Have used this a few times now to get an overview of the GFS forecast for a future trip. I have to go to Berlin in a couple of days and the following is really useful to know now:
While I know it could change, it certainly shows it could be a quick trip. And a check on the return shows that the strong winds at altitude abate by Friday.
All just GFS model based but really is a great visualisation tool.
So far I have found it to be pretty accurate within 24 hours and not bad at 48.
I've used the Ogimet graph in my flight planning for IR training last months. It seem to be quite accurate.
Last friday I had a QT (Qualifying Test), but the weather that afternoon wasn't too good. There was a: TEMPO 21017G27KT 4500 SHRA TSRAGS SCT025CB BKN040 in the TAF. The Gramet showed CB's with tops up to FL170, so I cancelled the flight.
Makes me wonder: Obviously there is a relation between the vertical size of a cloud and the violence one can expect inside that cloud. What vertical size would still be OK to fly through in a light aircraft?
I would try to avoid CBs generally. The last lightning strike I got was 2003 during an approach into EDDF. Tops were FL090 and the clouds were sprinkled all over the place no chance to escape. Some kind of weather you often see in April. When CBs are isolated you can ask ATC for a heading that keeps you clear. If you have no onboard radar it is sometimes better to fly VFR staying below the cloudbase in order to see the downburst of heavy precipitation and avoid. Experts say the most critical temperature range in clouds is 0° to minus 10°C.
Have used OGIMET for a few years now and have found them to be generally very reliable.
to see the downburst of heavy precipitation and avoid
That is indeed sound advice, did that a few times too. But even was surprised once when it was really almost impossible to escape flying under one. Since there was little downpoor I thought I would be OK. Well, not really. The downdraft forced me to go full throttle, and pitch up to a speed of some 90 knots in my DA42. That did not feel comfortable so I asked for a descent.
CB's are nasty. I really don't like the embedded stuff.
Does anyone know how often the GFS data gets refreshed at Ogimet?
If I look at a GRAMET now (20130527 1300Z time) The graph displays "GFS 20130527 00 UTC" in the top right. Does that mean that the GRAMET is based on 13h old data?
The GFS model runs twice a day, at 00 UTC and 12 UTC and Ogimet fetches the data as it becomes available. The GFS model uses balloon ascents as input data which also usually happen twice a day.
13h old data is not old 
Ogimet fetches the data as it becomes available.
Are you sure? There is usually a down period at around 7 where new data is downloaded. You get the message 'no grib data'.
EDIT: Found that on Ogimet site: los productos están disponibles en las páginas Ogimet aproximadamente 9 horas mas tarde. They have it available about 9 hrs later.
