Center for Analysis and Prediction of Storms (CAPS) Briefing by Ming Xue, Director CAPS is one of the 1st NSF Science and Technology Centers established in 1989. CAPS is the pioneer of radar data assimilation and storm-scale numerical weather prediction (NWP). CAPS leads OU’s efforts of CASA, an NSF ERC for Collaborative Adaptive Sensing of Atmosphere CAPS is part of the Model Development and Enhancement (MDE) Product Development Team (PDT) of FAA and contributes to the operational RUC/RR and NAM model improvement. CAPS developed and supports a complete high-resolution NWP and data assimilation system, the ARPS (Advanced Regional Prediction System). CAPS is one of the 4 initial partners developing the community WRF model CAPS works closely with NSSL/SPC via HWT and other joint projects.
50-m Grid Forecast v.s. Observation (Movie) Short-Range Radar Initialized Prediction of Thunderstorms, Strong Winds, Gust Fronts, Downbursts, and Tornadoes using NWP Model 50-m Grid Forecast v.s. Observation (Movie) Movie Forecast Low-level Reflectivity Observed Low-level Reflectivity 43 minute forecast 2
Tornadoes of 8-9 May 2007 10:54pm (0354Z) Minco Tornado El Reno tornado © Patrick Marsh Minco tornado 10:54pm (0354Z) Minco Tornado 9:54pm (0254Z) 8:30pm (0130Z) © KSWO TV 7:21pm (0021Z) Lawton Tornado Lawton tornado
2 hour forecast of a tornado using 88D and CASA radar data on a 400 m grid Minco tornado Observed tornado Predicted tornado Perfect timing and <10 km position error
Realtime high-resolution (400 m) wind analysis using obs from multiple Doppler radars and surface stations Analysis every volume scan. Ideal for airport-scale warning and nowcasting
Impact of assimilating polarimetric variables Blue: ZhZdr Gray: ZhZdp Green: ZhKdp Pink: ZhZZK Jung et al. (2008a,b,c MWR) 6
Prediction of Convective Initiation 12 June 2002 IHOP Case Prediction @ 3km Starting from 18 UTC IC 6 h assimilation at 1 h intervals Sat. obs (Liu and Xue 2008 MWR) 2120 UTC, from I.C. at 1800 UTC 7
Initiation of by colliding boundaries 6-7 h into forecast radar observation model forecast @0100, 13 June, 2002 8
Initiation of New Thunderstorms from Outflow Boundaries of Earlier Predicted Thunderstorms 5 hour forecast 9
Initiation of New Thunderstorms from Outflow Boundaries of Earlier Predicted Thunderstorms 6.5 hour forecast 10
Initiation of New Thunderstorms from Outflow Boundaries of Earlier Predicted Thunderstorms 6 hour forecast
Initiation of New Thunderstorms from Outflow Boundaries of Earlier Predicted Thunderstorms 7 hour forecast
10-minute time interval movie Hazardous Weather Testbed Spring Experiments (10 member 4 km ensemble and 2 km high-res forecast) 10-minute time interval movie (from 2007 experiment)
21 h ensemble forecasts valid at 18Z, May 24, 2007 CN N1 P1 N2 P2 PH5 Ph2 PH1 PH4 PH3 OBS 2 km forecast Probability of Z>35 dBz
Example for Spring 2008 HWT Experiment Assimilating 120+ WSR-88D Radars Initial Reflectivity Field without Radar Data With Radar Data Movie of 4 km forecasts with and without radar data
Other work and comments First ever study directly assimilating polarimetric radar data (Jung et al. 2008a,b,c, Monthly Weather Review) Working on assimilating clear air (refractivity and winds) data important for thunderstorm initiation Implemented a generalized cloud analysis package using NSSL mosaic reflectivity data in GSI for RR (Hu, Weygandt, Xue and Benjamin. 2007) Working on refining and testing Z and Vr assimilation in GSI for RR and HRRR Extensive work with Ensemble Kalman Filter data assimilation Assimilation of radar and other high-res data key to the short-range prediction of hazardous weather important to aviation. Ultimate goal: Explicit, thunderstorm-resolving, continuously cycled (~10 min intervals), ensemble-based probabilistic data assimilation and prediction, for strong winds, clouds and visibility, and turbulence from few minutes to several days - the FAA NextGEN goals also.