Presentation on theme: "The THORPEX Interactive Grand Global Ensemble (TIGGE) Richard Swinbank, Zoltan Toth and Philippe Bougeault, with thanks to the GIFS-TIGGE working group."— Presentation transcript:
The THORPEX Interactive Grand Global Ensemble (TIGGE) Richard Swinbank, Zoltan Toth and Philippe Bougeault, with thanks to the GIFS-TIGGE working group and other colleagues
TIGGE overview Introduction to TIGGE Objectives The TIGGE archive TIGGE-LAM Early results based on TIGGE data Intercomparison of ensemble forecasts Use of multi-model ensembles Towards the Global Interactive Forecast System Implementing THORPEX science The TIGGE User Workshop What users can learn about TIGGE this week
TIGGE A key component of THORPEX: a World Weather Research Programme to accelerate the improvements in the accuracy of 1- day to 2-week high-impact weather forecasts for the benefit of humanity The TIGGE project has developed a database of global ensemble forecasts collected in near real-time. Objectives: Enhance collaboration on ensemble prediction, both internationally and between operational centres & universities. Facilitate research on ensemble prediction methods, especially methods to combine ensembles and to correct systematic errors Enable evolution towards a prototype operational system, the “Global Interactive Forecast System” For more about TIGGE, see http://tigge.ecmwf.int
TIGGE infrastructure Data collected in near- real time (via internet) at central TIGGE data archives Could be implemented at relatively little cost Can handle current data volumes (approaching 300 GB/day) within available network and storage capabilities More information in Baudouin Raoult’s presentation tomorrow NCAR EPS 1EPS 2EPS n academicNHMSusers Predictability science Applications ECMWFCMA
Summary of TIGGE database Centre Ensemble members Output data resolution Forecast length Forecasts per day Fields (out of 73) Start date BOM331.50º x 1.50º10 day2553 Sep 07 CMA150.56º x 0.56º10 day26015 May 07 CMC211.00º x 1.00º16 day2563 Oct 07 CPTEC151.00º x 1.00º15 day2551 Feb 08 ECMWF51 N200 (Reduced Gaussian) N128 after day 10 15 day2701 Oct 06 JMA511.25º x 1.25º9 day1611 Oct 06 KMA171.00º x 1.00º10 day24628 Dec 07 Météo-France111.50º x 1.50º2.5 day16225 Oct 07 NCEP211.00º x 1.00º16 day4695 Mar 07 UKMO241.25º x 0.83º15 day2701 Oct 06
TIGGE-LAM: TIGGE for limited-area models The TIGGE-LAM panel, chaired by Tiziana Paccagnella (ARPA- SIM), supports the development of the Limited Area Model Ensemble Prediction System component of TIGGE. This Panel works in close coordination with the GIFS-TIGGE WG, in liaison with pre-existing LAM EPS initiatives and in coordination with the THORPEX regional committees. Aims: encourage a coordinated approach to LAM EPS for the mutual benefit of WMO members; facilitate the interoperability of the different modelling systems contributing to TIGGE; coordinate the archiving of limited-area ensemble forecasts – the three TIGGE archive centres have agreed to host a sub-set of high priority data; contribute to the definition of scientific issues related to LAM EPS and particularly to TIGGE-LAM and propose specific initiatives to address these issues to advance LAM EPS.
Early Results from TIGGE Ensemble forecasts Z500 Spaghetti Diagram Courtesy Mio Matsueda
Comparison of RPSS N. hem Z500 DJF07 (90c) JJA07 (84c) ECMWF UKMO NCEP MSC JMA CMA BMRC KMA ON07 (45c) AM07 (62c) from Park et al, 2008
Comparison of extra-tropical cyclone tracks Courtesy Lizzie Froude Ensemble mean error: Position (verified against ECMWF analyses) Ensemble mean error – Propagation speed Propagation speed bias
Comparison of multi-model ensemble with ECMWF Courtesy Mio Matsueda MGCE51 10 members from each of 5 centres MCGE168 All 12Z forecasts MCGE327 All forecasts for each day
Multi model M2 (ECMWF+UKMO): effect of bias correction (nbc vs bc) EC(nbc) UK(nbc) M2(EU,nbc) M2(EU,bc) RMSE RPSS M2(EU,nbc) M2(EU,bc) M2(EU,nbc) from Park et al, 2008 Z500 NHT850 Tropics
Multi-model combination: MSLP & T2m Brier skill-scores of multi-model combination of bias-corrected ECMWF, Met Office and NCEP forecasts. Mean Sea Level Pressure and 500 hPa (not shown) show only small benefits from combining different models 2m temperature shows more benefit (1-5 days improvement in lead time) More sophisticated weighting led to only small additional skill Courtesy Christine Johnson T2m > mean MSLP > mean T2m > 90 %
Similarity of ensembles D has small values (high similarity, blue) if the between model variance is small compared to the mean-square- error of the multi-model mean. mslptemp D+2 D+10 D+2 D+10
Applications: flood prediction Predictions with a 5-day lead-time for a flood event on River Jiu in Romania from Pappenberger et al, 2008
Towards the Global Interactive Forecast System (GIFS) The objective of the GIFS is to realise the benefits of THORPEX research by improving the delivery of improved forecasts of high- impact weather. As a first step, the GIFS-TIGGE working group set up a pilot project for the exchange of real-time tropical cyclone predictions using “Cyclone XML” format. Further plans for the development of GIFS will be presented by Zoltan Toth, and discussed later in the workshop.
The TIGGE User Workshop Overview of the TIGGE project The TIGGE archive and how to access it: Presentation during TIGGE-B session, Tuesday afternoon Hands-on user demonstrations in coffee/poster breaks Opportunity for you to tell us what you need Presentations of early results from TIGGE In TIGGE-A/B/C/D sessions and throughout the symposium. Plans for GIFS Presentation during TIGGE-C, Thursday afternoon Opportunity to get involved in developing improvements to operational forecasting of high-impact weather: focus group discussion during Thursday coffee/poster break.
Conclusions Since October 2006, the TIGGE archive has been accumulating regular ensemble forecasts from leading global NWP centres. The archive is a tremendous resource for the research community at large, and in particular the science working groups of THORPEX. TIGGE provides the basis for research and development projects targeted at specific applications of severe weather forecasts (health, energy, flood warning, wind storms, fire weather, etc…). Products to enhance the prediction of high-impact weather will form the basis of the development of the Global Interactive Forecast System. TIGGE website: http://tigge.ecmwf.int