Storage‐dependent drainable porosity for complex hillslopes

Abstract

In hydraulic groundwater theory the parameter drainable porosity f (a storage coefficient that accounts for the effect of the unsaturated zone on water table dynamics) is usually treated as a constant. For shallow unconfined aquifers the value of this parameter, however, depends on the depth to the water table and the water retention characteristics of the soil. In this study an analytical expression for f as a function of water table depth is derived under the assumption of quasi‐steady state hydraulic equilibrium, in this way accounting, in part, for the effects of the unsaturated zone on groundwater dynamics. The derived expression is implemented in the nonlinear hillslope‐storage Boussinesq (HSB) model (Troch et al., 2003) to simulate the drainage response of complex hillslopes. The model's behavior is analyzed by comparison to (1) the HSB model with a constant value for f and (2) measurements of water tables and outflow hydrographs on a 6.0 × 2.5 × 0.5 m laboratory hillslope experiment. The comparison is conducted for a pure drainage case on two different hillslope shapes (linearly convergent and divergent) and for three different slope inclinations (5%, 10%, and 15%). Comparison 1 is run in an uncalibrated and a fully calibrated mode, and it enables us to evaluate the effect of a dynamic, state‐dependent value for f on model output. Comparison 2 allows us to test the HSB model on several hillslope configurations and to analyze whether the concept of a storage‐dependent f enhances the model performance. The comparison of the HSB models to the measurements from the laboratory hillslopes shows that it is possible to capture the general features of the outflow hydrograph during a drainage experiment using either one of the HSB models. Overall, the original (constant f) HSB model, with one fitting parameter more than the revised HSB model, shows a slightly better fit on the hydrographs when compared to the revised (variable f) HSB model. However, the peak outflow values (the first few minutes after initiation of the experiments) are better captured by the revised HSB model. The revised HSB model's performance in simulating water table movements is much more accurate than that of the original HSB model. The improved match of the revised HSB model to piezometric measurements is worth stressing because the ability to model water tables is a key attribute of the model, making it possible to investigate phenomena such as saturation excess runoff. Also noteworthy is the good match between the revised HSB model and the outflow measurements, without any calibration, for the divergent slopes. The changing values of the calibrated drainable porosity parameter for the original HSB model as different configurations are simulated (slope angle, plan shape, initial conditions), together with the ability of the revised HSB model to more accurately simulate water table dynamics, clearly demonstrates the importance of regarding drainable porosity as a dynamic, storage‐dependent parameter.