Extended Abstract (308K)12.1
Real-time applications of the WRF model at the Forecast Systems Laboratory
Steven E. Koch, NOAA/ERL/FSL, Boulder, CO; and S. G. Benjamin, J. A. McGinley, J. M. Brown, P. Schultz, E. J. Szoke, T. G. Smirnova, B. L. Shaw, D. Birkenheuer, S. Albers, S. E. Peckham, and G. A. Grell
This presentation highlights scientific findings from a wide range of applications of the Eulerian mass-coordinate version of the WRF model by scientists at the NOAA Forecast Systems Laboratory over the last year. These applications include: (1) participation in the IHOP_2002 (International H20) Project , (2) a quasi-operational version of WRF at 5-km resolution running at a local Forecast Office under the NWS Coastal Storms Initiative, (3) a member of an ensemble of models supporting winter road maintenance operations in Iowa, (4) support for the NOAA New England High-Resolution Temperature and Air Quality (TAQ) program, (5) real-time physics testing against the 20-km RUC model over the CONUS region in preparation for the future "Rapid Refresh" WRF implementation at NCEP, and (6) a fully coupled "online" air chemistry version of the WRF model for air pollution prediction.
Some highlights are briefly mentioned here. The IHOP implementation included 12- and 4-km versions of the WRF model utilizing full volumetric WSR-88D reflectivity and radial velocity data and experimental GOES-11 satellite products in a diabatic initialization procedure known as the "Hot Start" technique. Verification results show decided benefits of using this technique for short-term (0-6 h) precipitation forecasts. Currently, scientists at FSL and Iowa State University are analyzing the IHOP model forecasts made by WRF, MM5, and Eta at 12-km resolution to determine the degree to which these models can properly represent the observed morphology of mesoscale convective systems. The TAQ implementation employed the Rapid Update Cycle (RUC) model to provide initial conditions for a 10-km configuration of the WRF model with the RUC Grell-Devenyi convective parameterization and RUC-Smirnova land surface formulation. These runs as well as the current 20-km CONUS runs have shed light on some difficulties with the Grell-Devenyi scheme as implemented in the WRF model, leading to the temporary replacement of this scheme with the Eta Betts-Miller-Janjic scheme. Overall, these WRF runs appear to produce more grid-scale precipitation than comparable RUC runs, particularly at 10-km resolution. The WRF TAQ forecasts also exhibited a somewhat diminished diurnal temperature cycle compared to surface observations, in large part the likely result of insufficient vertical resolution at lower levels. The WRF-Chem modeling system, running in retrospective mode, has been compared to a comparable configuration of the MM5-Chem model run during the TAQ project. These coupled models include simultaneous treatement of the chemical mechanism package based on RADM2, biogenic emissions, deposition, tracer transport by convection and turbulence, photolysis, and advective transport. Preliminary examination of the model results show that the MM5-Chem and WRF-Chem models are producing comparable results and have reasonable skill in forecasting periods of poor air quality.
Session 12, WRF Model Development and Applications (ROOM 605/606)
Tuesday, 13 January 2004, 3:30 PM-5:45 PM, Room 605/606
Next paper
Recorded presentation