Yeah the models have gotten a lot better at identifying the marginal boundary layers....but at 3+ days out, it obviously can still be off. The March 7-8, 2013 storm was like that....that's the one where he just got firehosed from the east and slammed 12-24" in mostly a 15-18 hour span (but we were spitting snow showers for 24 hours before that)....model guidance had initially kept trying to warm the BL, though a couple model runs were showing around -1C at 950mb, and then they kept cooling until it was like -5C....obviously that is an extreme example....it can snow at -1C. But the forecasts were very conservative at first....showing very little inside 495 and even only a sloppy 3-6/4-8 for ORH hills.