The high-dimensional geographic dataset revealed significant differences in the migration ability of cadmium from various sources in paddy fields

Cadmium (Cd) contamination in paddy fields and its subsequent transfer in soil–rice systems are of particular concern. Significant discrepancies exist in the transfer process of Cd pollution sources from soil to rice. Here, we proposed a novel hybrid framework to reveal the priority of controlling Cd pollution sources in soil–rice systems, based on a high-dimensional geographical database. We further defined transfer potential (TP) to describe the ability of Cd from soil to rice (TP(r) = Cd(r)/Cd(s)) and activated status (TP(a) = Cd(a)/Cd(s)), respectively, to reveal the priority sources of Cd pollution at the regional scale. The mining source has both high levels of TP(r) and TP(a), which should be a controlled priority. Followed by traffic sources with a higher value of TP(r), showing the risk to rice rather than the soil. The activated and enriched capacities of soil Cd are unequal in different sources that we attribute to the disparities of Cd transport in soil–rice systems. Cd contamination shows a significant spatial heterogeneity due to the difference in its transport performance. Our findings provide support for designing site-specific and pollution-targeted control priorities for suitable Cd pollution mitigation strategies at the regional scale.