Recorded presentation15A.6
The High Resolution Rapid Refresh (HRRR): an hourly updated convection resolving model utilizing radar reflectivity assimilation from the RUC / RR
Stephen S. Weygandt, NOAA/ESRL/GSD, Boulder, CO; and T. G. Smirnova, S. G. Benjamin, K. J. Brundage, S. R. Sahm, C. R. Alexander, and B. E. Schwartz
Since Sept. 2007, the Global System Division (GSD) at NOAA/ESRL has been running an hourly, convection resolving model, driven by the radar reflectivity assimilating Rapid Update Cycle (RUC), over a domain covering the NE U.S. aviation corridor. Known as the High Resolution Rapid Refresh (HRRR), the model utilizes a 3 km horizontal grid spacing configuration of the WRF-ARW model. Model initial conditions are obtained from the hourly cycled Rapid Update Cycle (RUC) model, which has an advanced mesoscale data assimilation system to accurately depict the evolving 3D mesoscale storm environment and a novel radar reflectivity assimilation procedure to specify ongoing precipitation systems. The RUC radar reflectivity assimilation procedure uses a diabatic digital filter (DFI) initialization to generate storm scale circulations in regions of ongoing convection and has significantly improved short-range precipitation forecasts (especially for convective systems) in both the RUC and HRRR. This RUC reflectivity assimilation (previously running in GSD parallel RUC cycles) was implemented in the NCEP operational RUC in Dec. 2008 and has been ported to the Rapid Refresh, which will replace the RUC at NCEP in 2010. The HRRR nest will be switched to the radar reflectivity assimilating Rapid Refresh later in 2009.
During the summer of 2008, HRRR forecasts initialized from RUC versions with and without the reflectivity assimilation were compared. 3-km reflectivity forecast skill scores were computed, and showed significant improvement for short-range (3-6 h forecasts) for the RUC radar reflectivity assimilating HRRR compared to the HRRR without the RUC radar assimilation. Smaller reflectivity forecast skill improvements extended out to 12-h, and as expected, significant diurnal aspects to the skill improvement were documented. Output from these HRRR forecast were used as input to the Collaborative Storm Prediction for Aviation (CoSPA), an experimental probabilistic convective guidance product produced jointly by ESRL/GSD, NCAR/RAL, and MIT/LL.
For the 2009 convective season, the HRRR domain is being expanded to cover most of the eastern 2/3 of the U.S. A more extensive demonstration / evaluation of its utility in providing convective storm guidance for aviation (in conjunction with CoSPA) and other applications (severe weather, precipitation) is planned. HRRR forecasts have continued in the cold seasons as well, and informal evaluation of HRRR forecast skill for mesoscale features (precipitation bands, lake effect snow, and terrain-related surface fields) is ongoing.
At the conference, we will provide an overview of the HRRR, illustrating the impact of the RUC reflectivity assimilation procedure on HRRR forecast skill and show real-time results from 2008 and 2009.
Session 15A, Data Assimilation: RTMA and Rapid Refresh
Thursday, 4 June 2009, 1:30 PM-3:00 PM, Grand Ballroom East
Previous paper