The 2021 NOAA Hazardous Weather Testbed Spring Forecasting Experiment
| What: | More than 130 researchers and forecasters convened virtually to test forecasting applications of convection-allowing models at lead times from an hour to two days, and evaluate various convection-allowing modeling and postprocessing strategies. |
| When: | 3 May–4 June 2021 |
| Where: | Virtual |
The 2021 NOAA Hazardous Weather Testbed Spring Forecasting Experiment (2021 SFE) was held virtually 3 May–4 June 2021. SFEs are co-led by the NWS/Storm Prediction Center (SPC) and OAR/National Severe Storms Laboratory (NSSL), and test new severe weather prediction tools, study how end users apply severe weather guidance to forecasting, and facilitate experiments for optimizing convection-allowing model (CAM) ensemble design informing Unified Forecast System (UFS; https://ufscommunity.org) development. Well-vetted virtual meeting tools enabled the most SFE participation ever: 133 forecasters, researchers, and students from around the world. Gathering a diverse community to learn from one another has been a hallmark of SFEs that has continued in virtual experiments. However, science-based discussions and establishing new collaborations are more difficult in the virtual environment. Nevertheless, lessons learned could benefit a future hybrid approach involving both in-person and virtual participation.
SFE 2021 forecasting activities involved SPC operational product and service improvements and “data denial experiments” with CAM guidance withheld from a control group. Model evaluations emphasized the 64-member Community Leveraged Unified Ensemble (CLUE; Clark et al. 2018), with CLUE experiments examining data assimilation methods, strategies for single-model CAM ensemble design, and impact of regional domain size on day 2 model performance. Other evaluations studied machine-learning applications for severe weather, mesoscale analyses, NSSL’s Warn-on-Forecast System (WoFS), and calibrated severe weather guidance from operational global and regional ensembles.
Some highlights of the 2021 SFE
To explore short-term forecasting applications of WoFS and other CAMs, participants used this guidance to generate mesoscale discussions and present the product to a group of participants. This was a popular activity pushing participants to synthesize CAM output. One participant noted, “I loved doing that! . . . As a model developer/researcher, I don’t get to sit in ‘the hot seat’ . . . It gave me a chance to be more creative and thoughtful in my forecasting process.”
In another activity, two groups generated Day 2 Convective Outlooks. One group used CAMs and all other available data (All Data), while CAMs were withheld from the other (No CAMs). There was little difference overall in the subjective ratings between the groups for hail and tornado outlooks, but the All Data wind forecasts were rated significantly better than No CAMs because of several cases where CAMs accurately depicted severe-wind-producing mesoscale convective systems that were not evident in the non-CAM guidance.
In one CLUE evaluation, configuration strategies for a Rapid Refresh Forecast System (RRFS) were examined. The RRFS is a rapidly updating CAM ensemble that will use the UFS Short-Range Weather Application (currently the FV3-LAM), and will subsume several operational regional models simplifying NOAA’s modeling suite. Encouraging subjective evaluation results indicated that a prototype RRFS from NOAA’s Global Systems Laboratory, which uses stochastic physics and initial conditions from the operational High-Resolution Rapid Refresh data assimilation system (HRRRDAS), performed almost as well as the current operational CAM ensemble, the High-Resolution Ensemble Forecast System.
More information on the 2021 SFE
In the second virtual SFE, progress was made in key areas to accelerate R2O for models and tools that improve operational severe weather forecasts. The full 2021 SFE Summary Report is at https://hwt.nssl.noaa.gov/sfe/2021/docs/HWT_SFE_2021_Prelim_Findings_FINAL.pdf.
Acknowledgments.
This work comprised regular duties at federally funded NOAA/NSSL for Clark, Burke, Wicker, Heinselman, Gao, Potvin, and Imy and NOAA/SPC for Jirak, Dean, and Marsh. Gallo, Knopfmeier, Roberts, Krocak, Skinner, Wilson, Vancil, Hoogewind, Dahl, Creager, Jones, Wang, Loken, Flora, Kerr, Yussouf, Martin, Guerra, Matilla, Jahn, Harrison, and Galarneau were funded by NOAA/Office of Oceanic and Atmospheric Research under NOAA–University of Oklahoma (OU) Cooperative Agreement NA11OAR4320072, U.S. Department of Commerce. OU’s Center for Risk and Crisis Management also funded Krocak.
Reference
Clark, A. J., and Coauthors, 2018: The Community Leveraged Unified Ensemble (CLUE) in the 2016 NOAA/Hazardous Weather Testbed Spring Forecasting Experiment. Bull. Amer. Meteor. Soc., 99, 1433–1448, https://doi.org/10.1175/BAMS-D-16-0309.1.
