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

36 minutes ago, Walsall Wood Snow said:

Do you mean by this, has of things stand, you're expecting January to be mild and unsettled?

Just to explain: The wording of the UKMO forecast is a nod towards latest GloSea, which (after the mid-late Dec shenanigans and slight bias towards blocking) then leans marginally in favour of zonality into early part of Jan. 

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

3 minutes ago, Singularity said:

Two different groups of outcomes in this case.

In the simplest sense, a bimodal distribution of observations finds something to either be one way or the other, with little or no evidence of anything in between 

 

Or in this case a clear split between the more mobile, milder outcomes versus those that are blocked/colder. As UKMO suggest, what may actually unfold is a hybrid of both: "At this stage, a return to a more blocked pattern is marginally favoured through the second half of the month; however this may manifest itself as slowly evolving transitions between more settled, colder and unsettled, milder periods."

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

1 minute ago, stodge said:

I was going to make the same point about not taking these four week average charts too seriously. Two very cold weeks followed by two very mild weeks could be an average four weeks - I think some people see an average chart and assume it's going to be four weeks of average temperatures.

Yes, but they are used seriously but NOT on pure basis of an anomaly chart. Inspection of the individual stamps is essential. Anomaly fields are effectively meaningless viewed in isolation.

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

UKMO point-out the increasingly bimodal distribution of members in ECMWF Monthly by week 4. This explains the sudden appearance of bland temperature anomalies into that lead time, through a process of cancellation. Interestingly, latest GloSea5 also goes bimodal into same period, with the result that both above and below avg temps are signposted as more likely than climatology and confidence remains very low.

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

1 minute ago, karlos1983 said:

Hmmm so heights to our NE is more than just a chance then!

Well, the issue we just discussed at work is how evidently *something* is driving the similar pattern to resolve between those models, but we watch and see...

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

5 minutes ago, booferking said:

I am just waiting for the there is better heights towards greenland in the run there i done it.

Hmmm. It's an interesting quandary, given how EC Monthly today goes ultimately (after a pathway effectively as presently signalled in op models) towards a solution mirroring seasonal output pattern in EC Seasonal & GloSea (by late Dec--early Jan, indicative of a broadly NW-SE axis to storm track west of UK, with +ve GPH/MSLP anomalies building across Scandinavia into same lead time, as +ve temp anomalies suddenly then fade to become indistinct... all indicative of mean SE-E flow between high to NE/E and low to SW/S). Stamps of course show wide spread. However, at least some parallel to the last crop of seasonal suites (albeit EC Seasonal updates again this week).

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

1 minute ago, nick sussex said:

I'm waiting for the extended ones to come out which include the wind direction.

Vast majority of members between S to W by 20th. Two clusters in final phase of run (as per 00z). Flip-flopping continues between these two very equally balanced broad solutions.

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

Notice however how NCEP's parallel 06z suite heads a very different direction, more in keeping with e.g. GloSea's continued - if somewhat less bullish - favouring of the blocky set-up resurrecting after mid-month (a return to blocking also the majority outcome via EPS products using the DECIDER tool run by UKMO).

There remains strong suspicion (and this is an old story) that current GFS remains very prone to going beserk with mobility at longer leads, thus must remain viewed with due caution.

Either way, confidence is of course very low further into the month and we await this evening's EC Monthly update with great interest to see what crossover (if any) it retains with GloSea.

[Model Output Discussions 12z 30/11/16]

1 minute ago, Ali1977 said:

Yep, I should have stated that. It can give a rough idea of what's happening in the NW of Europe though - although understand sometimes mainland Europe can be very different from our island. 

See my addition to post above.

[Model Output Discussions 12z 30/11/16]

35 minutes ago, Ali1977 said:

True, next one due out soon so will be interesting to see how it compares.

...erm, that's the output for De Bilt near Utrecht! 

EDIT The 00z EC ENS (for Reading) retain signal for switch back to avg temps around 15th, with trend then to below avg thereafter. So at least some inter-run consistency emerging in that respect.

[Model Output Discussions 12z 30/11/16]

Worth noting EC 12z deterministic and control sit in a cluster of.... just 5. Trend heads to slightly below average temperatures post-13th-ish onwards.

[Model Output Discussions 12z 30/11/16]

1 hour ago, Frosty. said:

The mean charts have been / are uninspiring too for coldies, not just the operationals with just occasional glimmers of hope from individual perturbations.

Yes, but there's the key points encapsulated together in your post: the mean fields will always mask the members that are less energetic/stray from climatology. Inspection of individual stamps/clusters is critical to appreciating the actual breadth of potential outcomes... and in this case, the reasons for such marked uncertainty being expressed!

[Model Output Discussions 12z 30/11/16]

23 minutes ago, Northern Sky said:

It could. It's actually quite interesting to know that nobody (including the models ) really knows what might happen beyond mid month. Could be good could be bad but it's fascinating watching, trying to pick out clues and trends in the far reaches of FI and waiting to see if the background signals and teleconnections fall into place.

The quandary that unfolded earlier this week (disconnect between operational models at longer lead time versus 30-d/seasonal models) is suspected to be a case where quicker, more volatile response in the former  - to solving solutions based on the relative dominance of initial starting values - reaches a crossover with the slower-responding climate-forcing factors in the other suites (GloSea5/EC Seasonal etc). Where we can't now be clear - as outlined by Matt - is whether the dog's dinner of solutions now offered by operational models is a sign of them coming around to realign with the (surprisingly steadfast) seasonal model output for 2nd half December, or instead resolving into foretelling a very different synoptic direction (mobile/zonal of sorts/milder).

Whilst we've seen a gamut of overreaction to recent operational runs by some folk, as Matt and Nick both stress: it's currently impossible to offer worthwhile clarity at that extended lead time, so just not worth the raised blood pressure!

[Model Output Discussions 12z 30/11/16]

1 hour ago, Ali1977 said:

Any more info on this comment anyone - I think he meant things are looking better towards colder Synoptics again !! 

I was referring to the original comment suggesting a de facto mild/2015-esque phase all December was signed and sealed. It's not. After a (in places notably) mild period, EC trends back to average temperatures, with latest ENS ultimately favouring no dominant synoptic signal with mean flow split almost equally from any quadrant between S'ly around to E'ly. This highly mixed/uncertain signal post mid-month has characterised each run since midweek. Although a colder, blocked outcome is still favoured at present, it's only marginally so and certainly considerably less than a few days ago.

[Model Output Discussions 12z 30/11/16]

1 hour ago, bluearmy said:

the hope is for the Russian low anomoly to spread sw but I agree that until it shows, would be a slog to advect cold uppers across the uk.  I'm intrigued that the back end of the eps has repeated and in line with the previous 46 dayer at that timescale. the weekends runs will be revealing although having seen low euro anomolys fail once, I won't be too keen to go overboard if they show again unless ian confirms glosea is showing similar  

assume u looked at the eps rather than the GEFS Crewe? 

 

7 minutes ago, igloo said:

grim stuff form the GFS looks like a repeat of december 2015 with storm after storm rolling in and in the later stages a euro slug to go with it nasty i dear hope its over doing things somewhat or it will be a long mild december by the looks of it

Yes, but 12z EC EPS are starting to lean back towards a different mode. Early days.

[Model Output Discussions 12z 30/11/16]

15 minutes ago, Glacier Point said:

Ian, could you qualify 'massive ramp up' as being return to climatology (as GEFS), or above tercile, and whether such similar signal exists for 70hPa or 50hPa layers ?

Return to climatology or exceeding that in some members (albeit with a minority going at times Dec-Feb the other way, into reversal). The persistence of that stronger signal (i.e. the sudden change coming after a period of exceptionally weak PV) was an underpinning consideration in how the initial 3-month outlooks were worded, when they started covering the early winter period and talked of a higher risk of colder/blocked set-ups for *early* winter but - as Adam Scaife blogged - a shift away from that sort of story could then dominate further into the season (latest outlook focuses this more towards Feb). I'm not sure re signals other than at 10hPa.

[Model Output Discussions 12z 30/11/16]

14 minutes ago, Jimmyh said:

What I am not seeing here is anything like the zonal @warm dross that we had last year....

My take on this really has not changed from middle of November. It can and will go either way.

...precisely so, and precisely echoing UKMO thoughts after some recent flip-flopping of output (including yesterday's ECMWF Monthly, which suddenly offered substantial weakening of the previously strong +ve MSLP anomaly north/NW of the UK that it yielded in the run only 3 days prior and in all previous recent runs, for later December. Is it wholly trustworthy, given marked swing? No, not until we see the sense of direction offered in runs on Mon-Thurs next week.

The situation currently is that whilst ECMWF now offers just a bland, weak +ve MSLP anomaly to the west of the UK, it equally - and quite starkly - offers no signal either to suggest a markedly zonal, +ve NAO set-up either....

In other words, whilst GloSea still remains more inclined to return a blocked story later through the month (and ever since output run way back in Sept, has not budged from indicating a massive ramp-up in stratospheric zonal winds later Dec, as now being suggested by GEFS), the forecast outlook by the UKMO remains quite reasonable in the current circumstances (not least as it would be easy to overreact to latest ECMWF in isolation).

Thus, a return to blocking still remains the more favoured outcome, BUT whereas this always had low confidence (based purely on lead time), we now can add a further tier of uncertainty given ECMWF yesterday and the fact that no clear synoptic regime is obviously favoured after mid-month.

Some reactions I've seen on social media have been bordering on hysteria, but there's nothing that suggests 'Dec 2015 Revisited' has suddenly manifested in model output. So it looks set to get milder through next week.... but does that really automatically, unquestionably characterise a whole winter ahead??? Deary-me ;-)

[Model Output Discussions 06z 04/11/16]

4 minutes ago, Nick L said:

How come you're confused? That the 850 temperatures are quite high?

In high pressure situations you'll often get temperature inversions, temperatures at the surface will be lower than higher up. This is why 850 temperatures in these set ups are pretty useless at predicting temperatures at the surface.

Yes, in fact you will actually see the 850hPa values rise under axis of the high cell... due to effects of subsidence at that level. This is not reflective of surface temps rising..   (yet)!

[Model Output Discussions 06z 04/11/16]

16 minutes ago, northwestsnow said:

Have..to...say...this...ecm..day10..looks...poor.insofar...as..any..retrogressive...signal..is...concerned..-

That..Euro-high...is...HUGE!!!!-...I'm...aware...things..can..and...do..change...quickly...but..i'm..struggling...to...see..how..we...would...rid...ourselves..of...that..Euro-high,and..that's..the..9thDecember...

Looks generally in keeping with the script of last 2 days however. Whilst there's been fine balance over what degree of mobility asserts into early Dec, the currently favoured outcome remains for the high to slip SE; some degree of mobility taking over initially, into especially NW/N (possibly more rapidly nationwide: but 30% chance high remains dominant more widely for a while)... This followed by a more concerted period of limited mobility, with greater W-E frontal progression, but this phase looking temporary before blocked patterns re-establish around mid-month onwards. That remains the broad UKMO sequence into medium range.

EDIT ie in keeping with Matt's explanation above

[Model Output Discussions 06z 04/11/16]

59 minutes ago, Man With Beard said:

I'm with you Crewe. I'm in total agreement about the background signals but making deterministic predictions beyond D14 for a small island is a very dangerous game. Now I'm sure our loved and respected posters do not mean to be so deterministic (and loved seeing Tamara post a 1962 chart!), but I fear a level of expectation in here that I've never seen before. Don't do it folks. The weather respects no-one!

But who is offering *deterministic* predictions?? It's one thing outlining a model signal. Quite another saying .. " and that's exactly what will happen by week 3/4 and the resultant sensible weather will be x/y...".  At that range we are merely interested in broad trend/probabilistic output; if folk then run off and interpret that as de facto/fait accompli outcome, then that's hardly a reflection on the model... they are merely guides, NOT gospel!

[Model Output Discussions 06z 04/11/16]

1 hour ago, KasimWeatherQuarnford said:

To be truthful, the current set up does have a heck of a lot of high pressure anomaly to the North West, so the EC was actually correct for the current time.

What was lacking was signal for below average mslp + 500gph anomalies to the South of the UK, with just bland old high pressure to the North, then we set up the current anticyclonic set up.

 

  Reveal hidden contents

 

Can't get rid of the spoiler

Actually I think the broad pattern match on GPH anomalies wasn't at all bad 3 weeks ago, considering lead time; inevitable weakening of signals at that range and the 'iffy' phase across all models that swung to and fro, as we all recall, as they modelled the late Nov into early Dec period. Here's EC Monthly 500hPa anoms from 3 weeks ago versus 12z EC today for same period (now-next few days). Note low heights towards Spain and +ve anomaly to our NW; plus -ve one Scandinavia-W Russia. 

[Model Output Discussions 06z 04/11/16]

8 minutes ago, CreweCold said:

All this talk of week 4....it has been stuck at week 4 for what seems like an eternity. Week 4 takes us to nearly the end of December....if this is to be a front loaded winter, that doesn't leave us much time afterwards to play with if it doesn't materialise at all by then...

??????? Eh?   The next 7 to poss more days are signalled as below average!! The next 2 nights look firmly set to be coldest of this autumn!! I suspect there's people so obsessed with snow forecasts that they're not seeing wood for the trees. No cold: no snow. And as I posted above, usual suspects in media aside, nobody sensible has given a snow prognosis 10+ days ahead... because the forecasting science to do so with any reliability doesn't exist. 

Honestly, compared to some recent winters(!), I fail to see anything in longer range models to make cold lovers moan. 

[Model Output Discussions 06z 04/11/16]

1 minute ago, Ramp said:

I tried the caution approach until lower down the page where GP was ramping a white Christmas

Nobody (sensible) should be automatically equating the signals with snowmageddon (or lack of snow for that matter). The broad pointers are to a cold set-up. That is very clear. Issues re snow are a matter for much shorter lead times.

[Model Output Discussions 06z 04/11/16]

EC Monthly consistent with Glosea and it's last runs. Strong resurgence of +ve MSLP/GPH anomalies north of UK further into Dec, especially Xmas week, with marked -ve temp anomalies developing across adjacent contiguous Europe as N'ly flow veers to mean easterly late Dec. Good to see both suites in such firm agreement, despite inherent issues of confidence at such lead time.

[Model Output Discussions 06z 04/11/16]

2 minutes ago, feb1991blizzard said:

And towards drier (very markedly) as well though?

Yes, re majority signal. Happy days. :-)