Comparison of Tornadic and Severe Non-Tornadic Storms Using Probability Matched Means of Radar Observations

Probability matched means combine the smooth mean state of a collection of events with the probability distribution function of instantaneous data. An advantage of utilizing probability matching is the ability to retain characteristic structure and variations in the data that are observed in individual events. This method can be used to investigate typical three-dimensional characteristics of a large group of storms observed by radar. Here, probability matched means of discrete tornadic and severe non-tornadic storms observed by NEXRAD WSR-88D radars are computed in a coordinate system that is rotated horizontally such that the individual storm motion vector points in the positive x-dimension. Comparisons between the two storm types are made at multiple altitudes between physical and dynamical variables, including the radar reflectivity at horizontal polarization, differential reflectivity, specific differential phase, copolar correlation coefficient, velocity spectrum width, radial divergence, and azimuthal shear. Preliminary results indicate that the most significant differences between tornadic and severe non-tornadic storms are found in low-level observations of velocity spectrum width within the hook echo, and that vertical alignment of low-level rotation (azimuthal shear) with mid- to upper-level rotation often precedes tornadogenesis and persists during tornadic periods.