Geostatistical analysis and multivariate assessment of groundwater quality

Appraising groundwater quality is critical for identifying health-related risks and establishing foundations for safe water supply management. Traditional statistical analyses of groundwater quality often overlook spatial information and autocorrelation, which are essential for comprehensive assessment. This study examines 76 water samples collected from Tehsil Jaranwala, analyzing 12 water quality parameters (WQPs) through multivariate statistical techniques and a non-parametric geostatistical kriging approach. We estimate the variogram parameters using three techniques: Ordinary Least Squares (OLS), Maximum Likelihood Estimation (MLE), and Restricted Maximum Likelihood (REML), selecting the method with the lowest mean square error for further analysis. These parameters are incorporated into kriging techniques to predict WQPs at unmeasured locations. Contour plots generated via the R software package geoR provide a visual representation of ungauged locations, highlighting the most critical areas. This case study revealed that eight significant WQPs, including Electrical Conductivity (EC), Sulfate, Total Dissolved Solids (TDS), Sodium, Chloride, Bicarbonate, Alkalinity, and Fluoride, were present in concentrations higher than the World Health Organization (WHO) permissible limits in certain regions of Jaranwala. Specifically, elevated contamination levels were observed in areas between the north longitude of 73.15°–73.20° and east latitude of 31.80°–32°. These findings underscore the potential health risks associated with groundwater consumption in this area and provide a foundation for informed policy decisions to mitigate these risks. These visualizations are intended to assist government agencies in making informed decisions and implementing necessary actions and policies.