For the snow lovers from about Seattle northward, don't write this off yet! The depth of the cold air over northern and central Canada is huge. It is interesting how easily the models develop the polar vortex westward from the main lobe over Hudson Bay and try to spill cold air over the mountains. Snow in western Washington will depend on a few things, how far southwest the cold gets into the sound (the dry should make it with Fraser outflow and the air mass aloft will likely be cooling significantly as heights fall with the approach of the trough from the NE) and the location and strength of low level development just offshore. Models have had a hard time determining if surface low pressure will develop off western WA or farther north. This diabatic development would occur as a result of the height falls combined with the warm sea surface. The southward location opens us up to overrunning early on Saturday or late Friday night - then into Saturday. The northern location amps up the upper troughing to the northwest, inhibits southward movement of cold air, and keeps more mild onshore flow going in western Washington. We are talking 100 mile errors in 48 hours (not bad from a model standpoint but crucial in the details here). Another point, the WRF seems too geostrophic at times (resulting in easterly flow instead of more northerly). I think this may be due to how it is handling the boundary layer. That could be why we are getting no precipitation showing up early in the WRF. The NAM has less of this issue and is more aggressive in bringing in the cold air, the more southward low placement, and the all day precipitation on Saturday (but has other issues). Maybe I'm wishcasting here, but I can see a 50-50 chance of snow on Saturday from about Seattle northward as the arctic boundary hangs up around Seattle, and it could persist into Saturday night as the pattern pulls up somewhat stationary for a while. The issue of overrunning on the other side of the event (later Saturday night or Sunday into early Monday) will depend on how much cold air gets in and the exact track of any southern stream lows. Negative tilt fronts, northward moving warm fronts and lows or triple points moving to the south are good for overrunning. Some arctic getting into western Washington helps the argument for overrunning if it occurs early on (we will know Saturday morning). I do find it amazing that the 12 km NAM is bringing the cold air farther southward than the GFS (lower resolution spectral model that allows cold air to usually spill too far southwestward). The GFS is internally consistent though seeing that the operational run is the warmest of the ensembles. The negative still is that this will likely be a marginal temperature event and that it is March (in a couple days). March 1, 1989 seems to be a good analog looking at historical upper air data. March 3-4 1960 somewhat fits the pattern. So it can happen, but is quite rare. This has been a weird year though! I make these observations looking at poor resolution pictures on my home computer (in some ways wish I had my work data). So my analysis may be all wet. Just my $0.02.