[Model output discussion - 5th December 2016 - Into Winter we go]

1 hour ago, Ali1977 said:

Is it looking like we could get something cold in by Xmas day on your models Ian? It looks a bit early on the usual suspects but the GFS p shows cold air from the SE in time for Xmas day . I see you mention 2 weeks so guessing it's not looking likely!!

Sorry, I mis-wrote part of my posting (flu!). What it *should* have said is we are seeing a period of strengthening SPV, which ultimately could manifest into a period of increased zonality/mobility with a lag time of 1-2 weeks. I didn't mean necessarily 1-2 weeks from now. The favoured period is late Dec-early Jan for this (which we suspect is why Glosea has been playing ping-pong with balance between mobile v blocked members into that period... the blocking ones ultimately win, *but* UKMO suspect the output from it & ECMWF could be masking a temporary phase of zonality into turn of month). I'm told a UKMO blog may be forthcoming re current/recent period of uncertainty leading into Xmas; this on the back of the Seasonal Team/Hadley meeting today at Exeter (which will be followed by a lengthier one on Monday to discuss J-F-M as a 'job lot' ahead of next contingency planner doc being prepared).

[Model output discussion - 5th December 2016 - Into Winter we go]

UKMO Seasonal Team meets this morning to discuss the situation at present and signals/output for rest of winter. I might be able to offer an update later. Meanwhile, current thinking is that downwelling from current phase of strat consolidation will manifest into higher likelihood of a zonal/mobile period ca. 1-2 weeks from now. This is suspected to be temporary: both GloSea and ECMWF products are keen on blocking resuming throughout 1st half of January (at least). 

[Model output discussion - 5th December 2016 - Into Winter we go]

3 minutes ago, karlos1983 said:

Not sure blizzard said anything wrong. Only what we must all be hoping as we have been here already!. Didn't see any dig at any pros. Peace !! Life's too short

Yup it's fine. There's every reason to have cynicism in longer-lead output currently, but equally there's scope for some optimism too (for the colder winter weather fans). 

[Model output discussion - 5th December 2016 - Into Winter we go]

ECMWF Monthly goes off-piste versus it's deterministic brother into & particularly beyond Christmas. The +ve GPH/MSLP anomalies out to our NE merely intensify again towards end of month, then show signs of retrogressing 1st week Jan to lie directly to our North, with mean E flow trending more E-NE with time. Glosea likes the idea of a more blocky Jan too... but the term 'deja vu' springs to mind! Other shenanigans aside, at least Glosea strat diagnostics have been sound last few weeks, so given emergent January signals the PV may yet yield a 'surprise from aloft'. We shall see.

[Model output discussion - 5th December 2016 - Into Winter we go]

1 hour ago, Bring Back1962-63 said:

A very warm, kind and popular man who will be sorely missed. I had the opportunity to work with him on one of the programmes in the BBC2 series "The Essential Guide to Weather" broadcast in 1999. Ian had just retired as a BBC Weatherman and hosted the series. I used to run the weather station in the Rickmansworth frost hollow and Philip Eden recommended that there should be a feature on this in one of the programmes. I did a short presentation by my Stevenson's screen while Ian did an experiment to demonstrate temperature inversions by digging a slanting channel in a flower bed and pouring dry ice down it - it was typically amusing! Wonderful memories.

Gosh. Amazing. And Rickmansworth too (I spent most my childhood there & in Chorleywood!).

Re modelling, a bit of context but I make no judgement on forecast confidence(!). The Dec-issue Glosea JFM signal is skewed by F-M. Jan actually looks quite blocked. Strat PV consolidation as per medium range may not necessarily prove a foregone conclusion all of next month... in short, any 'winter is over' postings based on present deterministic modelling need to wait a few more weeks to receive any final stamp of validity :-)

[Model output discussion - 5th December 2016 - Into Winter we go]

My retired colleague Ian McCaskill, RIP, would have relished discussing this Christmas Kerfuffle on-air with typical humour. 

GFS 12z deterministic very similar in final reaches of run to Thursdays ECMWF Monthly broadscale set-up for same period just after Christmas. Whether that will remain so tonight, we shall see....

[Model output discussion - 5th December 2016 - Into Winter we go]

16 minutes ago, Singularity said:

I suspect shorter/mid-term modifications to lower vortex perturbation (via poleward fluxes for example) may be having a large impact on the extent to which positive zonal anomalies are able to propagate down from the stratosphere, are you by any chance permitted to confirm or deny this i.e. how far does your non-disclosure go exactly?  

Regardless, thanks as always for your insights into Met Office thinking 

They note: "...an emerging signal for at least some lower tropospheric mobility to N of UK but with little mobility at upper levels...in the longer term, with a strengthened stratospheric polar vortex, this may result in increased mobility late December-early January, but latest GloSea products/anomalies are now less clear cut with this...".

[Model output discussion - 5th December 2016 - Into Winter we go]

6 hours ago, CreweCold said:

''a marked switch away from -ve NAO 2nd half of winter''

Well to see a switch away from a -NAO we'd have to have the -NAO first 

Not a chance this month will return a mean -NAO

Ha ha. I was, of course, referring to the *model output* and where it sat previously versus now.

Talking of which: yesterday's Glosea run has suddenly retreated from the stronger zonal/mobile theme into early Jan and is much less bullish now: more blocked members again therein. If today's run does likewise, anticipate some tweaking to wording of UKMO 15-30d outlook, albeit they note ECMWF tonight awaited for comparison anyway.

[Model output discussion - 5th December 2016 - Into Winter we go]

14 minutes ago, Sawel said:

Do the Met Office put more credence in GloSea5 than they do the ECMWF Seasonal model? Like, say they were the only 2 models that existed and they're showing completely different scenarios.

Good question. Both are equally valued but of course, they are ultimately merely *guidance* like all models. The added value comes when specialists in seasonal/sub-seasonal forecasting tease-out all the vagaries of forcing mechanisms/drivers, and in doing so cast a critical eye over the model output to weight any solution accordingly. The ECMWF J-F-M output (sort of a reverse of original seasonal expectations!) raised some eyebrows, certainly; but that doesn't mean it's discounted. 

[Model output discussion - 5th December 2016 - Into Winter we go]

Incidentally, after such a prolonged period of close broadscale agreement between runs of GloSea5 and ECMWF Seasonal, you'll note now how they finally bid each other farewell in their respective Dec updates and ride their own separate ways into J-F-M. Glosea fits with previous thinking (a marked switch away from -ve NAO 2nd half of winter); ECMWF continues similar to before. Of course, the combined 3-monthly means mask detail of how the journey unfolds, but it'll be intriguing to watch.

[Model output discussion - 5th December 2016 - Into Winter we go]

13 minutes ago, Catacol said:

Yes - thanks Ian - that feels a fair assessment. The extent to which a mobile or blocked pattern may evolve into early January seems to me to depend a great deal on finely balanced elements - propagation (or not) of stratospheric wind speeds to the trop, evolution (or not) of the GLAAM budget and strength of the MJO going forward. Day by day it is very difficult to pick a combination that looks coherent and, in that context, not surprising that the models are having a tough time. I note that none of the well regarded amateurs on here have made a stab yet at the first half of January. Even GP is quiet. I think this speaks volumes. Huge uncertainty. 

Yup, a few days ago the GloSea signal for mobile/zonal period into turn of month was very marginally favoured, but has gathered some pace since then (albeit of course with low confidence). It'll be interesting to see if it receives any ECMWF Monthly support tomorrow. 

[Model output discussion - 5th December 2016 - Into Winter we go]

19 minutes ago, Catacol said:

In terms of timing I don't think much has changed either. This process starts mid month and will evolve from there. The dice haven't quite rolled for us yet, but if this general prognosis can stay on track then at some stage, before too long, wintry potential may follow.

As (post-00z) UKMO assessment stresses, little has really changed over last 24hrs in terms of resolving the ongoing uncertainty further into December. The marked ensemble spread into mid-month onwards is where confidence begins to quickly unravel, with - as UKMO describe it - "larger scale diagnostics favouring a more mobile pattern developing towards the end of December and early January, with the Christmas period falling in the tricky transitional period between that and the current, slow moving regime. As such, we are no clearer as to the Christmas forecast at this time." 

How long this impasse remains, in terms of very low forecast confidence, is anyone's guess.

[Model output discussion - 5th December 2016 - Into Winter we go]

7 minutes ago, Man With Beard said:

Hmmm ... High Pressure over us or slightly to our east, slightly more likely to be south-easterly based (milder) but possibly also north-easterly based?

Actually, sounds ever so slightly like the conundrum posed by the OP runs tonight in the D7-D10 range?

Totally agree. This is why temperature outcome could be finely balanced, for reasons you mention. 

Or of course, it could be just a brisk, mild So'wester and all academic by then.....!!

[Model output discussion - 5th December 2016 - Into Winter we go]

5 minutes ago, Mucka said:

Oh no, just as we were beginning to quietly accept our fate, first we get the pub run and now this little snippet.

Well thanks to whoever it was that clambered into the wicker man. Very Christmasy of you - oh wait, not Christmasy, no, but thanks!

Can't reproduce entire meteograms from EC but here are ENS windroses for Reading, 22-23 Dec. Note (in major contrast to previous run) how whilst a SW flow quadrant (as per majority model climatology for that period) remains still the stronger signal by 23 Dec, we now have the 2nd largest - albeit oh-so-slightly - as an E'ly. Whether a wild goose chase (given recent ensemble roller coaster) or an incipient turning-point is now unfolding, we shall see. Jury out.

[Model output discussion - 5th December 2016 - Into Winter we go]

49 minutes ago, Gael_Force said:

Welcome! 

I'm afraid it is only the longer term models we cannot see which have built the idea of a cold early winter. Not one of the online (for all to see) models showed anything but a varying degree of positive for temperatures. The fact the UKMO had such confidence in their own products has allowed eyes to be taken off the ball and ignore the less palatable options from other forecast agencies.

Whoah.....hang about. That can't be allowed to pass unchallenged.

Do you *really* think Prof Adam Scaife - and the Hadley Centre scientists - sit in their monthly Seasonal Team meeting at Exeter and stare singularly and unscientifically at UKMO output and then dismiss everything else out of some sort of 'confirmation bias'?? 

The Contingency Planning 3-month document (which, stress, is a *probabilistic*, not *deterministic* outlook) is based on the consensus reached at those monthly meetings and written-up shortly afterwards. At those forums, the collective output across a number of respected seasonal modelling centres is viewed comparatively with weighting applied; and detailed areas of global forcings/teleconnections discussed. This includes output made available to UKMO from ECMWF, MeteoFrance, NCEP, plus Japan/Korea/S Africa and many others. Indeed you will often see reference made in those documents to where agreement exists between GloSea5 and "other seasonal forecast agencies/models". This isn't some cursory exercise of pulling bits of CFS off the web and comparing to UKMO products: it's a rigorous piece of scientific scrutiny with a duly critical approach applied.

The fact remains that as of mid Nov, excellent agreement on broadscale pattern existed between GloSea, ECMWF Seasonal and MeteoFrance-Climate. Not dissimilar output appeared in other (less highly weighted) models such as CFS.

Note that at NO point did UKMO probabilistic forecasts lean to 'very cold' categorisation: merely a higher than average chance of colder than average. Anxious that certain media types would blow it all beyond proportion to ape 2010-11, Adam blogged in detail on the UKMO website to offer context to the last two 3-monthly outlooks that were issued.

To suggest Adam and his team "took eye off ball" out of some deference to an in-house model firstly ignores the cross-model support available to them at the time. Secondly, it's actually quite insulting for a guy who has published or co-authored so many peer-reviewed papers in the field of seasonal to decadal forecasting; SSW's; etc etc. Whatever has gone pear-shaped so far in Dec, I'd wager Adam could explain unhesitatingly. 

And by the way: with cluster 1 of today's 12z ECMWF heading into building strong +ve GPH anomalies and rising MSLP above Scandinavia with a resultant easterly flow just before Christmas, it's also too early anyway to utterly dismiss what Glosea leaned to originally.

[Model output discussion - 5th December 2016 - Into Winter we go]

34 minutes ago, Mucka said:

Between a rock and a hard place but if there is necessary ambiguity in the forecast can people be blamed for misinterpreting it? 

For me the first half of the 30 day outlook is fairly clear, no chance of cold really with SW/W flow and high pressure SE, but the second half is very suggestive of the Atlantic trough being further South and high pressure being oriented further NW than the first period (not to our NW) with the battleground between colder and warmer air being over the UK and although we do get the caveat that orientation of the battleground will dictate conditions it is also suggested unsettled weather will be more likely to the W (not NW) and colder weather to the E (not just drier SE)

It is actually very well written  but I wouldn't blame anyone for interpreting it as being better for snow prospects than may actually be the intention.

Until long range forecasting becomes more accurate and reliable or the MetO decide  to shed their professional skins and go all Daily Express on us, I'm afraid the situation won't change.

What I do credit the MetO with is their lack of pretense and the integrity they display with regard to current long range forecasting capabilities and having people within the organization actually post to this forum is very much appreciated - even if we do shoot the messenger on occasion.

 

Personally, I would say blocking chances as the y relate to cold and snow possibilities are slowly being pushed away, there is very little talk now of Atlantic ridges or high pressure to our NW. Christmas is just beyond two weeks away so still enough uncertainty for hope of a turnaround but we need to see a cold signal proper appear within the ensembles soon to see any snow over the Christmas period. 

 

A good summary of the present situation, Mucka. They're conscious of the potential for (even fairly nuanced) changes to broadscale set-up to have quite marked impact in terms of UK weather type and temperatures. As you note, rock & hard place sums-up the difficulty well and none of it aided by erratic performance lately in EC Monthly output and another fickle/ambiguous run from it yesterday. History tells us that these scenarios manifest when the model is toying with two opposing outcomes at longer leads: given differences with GloSea5 (and for that matter, ditto lately between GFS & EC operational suites), it's perhaps unsurprising. 

The very bad luck for M-R forecast desks is this mess all focuses into Christmas period. Had this all been for late Nov prognoses, nobody would be paying much attention...!

[Model output discussion - 5th December 2016 - Into Winter we go]

1 minute ago, Banbury said:

Thanks Ian , much appreciated.

As I pointed out in a recent post , battleground on here will be taken as fronts stalling bumping into cold air, that's not how I would view it, it could well be a milder solution as you suggest

Sure, understood. The issue re how forecasts are worded is tricky and UKMO (and other commercial mainstream f'cast organisations) know that. One chief at Ops Centre rightly pointed out recently how the phrase 'winter hazards' (eg in 3-month outlooks) can get automatically read as *definitively* meaning 'snow' in isolation - an issue now dubbed, perhaps with reason, 'SMS' or 'Snow Myopic Syndrome' (!) - ie where media or cold aficionados seem to mentally discount how the phrase (as intended for contingency planners) actually includes freezing fog; hoar frosts; icy roads.. etc etc... as well as snow!  Cheers.

[Model output discussion - 5th December 2016 - Into Winter we go]

10 minutes ago, Banbury said:

Nothing I have read or seen in the daily update really suggests a battleground , I'm beginning to think its just folly following LRF ....of course just my opinion

There's no discussion presently of a 'battleground' set-up in sense of 'snowy'. UKMO forecast is worded to indicate the *slightly* more favoured set-up by Christmas (HP to E; LP to W) which - as they stress - may manifest in milder rather than colder conditions (as per EC ENS and EC Monthly). The difficulty at that range is compounded by very different solutions from GloSea and EC (GloSea5 remaining steadfast on a blocked Xmas period) and has no improvement in any cross-model agreement by early Jan. As I mentioned a few days ago, the route forward remains finely balanced, still with no real clarity and no one solution worth pinning hat on. 

[Model output discussion - 5th December 2016 - Into Winter we go]

Apols - Matt was indeed referring to Dec-Jan based on latest EC NAO plots. As you were...:-)

[Model output discussion - 5th December 2016 - Into Winter we go]

1 minute ago, mountain shadow said:

If we're going to see a DISTINCT -NAO in December there will have to be a quick and remarkable change in the model output soon. The NAO as far as I can see is mainly neutral or slightly positive for the foreseeable, I certainly cannot see a distinct -NAO, and by distinct I would expect a figure of at least -2 or below on the scale.

I think Matt Hugo meant Jan-Feb. The model update today doesn't cover Dec.

[Model output discussion - 5th December 2016 - Into Winter we go]

1 minute ago, blizzard81 said:

There is one possible explanation for this - both the EC and glosea seasonal models are wrong. Time will tell of course but quite telling that the EC 46 dayer has shifted away from the overwhelming northern blocking signal during the last 2 updates.

Ermm... but that's the point. Has the blocking signal actually now delayed to manifest in Jan-Feb, rather than Dec?? EC Monthly only takes us to early Jan. Bear in mind the Mon update was starting to show strengthening +ve MSLP anomalies over Scandinavia tail-end Dec into first days Jan. So it may yet dovetail with EC Seasonal but we can't be sure if or at what lead time. But no model shows roaring zonality into those lead times: Glosea weakly favours something zonal early Jan, BUT that's set against a wide synoptic spread in it's members.

[Model output discussion - 5th December 2016 - Into Winter we go]

13 minutes ago, Paul_1978 said:

Interesting stuff. So is this to say that if there is a disconnect between the strat and the trop, that if a SSW does occur, that this will have a negligible effect (due to the aforementioned disconnect)?

I'm unconvinced re SSW. What looks more possible from EC climagrams is a reversal of original expectations re this winter (going to back-loaded for colder risks Jan-Feb?) but it's impossible to fathom exact journey of travel from 3-month combined fields, that will mask month-by-month trends/extremes and also cancel out each other in eg resultant temp anomalies.

[Model output discussion - 5th December 2016 - Into Winter we go]

Given that today's EC Seasonal update remains steadfast on broadscale lack of zonality J-F-M, retaining it's eagerness for marked northern +ve GPH/MSLP and drier than avg anomalies, I'm not expecting any sudden change of tact from UKMO Seasonal Team based on current op modelling mash-up. What is now very striking is the growing evidence of apparent disconnect between impending and strongly-signalled upshift in stratospheric zonal winds, versus zero evidence in EC Seasonal for it to manifest in a marked change of mean tropospheric anomalies (towards zonality). I'm trying to discover via UKMO Seasonal boffins exactly *what* forcing mechanism could be overridingly at play here, because (despite the snipers/doubters on here) it is quite irrefutable that *something* is driving these two key high-end coupled models (EC & GLOSEA) to reach similar prognoses. Fascinating enigma; fascinating times...

[Model output discussion - 5th December 2016 - Into Winter we go]

1 hour ago, Frosty. said:

Tonight's Ecm 12z T+240 ensemble mean and operational have something in common, they both show high pressure firmly in charge.

....except where we've had just 2 clusters into and beyond that timeframe in successive recent runs, we now have...  5.

Given vast polarisation in the synoptics being output in 12z ECMWF EPS, with little to easily favour any particular outcome, it's effectively as clear as mud. No exaggeration.

Interestingly, latest GloSea5 continues (as it has signalled for weeks) to take 10hPa stratospheric winds up to above average westerlies (or majority of members do) by mid-Dec and remain there rest of month... but there appears, as yet, little sign of this connecting tropospherically in Hovmoller plots. Thus, a blocked phase remains quite conceivable later Dec-early Jan, but how this manifests here weather-wise is an imponderable at present.

What does seem highly likely is a protracted period now unfolding of model volatility and marked swings between differing extremes, given the very fine balance now abundantly clear in extended range output. 

[Model output discussion - 5th December 2016 - Into Winter we go]

41 minutes ago, mulzy said:

Short ECM ensembles for De Bilt illustrate the high level of uncertainty past 120 hours

 

Yes, and so it continues...

That period remains the key point of divergence still: 12z EC DET ultimately traces towards it's lower range of ensemble spread for SFC temps (taking Reading as representative) and a cursory perusal of stamps further out by T+168 suggests only around 12 members in unequivocal support of the truly anticyclonic deterministic run.

Essentially it's another of these will-it-won't-it chapters of low confidence, of which we've seen a few recently into similar model lead times. The bimodality of clusters further on through 12z EC's run is striking and has been a theme of recent output from that suite further into December (as with GloSea). Trend heads to below avg temperatures further into 15d timeframe, but (as yet) unconvincingly so and with scant sign of anything desperately cold anyway (not that we expected any sudden appearance of pronounced cold, based on GloSea and EC Monthly).