Something to break the pattern and maybe have a re-set. 2 weeks of an inversion gets us nowhere either. Maybe there a retrogression, but more often than not there isn't.
Probably are, but usually it’s just feed back from 500mb and blurry terrain. In my digging for CFSv2 data, one of the values of the CFS is that it’s something you can run for n years out and have it as a stable global model. As “dumb” as it is, it isn’t hydrostatic, so you see impacts from sea ice growth and snow cover too.