Wow, the first movie in the forum!
I think what is very complicated with your simulations is the triple junction between the three phases: The "diffusion lenth" there is near to 0, and I am not sure whether any model can exactly predict the speed of overgrowth with such high driving force! I guess, that this is the reason why you have this dependency on resolution... You will certainly need some form of calibration!
The most important factor, the wetting angle, is still not considered! If you ask Janin for her most recent optimized version, you can do the best possible with respect to that, but still I think it is a very tough task!
Bernd