Probabilistic QPF from a realtime multi-model storm-scale ensemble forecast system

摘要

@@ 1. Introduction Producing skillful and reliable probabilistic QPF is a challenging task, especially when high-resolution NWP is involved. In support of NOAA Hazardous Weather Testbed (HWT) Spring Experiment, in the past four years since 2007 the Center for Analysis and Prediction of Storms (CAPS), in collaboration with the Storm Prediction Center (SPC) and the National Severe Storm Laboratory (NSSL) and funded by the NOAA CSTAR program, had conducted unprecedented and highly successful real-time storm-scale ensemble forecast (SSEF) experiment (Xue et al. 2007, 2008, 2009; Kong et al. 2007, 2008, 2009). In 2010 spring, under the new three-year CSTAR grant and with newly added collaborations with DTC, AWS, and HPC, the CAPS storm-scale ensemble forecasting system (SSEF) has been brought to a new level. Major changes in SSEF for the 2010 HWT Spring Experiment include: The forecast domain has been expanded to cover the entire continental United States (Figure 1), increasing total computing grid points by ~40% compared to the 2009 domain; The total number of ensemble members is increased to 26, consisting of two WRF dynamical cores (ARW and NMM) and the Advanced Regional Prediction System (ARPS); New double-moment microphysics schemes coming with WRF V3.1 are included in some ARW members; A large set of ensemble post-processed products from a 15-member sub-ensemble that consists of multi-model, IC and LBC perturbation, and radar analysis members are generated and made available near real-time to the HWT participants.