It turns out that the circumpolar whirl over the Arctic is elliptical and hence has two nodes. One normally lies over Baffin Island; the other, northeastern Siberia (Yakutia? Chukotka?). This is contrasted with Antarctica's polar vortex, which is stabler with only one node over the Ross Ice Shelf; the best guess for why that is, is because the Southern Hemisphere has a landmass at the pole, and the Northern, an ocean.
The thing about an ellipse, though, is that it is less stable than a circle. Right now the Antarctic whirl looks to be between Ross and Graham Land. But look at that deep burgundy in Siberia!
Hypothesis time: Warm air is pushing the Siberian node of the Arctic whirl into the ocean; this in turn is displacing the Baffin Island node into eastern continental North America.
Hypothesis #2: The greater variability of the Arctic whirl (relative to the Antarctic's) has something to do with the ice age cycle of the past five million years. It is notable, for example, that the southernmost extent of the recent polar vortex broadly parallels, in latitude, the southernmost extent of glaciation historically associated with ice ages.
...And #3: Perhaps the proximate cause of the ice age cycle is the destabilization of the Arctic whirl? But if that is the case, what catalyzes it?