Ensemble prediction systems and forecasts of extreme events


Gregory Byrd, William Bua, and Timothy C. Spangler


Boulder, Colorado USA


In natural disaster preparedness, the longer the lead time before an extreme event occurs, the greater the opportunity to mitigate damage and loss of life. One tool that has become important in assessing the risk of occurrence for extreme weather events is the ensemble prediction system (EPS), which uses uncertainty in numerical weather forecasts and the initial conditions of the atmosphere to make multiple individual forecasts for a forecast. The subsequent EPS forecast can be used to quantify the probability of an extreme weather event, based on the percentage of forecasts in the EPS that predict the event.


To help the forecaster learn how to intelligently use EPSs in the forecast process, the Co-Operative Program for Meteorological Education and Training (COMET) has developed several training aids, which will be reviewed. These may include:


         A webcast (Introduction to Ensemble Prediction, found at http://meted.ucar.edu/nwp/pcu1/ensemble_webcast/) that provides an audiovisual introduction to the scientific concepts and forecast tools used in ensemble forecasts, along with warm and cold season case examples.

         A web-based module (Ensemble Forecasting Explained, found at http://meted.ucar.edu/nwp/pcu1/ensemble/) with an introductory section explaining ensemble prediction concepts, and subsequent sections discussing ensemble forecasts generation, important statistical concepts, tools used to summarize the ensemble forecast data and its verification, and finally some case applications.

         A table or matrix of information on current ensemble prediction systems available through the U.S. National Centers for Environmental Prediction (NCEP) that can be found in the EPS matrix (http://meted.ucar.edu/nwp/pcu2/ens_matrix).


We will apply the training information to the forecast of some extreme events that occurred over the past year, including a land-falling tropical cyclone and a heavy snow event.