Premium
An improved lightning flash rate parameterization developed from Colorado DC3 thunderstorm data for use in cloud‐resolving chemical transport models
Author(s) -
Basarab B. M.,
Rutledge S. A.,
Fuchs B. R.
Publication year - 2015
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1002/2015jd023470
Subject(s) - thunderstorm , flash (photography) , storm , meteorology , lightning (connector) , environmental science , radar , volume (thermodynamics) , convective storm detection , remote sensing , computer science , physics , geography , optics , telecommunications , power (physics) , quantum mechanics
Abstract Accurate prediction of total lightning flash rate in thunderstorms is important to improve estimates of nitrogen oxides (NO x ) produced by lightning (LNO x ) from the storm scale to the global scale. In this study, flash rate parameterization schemes from the literature are evaluated against observed total flash rates for a sample of 11 Colorado thunderstorms, including nine storms from the Deep Convective Clouds and Chemistry (DC3) experiment in May‐June 2012. Observed flash rates were determined using an automated algorithm that clusters very high frequency radiation sources emitted by electrical breakdown in clouds and detected by the northern Colorado lightning mapping array. Existing schemes were found to inadequately predict flash rates and were updated based on observed relationships between flash rate and simple storm parameters, yielding significant improvement. The most successful updated scheme predicts flash rate based on the radar‐derived mixed‐phase 35 dB Z echo volume. Parameterizations based on metrics for updraft intensity were also updated but were found to be less reliable predictors of flash rate for this sample of storms. The 35 dB Z volume scheme was tested on a data set containing radar reflectivity volume information for thousands of isolated convective cells in different regions of the U.S. This scheme predicted flash rates to within 5.8% of observed flash rates on average. These results encourage the application of this scheme to larger radar data sets and its possible implementation into cloud‐resolving models.