Sunday, 22 January 2012
The Evaluation of Enhanced Lightning Forecast and Warning Tools to Enhance Decision Support Services
Hall E (New Orleans Convention Center )
Lightning is one of nature's most dangerous weather hazards. During thunderstorms, cloud-to-ground (CG) lightning can frequently strike the earth's surface to threaten the lives of those nearby. At times lightning may strike miles away from the parent storm, bringing with it unexpected danger in a mere flash. The threat is especially high in Florida. Causality statistics show that Floridians account for nearly 11% of all lightning-related deaths in the United States from 1959-2010, and that during the same period, Florida experienced more than twice as many lightning-related fatalities as any other state. The National Weather Service Forecast Office in Melbourne, Florida, has been a leader in lightning applied research for decades, most recently developing an experimental lightning watch and warning system to help address lightning safety through the provision of enhanced products and services. During the summer of 2011, the accuracy and practicability of daily lightning outlooks were evaluated through a comparison against similar lightning (e.g., thunderstorm) forecasts made by computer models, as well as situational airport lightning climatologies. Then, during actual lightning events, variables known to be indicators of potential CG lightning development were monitored, using radar imagery, radar-derived products, and information gathered by the Lightning Detection and Ranging (LDAR) array based at Cape Canaveral Air Force Station. If a forecaster decided that the threat became appreciable for defined locations, the experimental lightning watches and warnings were issued for a 5-nautical mile radius around the specified latitude/longitude point. Ideal lead times for the watches and warnings were 30 minutes and 15 minutes, respectively, before the first observed CG strike. Verification scores (probability of detection, false alarm rate, advisory verification rate, critical success index) were calculated to evaluate the usefulness of both watches and warnings. During the project period, several of the tools used to produce these products were also evaluated, including the AutoNowCaster (ANC), a fuzzy logic convective initiation model. Another tool utilized throughout the study was an ambitious configuration of the local Weather Research and Forecasting (WRF) model, producing 1-km resolution output, which represents one of the finest resolution models currently used in operations at a NWS office. Finally, challenges were made to effectively communicate lightning-related risks through various mediums, including the implementation of a new collaboration and briefing tool, the Geo-Targeting Alert System (GTAS). The overall results of this project suggest that lightning forecast and warning services can greatly decrease public risk, but with a few caveats. When performed in close collaboration with emergency officials, the utility of lightning alerting systems can be increased significantly.