Spatiotemporal evolution and ecological risk assessment of heavy metals in agricultural black soils of Northeast China.

Black soils are globally important agricultural resources, yet comprehensive trans-provincial assessments of heavy metal contamination and its spatiotemporal dynamics under future scenarios remain scarce. This study assessed the spatiotemporal distribution, drivers, and future projections of soil heavy metal (arsenic (As), cadmium (Cd), chromium (Cr), copper (Cu), lead (Pb), and zinc (Zn) in Northeast China's black soil region. Data from 486 sites revealed significant spatial heterogeneity in the concentrations of As (4.17-38.59 mg kg-1), Cr (9.31-590.98 mg kg-1), Cd (0.05-5.89 mg kg-1), Pb (12.32-2099 mg kg-1), Zn (17.32-249.30 mg kg-1), and Cu (4.41-69.26 mg kg-1). Among these, arsenic, chromium, and cadmium exhibited extensive enrichment, with exceedance rates relative to the national background values reaching 69.55%, 50.00%, and 47.12%, respectively. Ecological risk assessment identified As, Cr, and Cd as the primary contaminants of concern, attributed to their high exceedance rates and toxicity coefficients (As: 10; Cd: 30; Cr: 2). The optimized extreme gradient boosting (XGBoost) model (via hyperparameter tuning) achieved good predictive accuracy (R2 > 0.65) and identified geographical, climatic, and anthropogenic factors as key drivers. Shared Socioeconomic Pathway (SSP)-Representative Concentration Pathway (RCP) scenario projections to 2040 indicated clear pathway dependence: high emissions (SSP5-8.5) drive rapid inland expansion of Pb and Cd, while sustainable development (SSP1-2.6) limits pollution spread. The findings provide a basis for source control and policy optimization, demonstrating the value of integrating machine learning with scenario simulation for environmental governance.