Source Apportionment-Based Assessment of Ecological and Health Risk for Potentially Toxic Elements in the Soil Around...

Uranium is a critical strategic resource. However, uranium mining can cause severe contamination of surrounding soils by potentially toxic elements (PTEs). Conventional approaches that separately perform source apportionment and risk assessment fail to establish a direct linkage between pollution sources and their associated environmental and health risks. Herein, we develop an integrated framework that combines the absolute principal component score-multiple linear regression (APCS-MLR) model, the potential ecological risk index, and Monte Carlo simulation. The established framework was used to try quantifying the potential sources of soil PTEs and their corresponding ecological and human risks in a uranium mining area. The results showed that the concentrations of U, Th, Cd, and Cr significantly exceeded local background levels, with Cd and U exhibiting higher spatial variability than the other PTEs. The APCS-MLR model identified three potential sources of soil PTE contamination: mining activities, mixed anthropogenic-natural, and natural sources. Mining and mixed sources were the dominant contributors to ecological risk, jointly accounting for 88.7% of the total ecological risk, with Cd identified as the primary ecological risk pollutant. The mixed sources also contributed the largest proportions of non-carcinogenic and carcinogenic health risks, accounting for 52.56% and 67.5%, respectively. Furthermore, children were found to face greater health risks than adults due to higher exposure levels. Priority factor analysis indicated that pollution management should continuously monitor Cd derived from mining activities based on statistical inference. Overall, the proposed integrated framework successfully established a quantitative linkage between pollution sources and associated risks, providing a scientific basis for source-specific and zone-specific soil pollution management in uranium mining areas.