Improving Tornado Warning Decision Making through New Conceptual Models and Interrogation Techniques with WSR-88D Dual-Polarization Products

Improving Tornado Warning Decision Making through New Conceptual Models and Interrogation Techniques with WSR-88D Dual-Polarization Products

John Stoppkotte^a and Aaron W. Johnson^b

^aNational Weather Service, North Platte, Nebraska

^bNational Weather Service, Dodge City, Kansas

ABSTRACT

Warning forecasters in the National Weather Service (NWS) have long struggled to consistently identify thunderstorms more capable of producing tornadoes from those that are not. Training has evolved in the last couple of decades to closely align with the theoretical outcomes of research from field projects such as VORTEX and VORTEX II (among others). But on average, statistical results from tornado warnings issued by the NWS have become stagnant or fallen somewhat over the last several years. The addition and integration of dual polarization (DP) within the NWS WSR-88D network has helped forecasters gain insight into hydrometeor character in various forecasting regimes, but its use in the tornado warning decision making process has been limited.

However, research by the academic community in the use of DP radar products in assessing a storm’s tornadic potential has been ongoing for well over a decade. As a result, new conceptual models and interrogation techniques have emerged within the research community in order to help identify those supercells that may be more conducive of producing tornadoes from those that are not. In particular, how use of differential reflectivity (Z_DR), specific differential phase (K_DP) and correlation coefficient (CC) can be used with the WSR-88D to identify storm-scale features associated with an environment more favorable of supporting tornadogenesis. Specific examples of these interrogation techniques for both tornadic and non-tornadic supercells are illustrated - with particular emphasis on a non-tornadic supercell case over south-central Kansas on 15 April 2017.

Clear evidence of a severe thunderstorm with this case exists in radar reflectivity and storm-relative mean velocity (SRM) indicating a robust midlevel mesocyclone. Nonetheless, a brief low-level SRM couplet creates a small window where a tornado warning may seem appropriate. However, hydrometeor size sorting by the storm-relative wind through a depth of the storm, as evidenced by a separation in the Z_DR Arc and K_DP foot, was mostly absent while almost no SRM enhancement was noted in the lowest tilts. Further, while a single 0.5° scan displays a Z_DR Arc and K_DP foot with some separation, the separation vector was nearly parallel to the storm motion, suggesting a storm dominated by crosswise vorticity. Additionally, both Z_DR and CC indicate large volumes of hail and/or rain mixed with hail near the updraft with this inferring a colder, more negatively buoyant rear flank downdraft (RFD). All of these characteristics imply little evidence of a surge in streamwise vorticity that may strengthen a low-level mesocyclone or the storm’s ability to stretch near-surface vorticity that are precursors for tornadogenesis. These combined signals also imply the low-level SRM couplet may be a false indicator of low-level shear via some form of contamination. Although the couplet appears to indicate strong low-level rotation, DP characteristics provide enough evidence to ultimately assist in determining this feature as velocity contamination emanating from a vertical sidelobe.