Identifying priority control factors for heavy metal pollution in urban agricultural soils of the Chengdu Plain: An integrated multi-model approach.

Heavy metal (HM) contamination in urban and peri-urban agricultural soils presents a critical threat to environmental integrity and public health, yet the contributions of specific pollution sources to these risks remain insufficiently quantified. This study investigates urban and peri-urban agricultural soils in the Chengdu Plain, China, through an integrated framework combining Positive Matrix Factorization (PMF), GIS-based mapping of factor contribution, Random Forest (RF), Geo-Big data, Potential Ecological Risk Assessment (PERA), and Human Health Risk Assessment (HHRA). Four major sources of HMs were identified: 30.0 % of natural sources (Ni, Cr), 29.5 % of mixed agricultural and transportation sources (Cd, Cu, Pb, Zn), 19.4 % of metal smelting activities (As), and 21.1 % of atmospheric deposition sources (Hg). Risk assessments revealed that Cd and Hg posed the highest ecological risks, while HHRA indicated unacceptable health risks for children, with non-carcinogenic and carcinogenic risk probabilities reaching 4 % and 10 %, respectively. Source-specific analysis showed that both mixed agriculture and transportation sources (37.6 %) and atmospheric deposition (37.9 %) contributed most to the ecological risk. In terms of health risks, metal smelting activities (30.4 %) contributed most significantly to non-carcinogenic risk, whereas mixed agriculture and transportation sources (42.7 %) were the leading contributors to carcinogenic risk. Cd, Cu, Zn, Hg, and As were identified as priority elements requiring urgent management. This study establishes a robust, multi-dimensional framework for tracing HM sources and evaluating associated risks, providing scientific guidance for targeted pollution control and sustainable urban and peri-urban agricultural soil management.