I agree Jan, real weather isn't feasible at this point in time. Maybe in a decade but probably later. My point is simply that the weather model (i.e. a simplification of reality) is a model and not a picture that has nothing to do with reality. For instance: It really bothers me that cumulus clouds are always just randomly distributed through the sky. No matter if the clouds are on the leeward side of a mountain, over a lake or whatever. This is not modelling after reality but just putting up some background pictures to make it look nice.
I would like to find cumulus clouds at the tops of mountains on the windward side where the moist air is and where the wind is mechanically pushing the moist air up forming clouds, not on the leeward side. There I want to see a blue gap (rain shadow I believe it is called? Sometimes extending 600! miles away from the mountains) I would like to find cumulus clouds just over the lake at the windward side and see a blue gap at the leeward side. This would make the it so much more realistic.
And yes, only real weather data as source of input for the weather model should be really nice, but I think not all would agree with this. Ipacs could make it optional: real weather input/slider input. The input data for the weather model is the same, only one is from data files, the other put in by sliders.
And yes, turbulence is not evenly distributed. That is my point. It would make flying in the sim much more challenging. Think of an approach during turbulent weather with a severe wind gradient. You really have to be on top of your game to cope with such conditions, not knowing if and when a wing will suddenly drop away or if you airspeed will suddenly decrease or increase as a result of wind gradient or a thermal that's changing your angle of attack.