Statistical and multivariate statistical techniques to trace the sources and affecting factors of groundwater pollution in a rapidly growing city on the Chinese Loess Plateau

Abstract

Groundwater quality is defined by various water quality parameters. The aims of the research are to understand the relationships among different groundwater quality parameters and to trace the sources and affecting factors of groundwater pollution via statistical and multivariate statistical techniques. The 36 shallow groundwater samples collected from shallow pumping wells in Yan’an City were analyzed for various water quality parameters. Correlation analysis, principal component analysis (PCA), hierarchical cluster analysis (HCA), and multivariable linear regressions (MLR) were jointly used in this study to explore the sources and affecting factors of groundwater pollution. The study reveals that the mineral dissolution/precipitation and anthropogenic input are the main sources of the physicochemical indices and trace elements in the groundwater. Groundwater chemistry is predominantly regulated by natural processes such as dissolution of carbonates, silicates, and evaporates and soil leaching, followed by human activities as the second factor. Climatic factors and land use types are also important in affecting groundwater chemistry. Cl^– is the greatest contributor to the overall groundwater quality revealed by the two regression models. The first model which has eight dependent variables is high in model reliability and stability, and is recommended for the overall groundwater quality prediction. The study is helpful for understanding groundwater quality variation in urban areas.