Comparison Between Mesocyclone and Updraft Area Proxies for Tornado Intensity Prediction in Supercells

Theoretical and modeling results have been used to argue that larger supercell updrafts, at both mid and low levels, are predictive of greater peak tornado intensity via circulation arguments and relationships among the sizes of the updrafts, downdrafts, mesocyclones, and tornadoes; the wider tornadoes are then assumed, generally, to be more intense. However, results from more recent simulations have been used to argue that tornado intensity is not related to updraft width, but instead to the size and rotational characteristics of the low-level mesocyclone. But observational studies using different remote sensing proxies for updraft size (overshooting top area and ZDR column area) have provided strong evidence that there is indeed a relationship between midlevel updraft area/width and peak tornado intensity. Given the aforementioned modeling results supporting mesocyclone area as the link to tornado intensity, it remains unclear why midlevel updraft area proxies show some skill in identifying storms more likely to produce significant tornadoes. One possibility to explore is that given that we would expect large OT and ZDR column areas to be more likely in higher vertical velocity updrafts, in turn the large OT and column areas may also relate to mesocyclone characteristics.

In this study, we use WSR-88D data from a large number (~200) of tornadic and non-tornadic supercells cases to isolate the importance of both midlevel updraft area and low-level mesocyclone width to tornado peak intensity (and formation). Midlevel updraft area is estimated using a novel algorithm that determines the area of ZDR columns and the low-level mesocyclone width is estimated via radial velocity data. With these radar proxy data, we use subsets of cases to setup different parameter spaces of updraft/mesocyclone size (i.e., small vs. large) and determine optimal tornado peak intensity prediction. Our approach ideally provides clarity as to the robustness of previous observational results. The implications of our results on using radar proxies of updraft and mesocyclone size for nowcasting peak tornado intensity also will be discussed.