Controversial insights into irrigation water quality in arid and semi-arid regions using AI driven predictions: Case of southern Gabès

Effective groundwater management is critical in arid and semi-arid regions, where water resources are essential for agriculture. This study assesses the Irrigation Water Quality Index (IWQI) of the Southern Gabès aquifer in Tunisia using a combination of traditional hydrochemical analysis and machine learning models—specifically, Classification and Regression Tree (CART) and Support Vector Machine (SVM). A total of 83 groundwater samples were analyzed based on five key parameters: Electrical Conductivity (EC), Sodium Adsorption Ratio (SAR), Chloride (Cl-), Sodium (Na+), and Bicarbonate (HCO3-). The results show that the CART model demonstrated superior performance with an R² value of 0.99 and a Root Mean Square Error (RMSE) of 0.43, while the SVM model achieved an R² of 0.87. These findings underscore CART's robustness in predicting IWQI, offering high precision even with limited datasets.

The groundwater quality was categorized, revealing that 62% of samples were classified as "satisfactory" for irrigation, while 31% were deemed "unsuitable" without treatment, highlighting areas of concern for agricultural use. The study also emphasizes the importance of continuous monitoring and adaptive management strategies to ensure sustainable water use in the region.

Overall, this research demonstrates the effectiveness of machine learning models, particularly CART, in accurately assessing groundwater quality. These insights provide valuable tools for resource managers to make informed decisions, ensuring the sustainable exploitation of groundwater in arid and semi-arid regions. The findings pave the way for future research and policy development in water resource management.