The UK Met Office has often been accused of producing ridiculously pessimistic forecasts, but nobody has been able to come up with any substance.
I think it's because the forecasts are almost always right by the time they throw in the PROB30 TEMPOs, which covers all posibilities 
OTOH one doesn't need to know anything about weather forecasting to realise that a lot of weather is not forecastable meaningfully. If you have say several adjacent lows which are just hanging around and moving about a bit, nobody will be able to forecast the actual surface conditions. It must be painful to generate a TAF that describes something like that, and anybody who tries to read something into it is going to get a hard time. That's why TAFs must be read in conjunction with the MSLP charts which show the wider picture.
With the UKMO not releasing its 3D model publicly, GFS is the best we have if you want a forecast.
Whether the forecast will actually be any good, depends. The game I used to play with Meteoblue was to check the planned route every day on the few days preceeding the flight, to see how the forecast varies. Usually, the forecast varied so much it was obvious it was rubbish - that far ahead, anyway.
But I also think that site had some serious software bugs. Often it produced a plot which looked amazingly good or amazingly bad, and running it again "sorted it". I think it sometimes used the wrong initialisation, or maybe the wrong flight date/time.
It may be useful to take the MSLP charts over the next few days and run the Ogimet site against them, and also get the Univ. of Wyming ascents to get the actuals. This is quite laborious - any volunteers?
If you want the conditions for a flight shortly departing, you cannot beat the tafs, metars, and the satellite IR images, and a plane with enough performance to outclimb the altitude uncertainty in the latter.
BTW AFAIK Meteoblue was a PhD project for a student at some German university, and Ogimet is the University of Granada. The Ogimet site is well known and has some interesting sources e.g. it is the only way to get a METAR for LGST!
Just an idea.. What if we start doing some actual notes on cloud tops (time, position, altitude) especially when breaking out on top, and then after arrival correlate this data to sat imagery (SAT24 comes to mind). Maybe we can come up with some sort of a grey scale that gives us a better indication of where we can expect to be VFR on top. I know that on sat images you cannot see possible VFR conditions in between cloud layers, but such a layer would be a bonus.
It is a good idea, but another way to do this would be to have airliners or other commercial traffic sensing when they are in and out of cloud and transmitting this data back to a collection of base stations working together to produce a global map. Of course airlines and commercial traffic generally follow certain routes or airways, but with some computer modelling and with some interpolation, surely could produce more accurate and real time cloud top maps.
Why would they choose to do this? A noble idea but will never happen.
This is Europe not US. Does anyone file PIREPs? Do you expect to get at PIREP? This is nearly unknown in Europe, but very widespread in US.
Yep, we can learn a lot from the Americans. Now, what about us just starting this? I mean, what about us asking controllers whether they could be so kind to check with an approaching/departing airliner to give some cloud top info before you take off? First get a feel for whether a controller is busy or not..
I don't think we'll get arrested for asking, will we? At least you'd know what you would be launching into. It could make you decide for an alternative plan, like to stay below (VFR) and shoot up further along the route, where SKC is predicted.
Peter,
Meteoblue is out of Basle university if I am not mistaken. I never worked with it, but heard good things about it
Re PIREPS: We (Switzerland) use them a lot and also ask for them. Mostly ATC comes right back to us after a few minutes if we request cloud tops or other data. So why not try it.
The other bit I have been trying to get into the system here is to get a properly done top forecast produced on a regular basis, either within the GAMET or elsewhere. I think that many underestimate the importance of the cloud tops in these bulletins. However, if there is a customer demand for them, the met offices will eventually react to it.
One way to check out forecasted tops is via prevision temps, however I am not sure if those are openly accessible. I´ll have to investigate this once I am back (currently near GCLP).
Best regards Urs
I tried asking for cloud bottoms in the front coming into Muenster last week. Radar controller had no interest.
Cloud bases are usually no problem - they are in the Metars.
The fact that busy ATC will not get you airport wx is a separate issue. That is why I like the satellite phone solution to get a load of tafs and metars myself.
Europe doesn't have the US-style PIREP system. There is no "institutional" support for GA in Europe.
I don't think one can do a simple "shade of grey" to cloud top height conversion, because the shade of grey, in the images which all come from EUMETSAT, corresponds to a temperature, not altitude.
In reality this is not a massive issue because, in the context of a plane which can climb to say FL180, it is obvious with a bit of experience what is say FL120-150 (grey) and what is FL300 (white). But if your plane can only get up to say FL100-120 then you need to be a lot more accurate, and there isn't anything around at present that I know of.
In stable conditions one could use the skew-t from Univ of Wyoming from 0000Z for a flight departing say 0800Z and that works well, especially if you look at which way the air is flowing (easy) and look for a skew-t which is 8 hours "upwind"...
This is a good site for winds aloft etc. It was developed by a friend of mine, and uses GFS (of course).
I was looking for bottoms enroute not at an airport as I couldnt get over the top.
Who could possibly get you enroute bottoms, away from official wx observation points i.e. airports?
