1 / 32

ECMWF long range forecast systems

ECMWF long range forecast systems. Dr . Tim Stockdale European Centre for Medium-Range Weather Forecasts. Outline. Overview of System 4 Some recent research results EUROSIP multi-model forecasts Forecasts for JJA 2013. System 4 seasonal forecast model. IFS (atmosphere)

royce
Télécharger la présentation

ECMWF long range forecast systems

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. ECMWF long range forecast systems Dr. Tim Stockdale European Centre for Medium-Range Weather Forecasts

  2. Outline • Overview of System 4 • Some recent research results • EUROSIP multi-model forecasts • Forecasts for JJA 2013

  3. System 4 seasonal forecast model • IFS (atmosphere) • TL255L91 Cy36r4, 0.7 deg grid for physics (operational in Dec 2010) • Full stratosphere, enhanced stratospheric physics • Singular vectors from EPS system to perturb atmosphere initial conditions • Ocean currents coupled to atmosphere boundary layer calculations • NEMO (ocean) • Global ocean model, 1x1 resolution, 0.3 meridional near equator • NEMOVAR (3D-Var) analyses, newly developed. • Coupling • Fully coupled, no flux adjustments • Sea-ice based on sampling previous five years

  4. Reduced mean state errors T850 U50 S4 S3

  5. Tropospheric scores Spatially averaged grid-point temporal ACC One month lead Four month lead

  6. S4 extended hindcast set Scores are smoother and systematically higher with 51 member hindcasts

  7. S4 extended hindcast set Gain over S3 is now stronger and more robust

  8. More recent ENSO forecasts are better .... 1981-1995 1996-2010

  9. QBO System 4 30hPa System 3 50hPa

  10. Problematic ozone analyses

  11. Land surface Snow depth limits, 1st April

  12. Sea ice

  13. Tropical storm forecasts

  14. Recent Research

  15. QBO A big reduction in vertical diffusion, and a further tuning of non-orographic GWD, has given a big additional improvement in the QBO compared to S4. Period and downward penetration match observations Semi-annual oscillation still poorly represented

  16. QBO forecasts S3 S4 New

  17. NH winter forecasts 0.371 0.319

  18. NH winter forecasts Even with 101 members, ensemble mean signal not always well defined

  19. NH winter forecasts New version has weaker signal, more noise

  20. NH winter forecasts Forecast skill is above perfect model predictability limit

  21. EUROSIP • A European multi-model seasonal forecast system • Operational since 2005 • Data archive and real-time forecast products • Initial partners: ECMWF, Met Office, Météo-France • NCEP an Associate Partner; forecasts included since 2012 • Products released at 12Z on the 15th of each month • Aim is a high quality operational system • Data policy issues are always a factor in Europe

  22. Recent changes: variance scaling • Robust implementation • Limit to maximum scaling (1.4) • Weakened upscaling for very large anomalies • Improves every individual model • Improves consistency between models • Improves accuracy of multi-model ensemble mean

  23. Revised Nino plumes

  24. Error vsspread (uncalibrated)

  25. Calibrated p.d.f. • ENSO forecasts have good past performance data • We can calibrate forecast spread based on past performance • We can also allow varying weights for models • We have to be very careful not to overfit data at any point. • Represent forecast with a p.d.f. • This is the natural output of our calibration procedure • Easier visual interpretation by user • Calibration and combination in general case • Ideally apply similar techniques to all forecast values (T2m maps etc) • More difficult because less information on past (higher noise levels) • Hope to get there eventually ….. .

  26. Nino 3.4 plume and p.d.f.

  27. P.d.f. interpretation • P.d.f. based on past errors • The risk of a real-time forecast having a new category of error is not accounted for. E.g. Tambora volcanic eruption. • We plot 2% and 98%ile. Would not go beyond this in tails. • Risk of change in bias in real-time forecast relative to re-forecast. • Bayesian p.d.f. • Explicitly models uncertainty coming from errors in forecasting system • Two different systems will calculate different pdf’s – both are correct • Validation • Rank histograms show pdf’s are remarkably accurate (cross-validated) • Verifying different periods shows relative bias of different periods can distort pdf – sampling issue in our validation data.

  28. Forecasts for JJA 2013

  29. ECMWF forecast: ENSO Past performance

  30. EUROSIP forecast: ENSO Past performance

  31. ECMWF forecast: JJA 2mT Tercile probabilities ACC skill (1981-2010)

  32. ECMWF forecast: JJA precip Tercile probabilities ACC skill (1981-2010)

More Related