prediction, Gradient Boosting, Sustainable Water Resource ManagementAbstract
Water-quality degradation threatens aquatic ecosystems, public health, and sustainable resource management, creating a need for accurate and interpretable predictive approaches. This study developed an explainable machine-learning framework for water-quality assessment and pollution-risk prediction using physicochemical, hydroclimatic, spatial, and seasonal variables. A total of 1,495 observations from 58 sampling locations across 13 rivers and streams were analyzed. Water Quality Index (WQI) was modeled as a continuous outcome using Linear Regression, Random Forest, Extra Trees, Gradient Boosting, and Support Vector Regression, while pollution status was classified into Good, Poor, and Very Poor categories using Logistic Regression, Random Forest, Extra Trees, and Support Vector Machine. Five-fold location-grouped cross-validation was applied to reduce spatial leakage. The overall mean WQI was 91.40, and 94.25% of observations were classified as Very Poor. Gradient Boosting achieved the strongest WQI prediction, with a mean R² of 0.9993, MAE of 0.1215, and RMSE of 0.3485. Random Forest provided the best pollution classification, attaining 0.9940 accuracy, 0.9061 balanced accuracy, 0.9394 macro-F1, and 0.9414 MCC. Permutation importance identified phosphate as the dominant predictor of WQI, followed by pH and dissolved oxygen. The framework demonstrates that explainable machine learning can provide accurate, transparent, and management-relevant support for sustainable water-resource monitoring.
References
Abba, S. I., Yassin, M. A., Mubarak, A. S., Shah, S. M. H., Usman, J., Oudah, A. Y., Naganna, S. R., & Aljundi, I. H. (2023). Drinking water resources suitability assessment based on pollution index of groundwater using improved explainable artificial intelligence. Sustainability, 15(21), 15655. https://www.mdpi.com/2071-1050/15/21/15655
Aderemi, I. A., Kehinde, T. O., Okwor, U. D., Ahmad, K. H., Adjei, K. Y., & Ekechi, C. C. (2025a). Explainable AI for Water Quality Monitoring: A Systematic Review of Transparency Interpretability and Trust. IEEE Sensors Reviews, 2, 419–443. https://www.researchgate.net/profile/Khalid-Ahmad-15/publication/397269056_Explainable_AI_for_Water_Quality_Monitoring_A_Systematic_Review_of_Transparency_Interpretability_and_Trust/links/690a5c4d9708d52f2da4e0cc/Explainable-AI-for-Water-Quality-Monitoring-A-Systematic-Review-of-Transparency-Interpretability-and-Trust.pdf
Aderemi, I. A., Kehinde, T. O., Ugochukwu, D. O., Ahmad, K. H., Adjei, K. Y., & Chijioke, C. E. (2025b). Beyond the black box: a systematic review of explainable AI for transparent and trustworthy water quality monitoring. IEEE Sensors Reviews. https://ieeexplore.ieee.org/abstract/document/11112533/
Aldrees, A., Jibrin, A. M., Dan’azumi, S., Mahmoud, I. A., Aliyu, U. U., & Abba, S. I. (2025). Explainable machine learning framework for assessing groundwater quality and trace element contamination in Eastern Saudi Arabia. Scientific Reports, 15(1), 45333. https://www.nature.com/articles/s41598-025-29598-8
Alshehri, F., & Rahman, A. (2023). Coupling machine and deep learning with explainable artificial intelligence for improving prediction of groundwater quality and decision-making in arid region, Saudi Arabia. Water, 15(12), 2298. https://www.mdpi.com/2073-4441/15/12/2298
Choudhary, R., Kumar, A., C, P., Naik, M. M., Choudhury, M., & Khan, N. A. (2025). Predicting water quality index using stacked ensemble regression and SHAP based explainable artificial intelligence. Scientific Reports, 15(1), 31139. https://www.nature.com/articles/s41598-025-09463-4
Chowdhury, M. A. H., & Debnath, P. (2025). Predictive Multiple Machine Learning Models of Riverine Water Pollution Index: A Comprehensive Data-Driven Study.. https://www.researchgate.net/profile/Md-Anamul-Chowdhury-3/publication/396063024_Predictive_Multiple_Machine_Learning_Models_of_Riverine_Water_Pollution_Index_A_Comprehensive_Data-Driven_Study/links/68dd69a902d6215259b7fe22/Predictive-Multiple-Machine-Learning-Models-of-Riverine-Water-Pollution-Index-A-Comprehensive-Data-Driven-Study.pdf
Doan, C. T., & Du Nguyen, H. (2025). Robust water quality prediction across multiple indicator formulations using an explainable ensemble learning model. Water Resources and Industry, 100329. https://www.sciencedirect.com/science/article/pii/S2212371725000538
Han, Z., Zhang, S., & He, L. (2025). Predicting and investigating water quality index by robust machine learning methods. Journal of Environmental Management, 381, 125156. https://www.sciencedirect.com/science/article/pii/S0301479725011326
Hellen, N. (2022). Explainable artificial intelligence and model calibration for water quality prediction.. https://dspace.bracu.ac.bd/items/22b3adac-3142-4bc0-94f7-f098df1e60a4
Hellen, N., Sabuj, H. H., & Ashraful Alam, M. (2023). Explainable AI and Ensemble Learning for Water Quality Prediction. Proceedings of International Conference on Information and Communication Technology for Development, 235-250. https://doi.org/10.1007/978-981-19-7528-8_19
Islam, M. J., Salekin, S. U., Abdullah, M. S., Zaman, N., & Khan, A. A. A. (2024). Evaluation of water quality assessment through machine learning: A water quality index-based approach.. https://assets-eu.researchsquare.com/files/rs-4370890/v1/6047f324-ea97-4edb-902f-64fb858786f1.pdf
Islam, M. J., Salekin, S. U., Anzum, A., Zaman, N., Khan, A. A. A., Sarkar, D., Rabbani, M. L., & Hossain, M. T. (2024). Machine learning-driven water quality index prediction: Enhancing accuracy with gradient boosting and explainable AI for sustainable water monitoring. Applied Agriculture Sciences, 2(1), 1–14. https://publishing.emanresearch.org/Journal/abstract/agriculture-2110031
Mallick, J., Alqadhi, S., Alsubih, M., Ahmed, M. F. M., & Abdo, H. G. (2025). Automated interpretation of deep learning-based water quality assessment system for enhanced environmental management decisions. Applied Water Science, 15(5), 110. https://doi.org/10.1007/s13201-025-02452-y
Mallick, J., Alqadhi, S., Hang, H. T., & Alsubih, M. (2024). Interpreting optimised data-driven solution with explainable artificial intelligence (XAI) for water quality assessment for better decision-making in pollution management. Environmental Science and Pollution Research, 31(30), 42948-42969. https://doi.org/10.1007/s11356-024-33921-7
Maurya, R., Saha, S., Misra, B., Chatterjee, D., Chakraborty, S., & Sarkar, P. (2025). Employing Explainable Artificial Intelligence (XAI) for Enhanced Water Quality Assessment. 2025 International Conference on Computing, Intelligence, and Application (CIACON), 1–5. https://ieeexplore.ieee.org/abstract/document/11189765/
Nallakaruppan, M. K., Gangadevi, E., Shri, M. L., Balusamy, B., Bhattacharya, S., & Selvarajan, S. (2024). Reliable water quality prediction and parametric analysis using explainable AI models. Scientific Reports, 14(1), 7520. https://www.nature.com/articles/s41598-024-56775-y
Politikos, D., Stefanidis, K., Varlas, G., Papadopoulos, A., & Dimitriou, E. (2024). Predicting the Ecological Quality of Rivers: A Machine Learning Approach and a What-if Scenarios Tool. Environmental Modeling & Assessment, 29(6), 1059-1077. https://doi.org/10.1007/s10666-024-09980-y
Shaheed, H., Zawawi, M. H., & Hayder, G. (2025). The development of a river quality prediction model that is based on the water quality index via machine learning: A review. Processes, 13(3), 810. https://www.mdpi.com/2227-9717/13/3/810
Sinha, P., Debnath, N., & Bhattacharjee, P. (2026). Water quality and hydroclimatic dataset for rivers and streams of Dhaka, Bangladesh (Version 1) [Data set]. Mendeley Data. https://doi.org/10.17632/xftpv7jvx3.1
Talukdar, S., Ahmed, I. A., Shahfahad, Baig, M. R. I., Rihan, M., & Rahman, A. (2025). Water Pollution Assessment and Management Through Interpreting Black Box Deep Learning Algorithms in the Loktak Lake. Advanced GIScience in Hydro-Geological Hazards, 359-375. https://doi.org/10.1007/978-3-031-76189-8_17
Talukdar, S., Shahfahad, Ahmed, I. A., Rihan, M., & Rahman, A. (2024). Integrating Explainable Artificial Intelligence and Machine Learning for Understanding Water Pollution and Its Management. Environmental Risk and Resilience in the Changing World, 115-128. https://doi.org/10.1007/978-3-031-62442-1_7
Tian, P., Xu, Z., Fan, W., Lai, H., Liu, Y., Yang, P., & Yang, Z. (2024). Exploring the effects of climate change and urban policies on lake water quality using remote sensing and explainable artificial intelligence. Journal of Cleaner Production, 475, 143649. https://www.sciencedirect.com/science/article/pii/S0959652624030981
Xue, Y., Niu, Z., Zhang, P., Zhang, R., Jia, L., & Duan, K. (2025). Optimizing water quality monitoring in an arid wetland using interpretable machine learning: From identifying dominant parameters to simplifying the water quality index. Ecological Indicators, 181, 114397. https://www.sciencedirect.com/science/article/pii/S1470160X25013299
Zalte, J., Rai, A., Shah, H., & Fulekar, M. H. (2025). Adversarial susceptibility analysis for water quality prediction models. Scientific Reports, 15(1), 37828. https://www.nature.com/articles/s41598-025-21797-7
Zhang, T., Wu, J., Chu, H., Liu, J., & Wang, G. (2025). Interpretable machine learning based quantification of the impact of water quality indicators on groundwater under multiple pollution sources.
