I am looking for a dataset of ensemble forecasts and the associated true observed values. For example, a dataset containing $n$ ensemble forecasts of the temperature in 24 hours and the observed temperature 24 hours later. The ECMWF datasets look promising, but as a mathematician, I am not familiar with the terminology. Clicking on a dataset, this page for example, shows that "Analysis" and "Forecast" data is available. Are either of these the ensemble members or the observations? Downloading the data hasn't helped either.

  • 1
    $\begingroup$ can you edit the question and clarify what is meant by true values ? $\endgroup$
    – gansub
    Apr 3 '19 at 14:43
  • $\begingroup$ @gansub You're right, I should have been more clear. Hopefully fixed now. $\endgroup$
    – Hasnep
    Apr 3 '19 at 16:28
  • $\begingroup$ reanalysis is very close to observations. You cannot do observations of entire atmospheric column from surface to top of troposphere for entire world at high resolution. So remote sensing plus surface obs + model output is the optimal way. $\endgroup$
    – gansub
    Apr 3 '19 at 16:44
  • $\begingroup$ Generally model forecast skill is not measured by comparing model forecast with observations because the resolutions of surface obs and satellite are different. You do a difference between Forecast(current period) and analysis. (when equations is NOT integrated). Please take a look at this Q/A - earthscience.stackexchange.com/questions/2105/… $\endgroup$
    – gansub
    Apr 3 '19 at 16:48
  • $\begingroup$ Part of your question is a duplicate of this one - earthscience.stackexchange.com/questions/2105/…. So edit the question to ask the information you want. $\endgroup$
    – gansub
    Apr 3 '19 at 16:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.