Mike Poole

nick sussex Yes, the strat modelling has been diabolical all winter. I suspect it is to do with the strat vortex never getting properly going after the Canadian warming in December, the trop seems to have been dictating things since, even up to 10 hPa at times, and this has meant that the modelling has been unreliable because the trop evolution is unreliable - usually the strat vortex is easier for the good models to predict over a longer timescale but not this year. ECM 46 getting us ready for the next embarrassing strat climbdown (number 4 if I haven’t lost count), here’s the zonal winds forecast today on left, yesterday on right: You can see a number of trends here, first the early peak is quite a bit higher, the plunge afterwards has a lot less going for it, the mean reversal is much less. Anyone else reckon, come the day, it’ll end up just touching the zero line again, or maybe just a little below? You have to laugh, after all, it doesn’t matter now, it’s all too late for snow…but this winter, Christ on a bike! The modelling has been dire, the teleconnections have been dire, the strat predictions have been dire, and - the actual weather has been dire! I honestly think I preferred the +IOD nightmare of 2019/20 - never any false hope there, was there?

Frigid The last two SSWs showed up as big reversals on some model runs before getting watered down to the absolute marginal, and the 3rd Jan event disappeared completely. I can see this one being different, partly because it is too late to bring a proper cold spell, but we’ll have to wait and see.

Rain All Night Interesting zonal winds chart: It looks like the second technical SSW is guaranteed, and just as marginal as the first! And then the ECM 46 has the garden path laid out once more for the first week in March! We may run into some complications as to whether that might constitute a 3rd SSW, it isn’t if it is also the final warming (which it might be) and it isn’t if there are less than 20 consecutive days westerly winds since the previous one (which also might be the case).

ECM clusters T192-T240: Good support for the UK=>southern Europe trough at day 8 - cluster 4 pulls in a northeasterly (contains the op run) which make the absolute most of what’s on offer, it brings the trough west into Spain, cluster 1 has it furthest east. Moving on the strength of the ridge variable across the clusters T264+: Cluster 1 swings the trough into the Atlantic, bringing milder air, cluster 2 maintains it to our southeast, so remaining colder, cluster 3 pushes the Atlantic ridge through and replaces it with westerlies.

TURNED OUT NICE AGAIN It may do in terms of synoptics - the concern I have is there is no real cold anywhere near the UK to easily tap into. GFS managed to draw some in from the far north but that would need some luck to land.

Purga Bournemouth isn’t over there!

A perhaps surprisingly cold plunge on the GFS 12z. At T270, uppers are a quite respectable -6 or below pretty much country wide:

Clusters T192-T240 have pretty high confidence in a northerly by day 10: Just cluster 4 not going with it. However, on none of the clusters is it a long fetch northerly, trace the contours north, they soon veer to the west, so the ultimate source is not going to be very cold, earlier in winter the same might have delivered better, I guess.

First, I think for clarity, we need to distinguish between the Canadian warming (which happened in early December) and the minor (as it was eventually) warming in early January. The Canadian warming was responsible in my view for the strat vortex failing to form properly, these things used to be more common apparently, but we’ve not had one for 30 odd years. These early winter Canadian warmings are, I understand, not strong enough to be technical SSWs. This set in motion a weak strat vortex from the get go - there was a similar Canadian warming in November 1962, so there was understandable interest in what might happen, how this might progress. That then brings us to the next event, which is the one you ask the question about - was the early January minor warming a benefit or a hindrance? I think what was a hindrance was that it downgraded from the event that was initially modelled, which was a technical SSW and a favourable split as I recall, to a virtual non-event. Had the split SSW event happened as initially modelled, there was a real sense that it might have locked in a genuine cold spell for the UK in the heart of winter. But it didn’t, it was dropped by the models and the rest is history. I’d be interested in the view of @Cheshire Freeze because he called the fail at the time as a significant moment in the evolution of this winter. From then on, it seemed to me that the trop was leading the dance, as might be expected with a weak strat vortex, and the (just) technical SSW about 16th Jan looked in a sense to come up from the trop, rather than the usual top down warming event. And it looks like we might get a further event which may be a technical SSW but only just. None of this has ever been enough to assist in developing a UK cold spell, and the problem it seems to me throughout is that the strat vortex has rather been ‘neither nowt nor summat’ all winter! By which I mean it has been so weak as to allow the trop vortex to run unfettered, but has never actually been able to actively drive the trop flow the other way (easterly) at our latitudes, i.e. we’ve never had any significant reverse zonal winds in the strat that might downwell to do that. A weak strat vortex alone is apparently not enough to deliver a UK cold spell. Well, that’s my take, anyway - but I’ve only been following this as a complete amateur for about a decade, which seems like a while but is only maybe 7 SSWs, we learn from each one, but they are all different, and learning is a slow process!

ECM clusters this morning, T192-T240: I’m struck by how deep into southern Europe the trough gets by day 10, on clusters 1 to 4. What’s at stake is how strong the ridge upstream of it is, with cluster 2 the strongest, cluster 1 the weakest. Cluster 5 is rather different with a UK high, cluster 6 fails to put up much of a ridge - these are very minority solutions though (4 members each). T264+: I’m sensing this timeframe very up for grabs, as we enter a timeframe where I’m not sure what is forcing the pattern (if anything). Opportunities to pull down some colder air in the mix though.

Blessed Weather ECM 46 (albeit of 0z data) actually still falls just short of reversal on the mean. The difference between today’s zonal winds and 3 days ago laughable really. Like January, it won’t make any difference if it is a technical SSW by the barest of margins (where’ve I heard that before?), it isn’t going to be delivering any BFTE this year. Absolute codswallop from the models (all of them) re the strat evolution this year. First the SSW that wasn’t (which did for winter in the view of many), then the January SSW which was watered down to a gnat’s fart, and now this equally pathetic fiasco!

Just to correct this point. One of the main benefits of the clusters presentation is that they are not mean charts - i.e. the charts are representative members best representing each identified cluster. So they are physically real, and not some averaged mess. While the shading is indeed an anomaly, the height contours are not - they are the actual height contours of an individual run. So the best way to read them is, yes take the colours as a guide (if they are dark colours) but bear in mind they are an anomaly, but look in detail at the height contours which give the upper wind flows which are real. For example, here’s today’s T264+ chart: Look at the top right one, cluster 1 T360, here is the representative member at that time: You can see that the height contours on the cluster plot match the height shading on the Meteociel plot. The clusters are a probabilistic representation - whether any of the clusters at day 10-15 have actually included the weather that eventually verified is a question that needs to be answered probabilistically (my view is sometimes, but quite often this winter they have showed blocking with quite high probability that hasn’t happened). But that is a judgement on the verification of the EPS, not because part of the cluster chart shows a shaded anomaly, as I say, when reading them it is best to focus on the contours. Original post: https://community.netweather.tv/topic/99814-model-output-discussion-22nd-jan-2024-onwards/?do=findComment&comment=5037908

LRD I agree with you about anomaly charts - particularly ones which are both an anomaly, and averaged over many runs, or indeed many days as well - they can be misleading. For two reasons primarily, first you can’t infer wind direction around an average anomaly very well. And second, a high anomaly may still mean low pressure/heights, just not as low as climatology - this can be particularly misleading in areas such as Greenland when climatology would have the vortex plonked over there, so a slight high anomaly does not mean a Greenland high. But as I have explained, the averaging problem does not apply to the cluster charts, and the anomaly issue is only relevant to the shading.

Just to correct this point. One of the main benefits of the clusters presentation is that they are not mean charts - i.e. the charts are representative members best representing each identified cluster. So they are physically real, and not some averaged mess. While the shading is indeed an anomaly, the height contours are not - they are the actual height contours of an individual run. So the best way to read them is, yes take the colours as a guide (if they are dark colours) but bear in mind they are an anomaly, but look in detail at the height contours which give the upper wind flows which are real. For example, here’s today’s T264+ chart: Look at the top right one, cluster 1 T360, here is the representative member at that time: You can see that the height contours on the cluster plot match the height shading on the Meteociel plot. The clusters are a probabilistic representation - whether any of the clusters at day 10-15 have actually included the weather that eventually verified is a question that needs to be answered probabilistically (my view is sometimes, but quite often this winter they have showed blocking with quite high probability that hasn’t happened). But that is a judgement on the verification of the EPS, not because part of the cluster chart shows a shaded anomaly, as I say, when reading them it is best to focus on the contours.

Cheshire Freeze I think now the signal for March will go the way of the signal for the rest of winter. Down the toilet. Shocking performance by the seasonals this winter, and don’t get me started on the ECM 46!

Clusters this morning, T192-T240: All clusters bar 5 have a northerly component to the flow by T240. Cluster 2 looks the most pronounced with a well placed ridge and trough. T264+: Clusters 1 and 4 develop a decent block, but it is not perhaps as far north as we would need for a significant cold spell, but both have promise. Clusters 2 and 3 look like more transient ridging. Still no obvious horse to back at this range!

johncam Difficult to say, I would say the signal has actually increased for something cold in the day 10-13 timeframe, GFS and ECM both pointing that way. After that, I’m not sure, just going by the extended EPS, the signal looks to falter a bit compared to this morning’s. But there’s obviously more uncertainty at that range anyway. I noticed that the ECM 46 had backed away from the second dip in zonal winds again today. In fact, post-SSW the winds now look more likely to recover, which might suggest a reduced chance of cold. On the other hand, some recent GFS runs have shown a more favourable split SSW.

ECM clusters T192-T240: Cluster 1 which contains the op, and follows the line of thought that the GFS has had recently, has 29 members, so the northerly by day 10 is definitely on the table. Cluster 2 remains flat. T264+: Cluster 1 with 19 members continues the blocked theme, although it diffuses by day 15. Cluster 3 also has some interest, but looks like the ridge is a toppler. Cluster 2 is westerly, and so is cluster 4 with heights to the south.

ECM just hinting at a change by T240, similar to what we’ve seen on recent GFS runs that go past that point:

GFS 0z is a good run for cold, and it is worth noting the zonal winds profile: Very much going with the double dip in zonal winds, with a deeper plunge about a week after the initial SSW - this was evident on last night’s ECM 46 as a strengthening signal and is worth watching.

ECM clusters this morning, with some earlier interest (surprisingly) in the T192-T240 timeframe: Clusters 1, 2 and 6 all show a ridge sufficiently to the west of the UK to bring in cold from the north by day 10. Clusters 3 and 5 edging that way too, just cluster 4 has the ridge initially through the UK and a mild flow. T264+: Cluster 1 really set apart with the big high lat block over Greenland and northerly winds, trough in Europe. The others are a bit nondescript by comparison, 5 and 6 showing some interest in a northerly too.

It does usually, when it’s on the back of a strong strat vortex, but there aren’t any significant westerlies there to flush down in this case, because of the weak strat vortex all winter, and of course the previous SSW in January. For example, GFS 12z zonal winds: If there were strong westerlies to flush you’d see a strong pink anomaly (right chart) in the lower strat moving down with time with the blue anomaly from the reversal above it. It isn’t there. Original post: https://community.netweather.tv/topic/99814-model-output-discussion-22nd-jan-2024-onwards/?do=findComment&comment=5036863

Where’s the pub run going? T276: About time the GFS turned out a decent run…

summer blizzard Big change on today’s plot compared to that: Much increased cluster going for the second dip a week after the first, and a big reduction in the mean extending over several weeks.

daz_4 Nah! It’s still autumn!