Mike Poole

I think a 2019 repeat is highly unlikely this year. My recollection is that there was one truly overriding driver that year - an extreme positive Indian Ocean Dipole - and the dire wet mild winter was well forecasted by the Met Office on the strength of that. This year the IOD is negative and expected to remain so, so that’s one thing in our favour!

Yes, good idea to get this thread up and running . It is looking like the strat vortex is getting off to a shaky start this year. I think how this progresses into early winter will be very interesting, although I remain to be convinced that a SSW is very likely this year. More on this in a minute, but I think a weak vortex going into winter is a distinct possibility which may give rise to cold chances in December, whatever happens thereafter. As far as a SSW is concerned, the combination of westerly QBO and La Niña does not bode particularly well: This figure shows QBO vs ENSO - and also two parameters relating to solar activity - I’m ignoring them for the moment as we are betwixt and between this year wrt the solar cycle. But for the middle part of the plot, QBO-W and La Niña 3 out of 14 years have had a SSW which isn’t great odds. One would think that the odds are slightly increased now, due to the weaker than average start to the strat vortex, all other things being equal.

I can’t see this as likely, to be honest, it goes against the predictions of all the main seasonal models for a dryer than average winter, and lots of blocking. From the modelling, while I do think there is a decent prospect of an early cold spell, I think the main hostage to fortune for coldies this winter would be a preponderance of mid-latitude blocking, not an Atlantic driven rain and wind fest. Which is why I will be looking for some better indication of the blocking setting up in favourable locations in the October and November updates of the seasonal models. Correct, there wasn’t one last year. Or ever looked likely to be, to be honest. I’m not sure we’re ‘due’ one, I don’t think it works like that! But they are slightly more common than average in Nina winters, I believe. On the other hand, the similarities in the models to last year may give the same result. We would probably need a SSW to avoid a milder second half of winter this year, I think. We will see…long way to go yet.

Yes, quite similar, but those 3 month average plots can hide a lot of detail, can’t they? Nonetheless, I will be making the same early prediction this year as last year. That of potential for a front loaded winter with cold for the UK in December, before the polar vortex can belatedly ramp up, and without prejudice of a SSW delivering something later into winter. It proved to be a bust last year, I’ll level with you, but I felt we were in with a good chance and the dice didn’t fall for us, while there was a chance before New Year. I think pretty much the same cards are on the table this year, surely we can draw a better hand? (Mixed metaphor overload, there). Well, in a way, the same cards would be on the table, it is a repeat La Niña year, like last year, so a double repeat, although I think the fact that it is a repeat Nina means more than the fact that it is the third in a row (although these are rare, data is short!). What has been clear throughout the other seasons for the last two years - both bookended by La Niña winters - is the total lack of significant weather from the Atlantic. And the current model output, and the seasonal models, seem keen on an Atlantic ridge going into autumn and so I am suggesting we go into winter this year with potential for blocking, and a subdued Atlantic. There is then a window for a cold pattern to take hold in December, before the polar vortex gets too wound up…it just needs luck this year that eluded us last year. ECM seasonal model, split into Dec, Jan and Feb: You can see what I mean about the 3 month average plots hiding useful stuff, because the pattern in December is very different to February, and I think the Atlantic ridge scenario in December gives a chance. Difficult to read too much into February (obviously it is different) because the runs will no doubt have ensemble members in which there has been a SSW and ones where there hasn’t and you have to factor that in. But that’s for later, I think the first chase will be a front loaded December evolution of the current blocked pattern, then maybe a SSW and all the drama that brings to follow…

I have to agree strongly with this, a single low is not necessarily a zonal pattern, lows can crop up in all sorts of ways on the models, like we’ve seen since the end of summer, a long lasting cut off low, ex-tropical storms threatening but eventually not delivering. Zonal means weather coming at us from the same latitude incessantly - and it does sometimes, but not this year, and actually not for most of the last two years. For us it means from the west on a jet stream straight across the Atlantic, but the term zonal could actually equally be used for a barrage of systems from the east like March 2018. What I find fascinating this autumn, is there is just no sign yet of a proper line of systems coming across the Atlantic, and there’s good reason to think that there won’t be with that super warm patch of SSTs in the North Atlantic, possibly a symptom not a cause of the current long lasting weather pattern: A ridge magnet, all other things being equal? Which is why it is not surprising to see the odd potent northerly turning up on the model output now, GFS 12z FI (T360) for example: I don’t see this general pattern shifting and giving way to normal autumn fayre, interesting synoptics to continue to show on the models would be my guess, with high pressure fairly close by for much of the time, but the position will vary…

Good idea to open this thread, we can have more of a discussion here than just commenting on the seasonal models in the MOD. I think it is potentially a very interesting winter season upcoming. I take a Bayesian approach to things connected with uncertainty, and with that in mind it is worth considering what the prior distribution for winter possibilities is for the UK in this warming world, before considering the specifics of this year, and how that might modify it. Like others, I take evidence from the last 10 years or so to reluctantly conclude a SSW is pretty much a requirement now for a memorable cold spell for the UK with one proviso. The proviso is specific synoptics that might arise to cause a front loaded winter - I think December 2009 and December 2010 are still recent enough to be relevant possibilities. For cold once the PV is established in the later part of the season, I think we are now reliant on a SSW - and when this goes well, a narrative evolution over weeks can be told either side of the SSW date that ends with a UK cold spell. But it doesn’t always go well, even if we get a SSW, there is only about a 2/3 chance of it causing cold in the UK, and even then it may occur too late. Now to this year. Since the SSW in early 2021, we have experienced very little weather dominated by the Atlantic, so that last winter I was giving more likelihood than usual to some kind of early season cold - but it didn’t happen (and there was no SSW either, or even a hint of one being likely if I remember). I think the same argument for some kind of December cold snap might be made this year, as there is no hint of a strong jet across the Atlantic into October now, and the seasonal models all seem to be showing a blocked pattern right into winter. I would like to see more evidence from the seasonal models as we approach the winter that the blocking predicted is in the right places, a worry is we might end up with some periods of mid latitude blocking that doesn’t really deliver cold to the UK. I think a raging Atlantic dominated winter is unlikely given the seasonal models, so one hopes we have a weak strat vortex - we need to see the precursors to this develop over the next few months. We have a west QBO, that is a negative. We have a negative IOD, that is a positive . We are likely in a strengthening La Niña, which gives the context for other indicators. And we have that anomalous hot patch in the north Atlantic, noted on SST anomaly plot: This looks quite unusual to me, might aid a ridge in the Atlantic, other things being equal. I would like to increasingly see low heights into Southern Europe on the seasonal model output. I am not of the view that years with warm weather in September gives a mild winter, one might as well ponder whether years with a change of Prime Minister precede mild winters! I also wonder if analogue years are very helpful given the changing climate, but I incline slightly to us being in a similar overall weather pattern to point to 1975/76. Well, pick the bones out of that - hopefully we’ll put some meat on them as we get closer.

The trend now seems clear once more with high pressure building in on the GFS 12z by T192, and then stays around until the end of the run! GEM has same by T192: GEFS mean at T240, 1025 hPa on the mean slap bang over the UK at day 10 - very clear signal: A retrospective word about this week - it was never an autumnal train of lows off the Atlantic for me, it was one rather odd cut off low which stuck about for a long time, then a possible incursion from ex-Danielle, then possibly another ex tropical storm in the mix, all in all a lot of uncertainty in model predictions, that uncertainty seems to have been resolved now in favour of a further settled period before any sign of Autumn proper, as often happens in September. RIP Queen Elizabeth II.

The August set of seasonal model runs looked very blocked to me also. I think there will be an interesting conversation to be had about the seasonal models this autumn about what they are showing for winter, and the nuances of how that changes over the next 3 months. Currently, a lot of blocking showing. So one might suggest a wet and windy Atlantic dominated winter is unlikely - just when you could argue we need one from a drought point of view. There is definite interest from a cold point of view. If I were looking for spoilers, I think that mid latitude blocking could prove to be the main hostage to fortune from what I’ve seen so far. But take the Atlantic almost out of the picture, the likelihood of blocking in a much more favourable place must also be significantly increased compared to ‘normal’. Long way off yet, be interesting to see if the model consensus this month remains blocked, when they have all reported…

Wasn’t this weekend supposed to be a rain fest? As that cut off low takes up residence close to the UK. High res models might help, here’s the cumulative precipitation to T48 (Sunday lunch) from HARMONIE 12z: Don’t forget your brolly! Little to show for it over most of the southern half of the country, the low set up camp too far west.

There have been the odd few that have shown up in FI, less than usual it seems to me. But that one there, and you can see why these features are a wildcard, but also why they often promote high pressure anomalies over the UK as that one would push up ahead of it. Not always, of course, they can come barrelling at us too!

So, this low then. 12z models at T168: UKMO has it weakest (this is still my bet), ICON makes a right meal of it, GFS somewhere in between, GEM tracks it furthest to our SW, bringing a clear SE flow to the UK, and it is no surprise seeing the heat following at T222: The UK direct hit not evident this afternoon on the ops. So far.

With so much uncertainty associated with the low affecting the UK over the next week, it doesn’t do to look too far ahead at the moment, so it is interesting to look at the ECM clusters T120-T168 timescale: Cluster 2 follows the ECM op with the low hanging out to the SW, while Cluster 3 is similar to the GFS op, and Cluster 1 similar to the GEM with a direct UK cement mixer… None of this should be viewed as the Atlantic breaking through or signs of a more Autumnal pattern taking hold in my view, it is just a stuck low milling about with very little apparently driving it. Worth a note on Cluster 4, which doesn’t have the green -NAO border, no clear high latitude blocking, and no low with any kind of identity. This is the kind of solution I was hanging my hat on a couple of days ago. Just 2 members in this cluster, so clearly not favoured now, but does suggest there is still the chance that this low is much less pronounced than most of the models are currently showing a week from now.

Yes, completely agree. Said as much yesterday. I think the models have had this low too deep and this afternoon’s ECM is likely to be leading the way in downgrading it.

Models in a mess at the moment in my view. None of the output looks to me to be particularly plausible after day 4. When the models start flirting with Greenland height rises, with nothing obvious driving it, you have to wonder. ECM mean, T240: My guess for what it is worth is the HLB is a red herring, the low encroaching to the NW of the UK is being overblown, and the result at this range will be a flabby ridge as the previous pattern tries to persist, but wanes just a little as we move into Autumn.

UKMO goes for a decent bank holiday weekend, here through to T168: The eventual lozenge of high pressure that stretches over the UK is not that high, 1020 hPa. So I expect still some uncertainty, and there is looking at GFS which is slightly weaker and messy over the same timeframe, here T168: GEM the best at T168, interestingly because the GEM has consistently tended to over egg the UK trough solutions during the last month or so - consistently wrong, I might add! We’ve seen some divergence in the op runs over the last few days, I think the outcome will be settled, not overly hot, for the bank holiday. Beyond that, my view hasn’t really changed for a while now, largely high pressure dominated into September, position of that - up for grabs.

Becoming a familiar picture, the resurgence of high pressure, here at the end of the ECM, and the heat marches towards the UK again…

Sea surface temperatures? They seem to be suggesting for midwinter, a return to La Niña or continuing ENSO neutral and a negative IOD. And right now, the North Atlantic looks very warm too:

Yes, I have said earlier that in terms of historical analogues from a drought point of view, this year resembles 1975, with the spectre of a repeat next year after a dry winter bringing the awful water shortages of 1976. The seasonal models are looking blocked indeed, here the mean 500 hPa height anomalies for December, January and February from the various model suites: Pick the bones out of that. Very little from the Atlantic there, the charts are mean over the month and over ensemble members - the first chart is the mean over all models. And, dry too, here precipitation anomaly for January, just the German model predicting wetter than average. Source, Copernicus C3S on Meteociel: Meteociel - Modèles de prévisions saisonnières de 1 à 6 mois de Copernicus C3S WWW.METEOCIEL.FR Meteociel propose les différents modèles de prévisions saisonnières mensuelles de ECMWF, Meteo-France, UKMO, NCEP, DWD, CMC, ECCC et JMA

Well, if GFS and ECM have it right tonight, by next Thursday T144 we are basically looking at an ex-breakdown. Edit, maybe a bit in haste there in the case of ECM for the north anyway, but gets to a proper reload of high pressure by T216:

ECM nosing the high pressure back in as soon as T192, stronger than on the 0z run: And with a partial backtrack from the GFS too, I definitely think the ECM looks the form horse here.

Yes, I agree, all we know for certain now is that pressure will drop into next week. Quite a difference between the GFS 6z and the ECM 0z but it isn’t the initial drop in pressure that is the issue, it is the next low arriving from the Atlantic, here’s GFS 6z at T186 and T234: GFS runs this into the UK. Compare ECM 0z at same times: High pressure puts up much more of a fight. I think ECM is more likely, and that the GFS is reverting to type in FI, we will see in future runs. So I would see next week as a low pressure transient between generally settled high pressure. And would like to see the evolution firm up and into the range of the high res models before guessing how much rain will fall, what type and where… Been a fascinating summer, and for the most part, with the spectre of La Niña hanging over it - I think that the equation La Niña = low AAM = Atlantic ridge UK trough in summer - needs a lot of provisos, and they were present in spades this year. But with the Australian BOM saying we are now ENSO neutral, have been for a while, with maybe La Niña retaking hold into winter, I wonder what the floor for AAM is going into September, should it fall from current values? Given September is often anticyclonic anyway, I think the chances of a 1976 style rain fest breakdown this year is now unlikely, and an extension to summer would be more likely.

Here we go! GFS has a hot continental flow at T186, week after next starting hot, might be some thundery activity as pressure drops over the UK. UKMO T168 has the pressure holding and brings the easterly flow.

Interesting…after last autumn, I think there is a very real chance that the Atlantic fails to make inroads this autumn again, and that would start to cause real issues regarding drought. On the upside, it leaves the tantalising prospect of a subdued Atlantic into winter, maybe that might deliver for coldies this year.

Lovely run from the GEM tonight, high pressure takes hold then gives way for the plume: …and real heat for the midlands south by the end.

GEM 12z goes for the hotter solution, here T192: High pressure to our east, and then brings the +20C isotherm into play, before a thundery breakdown for some T240: