Bridging Complexity and Efficiency in Vadose Zone-Aquifer Interaction Modelling

Recharge processes linking rainfall to groundwater and streamflow responses play a critical role in catchment hydrology, yet remain a major source of uncertainty due to complex interactions between vertical vadose-zone dynamics and lateral groundwater flow. In this study, a parsimonious, semi-distributed modelling framework is developed to evaluate how alternative representations of recharge influence catchment-scale groundwater dynamics by explicitly coupling a one-dimensional Richards equation solver for vadose-zone soil moisture dynamics with a Hillslope-Storage Boussinesq (HSB) model that simulates topography-driven lateral groundwater redistribution. Hillslope geometry is parameterised using geomorphological width functions, enabling efficient representation of subsurface storage and flow while retaining physical interpretability. Groundwater recharge flux is estimated using three conceptualisations of increasing complexity: (i) an HSB coupled linearised diffusion-based Richards equation, (ii) an HSB coupled nonlinear Richards equation with  van Genuchten soil hydraulic parameters, and (iii) an HSB-HYDRUS 1D coupled model wherein the Richards equation is solved using a linear finite element scheme with implicit time integration. These hierarchical approaches are applied to the well-instrumented Maimai catchment, New Zealand, using observed rainfall and groundwater-level time series. The results reveal that representation of vertical recharge dynamics exerts a dominant control on simulated groundwater response. The coupled linearised Richards equation  produces unrealistically rapid recharge signals, overestimating groundwater levels under wet antecedent conditions; whereas the coupled nonlinear Richards equation including HYDRUS-1D based coupled models could capture the critical vadose-zone buffering, yielding delayed and smoother groundwater responses with improved accuracy. These findings demonstrate that moderately- complex models can provide an effective balance between the physical realism and computational efficiency in modelling the subsurface storage–flow interactions at the catchment scale.