Effects of Model Fidelity and Mesh Resolution on Modeling Hydrological Exchange Flows

Hydrological exchange flows (HEFs) are fluxes exchanged at the riverbed between fast-moving surface water and slow-moving groundwater. Modeling HEFs requires to model both surface water and subsurface water. Because the multi-scale nature of river corridors, data availability of river topography and boundary conditions, and computational limitations, models with different spatial resolution and riverbed pressure estimation have been used at plot, patch, reach, catchment, and watershed scale. As previous studies show that HEFs predicted from numerical models are sensitive to mesh resolution and pressure boundary, this work aims to systematically evaluate how mesh resolution and riverbed pressure affect HEFs using 1D, 2D, and 3D modeling for surface water and 3D model for subsurface processes. Specifically, Mass1 (1D) and Mass2 (2D) are firstly used to obtain cross-sectional and depth averaged water depth and flow velocity in a 50 km × 55 km region; OpenFOAM (3D) is then applied to obtain more accurate flow and pressure boundary in a 20 km × 12 km subregion of the domain. With these results, the riverbed pressure can be either estimated as a hydrostatic pressure (for 1D/2D models) or a combination of hydrostatic pressure and hydrodynamic pressure (for 3D model), which is further used as a pressure boundary for PFLOTRAN (3D) to solve subsurface flow and reaction transport in a 7 km × 7 km subregion. As these models run with different topography information and mesh resolution, four combinations of these models and mesh resolution are used to study their effects on HEFs, including 1) Mass1-derived hydrostatic pressure coupled with 100 m resolution PFLOTRAN; 2) Mass2-derived hydrostatic pressure coupled with 10 m resolution PFLOTRAN; 3) OpenFOAM-derived hydrostatic pressure coupled with 10 m resolution PFLOTRAN; 4) OpenFOAM-derived total (hydrostatic + hydrodynamic) pressure coupled with 10 m resolution PFLOTRAN.