Yes. The AI models are initiated off the physics based ones, which is why the physics based ones can't be discontinued for foreseeable future, accuracy regardless.

Friday 9 Jan - Friday 23 Jan

Whilst confidence in details is low, high pressure will probably sit close to the UK initially, with periods of relatively settled weather interspersed with more changeable conditions. Temperatures will likely be near or slightly below average during this time, with an enhanced risk of wintry hazards affecting some parts at times. 

Physics-based models still provide the foundational scientific consistency that is crucial for long-term climate modelling and understanding rare, extreme weather events that AI models may struggle with due to a lack of training data for those specific scenarios.

Suggests unusual setups will result in poorer outcomes, which reminds me of GFS having a 'mare in a sliding/disrupting low set up a few years ago where Darren was confidently siding with ECM and UKMO which disrupted the LPs and kept cold going, where GFS collapsed it all. The end result being a comprehensive 'win' for the Euros. I assume the AIs would have similar issues in any setup where there's less historical examples, particularly where the limited data set has a lot of different solutions.

In terms of the 6z GEFS it's another cold set, with the control showing a lengthy cold spell. The op is a fair bit less interesting but not without it's moments - it's another run where it's milder solution has only minor support until it blends back in deep in FI. 

Not long for the next exciting instalment via the 12zs....

https://www.meteociel.fr/modeles/geme_cartes.php?&ech=174&mode=0 

There's still cherry picking to be done from the GEFS I suspect, but Im scanning the far horizon at t384hrs hours again., and not feeling this potential cold spell at all, happy to be wrong of course.

That’s the 00Z update, I think — certainly it hasn’t refreshed on TWO yet. That said, I’ve always favoured a flatter pattern as the most likely scenario, and I’ve discussed this elsewhere on the website and in the videos. Despite that, at this stage I wouldn’t rule out a potent northerly quadrant feed developing, at least for a time. As I said earlier, let's see where we are after tomorrow's 12Z updates.

The optimist would say that things could get interesting a few frames later with WAA in the western Atlantic, but I would see it just as easily flattening.

Nice pub run by days 8-10.  One for the virtual scrapbook. 

Generally looking to see where we are after the next 12z updates!😀

Seems to be the catchphrase of the modern era.  Failure of high pressure to establish at northern latitudes for long enough.  

GFS Op - following on FAX, the N-ly becomes much stronger but stays out to the other side of the N Sea. Then the HP cell moves S-wards and provides a mostly NW-ly spell until Thu 8th when it takes up a blocking position mid Atlantic and Britain is under definite N-lies which on FRi 9th convert to NE-lies for the SE. Then the Hp moves back to cover Britain 1025mb Mon 12th.

ECM - (early posting so download only to T+240 = Tue 6th) as GFS Op to Thu 1st, then the Hp is further west and a low from the Arctic brings quite strong N-lies for most of Britain, not just the E coast., weakening but still there Tue 6th

GEM - (also to T+240) more like GFS than ECM

AIFS - LOndon, maxima dipping from 5C now to a cold spell 3C with frosty nights for the first week of Jan (briefly mild Sat 2nd), milder from end of that week, some rain from Tue 6th. Edinburgh, similar temp trend but colder with ice days possible in first week of Jan, more persistent rain from Wed 7th.

GEFS - dropping from present 5C above norm to 5C below Wed 31st, slowly recovering but staying below norm (in the N  & W similar but not as much below norm) , ens agreement quite good until Mon 5th, very small amounts of rain from Jan 1st, perhaps a little more after Sat 10th.  Modest snow figures for the far east.

It's surely no surprise to see the return of blocking in the models late last night and into this morning. It's also no surprise to see aircraft data is back to normal over North America. And it's no surprise that a certain someone hasn't posted this morning - odd, that! 😁

Incidentally the usually rock-steady ECM-46 flipped to show a much milder week 2 onwards compared to the past couple of weeks' worth of output. I don't doubt that it too will flip back in the next output, as the most recent run is from 0z yesterday.

https://charts.ecmwf.int/products/extended-anomaly-2t?base_time=202512260000&projection=opencharts_europe&valid_time=202601120000 

For now at least we're back where we were - GEFS has a 50% chance of -10 850s over London in the new year period and I'd say that's a good summary of where we are. It's more likely than not we'll see the blocking continue for a while, IMO, but the million dollar question - will deep cold make it across the UK - remains to be settled.

It continues to be a fascinating period of model watching.

https://www.wetterzentrale.de/en/show_diagrams.php?geoid=50893&model=ecm&var=2&run=0&lid=ENS&bw=1 

No, it's not just you.  It's been another garden path special since around December 16th or so.  This morning's variety of jam tomorrow is focussed on the north a la ECM and UKMO.  Though huge respect to the GEM which has been steadfast (and you won't see much of a mention as its milder), and the 0zGFS Op was relatively snow free, again the GEFS provide some cherries to be picked.  

The other flies in the ointment are that the uppers are more akin to cold rain/ hill snow so could all be a cold damp squibb unless we advect some deep cold air -8 850 uppers or below. 

In summary I do think its all to play for around 2,3,4 5 January at present, but some upgrades rather than flatter GEM-esque downgrades are needed if this is to be more than a virtual affair.

That the most important factor we need to get here, the 00z GFS ensembles for here show only down to -5c meaning less cold temperatures and mild nights I do not want it. 

I'd respectfully disagree, I actually think the models have been pretty accurate with this spell of weather. With the exception of Christmas day there was never any snow in the forecast pre New Years and the Christmas day snow did happen just 200 miles to far south for the South /South East contingent.

Since about the 18th December the models have largely (with the occasional exception) forecast a cool / cold and dry period of weather up until the new year which is exactly what we've had so far. 

The Met Office long range forecasts have echoed this and only start to mention the risk of 'wintry hazards' as we go into January which again, I think, echoes the model output of the last 10 days or so. 

https://www.theweatheroutlook.com/twodata/ecmhrens.aspx?chartname=ecmwfenspreciprate&chartlocation=london