Validation of finite water‐content vadose zone dynamics method using column experiments with a moving water table and applied surface flux

Abstract

Abstract Data from laboratory experiments on a 143 cm tall and 14.5 cm diameter column, packed with Wedron sand with varied constant upper boundary fluxes and water table velocities for both falling and rising water tables are used to validate a finite water‐content vadose zone simulation methodology. The one‐dimensional finite water‐content Talbot and Ogden (2008) (T‐O) infiltration and redistribution method was improved to simulate groundwater table dynamic effects and compared against the numerical solution of the Richards equation using Hydrus‐1D. Both numerical solutions agreed satisfactorily with time series measurements of water content. Results showed similar performance for both methods, with the T‐O method on average having higher Nash‐Sutcliffe efficiencies and smaller absolute biases. Hydrus‐1D was more accurate in predicting deponding times in the case of a falling water table, while Hydrus‐1D and the T‐O method had similar errors in predicted ponding times in the case of a rising water table in six of nine tests. The improved T‐O method was able to predict general features of vadose zone moisture dynamics with moving water table and surface infiltration using an explicit, mass‐conservative formulation. The advantage of an explicit formulation is that it is numerically simple, using forward Euler solution methodology, and is guaranteed to converge and to conserve mass. These properties make the improved T‐O method presented in this paper a robust and computationally efficient alternative to the numerical solution of the Richards equation in hydrological modeling applications involving groundwater table dynamic effects on vadose zone soil moistures.