Development of a Real-Time Conditional Probability of Tornado Intensity Product

To better understand, evaluate, and communicate real-time tornado threats, the Storm Prediction Center (SPC) has worked toward calculating gridded probabilities of tornado intensity (conditional on the existence of a tornado) by deriving statistical relationships between tornado EF-scale ratings, manually-calculated radar-derived circulation attributes (rotational velocity and shear diameter), and environmental significant tornado parameter (STP) values. While the preliminary tornado intensity models have shown utility, rotational velocity (V_rot) and shear diameter cannot be objectively calculated in an automated real-time sense, preventing the probability products from operating in a real-time environment and effectively limiting their applicability in short-fuse warning-decision operations. To overcome this, SPC and CIMMS researchers are collaborating to develop conditional probability of tornado intensity (CPTI) products capable of producing real-time, CONUS-wide tornado intensity estimation probability grids. To do so, azimuthal shear —a gridded estimate of rotation strength that is calculated in real-time within the NWS Multi-Radar Multi-Sensor (MRMS) suite of products—replaced the manually-derived input of V_rot. Using a comprehensive SPC severe storm report database from 2014-15, a quantitative relationship between AzShear and V_rot was derived, effectively bridging the two measures of rotation and allowing azimuthal shear to be utilized in the automated real-time CPTI product. A technique was also developed to derive shear diameter from single-radar velocity data, which are then merged together from multiple single radars to form a merged gridded shear diameter product. Following the development and integration of these two real-time input products for CPTI, a logistic regression model was constructed using the same two-year dataset, providing a platform to generate real-time tornado intensity probability grids.

This presentation reviews the methodology and initial testing of the new AzShear-based CPTI gridded product, including discussions of AzShear-V_rot relationships, shear diameter calculations, logistic regression model creation, and the output of the prototype gridded CONUS probability products. Finally, plans for real-time testing of CPTI during Spring 2019 in the Hazardous Weather Testbed will be outlined.