Elucidating the spatial determinants of heavy metals pollution in different agricultural soils using geographically weighted regression

Abstract

Intensive human activities caused massive socio-economic and land-use changes that directly or indirectly resulted in excessive accumulation of heavy metals in agricultural soils. The goal of our study was to explore the spatial determinants of heavy metals pollution for agricultural soil environment in Sunan economic region of China. We applied geographically weighted regressions (GWR) to measure the spatially varying relationship as well as conducted principal component analysis (PCA) to incorporate multiple variables. The results indicated that our GWR models performed well to identify the determinants of heavy metal pollution in different agricultural soils with relatively high values of local R². Heavy metal pollution in Sunan economic region was crucially determined by accessibility, varying agricultural inputs as well as the composition and configuration of agricultural landscape, and such impacts exhibited significantly heterogeneity over space and farming practices. For the both agricultural soils, the major variance proportion for our determinants can be grouped into the first four factors (82.64 % for cash-crop soils and 73.065 for cereal-crop soils), indicating the incorporation and interactions between variables determining agricultural soil environment. Our findings yielded valuable insights into understanding the spatially varying 'human-land interrelationship' in rapidly developing areas. Methodologically, our study highlighted the applicability of geographically weighted regression to explore the spatial determinants associated with unwanted environmental outcomes in large areas.

Keywords: Agricultural soils; Geographically weighted regression (GWR); Heavy metal pollution; Spatial determinants.