The Performance of HIGH-RESOLUTION WRF Ensemble QPF During Heavy Winter Precipitation Events In California

Abstract

A series of annual winter forecast exercises over vulnerable watersheds of California and Nevada have been undertaken as part of the NOAA Hydrometeorology Testbed (HMT) to assess the potential for improved quantitative precipitation forecasts during heavy precipitation events. For the most recent exercise, a nine-member high-resolution (9 km) ensemble WRF forecast system produced forecasts during a week-long period of heavy and persistent rainfall in Northern California. An online verification system funded by the USWRP and implemented at the Developmental Testbed Center (DTC) concurrently provided quantitative assessment of the performance of the individual ensemble members, the simple ensemble mean, and (as baseline) a GFS deterministic run with nominal resolution of 40 km. The quantitative precipitation forecasts are evaluated in two ways. First, qualitative comparison of several different traditional verification scores are presented for selected episodes to illustrate the several models' individual characteristics in different meteorological scenarios. Second, the verification scores are aggregated over a several-week period to describe overall performance and to provide a reasonable basis for comparison of the WRF ensemble with GFS forecasts. For each set of results, attention is paid to the impact and implications of different spatial resolution to the verification results. The increased ability of the higher-resolution WRF forecasts to capture extreme precipitation events is among the conclusions drawn in this paper.