Looking at 12h GEFS ENS, at   384h only 3 of ENS are really cold.Even if December  is overall NAO + there is still the opportunity for cooler intervals.

When is that 3 month outlook from METO going to appear?

Next week isn't looking like the mildfest it was a few days ago with a number of ensembles briefly tipping sub -5c at 850Hpa at the end of the month and again just a few days later..wetter too suggesting the jet stream is favouring a more southerly track than was previously modelled. 

Very interesting stuff from Ian Ferguson on the t'other side. Looks like GloSea5 is starting to pick up a cold spell starting around the 21st December. Hope for the Cold crew.

Ian Ferguson Netweather forum

A bit of caution required however, given growing signs of amplification and equally anticyclonic bias for S'rn areas with time (and dropping temperatures in EC-EPS as a result). Note too that latest GloSea5 has -ve temp anomalies in all members across most of NW Europe into w/c 21/12. So we need to watch for any continuity in this emergent (v tentative) signal.

GFS can't wait until then. 

Goes to show coldies should be optimistic about this December anything could happen.

I wonder if something is stirring ? .

Sensational GFS0z op run. Would bin most winter forecasts with immediate effect. Even those in the DE would be too mild! 

The ECM looks very different though ?

Fantasy chart of the season so far - should be framed and kept for posterity.

But not the ECM outlook ?!

Unsurprisingly, a massive outlier when it comes to forecast pressure.

Yes it'll be interesting how it pans out. GFS has been quite consistent showing HP the dominant feature longer term, but can soon change that far out.

At least some models are starting to look interesting.

The published verification stats for ECM may look better than those for GFS, but when it comes to spotting future pattern changes, I trust GFS a thousand times more than ECM, which always appears to miss them until the last moment.

Perhaps ECM is better at the finer detail up the rhe end of reliable, but is useless when it comes to broad brush issues further out, where GFS appears to be much better.