Tuesday, 14 January 2020: 1:45 PM
258B (Boston Convention and Exhibition Center)
During the landfall of Hurricane Irma (Category 4; 2017) in the Florida Keys and the subsequent inland advance within the southwestern Florida mainland, the peninsula’s Atlantic coast was raked by numerous landfalling tornadoes contained within outer rain bands. Ninety-one tornado warnings were issued in the midst of this extreme tropical cyclone event which added considerable public angst and overwhelmed alerting systems. Fifty-one of these storm-based warnings affected east central Florida, with nine of the ten verified tornadoes yielding short and narrow damage swaths of EF0 to EF1 intensity. Post event interactions with County Emergency Management (Florida EM; Region 5) uncovered two customer-driven inquiries which serve as the foundation of this work. To reduce the number of alerts, the EMs first wanted to know whether it was possible to issue warnings only when the threat to life was (potentially) greater than the threat already being imposed by the greater hurricane wind field. To investigate, radar signatures of Irma’s suspect storm cells were assessed in simulated real-time to determine whether statistical tools could be leveraged. Methods described in Cohen et al. (2018) were used to gauge tornado probability, and with intent to differentiate potential tornado wind speeds. Associated scoring will be presented. The second inquiry was related to the length and substance of tornado calls-to-action rendered in hurricane situations which can be confusing. Interactive discussions were held among the EM group seeking consensus recommendation. Here, proposed safety messaging is made more deliberate according to the phase and intensity of the hurricane wind event. Sample products will be shown that are more concise and informative, further reducing the overall volume of information.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner