Assessment of four infiltration models for various soil experimental conditions

Abstract

Infiltration modeling in structured and heterogeneous soils remains a practical challenge, particularly where local soil properties have received limited systematic attention. This study reports on a controlled laboratory column experiment designed to evaluate four widely used infiltration models, Kostiakov, Modified Kostiakov, Philip, and Horton, across a range of experimental conditions involving soil layering, antecedent moisture content, and the presence of entrapped air. The soil under investigation is Hamri, a characteristic reddish soil from the Mehdia region of Kenitra, Morocco, for which comparative model performance data have been scarce. Across all conditions tested, the Philip model delivered the most consistent and accurate estimates, a result attributable to its grounding in the physical mechanics of water movement through porous media rather than purely empirical curve fitting. The Kostiakov and Modified Kostiakov models also performed creditably, particularly in homogeneous and pre-wetted soil configurations, with the modified form showing a meaningful advantage in conditions of greater complexity. The Horton model lagged behind in most scenarios yet proved to be the strongest performer under air-entrapped conditions, an outcome that points to the model's sensitivity to exponential decay dynamics when pore-scale air blockage suppresses infiltration capacity. These findings reinforce the value of physics-informed formulations for local soil characterization and draw attention to an often-overlooked variable: the effect of entrapped air on infiltration behavior. The results carry practical implications for hydrological model selection in semiarid Moroccan contexts, where soil conditions and irrigation practices can vary considerably across short distances.