Evaluating the Health and Ecological Risks of Toxic Elements in the Agricultural Soil of China using Traditional and Advanced Techniques: A One Health Approach

Abstract

Agricultural soil quality is critical for sustainable food production, yet contamination by toxic elements (TEs) may pose ecological and human health risks. This study assessed the concentrations, ecological impacts, and potential health risks of five TEs (Cr, Cu, Cd, Pb, and Zn) in agricultural soils across China by combining traditional pollution indices, health risk models, and machine learning. A total of 1131 georeferenced samples compiled from published studies, representing 6 plantation types, food crop, fruit, vegetable, tea, medicinal, and others, were analyzed. Spatial mapping, Pearson correlation analysis, and the United States Environmental Protection Agency (USEPA) health risk framework were used to characterize contamination levels and estimate exposure. Among the plantation types examined, food crop plantations showed the highest contamination levels, with Pb and Zn most prominent, whereas Cd contributed most to ecological risk. Pollution index classification identified several regional hotspots warranting further monitoring. Estimated hazard indices remained below the threshold of concern for both adults and children, although they were consistently higher in children, and carcinogenic risks fell within acceptable limits for both groups. Of the three models tested, Random Forest performed the best (R2 = 0.99, 0.98, 0.96, 0.92, and 0.89 for Cu, Cd, Pb, Zn, and Cr, respectively), outperforming CatBoost and support vector regression. Feature importance and partial dependence analyses identified pH, magnesium, clay content, and total phosphorus as the most influential predictors of TE distribution. Interpreted within a One Health framework, these findings support continued monitoring and targeted management of agricultural soils in China.