Cold Winter Night

1962-1963 had a Canadian Warming in late November, but that one was, like most CW's, no formal SSW with a 10 hPa reversal. Later on, at the end of January, there was a proper SSW.

Indeed the 192-240h cluster 4 resembles the GFS Op anomaly charts most. The De Bilt 2m plume is actually colder for Day 10-15 than the 0z EPS had it. A decent number of members going for serious cold in FI. Colder than the next few days.

UKMO 168h. Highs to the North not inclined to disappear yet. Atlantic Lows don't get past the British isles as they meet cold Europe, and no purple Greenland. Potentially milder than the upcoming weekend? Yes. But no Atlantic onslaught like the last few months, and lots of perspective later on.

That's what I was thinking. And if anything, the TPV is retreating West, with increasing heights near West-Greenland and Labrador. The 168h should be interesting to see what happens next.

Here are this morning's ECM SST output charts for 0h, 240h and 360h. The relatively shallow North Sea and the Channel are quite a bit cooler after 10-15 days. Original post: https://community.netweather.tv/topic/99428-model-output-discussion-mid-autumn/?do=findComment&comment=4958089

Here are this morning's ECM SST output charts for 0h, 240h and 360h. The relatively shallow North Sea and the Channel are quite a bit cooler after 10-15 days.

EPS 12z has three clusters between 120h and 168h that all move towards that NAO-/Euro Low combination. After that, from 192h to 240h, there are just two clusters, one with 35 members that contains the Operational, Negative NAO, and a smaller one with 16 members that reduces the amplification to our NW and has a more mobile look, but it keeps our area under higher than average pressure. I think enough to keep temperatures generally low, as indicated on the temperature graphs shown in replies above this one.

I'm quite surprised how rapidly the ensemble switched from little interest in Neg. NAO in previous runs to all Neg. NAO clusters tonight. Normally such a switch would be reason for a lot of caution, but given the developments we see on all the models, we can have a bit more confidence in this potential outcome, I think.

The EPS clusters strongly support High Latitude Blocking + Euro Low days 5-7. The Operational is definitely not on its own. Thereafter, all clusters now assigned to the Negative NAO category at Day 10. I'd say a much better ensemble than before.

Great setup, both short-term and long-term, putting the SPV under stress. I must say, it's fascinating to watch all this unfold, realising it's only November 22! And while I was away, I missed 10+ pages of comments...

What strikes me about the modelled developments is that despite seemingly massive differences between models and runs, they do seem to agree on lowering heights over Central and Eastern Europe, probably associated with the cold that has already taken hold of the Scandinavian/Baltic region. Many runs send a trough even as far South as the Italian Mediterranean. That alone is encouraging to see. It will help establish budding heights to our North and will form a readily available cold pool as well.

UKMO 168h Still lots of cold sent deep South on that trough into Europe. Not something we've seen a lot in the past few years. By the way, along the axis N-America-Siberia, it's not massively different from GEM at 168h, which does end up with a High between Scandinavia and Iceland.

Would be nice to have these as the first charts of the meteorological winter: One can dream! I know, I know, it's as FI as can be, but I do notice a lot of available cold around, even at T-0h. At least there is something we can potentially tap in to.

The most up-to-date ECMWF Forecast User Guide answers these questions: "The horizontal and vertical resolutions of the medium range ensemble (ENS) became the same as for the HRES after the upgrade to IFS (Cy48r1) introduced in June 2023. The HRES forecast and the unperturbed ENS control forecast are meteorologically equivalent and equally skilful on average. They have the same physical and dynamical representation of the atmosphere and use the same parameterisation of sub-gridscale effects. However, they can diverge on a day-to-day basis due to small technical differences and the chaotic nature of the atmosphere. The HRES will continue for the time being for ease of use by customers and users." So yes, they are basically identical, but they still run separately. The small differences in outcome are the result of the element of atmospheric chaos that is integrated in the model, and small technical differences. The separate Operational run still exists as a separate thing, because all of us, pro forecasters and amateurs alike, are so used to it. That is of course very considerate of them, but I assume we can expect it to be discarded in future updates.... Section 2.1.2.4 HRES - High Resolution Forecasts - Forecast User Guide - ECMWF Confluence Wiki CONFLUENCE.ECMWF.INT Original post: https://community.netweather.tv/topic/99428-model-output-discussion-mid-autumn/?do=findComment&comment=4951241

Absolutely. With an ECM Control that is at the same skill level as the Operational already was, it means that it is also equally superior to the UKMO, GFS, GEM operational runs. Plus having the EC Operational and Control running side by side gives us more insight in the influence of atmospheric chaos on the output.

The most up-to-date ECMWF Forecast User Guide answers these questions: "The horizontal and vertical resolutions of the medium range ensemble (ENS) became the same as for the HRES after the upgrade to IFS (Cy48r1) introduced in June 2023. The HRES forecast and the unperturbed ENS control forecast are meteorologically equivalent and equally skilful on average. They have the same physical and dynamical representation of the atmosphere and use the same parameterisation of sub-gridscale effects. However, they can diverge on a day-to-day basis due to small technical differences and the chaotic nature of the atmosphere. The HRES will continue for the time being for ease of use by customers and users." So yes, they are basically identical, but they still run separately. The small differences in outcome are the result of the element of atmospheric chaos that is integrated in the model, and small technical differences. The separate Operational run still exists as a separate thing, because all of us, pro forecasters and amateurs alike, are so used to it. That is of course very considerate of them, but I assume we can expect it to be discarded in future updates.... Section 2.1.2.4 HRES - High Resolution Forecasts - Forecast User Guide - ECMWF Confluence Wiki CONFLUENCE.ECMWF.INT

Indeed more and more indications that this snow thread about “snow chances over the coming days” may turn into a thread about “snow and cold chances over the coming weeks.” EC 240h now joins GFS and GEM by showing Scandi and/or Iceland heights: It seems like UKMO is heading in that same direction at 168h: I did not expect much anymore of the snow here in the West of Holland these days, but with a little snow fix today (one whole centimeter!!) my cold and snow excitement is a bit back again.

Usually everyone looks at the big hemisperic picture at long lead times and zooms in to regional charts at short lead times for events like we’re in right now, but for your personal experience and insight it’s good to take a mental note of today’s weather for your location (snow/rain/dry) with this synoptic setup at T+0h for the Northern Hemisphere: Might come in handy to draw from that when, say, two winters from now something similar to these synoptics pops up in some GFS run at 384h. Which takes me to verification. When today was still Day 10, EC did a very fine job at 240h amid all the post-SSW chaos: GFS and GEM were both far off:

Meanwhile, after the snow shenanigans of this week, EC is not at all interested in the UKMO 168h Northerly. Nevertheless, its 240h chart is interesting going forward with wedges to the NE and NW.

Yes, suggesting a more pM wave of colder air. In the extended, the ensemble remains at average, or below average with more sunshine (in De Bilt). In combination with some reappearing heights between Scandinavia and Greenland on the clusters, this may hint at the development of more Easterly/Northeasterly flow later on. (Bright skies are good to me, but I'd rather see warmth too in the second half of March, but we may have to endure a lot more useless dry cold).

The situation is complex, but I don't get why the mood is so downbeat. We all know that snow is difficult to forecast more than 48 hrsout, but it is not the odd FI run that shows snow. Every run there is a serious number of Operationals and ensemble members that is good enough for snow. With these snow charts for this EC run, all of the UK, Ireland and Holland gets to see snow, with the exception of Devon, Kent and the Isle of Wight (sorry, guys). You can complain later, but for this set most of you would have bitten off your hand two months ago. Plus, with this chart at 240h, you're not talking about a smooth return of the Atlantic either:

GEM has got the South coast covered. Literally! But seriously, we are all still in the game for snow over a period of at least 4 days, be it by way of snow showers form the North, or by way of a Low tracking East later in the week. Too early for anyone to be concerned.

Oh, ECM Op is indeed on the mild side of the EPS 12z for Day 9-10. Here's the EPS T850 plume for your location:

The 12z EPS follows the Op with colder Days 7-10. These complex SSW could be a time for the high resolution run to shine against its ensembles. This is De Bilt in Holland, but it's true for UK locations as well. The extended holds on to that warmup, but that's all even more uncertain than otherwise.

We just don't want the tPV to vacate the Western Hemisphere entirely. At least a shard, a lobe, or whatever you'd like to call it, must stay behind over North America to keep the Greenland High strong and in place. GFS has that. EC, GEM and UKMO don't. I think UKMO would be very similar to EC if it were to continue post 168h. It's a tricky situation, and I don't know which model, be it an operational or and ensemble, or which blend, gives the best guidance in this case.