A Seasonal Precipitation and Stream Flow Hindcast and Prediction Study in the Western United States during the 1997/98 Winter Season Using a Dynamic Downscaling System

Abstract

The authors present a seasonal hindcast and prediction of precipitation in the western United States and stream flow in a northern California coastal basin for December 1997?February 1998 (DJF) using the Regional Climate System Model (RCSM). In the seasonal hindcast simulation, in which the twice-daily National Centers for Environmental Prediction?National Center for Atmospheric Research reanalysis was used for the initial conditions and time-dependent boundary forcing, RCSM has simulated realistically the temporal and spatial variations of precipitation in California and stream flow in a northern California coastal basin. For the headwater basin of the Russian River in the northern California Coast Ranges, the Topography-Based Hydrologic Model (TOPMODEL) forced by observed daily precipitation resulted in a correlation coefficient of 0.88 between observed and simulated DJF stream flow. In the coupled stream flow hindcast, the authors obtained a correlation coefficient of 0.7 between simulated and observed stream flow for the same period. The coupled hindcast has generally overestimated (underestimated) low (high) flow events in the basin. Errors in the simulated stream flow were due mostly to the errors in the simulated precipitation. A seasonal hydroclimate prediction experiment, in which RCSM was nested within the global forecast data from the University of California, Los Angeles, GCM, has predicted well the season-total precipitation in the western United States. Temporal variations of predicted precipitation were affected strongly by the predictability of the general circulation model. The predicted DJF-total snowfall agrees well with the snowfall simulated in the hindcast, especially in the central Cascades and the Sierra Nevada, where snowfall was the heaviest.