Vertical source-risk mismatch of heavy metal(loid)s in stratified soils from an industrial-mining-agricultural area

Conventional source apportionment frameworks frequently emphasize surface soils, which can obscure vertical contaminant migration and may also conflate pollutant burden with health risk. We address this limitation by developing an integrated Stratified Positive Matrix Factorization (PMF)-Source-Oriented Human Health Risk Assessment framework, which is supported by the Automated Grouped Source Apportionment Tool. By analyzing 138 soil samples from three depth intervals and 46 paired crop samples from a typical industrial-mining-agricultural complex, we quantified both the vertical heterogeneity and the mechanisms of source-risk mismatch. The analysis showed that elevated Cd and Hg concentrations in surface soils were mainly associated with agricultural activities (26.86%) and atmospheric deposition (22.81%), whereas elevated As concentrations in deep soils were linked to historical accident-related inputs. A clear decoupling was observed: agricultural activities dominated mass flux, yet Factor 5 (Accidental Source) and Factor 2 (Natural Background) emerged as major contributors to the modeled health risks; dietary intake outweighed direct soil contact as the primary exposure pathway. Based on these findings, we propose a layered control strategy involving deep-source containment, surface-interface protection, and crop-process control to address the limitations of surface-focused remediation. Overall, these findings underscore the importance of incorporating vertical processes into risk management and provide a source apportionment framework for complex environmental settings.