Objective verification of high-resolution WRF forecasts during 2005 NSSL/SPC Spring Program

Forecast output from several versions of the WRF model will be evaluated during the 2005 NSSL/SPC Spring Program. These include a ~4km grid-spacing version of the WRF-NMM run at NCEP, a ~4km version of WRF-ARW run at NCAR, and a ~2km version of WRF-ARW run by CAPS at PSC. This output will be evaluated subjectively in real-time by teams of operational forecasters and researchers (see Kain et al. 2005, 21st WAF/17th NWP conf). In addition, several objective verification techniques will be applied to these forecasts, in particular, predicted and observed composite radar reflectivities will be compared. Object-oriented techniques will be used to compare the forecast and observed characteristics of a variety of rainfall systems. Other statistical techniques will be applied in order to measure biases in the forecast fields, structure, etc. This paper will report on ongoing research related to meaningful, objective verification of forecasts that contain realistic, high-resolution detail.