Groundwater contamination characterization using multivariate statistical analysis and geostatistical method

Abstract

The aim of the present study is to identify sources of groundwater contamination in Rupnagar district, Punjab, using an integrated approach of exploratory factor analysis (EFA) and ordinary kriging (OK). For this, a 13 physico-chemical parameter data set at 14 sampling locations for a period of over 25 years was assessed. The correlation was statistically examined amongst parameters. A five-factor model is proposed which explains over 89.11% of total groundwater quality variation. Three semi-variogram models, namely exponential, Gaussian, and spherical, fitted well for the data set and are cross-validated using predictive statistics. Spatial variability maps of all the parameters and factor scores are generated and are in good agreement with each other. The variation seen in groundwater quality is mainly due to various hydrogeochemical, anthropogenic, and geogenic processes occurring in the region. Thus, this study indicated that there is need to treat the industrial and municipal wastewater before discharging it (directly/indirectly) into nearby streams and pits and to encourage sustainable agricultural practices to prevent adverse health effects and minimize further environmental degradation in the study region.