Estimating water budgets of river basins in the western U.S. is a challenge because of the effects of complex terrain and lack of comprehensive observational datasets. This study aims at understanding the uncertainty in estimating cold season water budgets of the Columbia River (CRB) and Sacramento-San Joaquin (SSJ) River basins. An intercomparison was performed based on the NCEP/NCAR Reanalysis I (NRA1), NCEP/DOE Reanalysis II (NRA2), ECMWF reanalyses (ERA), regional climate simulations produced by the Penn State/NCAR Mesoscale Model (MM5) and NCEP Regional Spectral Model (RSM) driven by the reanalyses, and two precipitation datasets gridded at 2.5 and 1/8 degree for seven years between 1986 and 1993 to study the effects of spatial resolutions, model configurations and parameterizations, and large-scale conditions on basin-scale water budgets.

Results showed that overall, the regional simulations were superior in terms of simulating the spatial distributions of mean precipitation and precipitation anomalies compared to the global reanalyses. However, cold season precipitation was generally amplified through downscaling using the regional models such that basin mean precipitations were typically higher than the observed, while the opposite was true for the reanalyses. The amplification was the largest in the RSM simulation driven by NRA2, which showed the biggest difference between the large-scale and regional-scale basin mean precipitations. ERA and the MM5 simulation driven by ERA provided the best basin mean precipitation estimates when compared to the 1/8^o observational dataset.

This study shows that large uncertainties remain in estimating the water budgets of western river basins such as CRB and SSJ. In terms of atmospheric moisture transport, there was a 15-20% uncertainty in the global reanalyses. In terms of basin mean precipitation, differences among the reanalyses, regional simulations, and observations were as large as 100% of the mean. There were large differences in spatial distribution of precipitation between the RSM and MM5 simulations because of terrain representations and other factors. Runoff and snowpack were among the surface water budget components that showed the most sensitivity to model differences in spatial resolutions, physics parameterizations, and model representations. Better simulations of basin mean precipitation did not imply more superior simulations of runoff or snowpack.