Into February: Forecast and Model Discussion commentary

[Met4Cast]

 Wedgesmakesledges

I want to reply to this because it's clear you took a fair amount of time to write this up and I think it's worthy of a proper response in return.  To prevent a single post taking over an entire page I'll shorten the parts I quote but might reference things you said outside of the quote, if that makes sense;

3 hours ago, Wedgesmakesledges said:

Having reviewed the posts this week & taking our little Island to one side - whether it be GSDM posts or just general stratospheric ones - there has been heavy resistance from many indicating this could not happen with the background signals > More specifically 

-MJO being in the COD & no amplitude

I don't think i've seen anyone here say that it "can't" happen, but more the odds/probability of it happening are very much stacked against it happening, of course, short-term meteorology can and does throw curve balls and as you say, we are just a little island so small changes can make a big difference to what we see here in the UK. 

The GSDM is a diagnostic framework for the medium-long term timeframe, i.e 2-4 weeks. It encompasses the MJO and the GWO and what essentially it all comes down to is windflow budget, i.e how much momentum is in the atmosphere at any 1 time, this can then be used to "diagnose" broader global weather patterns in the 2-4 week range, naturally, with the usual caveats. It's one of the best tools we have in terms of longer term forecasting.

3 hours ago, Wedgesmakesledges said:

This leads me to believe that whilst the GSDM theory presents some great analogues fitting in with GWO & MJO - the reality is is that it cannot be used as a forecast tool because the people trying to decipher it ( the regulars ) are using its starting point + persistence assumptions to speak ( often as absolutes ) that specific outcomes cannot happen.

Having been studying the GSDM for 2-3 years now I fundamentally disagree with this. The GSDM absolutely can be used as a forecast tool very successfully. I'm going to quote myself here & then share some charts to show how the GSDM is used to pick out NAO & AO trends and how that is then factored into forecasts.  These quotes are from 10 days ago;

On 19/01/2025 at 12:33, Met4Cast said:

This up-coming unsettled spell looks rather grim, a cluster of cyclones barrelling into the UK driven by a very powerful jet stream c/o the strong temperature gradient across the USA. The MJO is moving into increasingly +NAO phasing and AAM tendency has tanked, strong westerlies are propagating poleward and I suspect this will help not only drive unsettled +NAO regimes but will help to couple the sPV with the tPV from further imprinting the +NAO signature. 

The NAO signature is now well established as expected, this was well sign-posted within the GSDM by (as mentioned in the above quote) poleward propagating +ve anomalies into  60-90N. This coupled with the MJO moving through the Indian Ocean and into the Maritimes, typically +NAO phases, it's no surprise we currently have the pattern that we have. We've also seen a well-coupled tPV & sPV helping to drive this pattern and anchor a strong vortex across Canada/Greenland. 

It's also very clear to see the impact this poleward propagating had on the AO, going from strongly negative when we had -ve anomalies around 60N to strongly positive now +ve anomalies have taken over with a reversal at 30N.

I went onto say this in the same post;

On 19/01/2025 at 12:33, Met4Cast said:

Teleconnectively it's looking about as bad as it gets in terms of blocking/cold weather patterns now. The MJO continues to progress east into February and pushes into more favourable phases around mid February looking at current forecasts, the usual caveats re: MJO forecasts often being unreliable in the longer term apply.

That said - Even if we do see the MJO forecasts play out as currently predicted, if we've got a strongly coupled tPV/sPV it's likely this will be the main driver of our weather patterns and the MJO (and I suspect by this point, rising AAM due to this progression via +FT & +MT processes) will likely struggle to produce much in the way of (if any!) sustained blocking. It now seems very unlikely that we'll see a major SSW this winter.

This has played out exactly as expected. he MJO has continued to make quick progress eastwards with a likely arrival in more "favourable" patterns around or just before mid-February, i.e phase 6. Interestingly, phase 6 during early-mid February does suggest Scandinavian heights with low pressure to the NW, similar to what we're seeing on a few outputs, albeit without any real sustained momentum currently.

The transition of the MJO into and through the Maritimes has begun to spike AAM tendency helped along certainly by the current developing +EAMT event, this is actually occurring a little earlier than expected.

Hopefully that shows there is predictability when it comes to the GSDM & the various processes, the strong sPV/tPV over Canada & Greenland is a direct result of those poleward propagating +ve anomalies, this has led to a strongly positive AO and a positive NAO, both of which we're seeing at the moment, we've also seen a coupling between the strat/trop at the same time further imprinting this pattern. Lets go back to the 16th where this was also being suggested, almost 15 days ago now..

On 16/01/2025 at 17:28, Met4Cast said:

It’s very clear that we currently have poleward propagating +ve westerly momentum which is absolute kryptonite for sustained blocking & very much supports the idea of a +NAO Atlantic driven pattern. 

Worth noting I made these posts at a time modelling, especially the GFS, was suggesting a full blown Scandinavian high for the end of January, this is an example of what was being shown;

Needless to say, that blocking didn't come to fruition and I was against the idea of blocking for the reasons highlighted in my above quoted posts & indeed in this one, further evidence that the GSDM framework can be used successfully and shows how it can be used in conjunction with NWP modelling to work out, probabilistically, what's more likely. Back then, teleconnections favoured +NAO and Atlantic regimes for the end of January, NWP modelling was throwing out very blocked runs for the end of January.. which one was more accurate? Which proved to be a better guide at that range? I think the answer to that one is obvious. 

Quick fire reply round!

3 hours ago, Wedgesmakesledges said:

1) The forecasting of things like MJO / GWO cycles within our existing spectrum of models is subject to the same error rate as the models we see as deterministic ones.

Agreed on the MJO forecasts certainly, GWO cycles not so much as that is the result of observational data rather than NWP model data.

3 hours ago, Wedgesmakesledges said:

2) Climate change - may infact be impacting our system so much that what used to be sound data / anaolgues & sequential links in terms of A+B =C that now because the anomalies are so distorted maybe have alligned to a different set of rules

Also agree - Climate change is absolutely making longer term forecasting more difficult because the baselines are changing. Applying the GSDM is never A+B=C and that definitely applies to the MJO & the often posted composite charts, the MJO is just one aspect of the broader windflow budget. Climate change is the reason I think analogue forecasting for seasons is now essentially useless. Anyway, hopefully this post demonstrates that the GSDM is a very useful tool in forecasting & how it absolutely does have predictability and can be used when viewing NWP output to provide further context to the "noise". 

I agree with your broader points re: talking in absolutes, nothing is ever "absolute", the atmosphere is constantly in a state of flux and whilst teleconnections can certainly provide a broader idea of expected patterns, short term meteorology absolutely plays a part and sudden changes can result in very difficult outcomes. 

Just to finish - I don't go against the idea of colder outcomes just for fun or to try and "stir the pot", I follow the data. I absolutely love cold/snowy weather and would love nothing more than for a strong Scandi high to pump bitterly cold air eastwards into the UK, snow is my favourite type of weather, but neither am I going to let that personal bias impact what I think is "likely" when it comes to forecasting.

25 minutes ago, Welsh_Weather said:

 Met4Cast I really don’t understand the term ‘model bias’. It’s bounded around in here way too often. There are tendencies for models to overdo certain scenarios but they are still only picking up on what is fed into them and how they deal with it based on history. 

Model bias is well documented, in fact, there have been numerous scientific studies on the topic and NOAA itself has a full lift of known biases within the GFS. There's a reason "bias corrected" models exist and are put into operation.

 

Edited January 29 by Met4Cast

[WYorksWeather]

It wouldn't be a GFS pub run without a ridiculous low blowing up somewhere deep in FI.

With that, I'm off to bed. Let's see what happens over the next few days.

[Met4Cast]

The lack of posts tonight on the 18z speaks volumes!

[damianslaw]

GFS18z in far outer reaches showing the type of synoptics it was churning out for the current timescales, alas deep bowling lows through the UK have not verified this week after last weekend storms, its not to say they won't happen..

Edited January 29 by damianslaw

[HarvSlugger]

 Met4Cast maybe it's a good sign that the pub run isn't good

Just like the JMA being bad

[Met4Cast]

 HarvSlugger

The 18z ensemble suite goes along with it, pretty poor and certainly a step away from the colder easterly idea, sadly. 

Mean barely dips below average throughout and fewer members dip below -10C compared with previous runs. Just one run of course. 

Edited January 30 by Met4Cast

[feb1991blizzard]

 HarvSlugger

Main thing is ens aren't too bad, i mean relative to what they were looking like a few days ago, obviously would still need a big improvement to be classed as a strong signal for a proper cold spell.

[kumquat]

Heck, I'd take the GFS 18z 384hr third time lucky runner from the Northwest and like it or lump it. Best split and pressure on the vortex for the whole run. 

We're getting to the snow cut-off date though lolz 

[kumquat]

Interestingly semi-decent Med low at 192 and a low marathon runner goes over the top of that (straight out of Novaya Zemlya) and ends up smack dab in the middle of the 384 vortex split. It's pulses. There's waves and pulses. We get hit by the backwash PV that come from the NW

 

 

Edited January 30 by kumquat

[Wedgesmakesledges]

@Met4Cast will have at your post today 

The overnight runs thus far - the UKMO / GFS / ICON all for once seem fairly well aligned-

The development of the initial pulse of heights to the NE remaining in situ & then building back west around 120 - with the deeper cold in central Europe.

A second attempt at amplification to the west of the UK day 6-7 & critically the upper air cold pool get another 'assist' west.

All we need is enough westward motion to get across the North sea.

A pretty good start to the day - UKMO especially better at 144 & 168 as well.

GFS 180 sums it all up quite well. 

 

What could well work in our favour is further trough activity diving south an the atlantic allowing for the Easterly to 'escort' that cold pool west. At this stage no sign of any azores high ridging NE killing it all.

Either way an impressive block.- we will need similar advection west a la 1991.

 

Edited January 30 by Wedgesmakesledges

[RJBingham]

 knocker Knocker is a retired Met Office meteorologist I believe, we're lucky to have input from the pros occasionally.  

[That ECM]

T168 ukmo and gfs. Gfs goes onto show what most coldies don’t want. Nothing yet sorted. 

[Chesil View]

I suggested a few days back that the basic situation is a stand-off between the Siberian block extention and the raging pv.

Nothing has really changed in the sense that all we are witnessing in the model output is the waxing  and waning of each of those features with the UK sitting on the borderline between the two.

Yesterday a bit of excitement in the modelling as the Siberian high extention waxed a bit and the pv waned a bit allowing hints of some brief easterly dominance.

What will today bring?

According to the 00z runs the waxing and waning has switched back in favour of the raging pv  and aligning with the meto nw/se split forecast.

 

Edited January 30 by Chesil View

[KTtom]

I was slightly bemused yesterday as I just coulnt come to grips with what some were seeing, or hoping for in the output. As far as I can see the output is as stable as ive seen it for a long time in terms of the next 10 days or so. An increasing high pressure dominating pattern for the UK, no storms in sight and no deep cold via an easterly either. Even when we have been shown a brief breeze from the continent on the odd run, its been nothing but that with no evidence of any real cold pool to the east heading our way.

UKMO 168 and Met text  sums up the outlook, mostly dry, temps around or slightly below average especially inland.

[That ECM]

@nick sussex a flabby high? Your concern yesterday? Not great. Gfs not great. Not done but not what coldies wanted to see unfortunately.

[Gowon]

Definitely downgrades this morning on the 850s front but the block is still there so hopefully things will change again.

 

 

 

[mulzy]

The key timescales of interest are still out in FI - so nothing has been firmed up.  The ECM run this morning is not as good as last night's but the ECM AI continues to be positive. 

Edited January 30 by mulzy

[mountain shadow]

Teleconnections are 2 nil up with 30 minutes to play, can't see the models coming back now.

Any Easterly will be about 1000 miles to far East for us.

[EML Network]

Were heading towards the final third of Winter now and it's that time, where time...is running out.

 

I can't see anything now in the various data sets that gives the snow starved amongst us, anything to be excited about out to as far as 10th February.

 

Which then leaves us with 3 or so weeks to try and squeeze something out of Winters tail..i'm not quite ready for a 'winters over' post, but its headed that way.

[blizzard81]

That's two poor Ukmo runs on the bounce now but much more concerning is that woeful ECM op this morning. If ever there is a model we want to stay on side in situations like this, it's the ECM. If it throws out another poor op this evening then it will be very ominous indeed. It's brilliant when it spots the milder solution at distance.

[northwestsnow]

Hoping we had some rogue ops this morning.

[Beanz]

So, winter will be signing off with a rather boring, but pleasant, still and settled high.  

How’s the LRF looking for a warm spring? 

[mountain shadow]

 Beanz Winter is not done yet. 

i think we'll have a cold shot towards the end of February. 

[Jacob]

On a more positive note, the ECM AI is great!

The WWA goes in a W-E angle and stays over Scandinavia

There’s then a 2nd push of heights, holding off the TPV for longer and the E’ly deepens

As a result, the cold uppers stagnate over the UK, with pressure thats’s low enough to bring winters showers across the UK. The west would get some very cold nights from surface cold as well 

Edited January 30 by Jacob

[Mike Poole]

ECM clusters this morning, T192-T240:

I must say, I was expecting worse given the comments this morning.  Cluster 1 (31) is the one containing potential for the easterly, cluster 2 (20) is the flatter solution with the high at lower latitude.  I should caution again that I expect only an unknown fraction of cluster 1 would deliver an easterly, because there will be some spread around that representative member and it wouldn’t take much drift east to kill it.

T264+:

Cluster 1 (16) the Scandi high rather disintegrates and we are left in a westerly flow.  Cluster 2 (14) continues the easterly theme and is cold.  Cluster 3 (12) asserts an Atlantic ridge with a northerly, trough in Europe and is cold.  Cluster 4 (9) looks a bit messy with heights remaining to our northeast.  All have potential apart from cluster 1.