7.4 Generating Calibrated Tornado Probabilities from NEWS-e Using Climatological Frequencies

Wednesday, 9 January 2019: 3:30 PM
North 221AB (Phoenix Convention Center - West and North Buildings)
Susan Beveridge, Ohio Univ., Bellville, OH; and P. S. Skinner, B. T. Gallo, K. H. Knopfmeier, B. T. Smith, and R. L. Thompson

Calibrated NSSL Experimental Warn-on-Forecast System for ensembles (NEWS-e) tornado probabilities using climatological tornado frequencies given a right-moving supercell were created for a case (1 May 2018) from the 2018 Hazardous Weather Testbed Spring Forecasting Experiment in which there were multiple tornadic supercells. The 2300 UTC NEWS-e forecast was selected as the base forecast for examining various parameters because storms were on-going, and it was approximately 1.5 hours before a supercell produced an EF3 tornado in Saline Co., Kansas. Prior to generating the tornado probabilities, 0-2 km and 2-5 km updraft helicity (UH), and surface-based and mixed layer significant tornado parameter (STP) were examined. Higher tornado probabilities were anticipated where a simulated supercell, as indicated by UH existence, was moving into a favorable STP environment. A novel aspect of this study is the ability to generate probabilistic tornado forecasts within a deterministic framework (i.e., a single member of the NEWS-e).

Tornado probabilities vary across individual members within the ensemble, between surface-based and mixed layer STP, and for varying percentiles of inflow STP used. It was observed that 1) larger surface-based STP values occurred prior to early evening transition, whereas larger mixed-layer STP values occurred afterwards; 2) tornado probabilities varied by >20% between members; 3) ensemble tornado probabilities varied by >15% between the ensemble mean and ensemble 90thpercentile. Given this variation, the use of reliability diagrams will help determine the most beneficial way to group and display tornado probabilities.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner