Construction of a process-driven 3D spatio-temporal model for precision soil remediation: Coupling dynamic retention with multi-objective cost optimization

Current static approaches for quantifying soil pollution damage often hinder effective environmental management due to their reliance on single-point thresholds, simplified 2D spatial assumptions, and linear cost summations. These limitations fail to capture dynamic pollutant behaviors or account for implicit lifecycle losses, leading to suboptimal remediation decisions. To provide a robust scientific tool for precision management, this study constructs a Process-Driven 3D Spatio-Temporal Quantification (3D-STQ) model. This model is underpinned by a novel Dynamic Retention Equation for Soil Pollutants (DRE-SP), which systematically couples multi-source input fluxes, biodegradation, hydrodynamic migration, and adsorption-desorption kinetics. By integrating adaptive 3D meshing with the NSGA-II genetic algorithm, the model optimizes the trade-offs between remediation cost, time efficiency, and environmental risk. Field validation at an electroplating site demonstrated high precision, with a pollutant concentration coefficient of determination (R2) of 0.96 and a negligible 0.1% deviation in crop yield loss prediction. Furthermore, application at a pesticide-contaminated site revealed that the model-optimized strategy reduced total lifecycle costs by 32.2% and shortened project duration by 41.7% compared to traditional methods, while maintaining 96.2% prediction consistency. This study demonstrates that the proposed dynamic framework significantly enhances the accuracy of damage quantification, offering a valuable decision-support tool for sustainable contaminated site remediation and cost control.