Sounds like it might have been rime ice or icing conditions. What confuses me is that the pilots apparently didn't even call out any type of distress. This must have happened pretty quickly. That said, 2300 ft isn't all that great of an altitude, from which I understand that was where they were at, on approach.
Looking at their ob trends, I'm seeing something intersting. Right at the time they crashed the airport reported freezing rain, but issued a correction minutes later changing it to snow. The surface temperature was to high for it to be freezing rain and the original observation was most likely made by an automated sensor that was corrected by the observer when he noticed the mistake. It is possible that if there really was liquid precip, and that could have been a bad issue once you got 1,000 feet up or so. The surface of the aircraft may have been cold enough for that liquid to freeze on contact, causing the moderate to severe icing. I'm looking at the sounding taken 4 hours prior to that though, and i'm not seeing it capable of supporting liquid precip. The freezing layer was only about 500ft up, with no warm layer. The tools just weren't there for freezing rain, or even for liquid precip higher up. Turbulence wasn't an issue, surface visibility was 4 miles, the ceiling was at 800ft, but that should not have been any issue at all since Buffalo's airfield minimums are 200ft.
So in my opinion there had to be another cause.
