Linking cadmium source uncertainty to crop accumulation and health risk in mining-impacted agricultural soils using...

Overlapping source signatures in mining-impacted farmland hinder source-to-risk interpretation of cadmium (Cd) contamination. This study integrated positive matrix factorization (PMF), stable Cd isotope analysis, Bayesian isotope mixing modelling (MixSIAR), crop monitoring, and pathway-specific health-risk assessment. The dataset comprised 240 soil and 48 grain samples from a 30-60-year coal-mining chronosequence in the North China Plain. Soil Cd ranged from 0.27 to 0.76 mg kg^-1 (mean ± SD = 0.51 ± 0.10 mg kg^-1), 5.40 times the regional background. PMF apportioned Cd to industrial/mining (50.8%), traffic-logistics/resuspension (36.6%), and atmospheric/geogenic (12.7%) factors. MixSIAR estimated posterior mean contributions of 30% from coal-related sources (95% credible interval [CrI], 3-66%), 27% from atmospheric fallout (2-62%), 24% from fertilizers (2-53%), 17% from secondary mining materials (2-40%), and 2% from irrigation water (0.1-6.8%). Wide CrIs reflected overlapping source signatures; together, the PMF and MixSIAR models identified coal/mining-related and atmospheric/dust-associated inputs as plausible priority sources despite uncertain fractions. Mean Cd was higher in wheat (0.14 mg kg^-1) than maize (0.09 mg kg^-1). Grain Cd exceeded the Chinese food-safety limit in 62.5% of wheat and 41.7% of maize samples. Multi-metal soil-contact screening yielded total hazard indices of 9.90E-02 for adults and 1.02 for children. Cd-specific soil-contact hazard indices were 1.17E-03 and 7.64E-03, respectively, compared with dietary Cd hazard quotients of 0.58 and 1.11. Thus, grain consumption was the main Cd exposure pathway. These findings support tiered management prioritizing industrial and dust-source control, paired soil-grain monitoring, pH management, and crop selection while acknowledging source uncertainty.