May 26 - July 31
A Comparison of Lightning Flashes Observed by Ground-Based Detection Networks and Video
James Coy Jr. and Kristin Calhoun
What is already known:
- The National Lightning Detection Network (NLDN), Earth Networks Total Lightning Network (ENTLN), and the Oklahoma Lightning Mapping Array (OK-LMA) uniquely recorded lightning flashes during the 31 May 2013 El Reno supercell.
- The El Reno Survey Project (El-Reno-Survey.net) was created to crowd-source the abundance of storm chaser video from this event and provide open-access to the scientific community of the data.
- An initial comparison of CG lightning flashes captured on the storm chaser videos with flash data from the ground-based networks revealed a disagreement on the total number of flashes recorded during the same period.
What this study adds:
- Data recorded by the networks can be plotted on Google Earth to create a 3-dimensional interactive map for each CG flash.
- The NLDN and ENTLN show different results when a 15 kA peak current filter is applied to the cloud-to-ground data, though the 15 kA does successfully remove a majority of the misclassified cloud flashes for both networks.
- Charge analysis of individual flashes from the OKLMA reveals that the ENTLN has a major flaw in determining the polarity of CG flashes.
Abstract:
Cloud-to-ground (CG) lightning flashes recorded by both the National Lightning Detection Network (NLDN) and Earth Networks Total Lightning Network (ENTLN) are compared with three- dimensional lightning mapping observations from the Oklahoma Lightning Mapping Array (OK-LMA) and storm chaser video recorded of the 31 May 2013 El Reno tornadic supercell. The El Reno Survey Project (El-Reno-Survey.net) was created to crowd-source the abundance of storm chaser video from this event and provide open-access to the scientific community of the data. An initial comparison of CG lightning flashes captured on these videos with CG data from NLDN revealed a disagreement on the total number of flashes, with NLDN recording many negative CG flashes at lower peak amplitude not apparent in any of the videos. For this study, the area of the comparison was expanded and data from both the ENTLN and LMA were used to compare the observations from each network in terms of timestamp, location detection, peak current, and polarity of each flash in the period 2230-2330 UTC. Initial results from the NLDN and ENLTN indicated a negative CG dominance, but, after a 15 kA peak current filter was applied, the NLDN indicated primarily positive CG polarity flashes while ENTLN still indicated primarily negative CG polarity. The average distance between the two networks for the same flash was more than 2 km and improved to approximately 1 km after the 15 kA filter was applied. When compared to video and the OK- LMA, both the NLDN and ENTLN had misclassifications of in-cloud (IC) lightning as CG flashes. Additionally, the charge analysis of OK-LMA revealed the NLDN-determined polarity as correct each time the NLDN and ENTLN disagreed. It is concluded that there is a major flaw in the ENTLN's ability to determine the polarity of CG flashes despite having roughly similar peak current magnitudes for most CG flash occurrences.