Variability in supercell motion

Recent studies (e.g., Bunkers 2018, Coniglio and Parker 2020) suggest that tornado-producing supercells often have motion vectors that deviate farther from the mean wind (being slower and more rightward) than non-tornadic supercells. This enhanced deviant motion could be either a cause of tornado production (as such storms would usually experience substantial increases in storm-relative helicity) or an effect of tornado production (due to internal storm dynamics that haven’t been fully explained). Cause and effect are rather elusive for real-world supercells, both due to observational limitations and because storms move through horizontally heterogeneous environments.  Here we isolate variability in supercell motion by studying a set of three existing ensembles of idealized supercell simulations (15 tornadic supercells from Coffer et al. 2017, 12 non-tornadic supercells from Coffer et al. 2017, and 25 tornadic supercells from Markowski 2020).  For each ensemble, all members were simulated using identical mean wind profiles, but with added random noise to produce spread.  Within each ensemble, the x- and y-components of storm motion both vary over ranges of 7-8 m/s. Among other questions, in this study we address the following.  How large are the accompanying variations in storm-relative helicity?  Are there lead or lag correlations between tornado times and changes in motion vectors in the simulations?  And, are the changes in motion vectors merely stochastic or can they be anticipated from observable processes?