An integrated model of log-normal ordinary Kriging interpolation-based source-specific human health risk assessment (LSR) for soil heavy metal pollution: insights from an abandoned industrial area in China

Abstract

Soil heavy metal (SHM) accumulation is a major global concern in environmental protection and public health, and accurate source-specific human health risk assessment is essential for effective risk control of SHM pollution. In this study, we developed an integrated LSR model that combines log-normal ordinary Kriging (LOK), source apportionment, and human health risk assessment. The model was applied to a multi-source contaminated site in southern Jiangsu Province, China. Statistical analysis shows that the mean values of Cd, Cr, Ni, Pb, and Zn are higher than the local background value (BV). Significant spatial variations of As, Cd, Cr, Pb, and Zn are also determined by their high coefficients of variation (CV). Cross-validation further shows that LOK performs better than ordinary Kriging (OK) methods in interpolation accuracy, especially under conditions of high variation coefficients of SHM contents. The reliability of source apportionment and human health risk assessment was confirmed by bootstrap and Monte Carlo simulation, respectively. The estimated 95th percentile total non-cancer (1.87) and cancer health risks (1.73 × E-3) for children were high, indicating elevated health risk for this population group. Source-specific risk estimates based on the LSR model showed that industrial and agricultural sources contributed 69.8% and 87.8% to the total non-cancer and cancer risks, respectively. These findings underscore that the LSR model could provide a reliable and cost-efficient way for quantifying sources related to human health risks, and support the decision-making in SHM pollution control.