Simulation of recharge for the Death Valley regional groundwater flow system using an integrated hydrologic model

A proof-of-concept study was conducted using the integrated hydrologic model, GSFLOW, to simulate spatially and temporally distributed recharge for the Death Valley regional groundwater flow system (DVRFS). GSFLOW is an integrated groundwater - surface water flow model that combines two modeling applications: the Precipitation-Runoff-Modeling-System (PRMS) and MODFLOW. Previous methods used to estimate recharge for the DVRFS include empirical models based on precipitation, applications of the chloride mass-balance method, and applications of a precipitation-runoff model, INFIL, which used a daily time step to simulate recharge as net infiltration through the root zone. The GSFLOW model offers several potential advantages compared to the previous methods including (1) the ability to simulate complex flow through a thick unsaturated zone (UZ), allowing for the dampening and time delay of recharge relative to the infiltration signal at the top of the UZ and also allowing for the redistribution of flow within the UZ, as enabled by the MODFLOW-NWT and UZF capabilities, (2) the simulation of rejected recharge in response to the dynamics of groundwater discharge and low permeability zones in the UZ, (3) a more explicit representation of streamflow and recharge processes in the mostly ephemeral stream channels that characterize the DVRFS, and (4) the ability to simulate complex flow paths for runoff occurring as both overland flow and shallow subsurface flow (interflow) in the soil zone using a network of cascades connecting hydrologic response units (HRUs). Simulations were done using a daily time step for water years 1980-2010. Preliminary estimates of recharge using GSFLOW indicate that the distribution of recharge is highly variable both spatially and temporally due to variability in precipitation, snowmelt, evapotranspiration, runoff, and the permeability of bedrock and alluvium underlying the root zone. Results averaged over the areas of subbasins were similar to results obtained from previous studies. However, estimates of recharge on the local scale of the HRUs indicate significant (greater than 100 percent) differences at some locations compared to results obtained using INFIL due to differences in (1) the geometry and scale of HRUs, (2) the layout of the cascading flow network and the location of stream channels, (3) the representation of the physical characteristics of the root zone, and (4) model processes controlling the simulation of evapotranspiration and the movement of water through the root zone.