Strengths and Limitations of Using NSSL’s Experimental Warn-on-Forecast System for Ensembles (NEWS-e) in Forecasting Cyclic Mesocyclogenesis
Kelsey Britt1,2,3, Patrick Skinner1,3, Pamela Heinselman2,3, and Kent Knopfmeier1,3
1 Cooperative Institute for Mesoscale Meteorological Studies/University of Oklahoma; 2 School of Meteorology, University of Oklahoma; 3NOAA/OAR National Severe Storms Laboratory
Cyclic mesocyclogenesis is the process by which a supercell produces multiple mesocyclones with similar life cycles. Supercells that exhibit cyclic mesocyclogenesis (i.e. cyclic supercells) have the potential to produce several tornadoes that could cause widespread damage. Therefore, having the ability to forecast the potential for cyclic supercells may be beneficial for forecasters when issuing watches and warnings. However, idealized simulation studies of cyclic mesocyclogenesis have found the process to be highly sensitive to both environmental and computational parameters. Thus, whether or not short-term, storm-scale numerical weather prediction models can actually resolve and predict cycling has yet to be determined. Therefore, this study performs three sensitivity experiments and analyzes their impacts on forecasts generated by NSSL’s Experimental Warn-on-Forecast System for Ensembles (NEWS-e) of four cyclic supercells occurring on 9 May, 16 May, and 18 May 2017. NEWS-e was created as part of the Warn-on-Forecast (WoF) initiative, and is a convection-allowing ensemble with 18 forecast members. The sensitivity experiments included changing the NEWS-e horizontal grid spacing from 3 km to 1 km, examining the effects of each forecast member’s different PBL and radiation schemes, and analyzing how changes in the environment across the different storms can affect the cycling process. Analysis of the individual ensemble members is conducted to assess the capability of a NEWS-e to resolve and predict cyclic mesocyclogenesis, and whether this process is physically representative of the current understanding of cyclic supercells. To provide a source of verification for the NEWS-e forecasts, a database of observations was created from manually analyzed WSR-88D radar reflectivity and radial velocities for each supercell.
Resolving cycling at 3-km was thought to be highly improbable from past research that suggested the coarsest grid resolution to resolve cycling was 1 km. However, seven 3-km, NEWS-e forecasts for 18 May 2017 show evidence of cyclic mesocyclogenesis-like processes occurring in both cyclic supercells observed that day. When the NEWS-e grid spacing was changed to 1 km, cycling was observed more frequently than at 3 km, but there was little skill in NEWS-e predicting the timing and frequency of cycles compared to observations. Also, PBL and radiation parameterizations showed no clear signal on impacting cyclic mesocyclogenesis. From a predictability standpoint, across the four supercells, NEWS-e shows promise in forecasting the probability that cyclic supercells can occur on a given day, but there is little to no skill in it predicting the timing or the frequency of those cycles.