fergieweather

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).

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).

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).

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.

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 :-)

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....

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...".

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.

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.

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.

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.

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.

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.....!!

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.

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.

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...!

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.

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.

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

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

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.

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.

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...

....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.

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).