Application of stochastic model to assessment of heavy metal(loid)s source apportionment and bio-availability in rice fields of karst area.

Mining activities and high geological background are considered the important factors causing heavy metal(loid)s accumulation in rice fields of karst area. In this study, the contents, main sources, and the factors influencing bio-availability of heavy metal(loid)s were determined using conditional inference tree (CIT), random forest (RF), and geostatistical analyses with 105 soil samples collected from rice fields in karst area. Contamination by Cd, Hg, As, and Pb in soil was relatively serious in the study area in which the compound pollution was highly similar to that in the flooded area. CIT and RF effectively identified the contributions of natural and anthropogenic inputs of soil heavy metal(loid)s. Concentrations of Pb, As, and Hg were closely associated with human inputs whose cumulative contribution rates reached 68%, 87%, and 86%, respectively. Industrial activities (28%) and geogenic characteristics (44%) were primary sources of Cd accumulation. The soil pH, soil organic matter (SOM), distance from city center, the contents of heavy metal(loid)s in soil, and industry type were the most important factors influencing bio-availability of heavy metal(loid)s. Combined effect of multiple metals could not be ignored, in which As and Cd contributed over 80% to total non-carcinogenic risks for adults and children.