Application of Lightning Measurements to Tropical Cyclone Rapid Intensity Change Forecasting

The next generation NOAA geostationary satellite system starting with GOES-R will include a geostationary lightning mapper (GLM). The GLM will provide nearly continuous measurements of total lightning activity with a location accuracy of about 10 km. A number of studies with ground-based lightning detection networks suggest that lightning activity in tropical cyclones is correlated with subsequent intensity changes. In this study ground-based lightning measurements from the World Wide Lightning Locator Network (WWLLN) are used in an experimental algorithm for prediction of rapid intensity changes of tropical cyclones. Although the WWLLN detection rate is somewhat low and primarily measures cloud to ground strikes, the coverage is fairly uniform over the global tropics making it well suited for tropical cyclone lightning studies. An updated detection algorithm developed by the University of Washington was applied to the WWLLN beginning in 2005, which provides a 5 year sample (2005-2009) for algorithm development. The lightning density is calculated in a storm-relative cylindrical coordinate system using data composited over six hour intervals. This data is available for the full life cycle of all tropical cyclones that occurred in the Atlantic, eastern and central North Pacific basins. A correction factor is applied to the lightning density to account for the somewhat low detection rate of the WWLLN data and to approximate the total lightning. The correction is a constant multiplicative factor that makes the annual average lightning density from the WWLLN data over large Atlantic and eastern/central Pacific domains equal to that from the Optical Transient Detector/Lighting Imaging Sensor climatology. Separate correction factors are applied for each year and for the Atlantic and eastern/central Pacific basins.

Tropical cyclone rapid intensification (RI) is defined as an increase in the maximum sustained surface winds of 30 kt or more in 24 h. Deterministic tropical cyclone prediction models have not provided very reliable forecasts of rapid intensification. However, the statistically-based Rapid Intensity Index (RII) has shown some skill in the prediction of RI. The RII is run operationally for all storms in the National Hurricane Center (NHC) and Central Pacific Hurricane Center (CPHC) areas of responsibility and uses a discriminant analysis technique to provide a probability of RI. Input includes eight variables determined from the NCEP Global Forecast System (GFS) model, GOES infrared brightness temperatures, Reynolds sea surface temperature analyses and oceanic heat content (OHC) estimates from satellite altimetry. An experimental version of the RII that also include predictors from the storm-relative lightning density analyses will be compared with the operational version. Two versions of the experimental RII will be evaluated, one which uses a discriminant analysis technique similar to the operational version, and the other which uses a neural network approach.

The experimental RII with the lightning input will be tested in near real time during the 2010 hurricane season as part of the GOES-R Proving Ground. For the real time version the lightning density will be calculated from the Vaisala GLD360 data.