Comparison and preference of source analysis models for heavy metal contamination of soil from alum mines in Anhui Province, China

Soil pollution is becoming increasingly serious due to resource exploitation in mining areas, which directly affects the stability of the surrounding soil and water environment, and the accurate prediction of heavy metal pollution source analysis has an important guiding role in scientific mine remediation and treatment. On the basis of the previous study, in order to improve the interpretation of the soil media model, positive matrix factorization (PMF) and absolute factor analysis-multiple linear regression (APCS-MLR) models, which have a high degree of maturity and accuracy in the evaluation of heavy metal source analysis in soils, were further selected to enrich the UNMIX model in the previous study. The results of the multiple receptor models analyzing the sources of heavy metal contamination and the contribution of the sources were combined with the ecological risk assessment method to determine the most suitable model in study area under quantitative conditions. The results showed that: 1) the ecological high-risk areas were mainly concentrated in the south and east part of the study area, and Cd was the main risk element of the mine;the mean value of igeoaccumulation index was 3.75, which was highly consistant with the mean value of 731.22 for the potential ecological risk index though latter provided a more comprehensive evaluation of ecological hazards. 2) Comparing the results of the three models, the PMF model resolved four pollution sources: coal combustion source, natural-traffic integrated source, natural source and atmospheric deposition source, and the source contribution rates are 38.15%, 20.62%, 24.28% and 16.95%, respectively. 3) The fitted variables R² in the PMF model reached to 0.96, which showed the best fitting effect. The PMF model refines source apportionment by accounting for sampling point errors and accurately ascertaining pollution sources and their contributions, making it highly suitable for complex mine-contaminated soils in realistic research settings. The findings of this study can provide a reference for tracing the origins of contaminants, which is essential for the remediation efforts of polluted soils following mining activities in the area.