Simulating water infiltration in the vadose zone is crucial for water resource management and environmental remediation, which is typically accomplished by numerically solving the Richards equation (RE). While existing research has acknowledged the limitations of the RE – notably its assumption of infinite air-phase mobility – a systematic assessment of its applicability across diverse hydrogeological conditions remains incomplete. This study evaluates the RE model’s performance in simulating infiltration by benchmarking it against a two-phase (TP) flow model under various boundary conditions, soil-water retention characteristics, and soil moisture states. The findings reveal that the RE model exhibits reliability degradation under saturated or ponding boundary conditions, as well as in scenarios involving high initial soil moisture, fine-textured soils (characterized by low van Genuchten - Mualem parameters , , and reduced absolute permeability ). Mechanistically, these scenarios restrict the air escape pathways by either reducing their effective conductance or by topologically sealing the exit routes. This restriction leads to elevated pore-air pressures that impede water infiltration, which the RE model fails to capture, thereby overestimating infiltration. In terms of practical applicability, the RE model likely remains robust wherever high-permeability channels form a percolating network connected to the surface within the region of interest. Conversely, for scenarios involving potential extensive confining layers, three conservative quantitative guidelines are proposed to assess the RE model’s suitability. • A comparison of the Richards-based (RE) and Two-Phase (TP) flow models is provided. • Conditions where the RE model overestimates infiltration are determined. • Mechanisms leading to RE model failure are elucidated. • Preliminary guidelines for when to apply the RE model are proposed.
Gao et al. (Sun,) studied this question.