This study focuses on assessing the hydrogeochemical characteristics and irrigation suitability of groundwater in the Ras El Aioun and Merouana districts, using an integrated approach that combines physicochemical analysis, machine learning (ML), and Geographic Information Systems (GISs). Thirty groundwater samples were collected in June 2023 and subjected to extensive analyses, including major ions (Ca²⁺, Mg²⁺, Na⁺, K⁺, HCO₃⁻, Cl⁻, SO₄²⁻), pH, TDS, alkalinity, and hardness. Hydrochemical facies analysis revealed that the Ca-HCO₃ type was dominant (93.33%), with some samples exceeding FAO limits, particularly for Na⁺, K⁺, SO₄²⁻, Cl⁻, Mg²⁺, and HCO₃⁻. Assessment of groundwater irrigation suitability revealed generally favorable conditions based on three key parameters: all samples (100%) were classified as excellent based on the Sodium Adsorption Ratio (SAR < 10), 70% showed good-to-permissible status by Sodium Percentage (Na% < 60), and 83.3% were within safe limits for Residual Sodium Carbonate (RSC < 1.25 meq/L). However, the Permeability Index (PI > 75%) categorized 96.7% of samples as unsuitable for long-term irrigation due to potential soil permeability reduction. Additionally, Total Hardness (TH < 75 mg/L) indicated predominantly soft water characteristics (90% of samples), particularly in the central study area, suggesting possible limitations for certain agricultural applications that require mineral-rich water. GIS-based spatial analysis showed that irrigation suitability was higher in the eastern and western regions than in the central zone. Advanced machine learning algorithms provide superior predictive capability for water quality parameters by effectively modeling complex, non-linear feature interactions that conventional statistical approaches frequently fail to capture. Three ML models—Support Vector Regression (SVR), Random Forest (RF), and Extreme Gradient Boosting (XGBoost)—were used to predict the Irrigation Water Quality Index (IWQI). XGBoost outperformed the others (RMSE = 2.83, R² = 0.957), followed by RF (RMSE = 3.12, R² = 0.93) and SVR (RMSE = 3.45, R² = 0.92). Integrating ML and GIS improved groundwater quality assessment and provided a robust framework for sustainable irrigation management. These findings provide critical insights for optimizing agricultural water use in water-scarce regions.

In the semi-arid Batna Belezma region of northeastern Algeria, groundwater is a vital resource for agriculture and drinking water. However, the climate leads to intense evaporation, which affects its quality. This study aims to identify the key hydrogeochemical processes that control groundwater composition in the Merouana Basin and to evaluate the predictive performance of machine learning (ML) models. A total of 30 groundwater samples were analyzed using multivariate statistical techniques, including Principal Component Analysis (PCA), and were modeled using PHREEQC to assess mineral saturation states. Additionally, ML-based regression models, including K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGB),were employed to predict groundwater chemistry. The results indicate that the dominant ion distribution follows the following trend: Ca²⁺ > Mg²⁺ > Na⁺ and HCO₃⁻ > SO₄²⁻ > Cl⁻. Alkaline earth metals (Ca²⁺ and Mg²⁺) constitute the major fraction of total dissolved cations, reflecting carbonate equilibrium and dolomite dissolution processes. In contrast, Na⁺ represents a smaller proportion of the cationic load; however, its hydro-agronomic significance is substantial due to its influence on sodium adsorption ratio (SAR) and soil permeability. The PHREEQC modeling showed that calcite and dolomite precipitation promote evaporite dissolution, while most samples remain undersaturated with respect to gypsum. The PCA results reveal high positive loadings of Mg²⁺, Cl⁻, SO₄²⁻, HCO₃⁻, and EC, suggesting that ion exchange and seawater mixing are the primary controlling processes, with carbonate weathering playing a secondary role. To enhance predictive assessment, several supervised machine learning models were tested. Among them, the Random Forest model achieved the highest predictive performance (R² = 0.96) with low RMSE and MAE values, confirming its robustness and reliability. The results indicate that silicate weathering and mineral dissolution are the primary mechanisms governing groundwater chemistry. The integration of multivariate statistics and machine learning provides a comprehensive understanding of groundwater evolution and offers a reliable predictive framework for sustainable water resource management in semi-arid environments. Geochemical model performance showed a high global accuracy (GPI = 0.91), confirming a strong agreement between observed and simulated chemical data. However, the HH value (0.81) indicates some discrepancies, particularly for specific ions or extreme conditions.