The detailed spatial distribution of precipitation in the Sierra Nevada in California is only given inferentially by sparse observations or rather coarse model output, but is crucially important for water supply and flood management. Forecasts of precipitation for various lead times are available from global and regional climate models; however, their spatial resolutions are generally much larger than watershed scale (e.g.,5-10kms) hydrological models. To estimate the spatial distribution of precipitation at finer scales, we have implemented a simplified 2-D orographic precipitation model(following concepts originally developed by Owen Rhea of the Sacramento River Forecast Center). This precipitation model was tested using a 5-km rendition of terrain over the Sierra Nevada, by driving it with 2.5x2.5 degrees gridded profiles of temperature, wind speed, and humidity from NOAA/NCEP daily reanalyses (1973-1996). This downscaled precipitation then was used to drive a distributed hydrological model(PRMS) of the Merced River, in the southern, high-altitude part of the Sierra Nevada. Comparisons of downscaled precipitation to observations indicate that the model performs well during large precipitation events, and generally replicates observed daily and interannual fluctuations of precipitation at representative sites. Comprehensive comparison of simulated streamflow to observed flows and flows simulated with observed precipitation indicates that this model is effective in downscaling climate model simulations and forecasts or sparse data observations to watershed scales.