XueJishan Boulder 2006 11

Slide1: 

Nowcasting-Oriented Data Assimilation in GRAPES Briefing of GRAPES-SWIFT Jishan Xue1 Feng Yerong2 Zitong Chen3 1, State key Laboratory of Sever Weather, CAMS, CMA 2, Guangdong Provincial Observatory , GRMC, CMA 3, Guangzhou Institute of Tropical and Oceanic meteorology, CMA Contributors: Wan Qilin3, Chen Dehui1, Liu Yan1, Liu Hongya1

Outline: 

Outline Motivation System structure GRAPES and its High Resolution assimi.-pred. cycle Severe weather integrated forecast tools Some tests and real time running Unsolved issues and plan for further development

Motivation: 

Motivation Combine the high resolution NWP products ( GRAPES) and nowcasting technologies (SWIFT) to improve severe weather forecasts within 6 hours Provide a new tool for the weather services for Olympic Games 2008 Beijing Promote the further development of meso NWP technologies driven by expanded application of NWP

Global-Regional Assimilation and PrEdiction System: 

Global-Regional Assimilation and PrEdiction System Schematic description of GRAPES Chinese new generation NWP systems Variational data assimilation: 3DVar-available, 4DVar-being developed; Non-hydrostatic model with options of global and regional configurations Used in various applications ranging from severe weather events, general circulation modeling, environmental issues,……

System composition: 

System composition Data input Cycle of Hourly Assi. Fcst. 6 hour NWP Id. of Conv Storm ( QPE ) TREC Wind ( Movement Esti.) Extrapolation and Forecasting Display and Validation GRAPES Sever weather integrated forecast tool (SWIFT)

GRAPES cycle of hourly assimi.-fcst. and Prediction: 

GRAPES cycle of hourly assimi.-fcst. and Prediction Non-hydrostatic model with spatial res. 13km (1km finally) 3DVar for analysis Digital filter controlling noisy oscillation 1 hour time window Data ingested: Temp Synop Doppler Radar AWS AIRep Wind profiler Two test beds: Beijing area (for BO2008) Pearl river delta

Cycle of Hourly Assimilation and Forecast: 

Cycle of Hourly Assimilation and Forecast IDFI

Test of Hydrometeors initialization: 

Test of Hydrometeors initialization model modelvar qcqr.dat ISI adjustment IDFI nudg model postvar 3DV Radar, Satellite Parameters to be nudged : qc , qr, qi, qs, qh, qg (skipped in this presentation)

Severe Weather Integrated Forecast Tool: 

Severe Weather Integrated Forecast Tool Radar based approaches Automatically monitoring data inflow and quick response High res. (1:5000) GIS coupled Meso scale precipitation systems as the essential objective to detect and predict Main components: Storm cell (SC) identification and QPE Estimation of movement of the cells (TREC wind) Extrapolation of SC, QPF

Main components of SWIFT: 

Main components of SWIFT Currently available: Identification of SC (storm cell) Potential of intense convection（tornado , hail, thunderstorm） TREC wind (estimation of SC movement) SC Tracking and forecasting Quantitative precipitation estimation（QPE） Quantitative precipitation forecast (QPF) To be developed: Potential of lightning Forecasts of storm-genesis and dissipation Urban water logging forecast Debris flow forecast

Slide11: 

Rapid Update VS Rapid Response DataSource Radar Data Mosaic Processor Mosaic Output TREC QPE QPF TREC QPE QPF output Triggered upon data arrival 数据流 １．触发机制 ２．统一调度

Nowcasting Algorithms: 

Nowcasting Algorithms SC identification: SC defined by a radar echo with reflectivity reaching specified thresholds Correlation between storm cell and observed severe weather events. Estimation of movement Spatial consistency check Special treatment for missing data area Adjustment based on continuity hypothesis Tracking radar echo by correlation

Slide13: 

Redar reflectivity Data of AWS GRAPES output FY2C TREC Wind Adjust. Based on cons. Of mass Z-R relation OI QPE Corrected TREC Adv. extrapolation of echo 1h QPF Corre. Of TREC and model fcst. 2 and 3h QPF Genes. Disp. Adjust. Extrapolation and forecasting algorithms

Extrapolation and forecasting algorithms : 

Extrapolation and forecasting algorithms TREC winds are used for extrapolation within 1 hour TREC winds are also used to find the model levels on which the NWP wind fits the movement of CS ( 500hpa or higher in most cases ) Forecast of CS with weighting mean of NWP and TREC Statistical approach with NWP products as predictors 1 hour Weight of TREC Weight of NWP

Slide16: 

韶关 梅州 阳江 广州 广 东 省 气 象 局 Guangdong Meteorological Bureau 汕头 深圳 Pearl River Delta Trials Radar

Auto weather stations: 

Distribution of auto weather stations(>=700) Auto weather stations

200608130710 case: 

200608130710 case 200608130710每隔10分钟外推 200608130710的2小时外推 200608130710的3小时外推

Quantitative Precipitation Forecast: 

Quantitative Precipitation Forecast QPF200608130710 预报

Slide20: 

Radar Mosaic --STS Bilis

Slide22: 

1-h QPF

Slide23: 

1小时后的回波

Slide24: 

2-h QPF

Slide25: 

2小时后的回波

Slide26: 

3-h QPF

Slide27: 

3小时后的回波

Further development: 

Further development Radar and satellite data ingested in real time system Data quality control Combine well NWP products with nowcasting technologies

Slide29: 

The end Thank you for attention