Autumn changes are less clear. Evidence suggests the trend probably depends on the latitude, with warming and delayed autumn being greater at high latitudes:
doi:10.1111/j.1466-8238.2011.00675.x
But autumn in general shows high variability which masks the signal.
That paper by Choi et al. is consistent to other published work. They note a decrease in the length of "full snow season" but see no such trend in "core snow season". I've seen other work that agrees that the core of winter hasn't changed much yet but the tail seasons are a different story.
As for the cause of all this... getting at any definable cause in the highly coupled climate system is nearly impossible. Arctic sea surface temperatures / lack of sea ice is a obvious and tantalizing suspect.
doi:10.1038/nature09051
Of course as already discussed, the reduced sea ice is could also be playing a role in stabilizing our winter temperatures but weakening the polar vortex.
