STATUS OF GROUNDWATER QUALITY IN LUDHIANA DISTRICT OF PUNJAB, INDIA

Abstract

The present study provides an integrated assessment of groundwater quality in Ludhiana district, Punjab, based on 152 groundwater samples characterized by 18 physicochemical parameters. These were synthesized into a Drinking Water Quality Index (WQI) to evaluate potable suitability. The WQI classification shows that 2.6% of samples are “excellent” and 57.9% “good,” whereas 32.9%, 4.0%, and 2.6% fall into “poor,” “very poor,” and “unsuitable” categories, respectively, indicating spatially heterogeneous degradation in groundwater quality. Hydrogeochemical characterization, supported by facies diagrams, ionic diagnostics, saturation indices, and multivariate statistics, indicates that groundwater chemistry is primarily governed by carbonate and silicate weathering, with active cation exchange. Saturation indices show that groundwater is generally supersaturated with respect to calcite, dolomite, and aragonite, and undersaturated with respect to halite, fluorite, and sylvite. Spatial variability in nitrate and other key parameters reflects localized agricultural and urban influences, while IDW interpolation in QGIS has been used to generate spatial distribution maps of major ions. In the present study, irrigation suitability has been evaluated using salinity, sodicity, and carbonate–bicarbonate hazard indices, including EC, %Na, SAR, and PI. In addition, an Irrigation Water Quality Index (IWQI) is computed using EC, SAR, Na⁺, Cl⁻, and HCO₃⁻ concentrations. Salinity classification indicates that 62.5% of samples fall within the “medium” EC range (250–750 µS/cm), while 37.5% fall within the “high” category. Hazard assessment indicates that 75% of samples fall within the excellent-to-permissible range based on %Na, 97.4% belong to the S1 class (low sodicity) based on SAR, and all samples fall within PI Classes I–II. However, IWQI based classification indicates irrigation constraints, with 21.7% of samples under “severe restriction” (IWQI < 40), 37.5% under “high restriction” (40–55), 33.6% under “moderate restriction” (55–70), and only 7.2% under “low restriction” (70–85). IDW based GIS mapping delineates zones of elevated irrigation risk, particularly in eastern Samrala and Khanna, parts of Ludhiana II and Machhiwara, and southern to western xvi Pakhowal and Sidhwan Bet, whereas low-restriction pockets are confined to limited villages such as Sherpur Kalan, Bardeke, Hathur, and Dholanwal. An ANN-based predictive model for IWQI is developed using the multilayer perceptron architecture in IBM SPSS with standardized inputs. The optimal network includes five input neurons (EC, SAR, Na⁺, Cl⁻, HCO₃⁻), two hidden layers, and one sigmoid-activated output neuron, trained using batch learning and gradient-descent backpropagation. The model achieved RMSE values of 0.09 for training and 0.07 for testing, indicating strong predictive performance and close agreement between observed and simulated IWQI values. Sensitivity analysis shows that Cl⁻ has the highest normalized importance (100%), followed by SAR (59.7%), Na⁺ (58.5%), HCO₃⁻ (53.1%), and EC (27.2%), confirming the dominant influence of salinity and sodicity on irrigation water quality. Additionally, nitrate distribution has been examined to distinguish regional hydrogeological controls from localized anthropogenic influences. The weak statistical association between nitrate concentration and hydraulic parameters indicates that regional groundwater flow is not the principal control on nitrate occurrence. The lack of a coherent downgradient nitrate plume suggests that nitrate distribution is governed mainly by localized recharge and anthropogenic inputs. A conservative nitrate-transport scenario simulated by coupling the calibrated transient MODFLOW flow model with MT3D-USGS shows a decline in mean nitrate concentration from 8.86 mg/l in Year 1 to 6.51 mg/l in Year 10 (~27% reduction), contraction of the >45 mg/l footprint from ~1.70% to ~0.88%, and a reduction in peak concentrations to a maximum simulated value of ~69.95 mg/l after 10 years.