Here are some notes on my talk….
It aimed to explain why, in the era of computers, there were still uncertainties in weather forecasts. It turns out that the evolution of the forecast depends on the initial conditions given to the model and that on some occasions the evolution is incorrect. The answer is to compare successive runs of the same model, so for one model, each run for a few days and compare models so the GFS v. the ECMWF v. the Met Office. Looking for agreement all the time at the particulat date and time that you are interested in.
If (when) there is agreement between the GFS and Met Office for example, then the wind guru and Met Office detailed forecasts for say, Calshot, ought to be very similar and therefore credible. You could try to find a wind guru or met Office forecast spot near Poole.
The page http://old.wetterzentrale.de/pics/avnpanel1.html shows a set of 8 charts from the American GFS model and allows you to see the model’s evolution on one screen.
I have looked at the Surface Pressure Charts link which goes to netweather.tv. The comparison charts where the GFS and ECMWF forecasts are laid side by side are very useful as they can be compared directly and it can be seen when the two models diverge and therefore can’t be trusted. I haven’t seen this website before with the comparison charts. They look very useful. Thanks for pointing them out.
The ensemble page I was using is in the http://old.wetterzentrale.de/topkarten/fsavneur.html link. Go to the top left hand corner and slide down to “diagrams” and click. Then click on “GFS 1 (7d) + GFS ENS…..(Europa)… then select “London” and “Ensemble t850 und Nds”. The chart shows the temperature over London at 850 mb or around 5000ft. Look for the point at which the ensemble members diverge from each other and the operational run (the thicker green line). The bottom scale is the date. You can’t place any reliance after they have diverged.
Hope all this is helpful and not too confusing,