Polygonology: A Realistic Look at the National Weather Service Polygon Warning System

Discussion of the National Weather Service's polygon warning system has been ongoing since it was introduced in October, 2007. As originally conceived, polygon warnings were intended to spatially enclose the hazards accompanying severe weather without regard to county or other geopolitical boundaries. In fact, statistics and visual evidence confirm that many storm-based warnings are issued from this perspective. However, NOAA Weather Radio All Hazards and most other warning dissemination systems remain county-based. This, coupled with the fact that polygons may have large dimensions in either space or time, can lead to the reality or perception of over-warning. Recent social science research indicates that the public's perception of false alarms is much greater than the actual number of warnings issued, regardless of event verification.

This presentation will introduce “polygonology” as an attempt to further the discussion by suggesting best practices and providing simple strategies to the warning forecast process. Adding polygonology to the warning decision process will allow forecasters to reduce the false alarm area, without reducing the more important Probability of Detection and lead time. The practice of not extending warnings into downstream counties until the event confidence is sufficiently high is one tactic that will immediately reduce the polygon warning area. This presentation will also show example polygons that, had they been drawn slightly differently, would have substantially reduced the false alarm area, resulting in better customer service and a clearer, more precise warning. The strategies and best practices that will be presented are intended to reduce the public's perception of false alarms, thereby increasing the overall confidence in the warning decisions.