At this point as we get in range of the mesoscale models the 12km NAM, 3km NAM and RGEM will be the top 3 models to watch closely as they will pick up on CAD better. The RGEM has a cold bias at the end of it's range and the NAM sometimes has a dry bias, keep those in mind. Having said that the signs I'm seeing from all 3 suggest a pretty icy situation not only for Western NC but parts of SC and northern GA possibly if trends continue. The 18z GFS also came in colder and it has a pretty big warm bias I've noticed in past CAD events, probably one of the worst models for it.