Key points are not available for this paper at this time.
Assessing heavy metal pollution in river sediments and identifying the key factors contributing to metal pollution are critical components for devising river environmental protection and remediation strategies to protect human and ecological health. This is especially important in urban areas where metals from a wide range of sources contribute to sediment pollution. In this study, the metal enrichment factor (EF) was used to measure the watershed distribution of Cu, Zn, Pb and Cd in sediments in the Wen-Rui Tang urban river system in Wenzhou, Eastern China. The Geographical Detector Method (GDM) was specifically evaluated for its ability to analyze spatial relationships between metal EFs and their anthropogenic and natural control factors, including densities of industry (DI), livestock (DL), service industries (DS), population (DP), and roads (DR), along with agricultural area (AG), sediment total organic carbon (TOC), and soil types (ST). Results showed that the watershed was highly contaminated by all metals with an EF trend of Cd ≫ Zn > Cu > Pb. The spatial distribution of EFs demonstrated high contamination of all metals in the southwestern region of the watershed where industrial activities were concentrated, and higher Cu and Zn concentrations in the northeastern region having a high density of livestock production. GDM results identified DI as the dominant determinant for all metals, while TOC and ST were determined to have a moderate secondary influence for Zn, Pb and Cd. Additionally, GDM revealed several additive and nonlinear interactions between anthropogenic and natural factors influencing metal concentrations. Compared to other correlation, multiple linear regression and geographically weighted regression, GDM demonstrated distinct advantages of being able to assess both categorical and continuous variables and determine both single and multiple factor interactions. These attributes provide a more comprehensive understanding of metal spatial distributions while avoiding multicollinearity issues when identifying significant contributing factors at the watershed scale.
Luo et al. (Fri,) studied this question.