fergieweather

Ha! The forecasting headaches of late certainly haven't changed...(epitomised below)!

No worries! I assume the misconception is simply because the primary deterministic global model we use on TV graphics is ECMWF (although we can, & some of us periodically do, also show GFS & UKMO-GM for comparative purposes such as the lead-up saga we endured into Christmas Day). The principal mesoscale/regional model we used was, until a fortnight ago, Euro4 (with its raw output edited as needed by DTN forecast team). However with that model now retiring, we've switched to UKVx as the default. None of this feeds into the app, however - both it, plus TV site max/mins & 7-day outlooks are MOS driven, as prior explained. Cheers

No, it isn't a replication of ECMWF deterministic (a common misconception). The BBC wx app - as with MeteoGroup/DTN's WeatherPro - is fed by DTN's proprietary MOS data, using statistical blend *including* ECMWF output. It can also be tweaked as needed throughout the day by DTN forecasters, who can 'group select' locations in a given area to update their respective site-specific solutions shown in the app (eg based on radar, obs etc).

Hi Tim, yup - absolutely. For those of us who accept the meteorological calendar as a mere statistical construct & instead pay attention to the astronomical version, winter '21-22 is in its mere infancy and where we've ended-up by March is intriguing - if currently imponderable - speculation. On face-value, so many pointers favour a colder outcome (broadly so, especially for parts of Europe but with UK periodically being in-play). However, to what extent that relies on outcome with the stratospheric polar vortex & the degree of offset (to otherwise cold historical analogues) from background climate warming is a learning curve we are now riding-on through the coming weeks!

Hi Tim, in all fairness to UKMO these updates you mention are just one of the many forecast duties performed by (I assume still - certainly was, quite recently) a deputy chief forecaster on a busy shift of many pressing tasks, some of greater immediacy. They're predicated on the 2x weekly updates of extended ECMWF output, plus - when/if relevant - any conflicting strong signal from Glosea6. Thus the continuity of overall emphasis won't necessarily change day-to-day, as that would become an unfathomable mess reacting run-to-run. In similar vein, the BBC Monthly outlook is also tied to ECMWF updates, albeit over many recent weeks they've (DTN) purposely leaned to caution over that model's solution(s) beyond a ~2 week lead due to poor performance, heading instead in favour of statistical models/analogues. So the UKMO longer-lead switches aren't down to forecasters being volatile(!), but instead the modelling. They're compiled in good faith from the available evidence as they see it, in simplified format, very much as are the BBC ones (ie translating into a public-facing style the more 'technical' sub-seasonal forecasts that DTN issue 2x weekly). It can be a poison chalice writing these, especially in winter. Cheers.

Hail, I think from the photos & videos I saw a few days ago of this event. Plus reportedly that camel photo is a prior one from another Middle East location a year or two ago.

Hi Michael, by the time I joined the BBC Weather fold in 2007 you'd moved-on from WX Centre days (albeit leaving the likes of Rob M, Peter G & a few others to 'carry the flag' from prior halcyon days!). Nonetheless it was always a pleasure to chat with you & hear your sage views on our daily regional UKMO conference calls when you were covering BBC SE at Tunbridge Wells. Like so many others have expressed, you were an inspiration in my younger days - and your memorable broadcasts through some key weather events have become inimitable parts of BBC history. You have been at the vanguard of UK weather broadcasting and I wish you all the very best for your well-earned retirement. Meanwhile, Merry Christmas & Happy New Year!

Actually it's an entirely competent mesoscale/European domain model (& indeed proved exceptionally useful on numerous occasions as a forecasting tool during the 2021 Formula One season, for various agencies). Alongside ARPEGE, EURO4 (RIP!) & others, always perused by e.g. UKMO duty Chief Forecaster in overall assessments. Lately we are seeing a repeated saga of various operational models expressing themselves run-to-run across a varied span of respective ensemble envelopes and indeed at longer lead times, not helped by ongoing poor performance of sub-seasonal models particularly beyond ~wk2. And winter (strictly) hasn't even started until tomorrow...

Interesting read Nick, and in many respects is replicating the thoughts of DTN/WCS, whose winter outlook discussion (with associated statistical modelling/analogues) was issued a week ago & discussed with their extended forecasting folk in a subsequent conference call. Like you, they stress how the colder risks post mid-Jan & especially into Feb are highly predicated on SPV outcome by/into that period - an imponderable (despite the strong analogue support, as you note) that simply won't be adequately resolved in seasonal models. However as DTN (& you) also stress, even a close analogue match to certain historical cold winters & associated key forcings (eg 1962) is no longer a viable direct parallel, especially re potency of cold etc., given the more recent backdrop of climate trends. It's certainly a fascinating winter ahead and not least, as WCS point-out, in assessing performance of the longer-lead dynamical models (including at sub-seasonal range) versus statistical & analogue assessment. Jury out! Cheers.

Detail tbc re scope of lowland snow potential in parts of south at least, but in broadest terms the cautionary point (& this is an annual one made by many of us on eg Twitter) is how eg ECMWF accumulates *all* solid ppn, ie sleet too; plus accumulation fields won't adequately represent depth temps (thus actual tendency of snow to genuinely accumulate on ground or not). What's seen falling from sky, as opposed to actually lying on ground (other than fleetingly), won't necessarily be adequately represented yet in model fields for lower elevations at least.

Yes, albeit I don't see this as some competition: it merely manifests an interesting dichotomy (but with some common threads) between the differing scientific approach of the two organisations. Also bear in mind that UKMO Seasonal Team will issue their 3-month update shortly for contingency planners; similarly DTN issue customers their updated UK prognosis for the D-J-F period imminently (this week). Those will of course assimilate factors that may have differed in outlooks prepared a month ago (eg ref scope of La Nina; SPV forecasts etc). How much overall faith is placed in latest suite of seasonal models from the main forecast centres is another matter, too. Cheers.

Not true - much as they do with Highways, DTN has various customers in the energy sector, including National Grid, and have forecast teams dedicated specifically to that industry. Latest DTN/WCS assessment for European-wide winter (published last week), with statistical analogues that also factor-in the currently strong stratospheric polar vortex, lean to temperatures being rather variable across Europe during the first half of the winter season, followed by growing potential for periods of high latitude blocking during 2nd half (sustained blocking considered rather unlikely until after mid-Jan); with Feb seeing strongest signal. Cheers.

It's specifically re modified fields. Not forecast rationale (we communicate that publically already). Cheers.

I know: but it's moderated etc. Facebook isn't.

The licensing agreement with UKMO, negotiated a few years ago, is for sharing selected items that help offer technical context to forecasts, but social media growth means a risk of items disseminated without permission or proper context and thus possibly not in spirit of original forecast thinking. For now, I'm afraid that means no further sharing of that material aside from more limited ones through twitter, given sensitivities. Back to model discussion.

Sorry - apparently it was copied to a Facebook group without permission. We had this issue a year or two ago. Thus removed to prevent further reproduction with risk of changed context.

Snap! Agreed Nick. It's clearly some modern mythology.

Meteorological trade runs 24-7 even over Christmas(!); the satellite data keeps churning; many aircraft providing data... etc. I've never heard such a 'theory' espoused anywhere before.

What's unsavoury? 18z GFS offers one heck of a chilly block. And effectively dry. Bear in mind 850hPa temps unrepresentative of surface temps in those set-ups (subsidence inversion/continued radiative cooling). Fog could be a real issue however, especially if there's any longevity of continental slack flow and resultant import of higher aerosol levels. NB Note v different sense of direction further into 12z GFS PARA. Fine balance... etc.

By T+192, 23 members support the flatter solution; 15 centre the high effectively over the UK; and 13 offer a more amplified solution with high pressure to the NW and a markedly colder NE'ly flow.

See edit above Steve (NB doesn't alter median temps - remaining below avg)

But 12z ECMWF DET sits well in the ensemble spread, with below average temperatures from start of Jan on throughout it's 15d run... why discount it? It's as plausible as any other solution at the moment. EDIT Whoops sorry, was reverted to 00z ENS - yes, 12z milder end of ENS.

We shall see how it unfolds. ECMWF Monthly also raises +ve MSLP/GPH anomaly to NW (S of Greenland) into week 2; interestingly it supports GloSea5 on broad idea of a colder than average spell much/most of Jan, with varying NW/N/NE flow (perhaps E'ly later in month); even by week 4 with unusually prominent -ve temp anomalies for southern UK (given the lead time). One way or another, model output looks an awful lot more interesting versus this time a week ago.

PS What is worth noting re UKMO is that the GM isn't modified at Ops Centre beyond T+132. That's considered a reasonable cut-off point for fair medium range steer, in terms of tweaking additional fields such as frontal positions/shortwave elements & associated upper forcing/ppn/cloud cover/ etc.

The 12z DET sat comfortably within scope of 00z clusters/ENS. Just one of a few potential outcomes retrospectively but let's see where it sits versus ensembles this time. Where ECMWF monthly heads tonight is of particular interest, given paucity of signals beyond Jan week 1 in last Thursday's outing... Incidentally, someone posted along lines of UKMO-GM being of no credence at T+144, but I've never seen evidence for that published or cited elsewhere.