Extended Abstract (256K)4.2
The Application of Climate Data Sets in Calibrating Ensemble Guidance for the Prediction of Hazardous Weather
David R. Bright, NOAA/NWS/NCEP/SPC, Norman, OK; and M. S. Wandishin, S. J. Weiss, R. S. Schneider, and J. T. Schaefer
The Storm Prediction Center (SPC) makes extensive use of ensemble forecast information in the construction of hazardous weather forecasts of thunderstorms and severe thunderstorms, freezing rain, heavy snow, and critical fire weather conditions. At the SPC, the ensemble probabilistic forecast is perhaps the most commonly utilized ensemble product. Yet despite improvements to ensemble systems, the envelope of possible solutions does not fully represent the complete range of all future plausible atmospheric states. This leads to probabilistic guidance that is not completely reliable, particularly when concerned with rare and extreme events.
Provided a correlation exists between the actual hazardous phenomena of interest and large-scale environmental predictors well resolved by the ensemble models, the application of climate data can be used to post-process and downscale ensemble probabilistic guidance thereby removing systematic biases and improving statistical reliability. The SPC applies a frequency correction technique to calibrate and downscale ensemble probabilistic guidance to aid in operational forecasting and real-time decision making. This post-processing relies on climatic databases of hazardous phenomena such as cloud-to-ground lightning, severe convective weather reports, and even winter road surface information.
This paper discusses the importance of maintaining real-time databases of hazardous weather phenomena across a wide range of operational programs to improve real-time ensemble guidance. The application of the SPC frequency correction technique to hazardous forecast problems is demonstrated and some verification results provided. Non-steady climatic data sets introduce additional challenges; trends in the long-term data are difficult to overcome when certain calibration techniques such as reforecasting are utilized. These challenges are also discussed.
Session 4, Prediction (Use of Climate Statistics in Forecasting)
Wednesday, 17 January 2007, 8:30 AM-11:45 AM, 206A
Previous paper Next paper
Recorded presentation