|Fig. 1. Northern Hemispheric weather patterns, Winter 2013-14.|
|Fig. 2. Black shows interglacial ice sheet extent; dark gray, glacial, over the past five million years or so.|
This corresponds to the Northern Hemisphere arctic whirl being elliptical. While correlation does not imply causation, let me advance a hypothesis: the ice sheets during glacial maxima are where they are because that's where the arctic whirl happens to be. Fig. 2 implies how that could happen: Notice how the Rockies (particularly the Cascades) are the only significant mountain range within range of the main ice sheet. The mountains have important weather effects, and among other things, they push equatorial air north. Trace the Cascades' spine and you'll notice that the net effect is a pushing of this air* through interior Alaska and across the Bering Strait onto the Chukchi and Kamchatka peninsulas and into the Yakutia interior; in fact, this whole region appears surprisingly livable during glacial maxima.
This, in turn, pushes the arctic vortex's American half southeast, into the Hudson Bay region, and its Eurasian half west-northwest, up in front of it, until the ellipse is stretched far enough** that it gets pushed back north-northwest, into the Baltic or Barents Sea regions. At this point, the Pacific warmth, flowing deep into Siberia, has pushed the permanent polar system into such a position that Greenland lies halfway between the nodes; the jet stream now strengthens into a pattern that shears the arctic whirl against a dominant Pacific ridge, on one side, and the (relatively unchanged) Atlantic ridge on the other, between the Rockies to the west and the Siberian steppe to the east.
Conditions are set for the reassertion of an ice age.
*Note that for this pattern to hold there needs to be enough force to push air over, or equatorial air high enough to clear, the Denali range. While this might not have been believable in the not-too-distance past, the Rosby pattern holding over Alaska for the past year shows that, if the jet stream is weak enough, this can indeed be the case.
**How so? One hypothesis would be that the link between the vortex nodes breaks, sending the Eurasian half spinning out somewhere over Central Asia (the Gobi, the Taklamakan, the Uzbek, Kazakh, and Ukrainian steppes) and decaying into a frontal system without its polar connection, sort of like an anti-hurricane/cyclone/etc.; cold air from the American node then pushes out east-northeast, parallel to the Pacific thrust, until it runs into the jet stream again, whose motion shears it into reforming the vortex.
Another hypothesis would be that the link between the two nodes is surprisingly elastic, and instead of seeing one node collapse and reform, it would sort of "boing" back once that connection reaches its stress limit.
I think Hyp. a. is the better one because the pattern this winter (Fig. 1) almost looks like a knife splitting the nodes apart from each other, and it's far easier to see how a low can expand and fill a bounded area than it is trying to see how a "bouncing" motion would work.