The impact of groundwater-land surface interactions on hydrologic persistence in macroscale modeling

Drought severity can be characterized in terms of runoff and soil moisture percentiles relative to their climatologies; however, spatially distributed and temporally continuous measurements of these quantities are scarce. Hydrologic models, like the macroscale Variable Infiltration Capacity (VIC) model, can simulate these quantities in an internally consistent manner, and also can simulate basin-integrated runoff (streamflow) which is relatively well observed. The University of Washington produces daily nowcasts and weekly forecasts of runoff and soil moisture percentiles over the continental U.S. at 1/8th degree spatial resolution. Although these products are generally consistent with independently produced estimates such as the National Drought Monitor, drought persistence, in some cases, appears to be underestimated by land surface models such as VIC. One possible reason for this discrepancy is that VIC does not explicitly incorporate the effects of groundwater storage, changes in which tend to occur over longer time scales than changes in soil moisture storage, which are represented by the model. To improve our estimates of the persistence of hydrologic states in VIC, we have incorporated the Simple Groundwater Model (SIMGM) of Niu and colleagues at the University of Texas. We compare the resulting evapotranspiration, runoff, and soil moisture fields from simulations performed using VIC with and without SIMGM over a set of basins selected to represent varying vegetation and hydroclimatic conditions across the United States. We also consider differences in the persistence of soil moisture and runoff deficits in each simulation.