Prediction of water quality in Jordanian dams using data mining algorithms

The evaluation of water quality constitutes a critical aspect of water management strategies, particularly in arid and semi-arid environments, where the use and protection of sustainable resources are crucial. This study focuses on assessing and predicting water quality in three Jordanian dams using advanced data mining techniques. Physical, chemical, and biological water quality parameters were collected and analyzed over a four-year period. The Weighted Arithmetic Water Quality Index (WA-WQI) was used to evaluate the overall water quality. Various data mining algorithms, including generalized linear models, decision trees, random forests, gradient-boosted trees, and support vector machine (SVM), were employed to predict WQI and understand the seasonal and annual variations. Key findings highlight significant fluctuations in water quality, influenced by parameters such as pH, conductivity, nutrients, and microbial contamination. The study emphasizes the importance of continuous monitoring and predictive modeling for effective water resource management. It also demonstrates the effectiveness of using SVM for water quality prediction in arid regions. The models were evaluated using different performance metrics. The SVM outperformed other employed models. This study provides a critical benchmark and a robust predictive framework for water resource management in Jordan and semi-arid areas, addressing a significant gap in regional environmental monitoring.