|
We encourage your correspondence, and although we cannot print all that we receive, we will include at least one short, perhaps edited, letter per issue. Dear WeatherZineAs usual Chuck Doswell's (August 1999 WeatherZine Guest Editorial) comments are provocative. I agree with him about the lead time issue. We all need to come to some sort of consensus about the optimal length of a warning. There are examples of 45 minute warnings where people have come out of their shelters after 30 minutes only to get whacked by a tornado. I take issue with Chuck's assertion about what it will take to reduce the false alarm rate (FAR) and increase the probability of detection (PODs) (and his characterization of the NWS goals). I think everyone should take a look at what happened at Sterling, VA in 1998 with the test of the new decision support system called "System for Convective Analysis and Nowcasting" (SCAN) which combines the NSSL Warning Decision Support System (WDSS) and the NCAR Autonowcaster into one system on AWIPS. This decision software PLUS TRAINING (I agree with Chuck on this point) allowed the forecasters in 1998 to increase the POD from .65 to greater than .80 and reduce the FAR from above .50 to below .30 for all severe weather warnings (and increase the lead time too). I am sure that Steve Smith (from TDL) and Steve Zubrick (from Sterling) would be willing to engage Chuck in a healthy debate as to what contributed to this success that gives hope for the NWS meeting its new strategic goals. My view is that the information inherent within the new observing systems and models is not accessed or utilized very effectively; and that the systems inherent within SCAN allow the forecasters to quickly access the data and associated information to make better and more informed decisions. — Louis Uccellini, NOAA/NCEP Comments? thunder@ucar.edu WeatherZine #18 Home Page
| Comments and Feedback
| Site Map |
