From DMX this afternoon: "Similar to the previous winter weather event, models are initializing poorly with 18z Sun temperatures from source region (Kansas)... The GFS and especially ECMWF, in particular, have initialized anywhere between 5 and 10 degrees too warm with these temperatures. The NAM and the Canadian models have initialized the coolest and the best, so have leaned towards them for hourly sfc temps from 12z Mon-00z Wed." Might be reasoning behind the differences in snow maps