Amelia Peeples ^* Reed Maxwell

• Princeton University, Princeton, United States

In hydrologic modeling, the assumption of homogeneity within a cell averages all variability finer than the model resolution. This loss of information can impact a model's ability to accurately represent hydrologic processes, especially in highly heterogeneous domains. This study quantified the impact of this loss of information on surface water fluxes by comparing the outputs of a high-resolution and coarse hydrologic model applied to an idealized domain. This study also presented a framework for including subgrid information in the surface water physics of integrated hydrologic models. Channel width was used as a representative subgrid parameter to better characterize surface water flow in cells containing subgrid channels. A new, nonlinear relationship between flux and calculated flow depth was derived based on assumed bathymetry and known channel width. This flux relationship was incorporated into ParFlow, an integrated 3D subsurface flow and 2D surface flow hydrologic model. In all scenarios tested, the subgrid channel formulation applied to a coarse-resolution model produced peak flows that only differed from the high-resolution model by more than 1% in 11/400 of scenarios and never differed by more than 5%. This is a substantial improvement from the baseline formulation applied to a coarse-resolution model, where peak flow differed by more than 1% in 213/400 scenarios and had a maximum difference of 78%.