A fully coupled numerical modeling for regional unsaturated–saturated water flow

SummaryIt is a long-lasting challenge in subsurface hydrologic modeling to develop numerically efficient algorithm for coupling unsaturated and saturated flow, especially in regional-scale modeling. In this study, a new scheme is developed for coupled numerical simulation of unsaturated–saturated water flow at the regional scale. The modeling domain is divided into sub-areas in horizon according to spatially distributed inputs, flow characteristics, and topography conditions. The unsaturated zone of each sub-area is represented by individual one-dimensional soil column. Water balance analysis method is employed to formulate the three-dimensional groundwater model. The unsaturated and saturated zones are implicitly coupled in space and time through the vertical flow between the unsaturated soil columns and the saturated aquifers in that the heads in the unsaturated and saturated zones are integrated in a single matrix equation. The coupling scheme is verified and computational efficiency is evaluated in several hypothetical examples by comparing the simulation results with those of widely used software, including Hydrus1D, SWMS2D, FEFLOW and HydroGeoSphere. In the real-world application, numerical results show that the coupling model can obtain satisfactory simulation results with fairly little computational cost. Compared with existing models, the new numerical scheme is more suitable to regional-scale modeling with complex domain geometry and alternating recharge or discharge fluxes. However, due to the assumptions involved in the method development, the coupling method has its intrinsic limitations and should be used with caution in cases where the lateral flow is predominant in the unsaturated zone.

► The methodology for linking unsaturated and saturated water flow is developed.
► A new 3D groundwater module based on water balance analysis is established.
► The simplification to the unsaturated zone greatly reduces the calculation cost.
► The model is accommodated to problem with complex layers and varying topography.
► The model is not suitable when the lateral flow is dominant in the soil water.