When the GFS starts to show more typical mild westerlies before the rest, experience dictates that it is worth noting. I really think though that the models are struggling at the moment to capture a definate pattern, which could be down to the unpredictablity of that massive, explosive low over eastern Canada at the moment. 

ICON 6z at v120 hrs has shifted its interpretation about 100 miles west compared to the 0z incarnation - goes to show how even at 120 significant changes will happen that will make a difference to the weather experienced on the ground. Im not wasting my time looking beyond 120 at present.  There is a rare 2020s style cold spell with no end in sight up to the edge of the semi-reliable time frame.

This is brilliant, and clears up all the debate over accuracy etc. ECM is king, but Has only two runs per day, and not bad for the old UKMO for its limitations, and GFS useful for its regularity of updates in rapidly changing situations. 

I guess thats one way to stay positive! At least we dont have all the nonsense here that you get on the other channel

I know some here don't rate it it, but it had a good summer and the last stats I saw it was leading the pack at least at a Northern Hemisphere level.

I mentioned last week, and indeed have said the same things before on numerous occasions as others have, that whenever the models are faced with blocked set-ups of any kind, they tend to struggle with them to one degree or another. In my view, that is what is playing out again at the moment. 

Even if there is a breakdown of sorts next week, there have been suggestions from some op runs in FI, such as the GFS 00z this morning, of pressure rising again over Greenland further into January. Therefore, there is always a chance that even if things do break down next week, we may not necessarily be faced with another prolonged unsettled period like we had until early last week.

Outside of the usual areas exposed to northerly convection that little wintry feature moving down the country may be the best chances some people have to see some snow, although ECM says mostly rain. 

Here's UKV's take on where lying snow will be by T120, was expecting to see a little down the east coast but winds maybe have too much of a westerly element? Unlikely to be 100% accurate of course and I'd not be surprised to see more for higher ground in the SW if wind direction steers showers inland, interesting though to see the piling up on higher parts of NW Scotland.

As an aside and very much an opinion/perception based one but generally I find when the models go resolutely Atlantic breakdown they invariably don't back off but do sometimes overplay the initial signal somewhat, with a delay of a day or so before it properly replaces any cold air in situ. Doesn't always happen and either way it still warms up but will be interesting to see how this pattern pans out (assuming, he says optimistically some of the ECM ens solutions which keep it chilly don't pan out!). 

I’d be amazed if that chart proved to be accurate as it has about a foot of level snow here!

The differences at a relatively early stage of the 06z GFS op run compared with the 00z show that there is still quite a lot to be determined in the medium to long term I think. That doesn’t mean extended cold necessarily just that the picture is still evolving.

This issue of ‘false’ surface high pressure over Greenland is a wider one that just ECM. It used to be said it was related to the height of the Greenland plateau and ice sheet and adjustment of pressure to sea level.  The solution is to look at the surrounding pressure patterns and evidence of an upper high.  If there are just a tight bunch of isobars around a tiny HP cell over Greenland I just ignore them unless other evidence suggests it’s real.

For what it’s worth it has actually marginally (maybe 18 hours or so) delayed the milder weather moving in on this op run compared with the previous (00z op run). I’m not sure that is too significant though.

Another GEFS suite and another dip in the lowest t850 mean. For here it now bottoms out at -11.1°C on Sunday on the 06z set.

https://www.wetterzentrale.de/en/topkarten.php?map=20&model=gfs&var=1&run=6&time=144&lid=OP&h=0&mv=0&tr=6#mapref 

Has weak Greenie blocking developing from around T300 (a touch before maybe) and cold air for a chunk of the country.  Given there was a cluster of colder ens long term in the 00z ens it's not overly surprising I guess. 

That LP has a warm sector which just produces a little transient snow for most places south of the track ahead of the system and back edge sleet or snow as it clears through.  The centre needs to be 150-200 miles further south to deliver for the Midlands and Home Counties, I think.

Yes, it’s a bit chilly by the end of the run. https://www.wetterzentrale.de/maps/GFSOPUK06_384_48.png

 By itself it is essentially of zero use beyond saying “look what could happen”. With the wobbles in both directions today, as has been said already by others, I think the best approach is to look at the next 5 days or so and treat things beyond that with caution (be they milder or colder).

Consistency (especially in the case of an AI model) will not necessarily mean it is right. The king could lose its crown at the blink of an eye, especially when referring to 9 days away. The AIFS resolutely held on to a cold solution (not supported by other models) run after run at one stage last winter, very consistently, and was completely wrong. 

I'm doing my best to tempt fate 😁

Cold bedded in till 7th, possible snow almost anywhere, and then uncertainty in FI (which WILL change post 7th Jan). I can't complain about any of that. 

Yep, but as we are all far more grounded here now (and I even include myself in that to a small extent), we won't get sucked into believing FI charts. 

Fear the worst, hope for the best 👌