The release of sludge-derived heavy metals (HMs) to soil and their subsequent migration into groundwater poses a significant challenge for safe and low-carbon sludge land application. This study developed a predictive framework to simulate 60-year sludge land application, evaluating the risk of HMs pollution in the soil-groundwater environment and assessing the influence of soil and water properties. HYDRUS-2D simulations revealed that highly mobile Cu, Ni, and Zn penetrated a 10 m soil layer over a 60-year period, contributing to groundwater pollution. In contrast, Cr was easily sequestered within the topsoil layer after 5-years continuous operation. The non-equilibrium parameter α could serve as an indicator for assessing their potential risk. Furthermore, the limited soil adsorption sites for Pb (f = 0.02772) led to short-term (1-year) groundwater pollution at a 0.5 m-depth. Bayesian Networks model outcomes indicated that humic-like organics crucially influenced HMs transformation, enhancing the desorption of Cd, Cu, Ni, Pb, and Zn, while inhibiting the desorption for Cr. Additionally, electrical conductivity promoted the release of most HMs, in contrast to the Mn mineralogy in soil. This study bridges the gap between the macro-level HMs migration trends and the micro-level adsorption-desorption characteristics, providing guidance for the safe land application of sewage sludge. This study introduces a framework integrating HYDRUS-2D simulations with Bayesian Networks to assess the risks of groundwater pollution by heavy metals (HMs) over a 60-year sludge application. Sludge-derived Cu, Ni, and Zn are found to penetrate soil up to 10 m and exceed safety limits, with the non-equilibrium parameter α serving as an indicator for pollution risk. The importance of nutrients from sludge-amended soil for the transformation of HMs in the subsurface environment highlights the need for enhanced sludge management, specifically through more detailed regulation of nutrient composition. These findings contribute to developing precise strategies for the long-term sludge land application.

Integrating HYDRUS-2D and Bayesian Networks for Simulating Long-Term Sludge Land Application: Uncovering Heavy Metal Mobility and Pollution Risk in the Soil-Groundwater Environment