113
Ensemble based probabilistic forecast and its verification

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Tuesday, 25 January 2011
Ensemble based probabilistic forecast and its verification
Yuejian Zhu, NOAA/NWS/NCEP/EMC, Camp Springs, MD

Ensemble based probabilistic forecast product has been generated operationally since 2004 in NCEP which include PQPF (probabilistic quantitative precipitation forecast) and RMOP (relative measure of predictability). From December 2007, NCEP global ensemble forecast system (GEFS) and North American Ensemble Forecast System (NAEFS) started producing a set of new probabilistic forecasts. The new product is from bias corrected and/or downscaled ensemble forecast. It includes ensemble 10%, 50% (medium), 90%, mean, mode and spread forecasts for about 50 variables which include near surface variables, such as 2-m (maximum/minimum) temperature, 10-m winds and surface pressure. How can we evaluate these probabilistic forecasts? For example, how can we verify 10% probabilistic forecast for 2-m temperature? A common probabilistic verification method may not be able to verify this very well. Therefore, this presentation will discuss a method to verify 10% or 90% ensemble forecasts, and demonstrate a useful value from NAEFS post process.