Identification of heavy metal sources in reservoir-adjacent soils and specific source risk assessment based on...

Heavy metal (HM) contamination in reservoir-adjacent soils is influenced by complex environmental factors and poses ecological and health risks. This study analyzed eight HMs (As, Cd, Cr, Cu, Hg, Ni, Pb, Zn) in 181 soil samples from Zhijin County, China, with mean concentrations exceeding regional background values by 1.06 to 7.00 times (except As, which was 91.4 % of the Guizhou background value). This study aimed to identify HM sources and associated risks by integrating GeoDetector (GD) and Positive Matrix Factorization (PMF) models in Zhijin County, China. The GD analysis revealed that river proximity (q = 0.544) had the strongest explanatory power for Cd spatial variation, while soil organic matter (q = 0.281) had the greatest influence on Pb distribution. The PMF identified five sources with quantified contributions and their dominant associated metals: a mining-agricultural mix (28.16 %; mainly Cd, Cu, Ni, and Zn), atmospheric deposition (20.39 %; mainly Hg), soil parent material (19.86 %; mainly Cr and Ni), agricultural activities (16.98 %; mainly As), and a transportation-mining mix (14.61 %; mainly Pb). Risk prioritization showed that mining and agricultural sources contributed 38.7 % of the ecological risk, while agricultural activities accounted for 41.2 % of children's non-carcinogenic risk, with ingestion contributing 89.7 % of non-carcinogenic exposure. The integrated GD-PMF framework improved the source resolution accuracy by 22-35 % compared to conventional methods. These results offer a novel source-oriented framework for prioritizing the control of Cd and Hg, which exhibited the highest ecological and health risks. Meanwhile, the framework also provides scientific support for the differentiated management of other heavy metals such as Pb and As, based on their source characteristics and spatial patterns.