S8 NWS Tornado Warnings: A Regional Analysis

Sunday, 6 January 2019
Hall 4 (Phoenix Convention Center - West and North Buildings)
Ronald Kennedy Jr., University of Oklahoma, Norman, OK; and H. E. Brooks and J. Correia Jr.

Sergeant JP Finley used a 2 x 2 contingency table to forecast tornado occurrences. This table allowed a user to obtain probability of detection (POD), false alarm ratio (FAR), critical success index (CSI), Success ratio (SR), and the Bias. These metrics help evaluate the value and quality of a forecast. Recent studies of the National Weather Service tornado warning performance for the Central, Eastern, and Sothern regions of the United States from 1986 - 2017. A significant decrease in false alarm ratio (FAR) and probability of detection (POD) starting in 2012 for the Central region, but the Southern and Eastern region show no change in false alarm ratio (FAR) or probability of detection (POD) until the following year in 2013. Considering a warning forecaster has lots of information to consider before issuing a warning radar, spotter reports, environmental conditions, historical data sets, etc. These elements all play a role in the weight of evidence required to issue a warning or not. If there is enough weight of evidence to issue a warning that goes above the threshold set in place most times the forecaster will issue a warning. This research will investigate a regional analysis of National Weather Service tornado warning performance system with a focus on the decrease in FAR and POD starting in 2012 for the Central region.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner