Improved timeliness of thunderstorm detection from mapping a larger fraction of lightning flashes

One application of lightning ground strike mapping systems has been thunderstorm detection. However, the climatological ratio of in-cloud flashes to cloud-to-ground flashes typically is greater than 2:1. Thus, systems that map either cloud flashes or all types of flashes will detect storms more quickly and reliably. The improvement typically is even greater than would be expected simply from the greater number of samples, because the first flashes produced by a storm usually are cloud flashes. However, the improvement obviously is affected by the fraction of flashes that a lightning mapping system detects. Because it costs substantially more to detect a larger fraction of all flashes, one would like to know how much lead time will be added by various levels of lightning detection. The U.S. National Lightning Detection Network (NLDN) being used by the National Weather Service is now capable of detecting roughly 10-20% of cloud flashes, in addition to a much larger fraction of cloud-to-ground flashes, over the contiguous United States. So far, this cloud flash option has been turned on only in a test region. The optical lightning mapper being planned for GOES-R is expected to detect 80-90% of cloud flashes. Some research mapping systems can detect essentially all but the smallest flashes throughout their coverage region. We analyzed the test cloud flash data from the NLDN network to see how much that system's cloud flash detection would improve thunderstorm detection. Furthermore, we analyzed data from VHF lightning mapping systems that detect almost all flashes, to evaluate how much the timeliness of thunderstorm detection can be improved over what is now achieved with ground strike mapping systems. In north Texas and Oklahoma, 50% of thunderstorms produced a cloud-to-ground flash within 5-10 minutes of their first cloud flash, but in roughly 10% of storms, no cloud-to-ground flash occurred within an hour of the first cloud flash. Behavior was much different in storms over the High Plains of northwest Kansas and northeast Colorado. There it required 30 minutes after the first cloud flash for 50% of storms to produce a cloud-to-ground flash, and 20% of storms produced no cloud-to-ground flash within their first hour of lightning activity. One might expect such a result on the basis of climatological studies showing that cloud flashes comprise at least 90% of all lightning over much of the High Plains and in a few other regions of the country.