Source apportionment of heavy metal pollution in farmland soils on a regional scale using geographically weighted PCA

Source apportionment of soil heavy metals is an challenge and urgent work as the result of the rapid development of industrialization and urbanization. The common approach is multivariate statistical analysis, such as PCA and APCS/MLR, which infers only a single pattern of sources of heavy metals in entire study area. Due to complicated pathways and processes, patterns of pollution sources in a whole region may include two or more. Hence, we developed an analytical framework based on GWPCA to explore multiple patterns of sources of soil heavy metals on a regional scale. Xiangtan county, an important grain-producing area in China, was taken as a case study, which suffers the problem of heavy metal pollutions. Our results revealed the pollution situations of five soil heavy metals(Pb, Cd, As, Cr and Hg) in farmland soils and suggested that there exists various pollution patterns of these heavy metals in Xiangtan county. In each pattern, the structure of contamination sources is different. Our study also indicates that the analytical framework considering the spatial heterogeneity of pollution sources can help take more precise practices to solve this vital problem.