1. WRF-based rapid updating cycling system of BMB(BJ-RUC) and its performance during the Olympic Games 2008 Min Chen, Shui-yong Fan, Jiqin Zhong Institute of Urban Meteorology, CMA, Beijing,100089

2. OUTLINE Introduction Performance Evaluation –Precipitation Verification –Proximity Sounding Verification Case Study Summary

3. Model Domain Configuration D1: –Grid distance: 27km –151×151×38 D2: –Grid distance: 9km –142×184×38 D3: –Grid distance: 3km –172×199×38 D1 D2 D3

4. Configuration WPS(v2.2)+WRFVAR(v2.1)+WRF(v2.2) Physics Package –WSM6 Microphysics Scheme; –Kain-Fritsch Cumulus Scheme (cu=99) (for both 27 and 9km domains), no cumulus scheme for 3km domain; –YSU PBL Scheme; –RRTM Longwave Radiation Scheme; –Goddard Shortwave Radiation Scheme; –Noah land-surface model; The Cycles run in cold start style at 1200UTC everyday and in warm start style for the rest cycles of the day. WRFVAR analysis and WRF forecasts are performed for 27, 9 and 3 km domains independently.

5. Distribution of the Data Assimilated by the BJ-RUC System SYNOP SOUND AMDAR METARGPS/PWSHIP Distribution of Ground Based GPS Stations in Beijing Area

6. 3DVAR Analysis Time 20:00 Wall-Clock Time (BJT) 23:00 21:50 02:00 00:50 05:00 03:50 08:00 06:50 11:00 09:50 14:00 12:50 17:00 15:50 20:00 18:50 20:0023:0002:0005:0008:0011:0014:0017:0020:0023:0002:0005:0008:0011:0014:0017:00 BJT 12 15 18 21 00 03 06 09 12 15 18 21 00 03 06 09 UTC GTS data, including SYNOP, conventional and intensive TEMP, AMDAR, SHIP, BUOY etc. Local AWS and ground-based GPS precipitable water in Beijing area Cold-start Run 00 and 03 UTC RUNs extend to 36 hours. Hot-start Runs Forecast Length: 24 hours Flow Chart of the BJ-RUC system (i)

7. Cold Start AVN/T213 Global Analyses and Forecasts GTS Data Local AWS and Ground-Based GPS PW in Beijing Area WRFVAR (Objective Analysis)WPS WRFV2.2 （ Forecast ） Variables for Station and Olympic Venues 2-m Temperature 10-m Wind 2-m Relative Humidity 1-hr Precipitation Diagnostic Composite Plots Surface Specific Humidity and Moisture Convergence Flux 0-2km Vertical Integrated Moisture Flux Convergence Contours of CAPE between -10~-30 ℃ levels, Height of 0 ℃ level and Vector Difference of Wind between Surface and 500hPa Forecasted DBZ on the -20 ℃ Isothermal Level and the Height of 0 ℃ Level 500, 750, 800hPa Wind, Geopotential Height and Temperature Contours Archived WRF Output Data Backgrounds for Warm Start Runs Provided by the 3-hr Forecasts of the Previous Cycle Lateral Boundary Generated from Global Data Background for Cold Start Runs Lateral Boundary Update Updated Lateral Boundary Uploaded to FDP Server to Support FDP Member Now-Casting Systems Diagnostics for Potential of Short-time Heavy Rainfall Diagnostics for Potential of Hail Storm Hot Start Conventional Products 1-hr, 3-hr and 12-hr Accumulated Precipitation Contours Sea-level Pressure and 2-m Temperature Contours 500, 750, 800hPa Wind and Relative Humidity Contours For References for Forecasters Flow Chart of the BJ-RUC system (ii)

8. Products Category Descriptions Short-time heavy rainfall potential forecast Composite map 1: Surface specific humidity, moisture convergence and wind Composite map 2: 0-2km moisture flux convergence Hail-storm potential forecast Composite map 1: CAPE of the column between -10C to -30C, wind vector difference between 500hPa and surface, and the height contour of 0C degree. Composite map 2: The forecasted dBZ at the iso-thermal layer of -20C, and the height of 0C degree. Surface Forecast Time series and text products of the forecasted precipitation, temperature, relative humidity and wind. Precipitation Forecast Hourly, per-3, per-12 hours precipitation forecasts. Conventional synoptic forecast in upper-air Composite map 1: Relative humidity and wind at 850hPa. Composite map 2: Relative humidity and wind at 700hPa. Composite map 3: Temperature, geo-potential height and wind at 700hPa. Composite map 4: Temperature, geo-potential height and wind at 500hPa. Composite map 5: Surface 2-m temperature and sea-level pressure. Real-time products from BJ-RUC system

9. Composite map: CAPE of the column between -10C to -30C, wind vector difference between 500hPa and surface, the height contour of 0C degree Composite map: The forecasted dBZ at the isothermal layer of -20C, and the height of 0C degree. Sample Products (1)----3-18hr forecast initiating from 2007080100 Indication of the potential of hailstormIndication of the strength and height of hailstorm

10. Sample Products (2)----3-18hr forecast initiating from 2007080100 Composite map: 0-2km moisture flux convergence Composite Map: Surface specific humidity, moisture convergence and wind Indication of the moisture convergence in low-levelIndication of the surface moisture condition and convergence

11. Performance Evaluation Data: –2008060100~2008093021(utc) –3-hr forecast interval –Only 3km domain results verified Verified against: –Precipitation Verification: 171 AWS stations in Beijing area –Objective Verification: sounding observations BJT (Beijing Time) = UTC + 8hr

12. Diurnal cycle of the accumulated 3-hr precipitation (mm) during the period from 20080601-0930 from AWS observations in Beijing area 00-03UTC (08-11BJT) 03-06UTC (11-14BJT) 06-09UTC (14-17BJT) 09-12UTC (17-20BJT) 12-15UTC (20-23BJT) 15-18UTC (23-02BJT) 18-21UTC (02-05BJT) 21-00UTC (05-08BJT)

13. Diagram of the 1-hr accumulated precipitation for 43 years in Beijing (Li, Yu and Wang, 2008)

14. Initial Time Valid Time (UTC) 8 runs with different initial time for each 3-hr time period For the thresholds beyond 1mm/3hr, each cycle had identical diurnal tendencies for precipitation forecasts, i.e. better skills in the afternoon and nighttime (09-00utc of the next day) than those in the morning. To forecast the convection occurring during nighttime, the runs initiating from 06,09,12 and 15 UTC are more significantly useful for forecasters than the other runs. Spin-up problem has been better resolved due to the updated cycling data assimilation style.

15. Evaluation of forecasted soundings in close proximity to convection Only one station (54511) verified 19 convective cases occurring in Beijing area during June to September, 2008 –Only 14 cases with positive CAPE value Verified against any available sounding observations, including conventional (00,12UTC)and intensive sounding(06,18UTC) 54511

16. Sounding Bias Analyses Bias –Temperature:  0.5C –Specific humidity:  0.25g/kg –Wind speed:  1m/s –similar magnitude to the measurement accuracy of the radiosonde observations Forecasted soundings with 3-hr and 6-hr leading time worse than analysis Evaluation of thermodynamics and vertical wind shear parameters derived from BJ- RUC analyses, 3-hr and 6-hr forecasts revealed same results. Analysis, 3-hr and 6-hr forecast all will be capable of indicating potential of convections. T QWind Speed

17. 2008081000 Observation Initial time ： 2008081000 t=0hr Initial time ： 2008080921 t=3hr Initial time ： 2008080918 t=6hr In this case, the soundings at 54511 derived from the analyses, 3-hr and 6-hr forecasts of BJ-RUC are quite similar to observation, demonstrating their ability of the system to indicate potential of convection with different leading time.

18. CASE1 ： 20080810

19. 06-09UTC 09-12UTC 12-15UTC 15-18UTC 18-21UTC 3-hr Accumulated Precipitation on 20080810 Initial time ： 2008081000Initial time ： 2008081003Initial time ： 2008081006Initial time ： 2008081009 The best forecast is the run initiating from 14BJT.

20. Initial time ： 2008081003Initial time ： 2008081006 Valid time:2008081009 Valid time:2008081010 Valid time:2008081011 Valid time:2008081012 Observed dBZ Forecasted dBZ

21. CASE 2 ： 20080814

22. Initial time ： 2008081403Initial time ： 2008081318 Observed 05-06UTC 06-07UTC 07-08UTC 08-09UTC Forecasted Precipitation Several cycles have forecasted the rainfall. Locations can’t be exactly match with the observed but the occurring time was well forecasted.

23. Observed Valid time:2008081406 Valid time:2008081407 Valid time:2008081408 Valid time:2008081409 Forecasted Initial time ： 2008081403Initial time ： 2008081318 预报的逐小时回波形状 与雷达拼图对应较好

24. Opening Ceremony-- 20080808 Initiating from 2008080809 Forecasted dBZ Forecasted PrcpObserved QPE 11-12UTC 13-14UTC 12-13UTC The Squall-line shape was forecasted. But the convection occurring in Fangshan was too weak to identify.

25. 10UTC 11UTC 12UTC Initiating from 2008082403Initiating from 2008082406 Closing Ceremony-- 20080824 Radar Mosaic The strong convective system in southwest of Beijing were forecasted 4- 6 hours ahead of their occurrence. But no impact on Beijing urban area and the closing ceremony --- same as what really happened later.

26. A failed Case---20080916, but a positive one. Initial time:2008091606 Initial time:2008091518 2008091521 2008091600 2008091603 3-hr accumulated observed precipitation Valid time: 16-19BJT 3-hr accumulated forecast precipitation Valid time: 16-19BJT The cycles initiating from 2008091518-2008091603 failed to forecast the convection occurring on 16-19BJT, 16 Sep. 2008. However, the cycle initiating from 2008091606 succeeded in forecasting the precipitation in the urban area only with the leading time of 1-2hr (almost now-casting). It can be found the assimilation of GPS-PWV at 2008091606 was the very crucial point for the forecasts of precipitation in the urban area!

27. Thanks for your attention!