GFS Op 0z: ignores the possibility of a storm on Friday and just keeps the SW-lies going between LP near Iceland (often 950mb) and HP in the Balkans (typically 1030mb). A lobe of the Icelandic LP dips S-wards 965mb N Ireland Fri 19th but rapidly rejoins the main LP to the north. A storm 970mb Wed 24th NW Scotland initiates a pattern change as it swings south into E Europe and pressure rises rapidly behind to cover Britain 1030mb Fri 26th.

ECM: similar to GFS at first, though the LPs passing Scotland occasionally link with LP over E Europe instead of just stirring up the Norwegian Sea. Then on Sun 21st the main LP regresses to Greenland while a weak depression drifts into Britain 985mb. The W-lies soon resume and in contrast to GFS bring a shallow LP 1005mb to Ireland.

AIFS: London maxima about 10C to Mon 22nd then 5C, one burst of rain Fri 19th. Edinburgh  becoming very mild (13C) Mon 15th, then about 8C for a week, then cooler; rain generally in w/b Sun 14th

GEFS: at T+180, a few perts have HP reaching towards Britain (e.g. P4, P14) but I think it's stretching a point to describe these as blocking especially when there are major storms in other perts (P8, P18). Line graphs show mild for early next week, then mean drifting down to norm (by Sat 27th op is very mild, control is quite cold). Rain  in S mostly from Mon 15th but in N at any time and in any case heavier in the west, chances of it drying up for the Christmas holiday itself at least in the SE. Inverness snow row figures for Christmas week like yesterday, no better than 5 or 6 /33.

For areas in Sussex the ban has been made more severe; from the SE Water website: "Some of our water resources in Sussex are in drought and we need to take further steps to protect supplies and the environment. Water resources are not quick to react to rainfall, it can take months in some cases. While we’ve had some recent rainfall, it’s not been anywhere near enough to counteract the very dry year we’ve experienced."

https://www.southeastwater.co.uk/about/updates/tubs-sussex/ 

Every single source for United Utilities (covers all of NW England up to the Scottish border) are at higher levels than last year, with all above average except 'Pennine Sources', which are a tiny bit below average.

We've had a wetter than average Sept & Nov, and an average Oct.

That's enough to refill all the reservoirs, lakes and rivers.

I think its algorithms need tweaking!!

It is scheduled for production next week. As I understand it, the GFS / Google GraphCast combination is running ahead of the European models in the verification stats.

Ah the beginnings of a 62-63 repeat 😂

The snag with these "AI" models - glorified pattern matching - is that they operate on what was, not what will be. Or, more simply, they're modelling the climate we once had.

It could be the AIGFS has fudged the issue, but as warming has been unequal across the planet it'd be a tough one to do with any great success.

I've saved the charts and will see how accurate they end up being on the big day. Probably not at all, but it's all a bit of fun!

Yes agree completely, would definitely back the actual models over the AI models, and unfortunately it's looking likely to be another very mild Xmas period. Hopefully January can deliver something cold as it has been the more reliable month for Winter cold in recent years.

Hmm, too early to be making any firm judgements about the Christmas period yet, IMO. GFS is only going as far as Boxing Day and we all know the futility of taking seriously any charts for so far ahead seriously.

https://images.meteociel.fr/im/95/23668/gens_20_1_336jpo9.png 

On a more serious note, it's good to see a decent showing of higher than usual pressure over Christmas on many of the GEFS members - that, allied with the ECM-46 ensemble means is a very good sign. Not necessarily in terms of snow, but it'd be refreshing just to not see temperatures well above average with a veil of cloud and drizzle!

You posted yesterday about a warm easterly that turned out to actually be a cold one (based on one selected member of an ensemble set) and now “looking likely to be another very mild Christmas period”, when the models show around average to slightly milder little above.

The reality is that Christmas is almost two weeks away and the only really useful guide at that range is the ensemble data which tend to show an average to slightly milder than average period. That of course could change to the very mild scenario you suggest is likely, but it is not what the majority of output shows currently.

I would say that a little milder than average, based on current output, is the more likely scenario as opposed to either “very mild” or “wintry”. I would also expect changes in the details over the next ten days’ output for the period over Christmas. I would also expect a few teasing straws to be grasped at to appear in random op runs in FI as has been the case for a few days.

I'd politely disagree.

Though I don't think they're performing at a high standard yet the suggestion that they don't work because they're modelling the climate we once had is incorrect. 

The algorithms are fed with all of the historical data and extrapolate based on that what might happen given the current state of the atmosphere at the time the model is run. That means they look for the most similar historical precedents to base that extrapolation on and that means pattern matching on temperature too.

One could argue that as the climate is warming rapidly then the data set of historically similar charts is too few but even given a warming trend of 0.3c  per decade that should give a wide set of precedents for most scenarios outside of the extremes.

Though the AI modelling is in its infancy I think it has enormous potential probably as a hybrid between the current approach and AI

That was based on the GFS 12z run which shows a very mild Christmas as does the ECM Opp. However I accept that In the GFS ensembles there is still a small chance of something colder showing. However my point is not to clutch at the straws as some will do, as often what looks like average or slightly mild ends up being 'very mild' and even average is still like 8c which I would consider mild in mid-winter.

It could also all change at this range as you point out but we have been waiting long enough now for some signs of something colder and there have been barely even any straws to clutch at so far.

I explicitly referred to the ensemble data given the timescales, however as you mention the ECM operational run, no it is not very mild for Christmas. For example the 25th, perhaps marginally milder than average in a few places but probably not.

https://weather.us/model-charts/euro/united-kingdom/max-temperature-6h/20251225-1200z.html 

Agreed anyone looking for finer details is barking up the wrong tree. A pattern change is potentially on the cards and with settled HP colder weather will come in time. Nice to see an easterly on Xmas day on ECM op run.

A beauty , but not quite cold enough. 

https://www.wetterzentrale.de/en/topkarten.php?map=1&model=ecm&var=1&run=12&time=336&lid=OP&h=0&mv=0&tr=24#mapref 

Good luck everyone, hope Santa brings what you want - cold, mild, semi-mild with a flake in it...

Yet the lack of deep cold over our corner of the world is telling, even if that is in FI. Best we can hope for is quiet seasonal type fare with frost and mist under any clear skies over the festive period. Would make a change from the usual storminess we tend to have at this time of the year.

LOL - but not entirely unsupported in some of the various output pert’s - still a healthy 20-30%  - so we can be ‘...dreamIEng of a white Christmas, just like the’ - odd  - ‘one we used to know’ (well 1970 at least) !! Ha ha ha.