Multiphysics modeling of soil freezing and thawing processes: synthesis, validation, and outlook for soil-system science
Abstract. Freezing and thawing transform soils through tightly coupled transfers of heat, water, vapor, solutes, and mechanical stress. These processes regulate frost heave, thaw settlement, infiltration, runoff generation, groundwater recharge, salinity redistribution, and the stability of cold-region infrastructure. Multiphysics models have advanced from empirical frost indices and thermal-hydraulic formulations to thermo-hydro-mechanical (THM) and thermo-hydro-mechanical-chemical (THMC) frameworks that represent phase change, cryosuction, unsaturated flow, vapor transport, ice segregation, poromechanical deformation, and damage. This review synthesizes freezing-soil modeling as a soil-system problem rather than only an engineering-design problem. We first organize the controlling mechanisms, including soil-freezing characteristic curves, soil-water retention behavior, generalized Clapeyron relations, frozen-fringe dynamics, ice-induced permeability reduction, and freeze-thaw damage. We then compare model families, primary variables, constitutive closures, numerical strategies, and validation requirements. Particular emphasis is placed on the gap between model complexity and data availability: temperature-only calibration can reproduce frost depth while failing to constrain water redistribution, deformation, and stress. We therefore argue that defensible THM/THMC modeling requires multi-observable calibration, conservation checks, out-of-sample validation, and explicit uncertainty propagation. The review concludes with a practical model-selection and reporting framework for soil scientists, hydrologists, and cold-region engineers. Priority needs include transferable closures for permeability and unfrozen water, shared benchmark datasets, uncertainty-aware inference, and hybrid continuum-discrete descriptions of ice lenses, interfaces, and damage.
I am not fully sure the correctness of equation (8). If relative permeability is already a function of liquid saturation only, which should mean that ice effect has been considered. Will the additional factor from the ice impedance function double count ice effect?