Integrated geostatistical and multiple linear regression modeling for assessment of groundwater quality parameters in the Burhi Gandak Basin

Groundwater resources in India’s northern plains are increasingly threatened by overexploitation and deteriorating quality, impacting aquifer dynamics, agriculture, and public health. This study integrates geostatistical techniques with multiple linear regression (MLR) to assess the spatial variability, controlling factors, and groundwater quality across the Burhi Gandak Basin. Groundwater levels were observed at shallow depths and exhibited a low hydraulic gradient (0.000168) with a predominantly unidirectional flow regime. Geostatistical interpolation of hydrochemical parameters revealed elevated concentrations of EC, TH, HCO3, Cl, SO4, Ca, Mg, Na, K, As, Mn, and Pb in the distal parts of the basin, whereas proximal zones exhibited more alkaline groundwater. A strong positive correlation between EC and major ions, together with MLR identifying Cl, Ca, and Mg as key TDS predictors, underscored the dominance of geochemical processes across the basin. Elevated ‘As’ concentrations in the central and distal regions, correlated with EC, HCO₃, SO₄, Ca, pH, and Fe, indicate mobilization through reductive dissolution of ‘As’-bearing Fe(III) minerals, with partial anthropogenic influence implied by its association with NO₃. The WQI and its MLR model identified Pb, As, and Mn as dominant predictors influencing drinking water quality, showing that 18.33% of samples fell within the poor to unsuitable category, particularly in Muzaffarpur, Samastipur, Begusarai, and Khagaria districts. For irrigation suitability, over 50% of groundwater samples fall within moderate to very high salinity classes (C3S1 – C4S1) in central and distal parts of the basin, indicating restricted use. SAR exhibited a strong correlation with Na and a moderate one with EC, with the MLR confirming Na as the dominant controlling factor. Overall, Pb, As, Mn, Cl, Ca, and Na identified as key influencing parameters governing groundwater chemistry. These findings establish a robust, cost-effective framework for spatial and predictive groundwater quality assessment, supporting sustainable monitoring and management in the basin and other alluvial regions.