Determination of the spatial correlation characteristics for selected groundwater pollutants using the geographically weighted regression model: A case study in Weinan, Northwest China

Abstract

Groundwater pollution is a serious issue in arid and semi-arid regions. In this study, the ordinary least squares (OLS) regression and geographically weighted regression (GWR) model were used to assess the relationship between hydrochemical parameters (NO₃-N, NO₂-N, NH₄-N, and F^-) and explanatory variables related to anthropogenic and natural factors, including elevation, slope, population density, groundwater electrical conductivity, groundwater pH, and land use in the Weinan region of China. The results showed that NO₃-N, NH₄-N, and F^- at 24, 4, and 54% of the samples exceeded the standard limits, respectively. Crop fields, grassland, and forest are the most common land use types in the study area, accounting for 62.84, 16.77, and 8.76%, respectively. The effects of explanatory variables on groundwater quality show strong spatial variation. Both positive and negative correlations were observed between groundwater nitrogen (NO₂-N, NO₃-N) and orchard, and between F^- and crop field. The water area has significant impacts on NH₄-N in Pucheng, Fuping and Linwei districts. The GWR model also suggested significant effects of water and orchard areas on groundwater NO₂-N concentration in western Fuping County and eastern Dali County, which was neglected by the OLS model. The research shows advantages of the GWR model in capturing local variation.

Keywords:

• Groundwater pollution
• land use
• nitrate
• geographically weighted regression
• water quality

Acknowledgements

We are grateful for the constructive and useful comments given by the journal editor and the reviewers. The English editor from Mogoedit is acknowledged for the polishing of the language.

Authors’ contributions

Fan Li conceptualized the core idea, analyzed the data, and wrote the initial draft of the paper. Jianhua Wu refined the idea, supervised the research, helped in the methodology development, carried out additional analyses and finalized this paper. Fei Xu, Yongqiang Yang and Qianqian Du contributed to the data analysis, writing and revisions.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The research was supported by the National Natural Science Foundation of China (42072286 and 41761144059), the Qinchuangyuan "Scientist + Engineer" Team Development Program of the Shaanxi Provincial Department of Science and Technology (2022KXJ-005), the Fok Ying Tong Education Foundation (161098), and the National Ten Thousand Talent Program (W03070125).