The trees at that altitude would still be alright I guess, but the terrain will need some adjustment for reality, I'm afraid. I don't like all one-way slanted peaks either. One thing you can do is mask the twist and shear, and just have patchy areas lean over, perhaps have another mask for other peaks leaning the other way.
And the snow is way to rough. Did you use displ intersection? Without a fractal noise before compute terrain (and larger minimum size of terrain fractal, say a few meters), snow can be smoother, and size of compute terrain also influences smoothness. You can then add detail in the rock after the compute terrain, as that will be smoothed out completely by intersect.
Interesting to see where you take this.
