Abstract :
The Scripps Experimental Climate Prediction Center (ECPC) has been making routine experimental, near real-time global and regional forecasts since Sept. 27, 1997 with the National Centers for Environmental Prediction (NCEP) global spectral model (GSM) and the corresponding regional spectral model (RSM), which is based on the GSM, but which can develop higher resolution simulations and forecasts for limited regions. Images of these forecasts, at daily to seasonal time scales, are provided on the ECPC World Wide Web site () and digital forecast products are provided on the ECPC anonymous ftp site to interested collaborators. The forecast skill of the GSM for the first two years was previously described in several papers. The purpose of this paper is to further compare two additional years of GSM and the RSM precipitation forecasts over the US, with special attention to various biases and errors, as well as the significant forecast skill at weekly to seasonal time scales. While these experimental forecast results suggest that there is precipitation forecast skill, out to at least a season, to be realized from both global and regional dynamical models, it was disappointing that the RSM did not seem to provide markedly more skillful precipitation forecasts than the GSM. Excessive forecast variance is a major problem with the current RSM.
Keywords :
climate , Forecasting , Seasonal forecasting , Precipitation