1 / 32

Storm-Scale Convection-Allowing Ensemble and Deterministic Convection-Resolving Forecasting

CAPS Realtime 4-km Multi-Model Convection-Allowing Ensemble and 1-km Convection-Resolving Forecasts for the NOAA Hazardous Weather Testbed (HWT) 2009 Spring Experiment. Ming Xue, Fanyou Kong, Kevin W. Thomas, Jidong Gao, Yunheng Wang, Keith Brewster, Kelvin K. Droegemeier, Xuguang Wang

sabina
Télécharger la présentation

Storm-Scale Convection-Allowing Ensemble and Deterministic Convection-Resolving Forecasting

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. CAPS Realtime 4-km Multi-Model Convection-Allowing Ensemble and 1-km Convection-Resolving Forecasts for the NOAA Hazardous Weather Testbed (HWT) 2009 Spring Experiment Ming Xue, Fanyou Kong, Kevin W. Thomas, Jidong Gao, Yunheng Wang, Keith Brewster, Kelvin K. Droegemeier, Xuguang Wang Center for Analysis and Prediction of Storms (CAPS), University of Oklahoma John Kain2, Steve Weiss3, David Bright3, Mike Coniglio2 and Jun Du4 2National Severe Storms Laboratory 3Storm Prediction Center 4Environmental Modeling Center/NCEP Contact: mxue@ou.edu

  2. Storm-Scale Convection-Allowing Ensemble and Deterministic Convection-Resolving Forecasting • CAPS/OU has been carrying out a project over the past three year to develop, conduct and evaluate realtime high-resolution ensemble and deterministic forecastsfor convective-scale hazardous weather, in collaboration with NOAA HWT coordinated by NSSL and SPC. • Supported by NWS CSTAR (Collaborative Science and Technology Applied Research) program, with leverage on NSF and other sources of funding.

  3. Motivations • Ensemble predictions are usually done at relatively coarse resolutions (e.g., ~40 km for SREF); • Ensemble prediction at the convection allowing/resolving resolutions is new; • High uncertainty and nonlinearity at convective-scale make probabilistic information at such scales potentially more important; • Optimal ensemble configurations at convective scale little studied; • Radar data assimilation for continental-sized grid over extended periods never tried before. • Post-processing/verification/evaluation for both deterministic and probabilistic products at convective scale has many challenges.

  4. Scientific Issues to Address • The values and cost-benefits of storm-scale versus coarser-resolution short-range ensembles versus even-higher-resolution deterministic forecast; • Suitable perturbation methods for storm-scale ensemble, e.g., IC, physics, and model perturbations; • Proper handling and use of boundary perturbations; • The value and impact of assimilating high-res data including those from radars; • The most effective ensemble processing and most useful products at the convective scales; • The impact of such unique products on forecasting and warning.

  5. HWT Spring Experiment Daily Discussions (pictures from spring 2007)

  6. Forecast Configurations of Three Years Spring 2007: 10-member WRF-ARW, 4 km, 33 h, 21Z start time, NAM+SREF ICs. 5 members physics perturbations only, 5 with Phy+IC+LBC perturbations. Single 2 km grid. 2/3 CONUS (2007 NWP conf.) Spring 2008: larger domain, 00Z start, Phy+IC+LBC pert for all. Radar Vr and Z data assimilation for 4 and 2 km grids! (2008 SLS Conf.) Spring 2009: 20 members, 4 km, 3 models (ARW, NMM, ARPS), mixed physics/IC/LBCs. Single 1 km grid. Radar DA on native grids. 30 h forecasts from 0Z About 1.5 months each spring season from mid-April through early June

  7. ARPS 3DVAR Analysis Grid 1 km grid: 3603 x 2691 x 51 WRF ARW (4 and 1 km) and ARPS forecast grid and common post-processing grid WRF NMM forecast grid

  8. Forecast Configurations – more details • ARPS 3DVAR+Cloud Analysis provide control IC • IC perturbations and LBCs from SREF for perturbed members • NAM forecasts for control LBCs • Level-2 radial velocity and reflectivitydata from over 120 WSR-88D radars analyzed. • Physics options: mixed for the ensemble. • 3D output every hour, selected 2D output every 5 minutes. • Hourly graphics posted on the web, and extracted 2D fields sent to HWT N-AWIPS • 1 km forecasts using ~10,000 cores on a Cray XT5 at National Institute for Computational Science (NICS) at University of Tennessee • 4 km ensembles using ~2000 cores on a Cray XT3 at Pittsburgh Supercomputing Center (PSC). • 4 million CPU-hours used. Over 100 TB of data archived. • Forecasts completed in 5-9 hours over night.

  9. efs_post

  10. Configuration of Ensemble Members PBL ARW NMM ARPS

  11. Forecast Examples(This talk emphasizes radar data impact and 1-km resolution – next talk emphasizing ensemble prediction) • May 26-27, 2008 multi-convection case, a case with weaker forcing, more complex evolution – the case used for 1-km testing in pre-realtime forecast period • May 8, 2009 MCV/Derecho case

  12. Sfc Chart for May 26, 2008 Case H H L L H L H L 00 Z May 26 03 Z, May 26 H L L H L H L 09 Z, May 26 23 Z, May 26

  13. 1 km with radar OBS 4 km with radar 4 km no radar t = 3 h

  14. t = 9 h 1 km with radar OBS 4 km with radar 4 km no radar

  15. OBS 1 km with radar SPC Severe reports 4 km no radar 4 km with radar t = 23 h Fig. 3. As Fig. 1 but valid at 2300 UTC, 26 May 2008, corresponding to 23 hour forecast time, and for a further zoomed-in domain.

  16. ETSs of hourly precipitation at 0.10 inch threshold for 26 May 2008 case 1 km with radar 4 km with radar 4 km without radar

  17. May 8, 2009 Bow echo/Derecho/MCV/Inland hurricane case 1820 UTC, S. Illinois Movie Missouri NWS WFO St. Louis MO

  18. IR images for May 8, 2009 MCV/Derecho Case MCC 1500Z, May 8 1215 Z May 8 2100Z, May 8 1800Z, May 8

  19. 1 km Prediction of MCV - Movie

  20. 18-h forecast wind speed (color), wind vectors, and reflectivity at ~30m AGL

  21. 1 km Prediction of Entire Domain- Movie

  22. FY WM TM TM WY TY FM TM WM TY Postage stamp view of 18-h forecast from 20 4-km ensemble members and 1 km ARW, valid at 1800 UTC, May 8, 2009

  23. Postage stamp view of 18-h forecast from 20 4-km ensemble members and 1 km ARW, valid at 1800 UTC, May 8, 2009 FY WY FM FM WM TY FY WM LT LT

  24. 18-hour ensemble forecast products Spread Spaghetti of 35 dBZ reflectivity Probability matched reflectivity Probability of reflectivity >35 dBZ Observed reflectivity 1800 UTC, May 8, 2009

  25. Precipitation verificationfrom 2008Hourly precipitation ETS scores against NSSL 1-km multi-sensor QPE 0.01 inch/hour threshold with radar No radar 0.5 inch/hour threshold with radar No radar

  26. Summary • Examples of 1-km CONUS-scale forecasts, and 4-km multi-model ensemble forecasts were shown. • Clear impact of radar data assimilation – up to 24 hours in some cases • 1-km forecasts provide more realisms, especially in detailed structures. • The results show great promise of explicit convective-scale forecasting – the warn-on-forecast concept (Stensrud et al. BAMS article) • Significant sensitivity to model physics. • Huge amount of valuable data waiting to be analyzed/exploited – some are easily accessible 2D form - collaborations very welcome!

  27. List of referred publications from collaborations • Schwartz, C., J. Kain, S. Weiss, M. Xue, D. Bright, F. Kong, K. Thomas, J. Levit, and M. Coniglio, 2009: Next-day convection-allowing WRF model guidance: A second look at 2 vs. 4 km grid spacing. Mon. Wea. Rev., Accepted. • Schwartz, C. S., J. S. Kain, S. J. Weiss, M. Xue, D. R. Bright, F. Kong, K. W.Thomas, J. J. Levit, M. C. Coniglio, and M. S. Wandishin, 2009: Toward improved convection-allowing ensembles: model physics sensitivities and optimizing probabilistic guidance with small ensemble membership. Wea. Forcasting, Conditionally accepted. • Clark, A. J., W. A. Gallus, Jr., M. Xue, and F. Kong, 2009: A comparison of precipitation forecast skill between small near-convection-permitting and large convection-parameterizing ensembles. Wea. and Forecasting, Accepted. • Clark, A. J., W. A. Gallus, Jr., M. Xue, and F. Kong, 2009: Growth of spread in convection-allowing and convection-parameterizing ensembles, Being submitted. • Coniglio, M. C., K. L. Elmore, J. S. Kain, S. Weiss, and M. Xue, 2009: Evaluation of WRF model output for severe-weather forecasting from the 2008 NOAA Hazardous Weather Testbed Spring Experiment. Wea. Forcasting, Conditionally accepted. • More coming.

  28. Future: A Dream System of Convective-scale NWP • Ensemble Kalman filter DA every 5 minutes at ~1 km resolution; • ~1 km ensemble forecasts launched every ~30 min up to 6 hours; • Reduced (~4 km) resolution ensemble forecasts up to 72 hours; • Deterministic forecasts from ensemble mean analyses at up to 250 m resolution; and at reduced resolutions beyond 6 hours; • Explicit forecasts for MCS down to tornado vortex scales, with probability estimates. • An NSF Track-1 sustained-petaflops machine (IBM Power7) with over 200,000 cores to come online in 2011 – are we ready?

  29. The Heroes NICS, Kraken (~66K cores) Thank you. Questions? PSC (4K cores)

More Related