J1.2 Verification of Supercell Motion Forecasting Techniques

Several storm motion prediction methods have attained widespread operational and research use, the most common being Maddox (1976) variations which apply a 30 rightward and 75 percent of mean wind assumption (hereafter 30R75), a 15 rightward and 85 percent of mean wind assumption (adapted from Johns et al. 1993; hereafter 15R85), and more recently the Bunkers et al. (2000; hereafter Bunkers) algorithm. This examination evaluates the performance of those three techniques using observed storm motions and Rapid Update Cycle 2 (RUC-2) model proximity soundings (Thompson et al. 2002a) for a nationwide database of 452 observed supercells (six cases were omitted from the full 458 supercell data set). Also tested was the assumption that the mean wind through the 0-6 km layer above ground level (AGL) represents storm motion.