We're starting to approach the main window of snow tonight between 10pm and 6am with the main chances in the valley after 2am.
My thoughts pretty much mirror the recent HRRR model runs for snow depth (below). The actual model snowfall output is misrepresentative of the actual accumulations with surface temperatures across most of the valley between 33-36F. For example, the 19z HRRR is showing kuchera snowfall of 1.2" in Knoxville with a 10:1 ratio showing 2.4" in Knoxville. This just seems like too much with surface temperatures above freezing and some rain mixing in reducing actual ratio significantly. The snow depth takes into account the melting at the surface and is more realistic with mostly just a few tenths of an inch of snow for most valley areas tonight.
Higher snowfall totals expected across higher elevations where surface temperatures will be a few degrees cooler. This trend is also shown in the SREF probs for 1"+ of snow with low probabilities in the valley.
View attachment 348943
View attachment 348944