I don't know why I didn't think of this sooner. I ran a computational fluid dynamics (CFD) simulation on my laptop, and it confirms that a ski can generate vortical flows:
(click image for a full size version)
Note: the color contours on the ski surface are pressure coefficient "Cp". Zero Cp is the freestream pressure, negative Cp is low pressure, and positive Cp is high pressure.
I decided to model my Karmas, since I want to run another simulation to look at the rooster tail behind twin tip skis. I picked 80mph as a high racing speed, though a lower speed would probably result in a nicer vortex. A 10 degree angle of attack seemed like a decent landing angle from a long/shallow high speed jump, though higher angles would make a stronger vortex (to a point, beyond which it would burst). And I did not put in a sideslip angle, but that would probably be closer to what the photo shows and also make a stronger vortex on the windward side. So this is a fairly conservative case.
The next step in the simulation would normally be to refine the computational model to better capture the vortex flow, but I think it's good to stop here since the basic existence of the vortex has been proven. This doesn't mean the original photo is 100% real/accurate, but at least the vortical flow mechanism on a ski is now known to be physically reasonable. With sharper tips on racing skis, I would expect a stronger vortex compared to what the rounded tip Karma generates.
Some interesting tidbits: the ski has about 2 lb of lift and 2 lb of drag at this flow condition. There is about 8 ft-lb of pitching moment on the ski (tip wants to rotate up and back).