It's not easy. I get annoyed with some people who say it's simply looking at a model. Well, no, not really, because models aren't 100% accurate. You actually have to know the various elements of meteorology and determine whether the models actually make a lick of sense...in addition to looking at elements that models overlook (fresh snow cover, for example).
Just finished my last flight of the semester: a cross country from Grand Forks to Park Rapids to Bemidji and back into North Dakota. It was kinda weird to take off from Grand Forks, where there was no snow and the temp was around 40 (HOT by our standards at this time of year), to going to Minnesota and seeing snow on the ground, and temps barely at 30.