Predictive Modeling and Spatial Analysis of Irrigation Water Quality in a Key Agricultural Region: An ANN-Based Approach.

Abstract

The quality of groundwater plays a critical role in ensuring effective irrigation, directly impacting crop productivity and soil health. This study was carried out to assess the suitability of groundwater in Ludhiana district of Punjab, India, for irrigation quality. Salinity hazard for the water was categorized by using EC values, which, for 62.5% of the samples out of 152, falls between 250 and 750 (μS/cm), that is, in the "medium" category. The remaining 57 samples are categorized as having a "high" salinity hazard. High values of salinity hazard can create a physiological drought condition for the crop. Sodium hazard was analyzed using %Na and SAR values. For %Na values, 114 samples fall in excellent to good category, whereas, for SAR analysis, 148 samples fall in low sodicity category. Also, all the samples fall in class I and II for PI value. However, overall quality of irrigation water has been determined by calculating Irrigation Water Quality Index that aggregates EC, SAR, Na⁺, Cl^-, and HCO₃ ^- values. Based on this analysis, 21.7% of the samples fall in the "severe restriction" category whereas 57 samples, that is, 37.5% fall into the category of "high restriction." The remaining samples fall in moderate to low restriction. The maps depicting spatial distribution of contaminants and index values were prepared using IDW interpolation technique in QGIS. An optimized model for the study area was also created using ANN to estimate IWQI. The model was created using IBM SPSS software using multilayer perceptron feed forward mechanism. The calculated RMSE value of the proposed model is 0.09 and 0.07 for training and testing data, which suggests that the model's predictions are quite close to the actual values. This implies that the proposed model fits accurately and can be used for future IWQI prediction. This study advances SDG 6 by encouraging the responsible management of water resources. It also assists policy makers in developing sustainable irrigation strategies and provides researchers with important tools for predicting water quality.