Singularity

I can’t get over how strange it is what the two GFS versions are now doing with the low that starts out south of the UK next week. A broad cyclonic arc, with a peak well north, when there’s extensive high latitude blocking? Even allowing for the highs being split it just seems utterly bizarre to me. This may yet be rendered irrelevant anyway, if ECM has things right. If it is due to stratosphere resolution then GFS simply won’t be able to resolve the outcome correctly for a whole longer yet. If.

When you look at it this way, UKM (middle) is a lot closer to ECM than GFS. More so than the 00z was, too. GFS pinches off the Pacific connection to the cross-polar bridge of high pressure a bit quicker & that has a massive impact going forward. MJO-wise, a big issue we face when considering upcoming behaviour is that diagnosis tools can be well off the mark sometimes. These days, the 'traditional' phase space plot that's used on the NOAA site seems to go awry very often, perhaps due to contradictory low frequency (La Nina) and high frequency (MJO) states (they interfere with one another when the MJO is hanging around the Pacific). I've found Mike Ventrice's own analysis to be more reliable overall this year. It shows the swift return to the Maritime Continent that A.M. has put forward. I wish we could see a version using ECMWF data!

Anybody know if ECM still has much of an edge in terms of number of vertical layers? Higher vertical resolution in the stratosphere would be a very plausible reason for the model being so resolutely different to GFS & GEM, to a lesser extent UKM. Reason being, it’s a troposphere-lower stratosphere-troposphere interaction that drives the polar vortex split. The more efficient and fluent that process is, the greater and more sustained the split.

Story of the 12z GFS & GFS // runs is an even slower focusing of the blocking high over Greenland. GEM is a little quicker than its 00z run but still at the slow end of the spectrum. Not sure about UKM as of +144. Nor can I really say whether to take the trend seriously or not. It sure makes the past few ECM runs stick out even more with their fast HLB focusing via the 'trigger low' setting up near Svalbard and interacting with a well-defined cross-polar flow.

GFS 18z is the outcome of moving toward the ECM 12z without fully committing to that drop-down of a cold low from the Arctic. To be fair to it, the 00z ECM wasn’t that far beyond & could have unfolded similarly past day 10 although probably not quite so far west with the main low relative to Europe.

That’s a great idea, I don’t have the setup to go that far yet but could build it… time allowing. Model shenanigans continuing apace, so many unknowns yet to be resolved.

Ah. Forgot to account for this being GFS - unrealistic compact low development from initially very weak positive vorticity. If we look past that, the shape of the high to our E and NE is much more conducive to supplying some cold air from the east... but this is now over-analysing for such a long lead time. Edit: It may not seem to aloft but the surface feed does turn out pretty cold come 8th Dec. Just to prove that I'm not writing nonsense

For the heck of it as I don't usually find much time for this these days... here's more detail on the key 06z v 12z GFS run differences for what happens next weekend to the west & southwest of the UK. The 06z slide a low SE from the S tip of Greenland to merge with another to the W of the UK. The 12z slides the low later and less emphatically, preventing that merger and allowing high pressure to bridge the gap. Meanwhile, the low to the SW of the UK is far less consolidated on the 12z run. This appears unrelated to what goes on to its north - a bit of internal variability from GFS - it just happens to be a less efficient baroclinic process on this run. Reality does tend to be toward the less efficient end of the spectrum (as many will attest to in their daily lives).

Just tried isolating runs of days since Jan 1958 (start of records) with sustained high AAM in Nov (as we've seen lately) & the result was... interesting. I'm now waiting to see what's produced for Dec. A less organised low to our southwest on the 12z GFS compared to 06z as of Sunday. There's also less of a break being maintained in the high pressure to our west. Both should mean less northward advance of milder air across the Low Countries going forward.

While we can't compare the available cold air to Nov-Dec 2010 (not even close until at least the end of this week), the high-latitude blocking sequence has enough similarities that I'm really interested in how well the various forecast models perform this time around. I'm relying on memory here so please don't take this as for sure how it went down, but I recall that in mid-Nov 2010, the forecast models did a decent job picking on the broad-scale change to a prolonged HLB setup, but struggled to resolve two main features: The nature of lows tracking well south of usual (just how far south and how well organised i.e. rounded as opposed to flattened shape) and, perhaps more importantly, the cut-back of the polar jet down the eastern flank of the blocking high. If I'm recalling this correctly, the magnitude of cold air advance west across Europe was underestimated, though I can't recall at what lead times. There has since been a dozen years of model upgrades (well, mostly upgrades...), during which we've had a few strong HLB episodes, of which 2018s was the most comparable but was driven by a split-vortex SSW, limiting relevance. As far as I'm aware, this upcoming event is the first strong directly troposphere-led one since 2010. So, who knows whether the old biases are still applicable now? Well, before long, we'll find out!

About time we had a deterministic run that has the west Euro low respond to a steep thermal gradient to its east & head that way. Such outcomes have been surprisingly sparse in the ensembles pre +10 days.

GFS & ECM develop a break in the high to the NNW of the UK which allows the Iberian low to drift north a bit. UKM & GEM keep the break further west which keeps that low further south. All but GFS are on the 'set up impressive blocking across Greenland' path, though. Experience tells me that something like UKM or GEM has tended to happen more often, but also that we're not talking about a strong skew in that direction; maybe 3 in 4 times it goes that way. Hoping (forlornly, I expect) for some alignment one way or the other in the 00z runs, as I have a site specific forecast to put together first thing.

I suppose it’s worth noting that the 00z GFS & ECM runs were quite a way short of the ensemble means for heights over Greenland 10 days from now. GFS 06z much more representative.

Slight differences in the polar jet have large impacts as soon as 8 days from now on which then carry forward via feedbacks. It’s a much more sensitive situation in that regard compared to typical westerly regime patterns.

Notably colder air making its way to our near-east by Friday in the 00z GFS & especially ECM runs. Enough to give me pause for thought. What if the low to our south corrects weaker and more squashed in shape as we’ve so often seen in similar situations past? That would shift the boundary with milder air southward and add more or a northerly component to the easterly flow. Something to watch out for while the retrogression stage is still at such long lead time (details not worth studying yet).

However it pans out, it will be educational to see how the various models fare with the relative strength of the Scandi-Russia & Greenland highs & now those evolve during next weekend into the week after. The theme of the 12z deterministic runs has been to keep the former high stronger for longer, making for a slower transition to a drop of cold air toward the UK from the N or NE... but let's not forget what the GFS 06z did & the parallel version has done so often in recent days. Faster outcomes remain on the table, although I sense they're the less likely ones as one of the most dependable modelling biases at the 7+ day lead times is toward carrying out pattern shifts too quickly. I can see this being a very long game indeed, a big test of the nerves. In my view the chances of a snow-sufficient cold spell affecting the UK by mid-Dec has edged up a notch today but I still wouldn't hang my hat on it.

Frontal rain reaches Ireland & fringes into the far west of Wales & Cornwall before stalling. Low would indeed slip a little south & gradually fill in (weaken) as it has no support from upstream (west or northwest). Space for high pressure to firm up around Iceland too.

Models usually struggle to resolve a cut-back polar jet on the eastern flank of a blocking high until it’s working at least a week’s lead time, so I’d not worry about the details of just how cold or not the air is at any given stage once the high has shifted to our N or NW. Also, localised cold pooling in slack areas tends to be better picked up once within 5 about days range. No denying it will take a lot to overcome the relative warmth of the North Sea and bring snowfall to lowland areas, though. Hence I’m excited only by the unusual weather patterns at this stage, not the surface outcome which could fall within a wide range of possibilities.

The broad scale progression has an unusually tight spread thanks to a combo of MJO forcing & the lower stratosphere offering little resistance. It may even be adjusted by the troposphere pattern to align with it, prolonging the HLB episode. So we have a rare level of confidence in a general retrogression of blocking high pressure toward Greenland, just a bit or uncertainty on how fast that happens. Trouble is, the blocking high looks to be so vast that it’s not physically possible for the whole thing to head west. It must split in two at some point with low pressure dropping through Scandinavia or thereabouts. Consequently we could well see a ‘milling about’ interlude such as the ECM 00z shows, should the split of the high place a lot of it to our east. It’s then a bigger ask, especially so early in the winter, for a lowland snow-conducive level of cold air to reach much of the UK. Still feasible, though. Or, there might be less of a high to the east enabling a swifter transition with more efficient cold air import. Or, the transition might be delayed but the high orientate such that we see something notably cold from the east for a time. Or, the eastern high could dominate and orientate such that cold air gives the UK a miss. All these options have featured in recent modelling & we can only guess which will actually unfold. We can see the wood but the trees are shrouded in fog.

My eyebrows have raised this evening as GFS has produced a rare 'bottom-up' stratospheric polar vortex split via a trough splitting against Greenland & ECM has something similar by Greenland on the same day (next Sat).. As far as I'm aware, this is what happened in late Nov 2010, so it's worth keeping an eye on future runs for any consistency regarding the trough split against Greenland.

Astonishing that GFS has reverted to not really sliding the low during Mon-Tue, which makes for a messier prelude to the next round of amplification that sets up the easterly. A higher chance of something happening that prevents that easterly ever really taking off. UKM seems resolute at a glance, but actually has become a little less clean with the Mon-Tue slider. Eyes to ECM later, which has been the most insistent on the slider being clean, passing the low fully southwest of the UK. As far as I'm concerned, the cleaner outcome for next week lowers the odds of a stalemate or Atlantic return scenario the week after, as high pressure is able to establish a stronghold as far west as the UK or even Iceland while we wait for whatever happens to the east to play out.

How about that, GFS found the Monday slider outcome that the ECM 12z went with. This alone shifts the balance strongly in favour of the easterly flow for Wed onward. Such a shift at short range does suggest that the ECM run was picking up on a crucial detail that only half the EPS caught onto… but this isn’t for certain, so I’ll await the 00z runs before deciding whether the first step is now set in stone. What happens afterward will still be open to much speculation regardless, as it will be sensitive to the precise orientation of the flow and whether it becomes slack enough upwind over the continent to enable cold pooling. Influence of mountain ranges must also be considered (foehn effect can hit hard as some recent ECM runs have illustrated).

ECM 12z sneaked in an extra trough disruption for next Mon, interesting at such short lead time given the model’s relative strength at that range. It serves to bolster the height rise through the UK on Tue which makes for a swifter, further west trough disruption on Thu. It’s all connected. GFS seems way out on a limb until you consider the EPS which are about 50/50 between it and ECM. Usually the deterministic has a valuable edge at 4 days lead time, this will be a good test of how reliable that is.

It seems we have a classic 'slidergate' to enjoy/suffer, depending on what you're looking for. UKM slides the little low cleanly and swiftly toward the Low countries to support a brisk, chilly easterly by Thu, while GFS has it mostly departing northward while a tiny bit breaks off to linger just east of the UK and interfere with the easterly flow there. We had a similar disagreement with respect to last weekend except that never looked to slide fully SE - but GFS did have to correct the position of the low a fair bit to the southeast. So it could be that the model still has its old Achille's heel with handling trough disruption... we'll see.

The important first step is now at a weeks lead time - the one that determines whether we start out with an easterly drift or westerlies prior to whatever advances north/west the high makes down the line. For what it’s worth I believe even the UKM 00z is just about slowing the polar jet sufficiently for the continental drift, judging by the low dangling down west of the Azores. The deep one west of Iceland therefore looks to be a stalling, amplifying feature rather than the sort that steamrollers the downstream pattern. Past experience tells me to wait a couple more days before calling it for the first step. Regardless of how that turns out, longer-term possibilities should keep us on our toes as the blocking high looks well placed to drive strong wave breaking on the polar vortex. I’ve yet to see enough of a wave-2 imprint in the stratosphere to really raise my interest on that aspect yet, but that *may* start to show up in due course for sometime 2nd half Dec. Much water to pass under the bridge yet.