A quantitative examination of lightning as a predictor of peak winds in tropical cyclones. Issue 9 (8th May 2015)

• Record Type: Journal Article
• Authors: Whittaker, Ian C.; Douma, Emma; Rodger, Craig J.; Marshall, Timothy J. C. H.
• Is Part Of: Journal of geophysical research. Volume 120:Issue 9(2015:May)
• Page Start: 3789
• Page End: 3801
• Publication Date: 2015-05-08
• Subjects: Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2014JD022868
• Languages: English
• ISSNs: 2169-897X

Abstract

<abstract abstract-type="main" id="jgrd52123-abs-0001"> <title>Abstract</title> <p id="jgrd52123-para-0001">We use the World Wide Lightning Location Network to investigate lightning strike variations in 8 years of categories 4 and 5 tropical cyclones. A cross‐correlation analysis is performed between the lightning and maximum sustained wind variations, giving lag and lead times related to the peak linear correlation for each tropical cyclone. A previous study of 58 cyclones by Price et al. (2009) is reexamined using the International Best Track Archive for Climate Stewardship database for the maximum sustained wind speeds of each tropical cyclone showing a moderate to strong correlation between lightning and wind variations. An 8 year data set of 144 tropical cyclones are analyzed in the same way, with a 10° square window, giving similar results to the smaller data set. Using a radial lightning collection window of < 500 km, we confirm the general results of previous studies that lightning can be used on a ∼1 day timescale to predict the evolution of the winds in tropical cyclones. Investigation of different lightning collection window sizes indicates that the lightning lead times are highly dependent upon the window size. Smaller collection windows have modal lightning lead times of ∼2.75 and 0 days, indicating that the lightning location inside the cyclone is as important as the total lightning variation. We have also performed a fixed time lag correlation which shows that preexisting knowledge of what time lag to use is needed in order to use this approach as a predictive tool.</p> </abstract>