P2.1 A closer look at verification statistics for two tornado outbreaks in 2006

Monday, 6 November 2006
Pre-Convene Space (Adam's Mark Hotel)
John T. Ferree, NOAA/NWS, Norman, OK; and D. McCarthy and H. E. Brooks

A major outbreak of tornadoes occurred from the Central Plains through the mid-Mississippi Valley on Sunday March 12, 2006. Another outbreak of tornadoes struck the mid-Mississippi Valley region on Sunday April 2, 2006. Both outbreaks were well anticipated, and the watch/warning services provided were described as exemplary. The authors noted that even though warning performance by NOAA's National Weather Service was viewed as excellent by emergency managers and the media, preliminary calculations of the False Alarm Rate (FAR) exceeded established goals.

A closer look at these two major outbreaks showed that many tornadic storms traversed large areas. Tornadoes occurred in counties upstream and/or downstream from falsely warned counties. Often times these warnings gave the public in the impacted downstream counties even greater advanced warning. Thorough post storm analysis by Science and Operations Officers indicate that most of the falsely warned counties fall into the category of strong indications of a tornado, and given the same conditions these experts would issue a tornado warning the next time. Additional, “false alarms” can also be attributed to the county-based verification method. If a warning covers portions of four counties, and the tornado only touches down in one of the four warned counties, three “false alarms” are counted. The False Alarm Rate (FAR) is impacted by several factors that do not relate to warning performance.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner