Evaluating the Three-Ingredients Method for Nowcasting QLCS Tornadoes

Quasi-linear convective system (QLCS) tornadoes are a significant operational forecasting and nowcasting challenge, partly due to knowledge gaps surrounding processes that lead to QLCS tornadogenesis, or how those processes are linked to the QLCS near-storm environment. These gaps are due to a lack of quality QLCS observations, the small size and transient nature of most QLCS tornadoes, and the fact that most QLCS tornadoes are weak. This is an issue when you consider QLCSs produce a high percentage of tornadoes across some locations of the US, and QLCS tornadoes have much worse tornado warning skill scores (i.e., lower probability of detection, higher false alarm rates, and less warning lead time) compared to supercellular tornadoes. Improved understanding of QLCS tornadoes and their environments is critical to improving tornado warning performance and thus advancing the National Weather Service (NWS) core mission of protecting life and property.

NWS forecasters commonly use the “Three Ingredients Method” (3IM) to anticipate QLCS mesovortex- and tornadogenesis. This includes finding balanced or slightly shear dominant regimes of the QLCS, coincident with line normal 0-3 km bulk shear ≥ 30 kts, and surges or bowing segments in the QLCS. Locations that meet all three ingredients are often favored locations for severe storm reports, either through damaging wind and/or tornadoes. Although the 3IM is widely used to anticipate QLCS severe weather, the method has received little attention in the literature. Using six years of historical case data, RUC/RAP 0-hour environmental data, and Doppler radar data, this presentation will provide results from an ongoing project evaluating both the QLCS tornadic environment and the 3IM. Results are generated for the entire dataset and partitioned by differing climatic regions across the United States. To date, over 1000 QLCS cases have been analyzed. Preliminary results indicate that lowering the 3IM shear threshold to ≥ 20 kts improves the 3IM tornadic “probability of detection” by approximately 20%, but also increases the number of false positive cases by 9%. Additional analysis shows the 3IM has different verification statistics across the different regions, suggesting the 3IM is not a one size fits all model for nowcasting QLCS tornadoes. This presentation will also discuss other environmental parameters, and suggestions for utilizing these results in the operational environment.