As Mr Murr points out, owing to the smaller number of vertical levels in the GFS model we won't be seeing any long-range effects of a SSW descending from the strat to the surface in either the GFS operational run or its ensemble until after it's already shown up in the ECM output.
I found a paper by some Chileans who compared the accuracy of the ECM, the GFS and the WRF at three different levels in the Andes (140 m, 2635 m and 5104 m):
Results for the three sites show that the lowest correlation and the highest errors occur at the surface. E C M W F model presents the best results at these levels for the two hours analyzed. This could be due to the fact that the E C M W F model has 91 vertical levels , compared to the 64 and 27 vertical levels of G F S and W R F models, respectively. Therefore, it can represent better the processes in the Planetary Boundary Layer (PBL). In relation to the middle and upper troposphere, the three models show good agreement.
http://www.astroscu.unam.mx/rmaa/RMxAC..41/PDF/RMxAC..41_lcortes.pdf
As it happens, ECM now has 137 vertical levels, while GFS still has only 64 (when they increased the horizontal resolution, they left the vertical resolution alone, so in the current situation it's not become more accurate - just more precisely wrong).
So, it is possible that the relatively exciting ECM op run we saw this morning is indeed the trendsetter, and we'll be waiting a while for the GFS to catch up.
2 miles west of Taunton, 32 m asl, where "milder air moving in from the west" becomes
SNOWMAGEDDON.
Well, two or three times a decade it does, anyway.