Since at least 1955, meteorologists have periodically looked at whether there is a reliable correlation between tornadoes and lightning and whether lightning can be used to predict where tornadoes might occur.
Above is a video of lightning associated with the record tornado outbreak on April 27th. Here is the description from NOAA:
On April 25-28, 2011, a large outbreak of over 300 tornadoes wreaked havoc in much of the Southeast United States. It is listed by NOAA as the fourth deadliest outbreak in United States history, and as the year with the highest EF4 and EF5 tornadoes to date. (Read more about the events of April 27 here.) Due to intense updrafts inside tornado-forming supercell storms, the lightning that occurs within clouds that are high above the ground tends to increase and the cloud-to-ground lightning strikes tend to decrease before the tornado actually forms. Knowing that a storm has a rapid increase of in-cloud lightning activity is a warning sign of severe weather, such as tornadoes, and is used to predict the onset of such events. Currently, weather forecasters can make use of the SPoRT Lightning Mapping Array systems to measure lightning activity. SPoRT is also being used as a research and demonstration testbed for the future NOAA GOES-R satellite. When GOES-R launches in 2015, it will contain the Geostationary Lightning Mapper, a new instrument capable of monitoring total (in-cloud and cloud-to-ground) lightning over the entire U.S and adjacent oceans — a major advance over ground-based lightning detection systems that, like radar, are limited in coverage.
Imagine connecting the dots of a constellation of stars. The data visualized here are much like those stars, and represent 10,405,546 points that make up thousands of individual lightning bolts from the ground up to an altitude of 9 miles. Each frame of the animation shows all of the measured lightning points in a one minute interval, and the total time span of the animation is 24 hours, starting from 7:00 pm CDT on April 26th. Notice that the large clusters of data points are extremely dense and are located at higher altitudes. These points coincide with observations of almost no lightning ground strikes during severe weather. The storm bringing the largest set of tornadoes during Alabama’s deadliest tornado outbreak of all time begins around 40 seconds into the animation, including the Tuscaloosa/Birmingham EF5 and the Hackleburg/Madison EF5 tornadoes. A vertical rod marking Tuscaloosa has been added to provide a reference for the height of the lightning points.
I, too, have noticed that there is often a lack of cloud-to-ground lightning during major tornadoes such as during the 1991 Wichita-Andover tornado (one of the most photographed in history).
However, the Greensburg tornado had nearly continuous vivid lightning. We saw the “polarity” of the lightning flip from almost completely negative electrical charge to positive charge about the time the tornado touched down. Could the flip to positive charged-lightning be a signature of a major tornado?
So, I looked at the Joplin tornado, hoping to see a “signature” like Greensburg’s and … nothing. It stayed primarily negative the whole time.
My point is that, while I certainly support further research, there is little reason to be hopeful of a consistent lightning predictor for tornadoes in the near future.