That's doubling the mean temperature accuracy, Azelor. The same method can be used, but I worry about the increment in the amount of work needed.

On the other hand, I see your problem with the very large savanna area. I think the problem with that was the original temperature+rain to climate that we are trying. Adjusting from a "rain pattern" map to a "available humidity" one will probably yield a larger desert/steppe in the center of the continent and it will also make the regions closer to the poles less dry.
I am posting such a "available humidity" map in the original thread for the tutorial-to-be on this.