This report describes the construction, calibration, evaluation, and results of a steady-state numerical groundwater flow model of the Great Basin carbonate and alluvial aquifer system that was developed as part of the U.S. Geological Survey National Water Census Initiative to evaluate the nation’s groundwater availability. The study area spans 110,000 square miles across five states. The numerical model uses MODFLOW-2005, and incorporates and tests complex hydrogeologic and hydrologic elements of a conceptual understanding of an interconnected groundwater system throughout the region, including mountains, basins, consolidated rocks, and basin fill. The level of discretization in this model has not been previously available throughout the study area.

Observations used to calibrate the model are those of water levels and discharge to evapotranspiration, springs, rivers, and lakes. Composite scaled sensitivities indicate the simulated values of discharge to springs, rivers, and lakes provide as much information about model parameters as do simulated water-level values. The model has 176 parameters and little parameter correlation. The simulated equivalents to observations provide enough information to constrain most parameters to smaller ranges than the conceptual constraints, and most parameter values are within reasonable ranges.

Model fit to observations, comparison of simulated to conceptual water-level contours, and comparison of simulated to conceptual water budgets indicate this model provides a reasonable representation of the regional groundwater system. Eighty-six percent of the simulated values of water levels in wells are within 119 feet (one standard deviation of the error) of the observed values. Ninety percent of the simulated discharges are within 30 percent of the observed values. Total simulated recharge in the study area is within 10 percent of the conceptual amount; total simulated discharge is the same as conceptual discharge. Comparison of simulated hydraulic heads with the conceptual potentiometric surface indicates that the model accurately depicts major features of the hydraulic-head distribution. The incorporation of new recharge estimates and of mountain springs and streams as model observations creates higher simulated recharge mounds under many mountain ranges and highlights that in many cases, the regional flow paths go around, not through (or under) mountain ranges. Results from the model show that much of the flow in the groundwater system occurs in deeper layers, even though about 86 percent of the discharge occurs in layer 1. Over 95 percent of the recharge moves down from layer 1, and about 25 percent moves down to layer 8.

The model was used to delineate six simulated groundwater flow regions that connect recharge areas to discharge areas. The eastern Great Salt Lake and Great Salt Lake Desert model regions contain 75 percent of the groundwater budget, but only 42 percent of the study area. In contrast, the more southern Death Valley and Colorado model regions contain only 12 percent of the groundwater budget, but 37 percent of the study area.

Examples of potential use of the model to investigate the groundwater system include (1) the effects of different recharge, (2) different interpretations of the extent or offset of long faults or fault zones, and (3) different conceptual models of the spatial variation of hydraulic properties. The model can also be used to examine the ultimate effects of groundwater withdrawals on a regional scale, to provide boundary conditions for local-scale models, and to guide data collection.