Model Output Discussion 29th December - Into mid-Winter.

[I remember Atlantic 252]

41 minutes ago, Skyraker said:

 

Might get a polar low out of this one, worth checking the visible sat images while this is going on as they materialise out of thin air.

not a polar low, but being modeled on GFS as a boxing day 2014 style event, or even a Nov 19th '96 event 

UKMO may go same way

[northwestsnow]

Actually having looked at ukmo 144 im not convinced.That low moving across the Atlantic has spoiler written all over it...( imho)

That said it will still be cold until sunday at least so its still brill, esp for those getting some snow through Thursday.

Hoping for an EC victory..

Edited January 8, 2017 by northwestsnow

[Frosty.]

7 minutes ago, marksiwnc said:

Tbh you can almost laugh at these charts.  I'm hoping the ECM and UKMO is correct but I can't see the GFS being wrong when showing mild.  It almost guaranteed to be correct knowing our luck.  Here's to hoping...... Forecasters nightmare every week 

Just enjoy the cold blast later in the week, it will be the best wintry spell so far..most of the uk has a chance of snow, too often on here we end up worrying so much about the end of the cold that we don't appreciate it when it's happening!

Edited January 8, 2017 by Frosty.

[Man With Beard]

5 minutes ago, northwestsnow said:

Actually having looked at ukmo 144 im not convinced.That low moving across the Atlantic has spoiler written all over it...( imho)

That said it will still be cold until sunday at least so its still brill, esp for those getting some snow through Thursday.

Hoping for an EC victory..

Yes I have to agree with this. Agreement for a snowy blast now looking good up to Saturday (southern areas will have to wait and see if they actually will join in until T48 I think, in case of a last minute pattern shift to the east).

But beyond this, we have too many different solutions. The ECM may be consistent, but no other op appears to be backing it. The UKMO has this strange low in the Atlantic that might prevent ridging later on - it looks wrong but the UKMO has called odd patterns right before. And you wouldn't want to trust the GFS steamroller, either. I wouldn't like to call things beyond Friday right now.

[Ali1977]

If the EPS follow the Op I'd be surprised, they didn't the last few times the Op toyed with an Easterly - if it does follow them game on and lets hope the next GFS starts heading that way. Not convinced!!

[bluearmy]

4 minutes ago, northwestsnow said:

Actually having looked at ukmo 144 im not convinced.That low moving across the Atlantic has spoiler written all over it...( imho)

That said it will still be cold until sunday at least so its still brill, esp for those getting some snow through Thursday.

Hoping for an EC victory..

At last someone mentioned it.

its a real wildcard feature. Fwiw, the three GEFS members that go on to build a scandi ridge ala ECM all have it to a degree. ukmo seems over developed on it at day 6 in comparison and if it just toddles along in the general flow of the jet thereafter then it's actually a potential snowmaker for areas to its north and east as it cross the country. but it's a day 6 ukmo chart so little developments like this need to be treated with caution. if it's there on the 12z run then we need to start looking at potential effects re what it could do to the evolving pattern. The Azores shortwaves are fresh in the mind as we watch e Europe freeze in our trough!  (Bit simplistic I know but you get the gist).

excellent continuity from ECM and the GEFS, whilst not looking good on the mean and uppers,  do head the ECM way re their anomolies and the mean anomoly 9/12  is a suppressed scandi ridge with anomalous cold surface temps across s half uk. 

Anyway, eps due in thirty mins and whilst I don't expect them to scream easterly, the spreads will be important to see if support for the op solution is gathering pace or has peaked.

 

[bluearmy]

7 hours ago, Trom said:

Excellent post - the obsession with mean runs (which I haven't noticed in prior years) in FI seems very odd to me.  As an Economist looking at data distributions and describing them by mean without a dispersion measure seems pointless.  Even with a dispersion measure you would still want to see skew.  From my perspective all these charts give you is a very broad brush trend of all the ensembles.  When you look at the dispersion patterns in FI they are so vast. Median would be a better measure given how most people use them.  There are dispersion measures using standard deviation which give a good idea of how well the op is supported relative to ensemble mean.  Nick F posted one back on page 168.  I'd like to see much more of this data being posted on the model thread.  The charts he posted showed exactly where the models had the greatest variability in the northern hemisphere.  We have a lot of data produce by the models but this measure goes under reported on the forum.  Essentially it's telling us where the models are solid and where they are uncertain.  Nick F if you could post more (or let me know where they can be found) that would be great.

Ecmwf.int and go on forecasts (medium range). The ensemble is what you are after. Has a great explanation on there too. Maybe some of you should read it ! 

[SMU]

Morning All

agreed with all the comments above - absolute crackerjack of an ECM however UKMO not on a par today because of the atlantic low & indeed by 168 appears pretty poor ( athough behind the low there is NE ridging so very difficult to tell )

That makes me uneasy as the ECM doesnt show this profile or evolution

So all in today fantastic charts up to 144 thereafter a lot of divergence - but not without potential- at least its UKMO 144/168 so room for change... ( as is the ECM )- remember the other week 38 ECM ensembles clustered to cold V UKMO... we know who was the winner

Thursday overnight into Friday still sees the biggest window ATM...

S

Edited January 8, 2017 by Steve Murr

[IDO]

Well no conclusion just yet, though developments towards a middle ground, maybe?

Anyway the GEM has flipped to yesterday's GFS take on things, but we know at that time there was a split cluster, so probably not decisive.

GEFS D6-10 have moved towards varying degrees of a ridge building towards the UK, with a good cluster (>33%) trying to build NE further towards Scandi, but not really close to the ECM take. ECM in these scenarios tend to give us synoptically perfect charts and that is one reason I discount them without good support. No point guessing but I would lean towards a continuation of the winter pattern with the mobile PV and waves of heights building from the Atlantic/Azores, with maybe the jet closer to the UK due to the strengthening PV, so a high close to the UK? FI on the GEFS are lame if cold is your thing, so really even a half-hearted attempt towards the ECM take is not going to help; we need the ECM to be very close if a pattern change is forthcoming to a colder setup.

D10 means on the GEFS: 

They highlight the fine margins there are to get anything prolonged out of this.

As for the cold spell the Arpege has significantly lower uppers arriving so not as snowy as the GFS run:

T108: 

This should resolve in the next day or so as the GFS has a cold flow pushing hefty showers into western Northern regions and more likely inland, where as the ARPEGE will restrict more to coastal regions.

[IDO]

7 hours ago, Trom said:

Excellent post - the obsession with mean runs (which I haven't noticed in prior years) in FI seems very odd to me.  As an Economist looking at data distributions and describing them by mean without a dispersion measure seems pointless.  Even with a dispersion measure you would still want to see skew.  From my perspective all these charts give you is a very broad brush trend of all the ensembles.  When you look at the dispersion patterns in FI they are so vast. Median would be a better measure given how most people use them.  There are dispersion measures using standard deviation which give a good idea of how well the op is supported relative to ensemble mean.  Nick F posted one back on page 168.  I'd like to see much more of this data being posted on the model thread.  The charts he posted showed exactly where the models had the greatest variability in the northern hemisphere.  We have a lot of data produce by the models but this measure goes under reported on the forum.  Essentially it's telling us where the models are solid and where they are uncertain.  Nick F if you could post more (or let me know where they can be found) that would be great.

Not really true. With the ECM we have no public record information as to clusters, so we are relying on what we get, the mean being one of these. The fact the mean has around a 50% better verification performance level means it is far more useful than an op run, yet so much chatter is on D10 ECM  wonder charts!! You should maybe say that the op charts should be ditched after D9, not the mean charts. These are the D10 verification for GFS:

mean  OP 

Clearly for statistical analysis on the data in the public domain the mean is far more relevant than any other information, including the op after around D9, if like the ECM you do not have cluster charts? 

[TEITS]

Thought I was looking at different charts this morning because the UKMO isn't anywhere near as good or like the ECM. The UKMO would only bring a weak ridge of high pressure following the low at +144 before a return to W,lys like the GFS.

If I put my unbiased hat on then in my opinion none of the models are correct. I feel the GFS is too quick in bringing a return to milder W,lys. The high pressure building on the ECM bringing the E,lys is more likely to be further S bringing E,lys to the Med rather than the UK.

[knocker]

Actually, unless I'm very much mistaken (which has been known ) the UKMO low was also on the GFS hanging around with intent at 120 and it's this low that nips up the channel and produces copious snowfalls over Dartmoor and Exmoor.

Edited January 8, 2017 by knocker

[Frosty.]

Let's enjoy what the Ecm 00z shows and HOPE its nailed it..it's a beautiful run for coldies with so much potential beyond day 10 and its been hinting at this on previous runs recently!

The good news is a cold outbreak is set in stone, the longevity is still unclear but I'm cheering on the ecm!

[swfc]

12 minutes ago, TEITS said:

Thought I was looking at different charts this morning because the UKMO isn't anywhere near as good or like the ECM. The UKMO would only bring a weak ridge of high pressure following the low at +144 before a return to W,lys like the GFS.

If I put my unbiased hat on then in my opinion none of the models are correct. I feel the GFS is too quick in bringing a return to milder W,lys. The high pressure building on the ECM bringing the E,lys is more likely to be further S bringing E,lys to the Med rather than the UK.

good points 

id agree on that esp regarding the ecm.it would appear to be a good evolution but looking to the nnw and the pv plus the general trend for heights to sink id agree that any easterly would go to our south

it has been shown earlier but imo looks unrealistic altho i hope im totally wrong!!!

 

[Weather-history]

Looking at the Manchester GFS 0z ensembles is how anticyclonic they are after next weekend. Whatever happens after next weekend, it doesn't look like we are going to go into a stormy pattern. It's the story of this winter thus far, anticyclonicity never too far way.

[IDO]

Looking at the D10 mean from ECM, that is clearly moving away from support for the op, and another ECM backtrack expected this evening. In fact the GFS and ECM d10 broadly on the same page:

 

[Captain Shortwave]

Consistency with the ECM op and in fact ensembles, the mean again flatter but the evolution looks similar with the Azores high ridging into Europe over the top of low heights in the Mediterranean.

      

So issues earlier on.

Well lets take the day 5 charts 

ECM/GFS/UKMO

     

Differences in the Atlantic here, the ECM shows a broad ridge (A slight feature in the middle but insignificant), the GFS develops a very shallow trough in the middle of that ridge which runs across the south of the UK later on which does bring snow away from the far south coast. The UKMO makes a much bigger deal with this which actually amplifies the northerly downstream further initially. It looks to well developed to give a snow event like the GFS and could flatten the patter later on. 

So the ECM looks solid in its evolution, the GFS looks to be finding complicating secondary features to break the pattern down but could very well give some snow events before we get there, the UKMO looks iffy but we should see a good few days of cold with the risk of snow for many places hopefully.

Edited January 8, 2017 by Captain Shortwave

[northwestsnow]

Ec mean looks goid to me again keeping the UK cold through the weekend

Im still thinking the fun could start wed night into thur morning for elevated nw britain..

[Polar Maritime]

17 minutes ago, knocker said:

Actually, unless I'm very much mistaken (which has been known ) the UKMO low was also on the GFS hanging around with intent at 120 and it's this low that nips up the channel and produces copious snowfalls over Dartmoor and Exmoor.

Yes Knock the South certainly not out of the question for Snow, Still 6 days away so plenty to keep tabs on re detail. As shown by the Net/Wx M-R Model.

Edited January 8, 2017 by Polar Maritime

[Penrith Snow]

The ECM run is straight out of the 1960's with a north westerly bring snow to the north west, then a northerly bringing snow to the east and the Midlands, then an easterly bringing snow to the south east.

Unless you live on the Scilly Islands it has snow for everyone! Set ups like that in recent years or even decades have been very rare so I do very much doubt it will verify, a solution that brings high pressure across England and another easterly plunge into the Med is the most likely outcome IMO.

However Weather Porn doesn't get any better and I will spend the day looking at it before it vanishes at 7pm.

Andy

[Frosty.]

1 minute ago, northwestsnow said:

What do you mean another ecm backtrack.seriously, its boring now.

Agreed, we have a cold outbreak in the bag, hopefully it will last into week 2! The Ecm 00z op could be right this time.

[knocker]

The EPS 8-13 mean anomaly which I can't post is not hugely different to last evening. The main player is the Canadian vortex with a deep trough over Alaska and a lobe over Greenland with associated trough in mid Atlantic with ridging over the UK. Thus WSW upper flow with temps around average.

[SMU]

@Trom

here are the ECM spreads

144 - shows a small cluster in the atlantic relating to what UKMO shows - ( the low lifting out from under the high & moving E towards the UK )

NB : GFS op @ 168 is the fastest east of the 3 models showing a band of PPN crossing the UK- expect that to trend west in the following runs

ECM spreads 168

clearly some divergence around the amplitude of any wave coming off the states

ECM 192

The uncertainty spreads NE & magnifies with time relating to the possible high pressure... 

S

 

Edited January 8, 2017 by Steve Murr

[Stephen W]

Mmm...I'm not liking the shift in the GFS ensembles this morning after 144.  Whilst the ECM eps still show a cold cluster, they've been known to flip on mass. UKMO at 144 is also picking up on something with that low out to the west and doesn't look as good.  I'd be very surprised to see the easterly develop after the initial cold shot. GFS has a lot of energy in the northern arm of the jet which results in a MLB again which could give us mild sw's or a cool continental feed depending where it settles. That seems to be the form horse this winter but hoping for something different and more interesting of course to verify. Just have to hope the intensity of cold from this initial shot at the end of the week remains to give a taste of winter for some.

[IDO]

^^^ Steve

Looking at the D9 spread on T850s and it is clear the ECM is clueless upstream :

That sort of entropy will produce synoptic outliers and the op could well be in that mode.