Singularity

Two distinct pairings here, UKM & ICON then GFS & GEM. Yet while GFS & GEM are still somewhat similar for a few days later, UKM & ICON drift apart with UKM significantly quicker: Never thought I'd be using ICON to illustrate why I'm skeptical of a UKM run but here we are!

Speaks volumes that it still tracked the low so far south, in my opinion. Even looks well capable of bringing the cold continental easterly in the low's wake across all areas. Still though, the model is either having an enormous case of zonally progressive bias with that low, or will soon make the other models look very poor indeed.

Just hit me that this fast possible (or I daresay probable?) bias from UKM with the low across the Azores was also there on the 00z run. Even compared to the ECM 00z it was way quicker. Either on for a massive win or humongous fail versus the other 'main' models (ECM, GFS, GEM).

Compared to the GFS and GEM 12z runs the 12z UKM is both much quicker to move the low crossing the Azores east and keener to peel off a low from the trough across the UK and send that west. The resulting situation at +144 is so very complicated... what happens next will depend on whether the little feature northwest of the main low his sufficient to out-muscle the Scandinavian high and pivot that around, sending it ENE. That high will also be attempting to head a bit west as the smaller northern low heads W or SW. I really don't feel inclined to trust it with the speed of the low across the Azores, though. For a while now we've been looking at a Wed-Thu closest pass to the UK but this run has it knocking on the door as of noon Mon! I'll give it more credence if ECM turns out even loosely similar later, though.

Nice of ICON to show us which way the cold air heads next when the low scoots east so quickly while staying south of the UK. It was critical on that run for the low to become less organised with secondary features preventing a well-rounded circulation that would have lifted the cold air out a bit more. Yes, those 'shortwaves' aren't always working against cold weather in the UK.

For me this is a significant unknown with respect to what goes on next week. GEFS have been upgraded since 2018 but how much of a difference has that made to their ability to resolve trough disruption? In times past, if there's going to be a classic southward trend to the northern limit of a low running into cold air, it's tended to begin when the event is at around 5 days lead time. So we've still got 2-3 days to wait in that regard - though any persistent ensemble trends in the meantime are worth noting. The EPS behaviour has my attention for that reason - if the 12z again shifts the boundary north, that will ring a distant but undeniably present alarm bell for me. I have mixed feelings about what we get in the runup to this possible event - on the one hand, terrible timing from a heating bills perspective, but on the other, a scenic run of weather with plentiful sunshine, lingering frosts and some icy developments to admire, with the chance of a bonus dusting of snow. At least it's not looking to be a week of very cold anticyclonic gloom. Some houses will benefit from the sunshine on sun-facing sides, providing a bit of natural heating.

Comparing the 00z and 06z GFS runs for this Sunday, it looks important what goes on over NE Canada with the exact shape of the cold high there. A narrower high drives some very cold air further down the western flank of the low by the Azores, which re-orientates it away from an initially zonally elongated form. That leads to the stall and subsequent reduced encroachment on the UK. Can't say what's more likely at this stage. In any case, I sense we could yet see the low by the Azores toned down at least a little, as that so often happens once intense systems move within about 5 days lead time. That would shift southward the most probable track of it - or any lows splitting away from it - across Europe.

So now we know the next modelling saga - a an easier-going one for many: How squashed south will the low across the UK be this weekend? GFS has its own trend toward more of that but other modelling hasn’t been so keen. Least of all ECM. Yet all of those kept all of the UK very cold at the surface through next week. What a situation to be in.

Ha… I know my onions… from shallots to good old reds Cheers! For anyone wondering why little disturbances tend to be missed at longer range, it’s to do with the way global scale models handle convection: They parameterise it to some extent or other (varies with the model). That means using equations that approximate cloud changes to the conditions at hand, which tends to work alright on the larger scale but not the small scale such as individual cell development. This effect seems most pronounced at the 3+ day lead times so there may be a reduction in how much parameterisation is used at the shorter lead times (I don’t know that ‘onion’ for sure…). Regional scale models attempt to resolve most or all of the convective processes, with varying degrees of success. Unfortunately it leads to more ways in which predictions can go awry, such as overdone core intensity or excessive concentration into too few cells (not enough spatial coverage).

Sure could. It happened in the far south to begin March 2018, resulting in a thin ice layer on top of the 10-18 cm of powder snow.

It's feasible that less cold air could arrive at the 850 hPa level (usually 1.5 km above mean sea level +- a few hundred metres; positively correlated with sea level pressure) while it stays about as cold at the surface. Such 'riding over the top' is not uncommon with entrenched cold airmasses. This can be spotted by model runs via comparing the surface flow (if such charts are available) with how the 850s move aloft. Quite a few runs lately have featured a spell where 850s slowly increase from the southwest while the surface flow is from somewhere between the northwest and east.

Yeah, confidence is low on where we'll be to begin next week. Take the UKM 00z (left) v 12z (right) for example: A small low dropping through the Greenland Sea (upper-left of image) kicks off a chain reaction that sends the cut-off high eastward to Scandinavia. Hence the situation at run's end looks a lot less finely balanced than it did on the 00z. GEM 12z is along similar lines to the GFS 00z and to a lesser extent 06z runs, in that it stops just short of feeding the low by the Azores some extra positive vorticity via a breakaway low Mon-Tue. The GFS 12z threw that right across to the big low which allowed it to sustain as a highly organised system for a bit longer. That a significant snow event was still modelled for central UK speaks for itself. Overall, another very interesting set of runs so far, with a wide variety of details for the UK that affect just how cold conditions are and how much snow falls (...which aren't always positively correlated). Clear that a lot of key moving parts are still to be resolved.

Fascinating that GFS has now come full-circle on what it does with the low near the Azores (the stall last featured on yesterday’s 00z run)…. but in that time the position and orientation of the trough across the UK has become significantly colder from our perspective. The UKM 12z was further west with that low near the Azores all the way to about +132 hours compared to its 00z run, before a large acceleration that meant it ended up being east of the 00z by +168 hours, which I did find a bit suspicious. Anyway, I’m now wondering whether that shift west was a nod to what GFS has now pulled out of the hat. I also wonder just how long it’s going to be before this hugely challenging system is reliably resolved. If it goes down to +72 hours there are going to be a lot of fatigued forum members... I’ve been there and it’s tough going. What I would advise is to take individual runs lightly, consider them mere suggestions for anything past 4 days from now. It really helps the mentality.

ECM's found the 'nose' that the GFS 12z added to the low by the Azores. Caused by a section of jet stream that heads across the Mediterranean, drawing some of the positive vorticity with it. For some reason the models tend not to pick up on such features at more than a week's lead time. It barely existed on the UKM 12z, mind. Keeping us on our toes!

A chance to see what's changed with ECM over the past 24 hours. Broader cold trough by the UK & low by the Azores a bit further west / slower. Latter shift could be significant if it's part of a trend going forward. 12z is much closer to GFS 12z than UKM 12z... let's see if that means a similar outcome down the line. It's not quite as far west & seems slightly broader, which might prove a critical difference - at least as far as the model's concerned.

00z ECM for reference v 12z UKM & GFS runs for noon Sunday. Note how much larger and more rounded the circulation of the low by the Azores is on the ECM run. Enormous warm sector too - quite strange really for a model that usually has a thing for producing overly compact systems. Looking for a more squashed appearance to the low on the 12z ECM run to put it at least close to UKM and GFS which both send the low into mainland Europe (extrapolating for UKM based on how it moves in the 24 hours up to +168, about 75% confidence).

Yet another twist as GFS brings back the scenario where the tropical low deepens & upscale so rapidly in the next 72 hours that it becomes entangled with the low to its west, stalling it. Its almost on par with the GFS runs up to & including yesterday’s 00z - just not quite as far west with the low as it stalls. This adds another at least 24 hours for the low to peak and start coming apart before any approach to Europe. The low nipping across S Greenland is there again on this run. Despite other det runs not showing any interest in that yet, will be interesting to see how it interacts with a further west and slower-moving low to its south.

Feels significant to me that the approach of the low from the Azores is tending to be delayed. That gives it more time to wind down & disrupt, barring any new low dropping in on it from the northwest GFS-style. Not everything a NWP model predicts is plausible, by the way. There are systematic errors which lead to feedbacks capable of producing unrealistically intense low pressure systems.

For me the focus of todays 12z runs will be how EPS behave, to gauge whether they are now partially lagging behind on a correction, or correctly keeping the door open on something more like yesterday’s ECM runs.

Yep, in fact as others have stated, if the flow is slack enough radiative cooling can lower surface temps overall away from the coasts, mainly by night but with some knock on effect by day. A near-coast complication can sometimes arise if there’s any onshore flow though - a warmer layer a little above the surface. Then there’s a risk of freezing rain if any precipitation occurs.

Without any further cold inflow it would gradually warm via vertical mixing above the seas then horizontal mixing aloft… but this works both ways so it’s possible for the vertical temperature gradient to become very small until very near the surface over sea and all the way down to the surface over land. Hence it can be not much above freezing early afternoon at the surface with 850s of -2°C. Keep a cold feed into the low and of course that can keep it as cold or even turn it colder overall. Either way, heat will be drawn out of the seas near the UK at an above average rate, which should soon bring SST anomalies down to the long term average (then below if the cold spell keeps going).

Only three of those have a sustained climb of 850s above -3°C & most are predominantly -4°C or belle even 15 days from now. Serval have transient milder incursions as lows track further north but in most of those cases cold air remains close to the north or north returns south. It’s a typical extended cold weather GEFS set, with reasonable scatter to account for possibilities (however remote). Also, 19 out of 20 times is a bit much, I’d day more like 9 out of 10 when there’s a typical zonal wind in the stratosphere above the N Atlantic. When the zonal flow is disrupted & displaced away as we have in the coming weeks, the odds of an extended cold spell are inherently increased across the 5+ day lead times, however the models behave on any given day.

Hands up if you figured that the first low moving into Europe from near the Azores would step up to keep the UK in colder air via becoming the dominant feature of the trough, drawing the low over the UK toward it as it passes south of us? That's the sudden deterministic run consensus between UKM, ECM, & GFS. GEM is more like a hybrid of Wed & Thu’s GFS runs which is colder than its previous few runs. Fascinating stuff. Not a done deal yet of course!

During the Atlantic hurricane season, I read quite a few comments from mets focused on the tropics talking about ECMWF having a problem with stringing out areas of vorticity leading to lots of small, weak disturbances that shouldn't be there. I'm not sure if that has any relevance outside the tropics but thought I'd put it out there just in case the model drastically changes its tune some point this weekend. Rare that we see such a stark difference between ECM/EPS & UKM/MOGREPS at the 4-6 day lead times - it will be very informative which way this goes.

That at least partly tropical low and how it interacts with areas of vorticity to its northeast is being handled more than a little differently by ECM compared to other modelling, with far more progressive results. Ones we’ve seen GFS drift away from today while UKM has shown no interest. Therefore I see it as a separate source of uncertainty to the trough arriving from the northeast and how far south that is. We have a way in which the cold spell might not quite manage snowfall in the lowland south (trough too far north), and a way in which it might get blown away (very progressive low from near Azores). ECM has both ways combine for the most extreme result. I find it hard to take seriously despite some EPS runs being similar (same model dynamics, after all), but only due to the progressiveness of the low from near the Azores. So on balance, I’m still seeing a continuation of cold through next weekend for most or all UK, but perhaps moderated a bit in the south, to much gnashing of snow seeker teeth. Still time for that trough from the NE to adjust back south but it’s getting tight now for southerners.