Earth and Atmospheric Sciences, Department of


Date of this Version

Summer 8-2014


Obermeier, H.M., 2014: Verification and analysis of impact-based tornado warnings in the Central Region of the National Weather Service. Master's Thesis, University of Nebraska-Lincoln, Lincoln, NE, 92 pp.


A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Earth and Atmospheric Sciences, Under the Supervision of Professor Mark R. Anderson. Lincoln, Nebraska: August, 2014

Copyright (c) 2014 Holly B. Obermeier


Tornado warnings are one of the most critical products issued by the National Weather Service (NWS), and favorable verification statistics are desirable. The 2011 NWS statistics for traditional tornado warnings indicate that the probability of detection (POD) is 70%, while the false alarm rate (FAR) is 76%. The recent Joplin, Missouri EF5 tornado event on 22 May 2011, which resulted in massive devastation and loss of life, prompted the NWS to re-evaluate the current tornado warning format. After the Joplin, MO event, the Central Region of the NWS implemented the impact-based tornado warning (IBTW) experiment in 2013. IBTWs consist of tiers including damage tags and impact wording which convey increasing levels of damage. The damage wording within an IBTW is shown to relate to the Enhanced Fujita (EF) Scale. Wording included in non-tagged IBTWs corresponds to EF0-EF2 tornado damage, while the damage wording for tagged IBTWs corresponds to EF3-EF5 tornado damage. This study investigates the accuracy of IBTWs by examining if a tornado occurs during the warning time frame, and whether the resulting damage matches the damage wording in the IBTW. All IBTWs from 1 April 2013 through 30 November 2013 are collected, as well as tornado survey information, including EF Scale intensity, for every tornado which occurred in the Central Region during the same time period. Using these survey data, IBTWs are verified by the intensity of the tornado, if one occurs. POD and FAR statistics are calculated through 2x2 contingency tables for both non-tagged and tagged IBTWs. Results indicate that the majority of both non-tagged and tagged IBTWs are false alarms, and tagged IBTWs have a very low POD. Case studies of several events explore successful and unsuccessful implementation of damage tags, revealing that limitations in current technology and scientific knowledge may contribute to false alarms and missed detections. These findings suggest that more advances in technology and the understanding of tornadogenesis are necessary for more successful implementation of IBTWs.

Adviser: Mark R. Anderson