by Jeff Trapp

In a previous issue of NSSL Briefings, I reported on research Bob Davies-Jones and I did on hypothesized modes of tornadogenesis. In what is considered by many to be the archetype of tornadogenesis, the embryonic tornado develops within the mesocyclone, several kilometers above the ground, and gradually descends to the ground, perhaps through a process known as the "dynamic pipe effect." We showed, however, that tornadoes also may form somewhat uniformly over a several-kilometer vertical depth, or appear only at the lowest altitudes and then "ascend."

Our study provided theoretical explanations for these two primary modes of tornado development but could say nothing about their occurrence in the atmosphere. In particular, knowledge of the relative frequency of each mode is important because, theoretically, the "non-descending" tornadoes tend to develop much more rapidly than do their "descending" counterparts. This has obvious implications on the issuance of timely warnings by operational meteorologists.

A tornadic vortex signature (TVS) in Doppler weather radar data is a degraded image of an embryonic or fully-developed tornado, as shown by Rodger Brown, Les Lemon, and Don Burgess in the 1970's. The opportunity to examine a large variety of tornadoes via TVSs has only recently been provided by the implementation of the WSR-88D network. Thus, with support from the Cooperative Program for Operational Meteorology, Education and Training, we (DeWayne Mitchell, David Andra, Dan Effertz, Greg Tipton, Irv Watson, a few others, and I) sought to determine the frequency of descending versus non-descending TVSs, and hence tornadoes, by analyzing signatures of radar-detected tornadoes.

A geographically-diverse data set, comprised of 52 events varying from southern Great Plains supercell tornadoes to landfallen tropical cyclone-spawned tornadoes, was considered. Upon classifying each by an objective means, we found (with a standard error of 7%) that 52% of the sampled tornadoes had descending TVSs, and 48% had non-descending TVSs.

The results were stratified according to attributes of the tornado and TVS. For example, the descending TVSs in our sample were associated with greater differential velocity (the change in Doppler velocity across two adjacent radar beams) and greater tornado lead time. Tornadoes within squall lines and bow echoes tended to be associated with non-descending TVSs (see figure), an identification which provided a mean tornado lead time of 5 min.

Based on the results of this study, radar operators should recognize, while interpreting Doppler radar signatures for tornado warning decisions, that a large percentage of tornadoes form in a manner different than had once been thought.

For more information contact Jeff Trapp at: trapp@nssl.noaa.gov