FI Charts ( T120+ onwards) Winter 2025/2026 **READ MOD NOTE IN FIRST

All of the ensemble probabilities detailed by the UKMO point to one thing, the Ireland is going to be at the weakest extremity of any outcome, and that cold muck is heavily favoured in a slack to moderate flow, above energetic instability and snowfall.

But one very high probability is that the Alpine states are smack in the centre of the peak conveyor and that if the Italians were complaining last week that the Winter Olympics site at Cortina had next to no natural snow, that they will have dangerously disruptive amounts of it when the Games kick off in a fortnight. So much in fact that inundation of transport links and instability of snow on the ski slopes could cause them more problems than a lack of snowfall ever would.

Appreciate that model run by model run analysis is a mugs game but the 6z GFS sees all the cold sink south rather than travel west

The AI models have been the most keen on an easterly. Surely the amount of failed easterlies would be part of their machine learning, so the idea that they would have difficulty modelling an unusal setup( easterlies) is a bit of a red herring. It would be good if systems over North America instead of spilling out into the Atlantic decided to travel north as that might aid a Greenland block down the line.

Are we really looking to Greenland for now re: the block? I feel like that is far down the line of obstacles to get through. Or are you specifically talking more long term?

It's a Scandi-Euro block building, which eventually may split and then we get a chance at a Scandi block developing on it's own that is far enough north to start sending cold air west. Then yeah, at that stage a push west by the heights into Greenland to prolong or sustain things would be nice.

We're not going to get any decent cold going off the Scandi-Euro block initially while it's closer to a single entity, even though it may steer in a SE/E flow for a time.

It's going to be the second trough disruption SE with the Scandi-Euro splitting and moving north, that delivers the goods. The AI models are keen on this scenario happening and bringing in an Easterly. Mid-Atlantic reinforcements thereafter would be nice.

ECM 6z at Day 6 was a nice upgrade on the positioning of the second low, just to our SE. Looked like a cert for sliding and bringing in an easterly quicker than it's AI counterpart. Ideally we aren't sitting waiting for multiple sliders as the more we delay, we all know how it goes.

GFS 12z as I type is another upgrade on the positioning of the low around Day 6, much further south than the previous run. EDIT: doesn't make it straight to cold from there, looks like it needs to wait for the next slider…

Hoping for big changes in coming days as commodity markets are running away with themselves. The price of gas is skyrocketing due to this. Europe will be in a bad place come spring if it stays this cold and a few nice snow pictures won't mean much then

Looks like the deep cold will be hitting Boston down to Pennsylvania and Washington

On the plus side,there doesn’t seem to be any low pressure system developing that would spirit that cold plunge into the Atlantic

There is high pressure to the west of California which in fact is blocking pacific systems from routing in to become the ‘northeaster’s’ that inject that cold into the warmer Atlantic and balloon the jet stream,so that’s a help for our easterly’s to reach Ireland?

While I'm certainly no expert in this space the company I work for have been working with predictive AI tech for the past 18 months and if the ECM and GFS teams are managing it properly, which I'm sure they are, then the issue we tend to see with broader AI hallucinations and their compounding impacts won't be a factor. The key thing to this predictive modelling is to constantly correct and retrain the models to work out the kinks. For example, model predicts scenario A based on the data given to it but scenario B actually happens. You tell the model that we actually had scenario B in this case, so it refactors it's prediction algorithm based on that.

I won't say too much here due to commercial sensitivities etc but our prediction models are incredibly accurate, even more than we thought they would be initially. Prior to this we had a team who were crunching numbers to try to produce this same data manually and they never came close.

Weather I suspect is a hell of a lot more complicated and variable than the data we're dealing with so it's going to much more challenging but I do think these AI models are going to far outpace the traditional weather forecasting approach in the near future.

No rollercoaster or pub run ☹️

Tbh I think the ai models have performed quite poorly in our part of the world this winter. The long term trends they seem to be able to sniff out quite remarkably during Summer doesn’t seem to apply during Winter. Maybe they will shine in the coming fortnight but I have my doubts about the evolution of how easily they are disrupting heights in the Atlantic southeastward towards Europe

JMA beautifully primed with that LP coming out of the Baltics. ⛄️

Sorry mods if this has gone off topic! So when it comes to these predictive models it won't come as any surprise that the more often a particular pattern appears in the real data then the more successful the AI model is likely to be in predicting that in future. So when it comes to our winter weather, if we were in our usual Atlantic onslaught then the AI models would do well at forecasting the development and track of low pressure systems because it has a wealth of data backing up its predictions. It will have seen the patterns and parameters influencing that a thousand times.

However with the cold scenarios we're seeing now it's likely going to be much less accurate in its forecast because there's much less historical data for it to model its forecast on. In these scenarios it will make a best guess based on the parameters it can match. I suspect that we'll see a drop in the verification of the AI models in the current setup we find ourselves in.

Decent upgrade on the ICON this morning. The Jet Stream profile is pretty good, LP diving to the continent and the really deep cold air well poised coming out of Scandinavia, the Balkans and Eastern Europe. This is what we want to see.