Using polarimetric radar data to identify tornadoes in near real-time

Many automated near real-time tornado detection methods rely on estimates of rotation computed from radial velocity to identify and represent tornadoes. While these methods often capture tornadoes, they have a high false-positive rate since not every rotation detected by radar is indicative of a tornado. Conversely, identification of features such as debris signatures can be biased against weaker tornadoes or those that do little damage. The Verisk tornado probability algorithm leverages polarimetric radar data in near real-time to capture tornadoes across the intensity spectrum while reducing false positives, providing more accuracy in conveying the likelihood of tornado occurrence. The result is a robust relationship between tornado probability and damage survey paths compared to either rotation tracks or debris signatures alone. Future updates to the algorithm will improve detection of tornadoes in a wider range of storm morphologies and environments while further reducing false positives.