Right now the hi-res snow footprint is much larger than the global footprint. The NAM snow footprint there on the 12k and 3k is 300+ miles north to south. The globals are about half or less of that.
3k is a more robust version of it's 12z run. Tons of mixing issues in the far eastern areas.
12k is decently SE of it's 12z run and has more qpf. Knox County is a battle ground on there with snow and freezing rain.
I wouldn't be shocked if the SE trend continues. The map is very similar to the SRF mean map.
East Tennessee looks really cold through the entire HRRR. There is probably downsloping showing up along the foothills. Looks like Knoxville is 29 at the end of it's run. Other models are showing more of a warm push though.
The Tennessee Georgia game was like the models. We started out way ahead. Then against all logic, we fell behind, now let's hope we come back at the end.
I foolishly thought we finally had a simple set up that would work for most of us. But it's Tennessee, so of course not. Models can nail snow in Texas or Oklahoma a month out. Here 6 hours out they struggle.
There's a low in Canada over the top. When it's weaker the Euro is further north. When it's stronger, further south. Yesterday at 12z it was 986 and most of Tennessee got a good run. Today it's 981 and half of Tennessee gets blanked.
We are approaching the "something's gotta give" stage of modeling. These tracks for the precip shield are well over 150 miles apart 36 hours before it's starting in the western areas. They'll either converge or one set is riding off the cliff into oblivion.