The AI models may be “trained” on previous data but any analyses I have read shows their verification stats are very impressive (and that is when compared with models such as the ECM).

I think however in our little island at the boundary of a potentially cold continent and mild ocean things can be a bit different to the global NWP picture and such verification can reflect overall patterns rather than relatively “micro scale” synoptics. Having said that, I would in no way dismiss the AI models as simply pattern matching given the significant development that has gone into them by the world’s leading NWP organisations. If we can see the changes in weather patterns and climate in recent decades, I doubt somehow that the global experts would simply have ignored that when developing these new tools - this will be an ongoing process but probably will still miss the more extreme situations that seem to be occurring recently. Models such as GraphCast and NeuralGCM may have a better handle on climate change related issues.

They are probably, like the “traditional” models, still being refined and improved and I suspect they will incorporate the traditional models in the relatively near future where there will not be an ECM run and an AI ECM one, rather the latter version’s benefits incorporated into the former and probably just named the ECM. I see the current situation as a wide ranging and public pilot study.

If there’s any learning to be done it is that Weather and climate are different things. It is a fact that the climate is getting warmer. We have had plenty of statistics from Brian to demonstrate the impacts to UK temperatures, and the warmer it gets, the harder it is to get cold spells of the type we once did: For those of us like me that are old enough and have a non-biased memory, winters were much colder and snowier in the 60s-80s than in the decades after, that is just fact. There will always be extreme weather events, cold, wind, rain and heat, and GW has increased these extreme events, but there are fewer cold extremes, and more heat Or rain and wind extremes. 

I did have a quick look at this last night after Neil and Darren's posts. ECM AI ens are trained using both the 1979-2017 data set and the 2016-2023 main model operational analyses, not sure if the main AI run is too or they run the main run just using older data set to get a feel for any biases. 

There is in general an acknowledged bias in some scenarios from some of the AI models due to the warming climate but not sure how much that applies to the ECM version, given the above (and it has in some extreme events modelled warmth better than the non AI version). 

In terms of model output the general pattern remains consistent in the mid term - HP establishing towards Greenland, likely but not guaranteed to bring cold air across Scotland and maybe northern England, much more uncertainty about whether that cold air makes it over the whole country.

I'd say the ECM ens favour countrywide cold for a couple of days at least but with a fair bit of uncertainty, GEFS much less certain and you'd err on it not happening from those. MOGREPS this morning showing signs of colder air IMBY although it's still at the edge of its range. 

Whilst there's plenty of interest in the models anyway, for fans of that kind of thing the ECM strat winds forecast has been creeping towards a possible reversal for a while now and latest set are I think the closest yet:-

https://charts.ecmwf.int/products/extended-zonal-mean-zonal-wind?area=nh&base_time=202601300000 

Usual caveats apply but it's been interesting looking at that signal appear and grow - take a look back through earlier runs and it wasn't really an option. For fans of early spring if it does happen and it has a trop response, maybe a chance of unseasonal early warmth or for oddballs like me, maybe a bit of late snow and cold just as the sun is really getting going (or it might do nothing or not happen)🤪

Edit - should caveat early March IMBY is essentially spring but far less so as you head up the country!

The old saying goes unless UKMO is on board, there will be no easterly.  We need to see that shift before getting interested, besides there is no cold air anyway. utter waste

I know it is not relevant to our weather, but to the best of my recollection there doesn't exactly seem to have been a lack of severe cold events in recent winters in the USA, including the current one. 

My understanding has always been from any dictionary I have ever consulted is that the term "climate" refers to what are considered to be the prevailing weather conditions in a particular region. I get that there is a distinction to be made between weather and climate, but I don't think it is correct for anyone to make out that one has nothing to do with the other (not that you have done that, for the avoidance of doubt).

On topic: As has been noted above, ECM 00z op run looks reasonable for a cold spell of some degree this morning but GFS less so. I'm sure those living in the areas affected by flooding in recent days will be praying that the ECM and GEM are proved to be correct about a weakening of the atlantic as we head further into February; they badly need a drier spell.

I vaguely remember reading  there was a Maunder minimum in this period and something about sun spots which helped reduce the jet energy. could be wrong though

Think that probably replies to your comments adequately 🙃

The Maunder minimum was over 300 years ago so probably didn’t play a role in the 2009-2011 period. It was however around the solay cycle minimum (between cycle 23 and 24) but I’m not sure there was any direct impact.

As Dave has clearly illustrated there is cold air around and I am of the kindest the chance of colder weather from the second week of February is as high as it has been this winter (and I don’t just mean for here). Whether that chance diminishes in the coming days remains to be seen, but as yet it has it has not done so.

Spot on. 👍

One of the things I find most annoying is when people pompously assert that “Britain has always had a maritime climate”, making it sound as though they are revealing some great, unappreciated truth. You don’t need to have studied at Reading to know that this is the case, or to understand what it means. However, the facts are that average temperatures have climbed in recent decades in the UK, and the odds have tipped in favour of hotter spells in the summer months and away from colder ones in the winter months. This is a basic truth which anyone trying to objectively consider the medium range weather prospects should always keep in mind.

https://www.theweatheroutlook.com/twocommunity/Posts/t24393-Breakout-thread 

You don’t need to “pompously assert “ that Britain has always had a maritime climate, it is just stating a fact!

The odds may have tipped in favour of fewer cold spells in winter and more hot spells in summer and that doesn’t need really to be asserted either as it is apparent from the data.

The medium range output is based on the current input and synoptics and should be viewed accordingly not just through a “climate change lens” as that will already be in the input data. The laws of thermodynamics mean that cold and warm air masses will have the same properties as they did decades ago when they are shown in the models, it’s just the former may be less likely to occur. When they do occur they can deliver big time, like the start of January showed here.

None of this really needs to be stated as the bottom line is that easterly weather has been dominating for a considerable time and cold air has been around it has just not really reached here. A few wobbles and trends have come and gone showing the colder air reaching our shores, and that continues to be the case. It may do, or it may not and the balance of probabilities is probably slightly in favour of colder conditions in the medium term. Will it last? Who knows and I’m not going to try to second guess any tweaks that may occur in coming days. Given the range of options in the ensembles it could once again be close but no cigar, but that’s by no means certain and anyone who suggests otherwise is simply guessing.

Well said, Michael. 👍

Odds on a reversal for second week of Feb are shortening 

Strat temps point to a split vortex

Lots to keep an eye on in February whether it give UK cold weather is the eternal question. Perhaps we could see a repeat of 2018…..or a March close to 2013. Better model watching than Feb last year!

The iPhone app has at least changed its endless run of 8s and 9s and gives a max of 3C on day 10 now. But then a couple of weeks ago it did the same, even having a max of 0C 10 days out, before flipping back to the milder picture. 

https://modeles3.meteociel.fr/modeles/ecmwf2/runs/2026013106/ecmwf-0-144.png?6 

Looks reasonably like a decent chance of a lengthy colder spell to me 😜

I suspect you may be broadly correct for YBY/MBY as it stands, assuming the colder uppers do make it as far as here, although in the ECM ens at least there's a cluster bringing a longer spell of colder uppers and the op for example has 6 days (ish) with sub 5c maxes for London. Not exceptional of course but chilly. 

Past performance doesn't always indicate future results but it does feel like the models are generally getting the balance between blocking and jet strength wrong in the mid to long term this winter. Last year I'd have said it was the opposite, with zonal spells toned down and some longer dry and chilly periods. 

As long as the colder signal is shown though I'll keep checking the charts and who knows, we may even get a few days respite from the rain.