Rapid-Scan, Mobile Radar Observations of Tornadogenesis: Warning and Forecasting Implications

Over the last decade, high-resolution numerical weather models and rapid-scan radar platforms have shed new light on the spatio-temporal evolution of rotation associated with nascent tornadoes (e.g., Mashiko et al. 2009; French et al. 2013; Dahl et al. 2014; Markowski and Richardson 2014; Houser et al. 2015; Mashiko 2016; Orf et al. 2017; Yokota et al. 2018; Bluestein et al. 2019; Wienhoff et al. 2020). The results from such recent studies have concluded that tornadogenesis, which historically used to be thought of as a top-down process, commonly follows a sequence of events more congruent with a non-descending process. While numerical models have become substantially more sophisticated and capable of resolving near-ground processes, to date, there has been a paucity of high-quality observations having sufficient spatio-temporal resolution to truly resolve real-world tornadogenesis events, particularly at heights < 100 m above ground. Only a handful of such studies have analyzed such processes, and most of these document the evolution of significant tornadoes (e.g., French et al 2013; Houser et al. 2015; Bluestein et al. 2019; Wienhoff et al. 2020). Furthermore, near-ground observations (z<100 m AGL) are missing in most of these studies, owing to the inherent limitations of working with mobile radar data.

This study uses rapid-scan, mobile radar observations acquired over multiple spring field campaigns in the Central Plains of the US from the rapid-scan, X-band, polarimetric (RaXPol) radar (Pazmany et al. 2013). It analyzes high spatio-temporal resolution [volume updates occurring on O (20 s)] observations of eight tornadogenesis cases for tornadoes varying in intensity from EF0-EF3 by tracking 1) the evolution of the tornadic vortex signature (TVS), 2) the maximum velocity differential across the TVS, and 3) the pseudovorticity for five minutes prior to tornadogenesis through the first 2 minutes of the tornado. Six out of the eight cases collected near-ground data (within the lowest 100 m AGL). Results for seven out of the eight cases clearly indicate a non-descending evolution of the TVS. All cases observed tornadogenesis occurring on timescales on the order of 1-3 minutes. In addition, it was discovered that, many of these tornadoes were very shallow, being confined to heights < 1.5 km AGL. Lastly, the implications of these results on operational data availability, nowcasting, warning and forecast strategies is discussed.