College of Engineering | Apply to CSU | Disclaimer | Equal Opportunity Statement | Privacy | Search CSU

Prediction of Total Lightning Behavior in Colorado and Alabama Thunderstorms From Storm Dynamical and Microphysical Variables

You must be on the CSU network—either physically or using VPN—to watch this or any of the videos on this site.

November 3, 2014
Brett Basarab
Hosted by Steve Rutledge (advisor), Sonia Kreidenweis, Wiebke Deierling (Affiliate), Steve Reising (Electrical and Computer Engineering)

Abstract

Thunderstorms impact their environment in a variety of ways, including the production of nitrogen oxides (NOx) by lightning (LNOx). Accurate prediction of total lightning flash rate in thunderstorms is important to improve estimates of LNOx from the storm scale to the global scale. New flash rate parameterization schemes have been developed based on observed relationships between lightning flash rate and storm parameters for Colorado thunderstorms during the Deep Convective Clouds and Chemistry (DC3) field experiment. Storm total flash rates are determined using an automated flash counting algorithm that clusters very high frequency (VHF) radiation sources emitted by electrical breakdown in clouds and detected by lightning mapping arrays (LMA). Storm parameters, including hydrometeor echo volumes and ice masses, are calculated from polarimetric radar variables retrieved by the CSU-CHILL and National Weather Service WSR-88DP radars. Measurements of updraft strength are obtained by synthesizing radial velocity retrievals from the CSU-CHILL and CSU-Pawnee radars to determine three-dimensional wind fields.

Bulk storm parameters including the graupel echo volume, 30-dBZ volume, and precipitating ice mass are found to be robustly correlated to flash rate (R2 ~ 0.8). It is shown that simple flash rate parameterization schemes based on these quantities predict gross flash rate behavior reasonably well. However, these flash rate schemes sometimes do not predict small fluctuations in flash rate, possibly due to fluctuations in storm updraft intensity. Updraft intensity-based flash schemes are therefore developed, but updraft parameters were more weakly correlated to flash rate than storm volume quantities. The use of multiple storm parameters to predict flash rate is also investigated, since flash rate may be sensitive to multiple processes or characteristics within thunderstorms. A simple approach is found to be most effective: storm-total graupel and dBZ volumes were split up into representative area and height dimensions and multiply regressed against flash rate. The combined quantities predict flash rate variability somewhat better than simpler single-parameter flash schemes. All new flash rate schemes are tested against observations of Alabama thunderstorms documented during DC3 to examine their potential regional limitations. The flash rate schemes developed appear most applicable to strong storms with sustained high flash rates occurring in a strongly continental environment. Finally, relationships between total flash rate and flash size are discussed, with implications for the improved prediction of LNOx.