The Vaisala National Lightning Detection Network (NLDN) has been in operation since 1989. The Vaisala Long Range Lightning Detection Network (LLDN) is a much newer capability utilizing NLDN sensors and Canadian Lightning Detection Network (CLDN) sensors to detect cloud-to-ground lightning flashes up to two thousand kilometers off the coasts of North America. In addition, Vaisala's LLDN includes several long range sensors located in the North Pacific Ocean. The National Weather Service receives both datasets (NLDN and LLDN) in real time and disseminates them to its local weather forecast offices (WFOs). Both datasets are also available to the National Centers for Environmental Prediction (NCEP), of which the Tropical Prediction Center/National Hurricane Center (TPC/NHC) is a part. The LLDN is of particular interest to TPC/NHC since it provides some detection capability as far offshore in the Atlantic basin as the eastern Caribbean Sea, and well into the eastern Pacific off the coast of Mexico. TPC/NHC began processing and displaying the NALDN and LLDN data in its operational environment in April 2007. TPC/NHC tropical cyclone and marine forecasters have been evaluating the data since then, making great use of the ability to overlay the lightning data on conventional geostationary satellite imagery.

During the past couple of years, TPC/NHC and Vaisala, Inc. have preliminarily examined how these data might be beneficial for operational hurricane analysis at TPC/NHC. We examined the overall amount of lightning activity in several tropical storms and hurricanes during 2004-2006, including Katrina and Rita. Nearly every cyclone we examined produced some lightning in the inner core at some point in its life cycle, but Katrina and Rita appear to have been the most prolific lightning producers we examined. Overall, lightning outbreaks seemed to occur when inner-core organization was significantly changing in some way. For example, abundant inner-core lightning was often observed in strong tropical storms just before they became hurricanes, and in major hurricanes just before the inner eyewall collapsed (with much less lightning thereafter). Based on these preliminary results and the work of other researchers, the lightning data seem potentially useful in operationally diagnosing some aspects of current hurricane structure, especially when the cyclone is outside the range of land-based radar and is not being investigated by aircraft reconnaissance. Since the lightning data are now displayable on TPC/NHC operational workstations, we are evaluating the utility of more precisely knowing the spatial and temporal distribution of the lightning strikes. Future work will also examine how much, if any, short-term predictive value there might be in the data for hurricane intensity.

TPC/NHC is already making routine use of the lightning data in our Tropical Analysis and Forecast Branch (TAFB), where marine forecast operations function 24 hours per day, every day of the year. Lightning data have been utilized in the preparation of several operational TAFB products, including surface analyses, tropical weather discussions, and offshore waters and high seas forecasts. The data enable TAFB forecasters to more accurately differentiate between showers and thunderstorms, and between areas of true deep convection and areas of cirrus on geostationary satellite imagery. Our extended abstract will describe specific cases in which the LLDN data provided TAFB meteorologists with greater confidence in their analyses and/or forecasts.

Challenges for TPC/NHC in using the LLDN data include low detection efficiency (especially during the daytime), potentially mis-located events, and little or no coverage over large portions of our areas of responsibility that are far from North America. In addition, due to sensor location limitations, some geographic areas (e.g., south of Florida) are prone to false and/or linear returns. Planned LLDN expansion and planned improvements in data quality by Vaisala will help mitigate TPC/NHC challenges related to geographic coverage, low detection efficiency, and potentially mis-located events (false and/or linear returns).