A Look at Some New HRRR Model Variables For Diagnosing Lower Level Storm Rotation and Their Relation to Tornado Potential

The High Resolution Rapid Refresh (HRRR) model, developed at NOAA/ESRL/GSD and run operationally at NCEP, is on track to be updated to its next version (HRRRv3) in July 2018. The HRRR is the only operational NCEP convection-allowing model with a new run every hour, currently run out to 18 h operationally, but extending to 36-h at 6-h intervals (00, 06, 12 and 18 UTC). In addition to a number of improvements to the HRRR and its parent model, the RAP (Rapid Refresh model), which will also be updated at NCEP at the same time, there are several new variables that are being output. While some of these will not be available as operational products delivered to WFOs and AWIPS2 via the SBN, they will be available on the HRRR GSD website. Some of the new variables are related directly to rotational characteristics of supercell storms, expanding on the 2-5 km AGL Updraft Helicity (UH) that has been a variable in the current HRRR version and generally evaluated for potential tornado threat operationally and at the Hazardous Weather Testbed Spring Experiment for many years. The new variables include UH over varying depths (1-6, 0-2 and 0-3 km AGL) as hourly output and as run-total output to better display the overall supercell lifetimes. In addition, Vertical Vorticity (VV) fields will be available for 0-1 and 0-2 km AGL. While the HRRR horizontal grid resolution of 3-km is too coarse to directly consider the VV to be a realistic representation of a tornado, there may be an association between these fields and the lower UH depths and the likelihood of tornado production. Indeed, one of the aspects of this study will be to examine the new UH and VV fields for a variety of tornado days across the CONUS to get a better understanding of what model values are associated with these new variables, but also to consider days where there is a clear distinction between supercells producing only severe weather but no tornadoes and days where tornadoes were more prevalent. We will look to determine if there is any potential in the value of these new variables to differentiate between such days.

We will closely coordinate this study with a separate study (and abstract submitted to SLS) by Logan Dawson and co-authors from NCEP, who will emphasize a more detailed look at some of the cases reviewed by NCEP/EMC’s Model Evaluation Group and presented at MEG meetings from spring 2018.