Model output discussion - cold proper on the way?

[Met4Cast]

The 00z models have seemingly struggled with the US storm which is still causing issues within the model outputs, it’s really not surprising to see such wide variation. 

Could the ICON/ECM be right? Sure, could the GFS/UKMO be right? Yes. 

Calling this one way or another isn’t possible right now, so being so reactive to a ECM Det run doesn’t make a whole lot of sense

[Roger J Smith]

The accuracy debate is rather pointless when you consider that the performance of almost the entire set falls within a narrow range, so it's a case by case decision that forecasters have to make -- the European model is not infallible and the GFS or even the GEM are not always wrong when there's a difference. I tend to the view that it's maybe a 60-40 proposition when these camps form (we sometimes see different combinations, I can recall a Euro-GEM vs GFS-Icon split earlier this winter). I do give top marks to the reality model, it verifies at 100% at the zero hour time frame. 

 

[The Eagle]

Just now, smith25 said:

I've seen this drama happen a few times with the models flip flopping just before a cold spell and usually over a weekend.  This drama wont stop until late sunday and i bet things will be much better for coldies then... Try not to panic!

 

Maybe. UKMO is not bad for example. I can see how we get deep cold from 144. The rest...

We need a small miracle today.

[Captain Shortwave]

Two areas of frustration here;

First the ECM, a whopping deep and broad trough forming over north America around day 5 which seems broader than the other models, that is where the Atlantic pattern flattens out and the trouble is once that initial chance goes the ECM locks the pattern in with that Pacific ridge/US trough combo which leaves us with a fairly modest westerly flow (sort of like winter 2014/15)

The second issue is even if the initial ridge works we still have the finer resolution of the cols pool to our north east, note how the UKMO and GEM are holding a cold pool over Norway where the deepest cold is held and forming its own little circulation which will cause trouble for advecting the cold in our direction.

Ah the perils of easterlies, last year was less stressful because you already had the Scandi high in situ before things really kicked off.

Still could all change, but i must stand back in awe at what could happen across parts of the US in the next ten days, to be hit with such brutal cold (sub -20c 850s) not once but possibly twice in that period.

It could still change, you have two solutions at day 4/5 from models which should be close to getting the pattern right at that timescale so fingers crossed for better output later.

Edited January 19, 2019 by Captain Shortwave

[Banbury]

Can someone tell me 'if' ICON is so  good why do people say " the big 3" 

Edited January 19, 2019 by Banbury

[booferking]

10 minutes ago, Daniel Smith said:

It goes wrong at 96hrs. 

The trough sinking South of the UK has a better disconnect from the system exiting the US which allows ridging further NE thay eventually leads to the Easterly. Not the subtle difference on the tip of Greenland  

The ECM keeps the two systems “attached” and thus we see less ridging & no Easterly 

The difference between full blown Easterly & flat Westerly is incredibly small, the fact we haven’t seen far more wobbles in the output is surprising because of this. 

UKMO/GFS both make it, ECM/ICON doesn’t. 

Ensembles aren’t really going to be any help whatsoever because their lower resolution will make it even harder to resolve the situation so for that reason ensemble means likely to be fairly poor (reliability wise) until this is resolved. 

96hrs.. we have to see resolution very soon, one way or another

Yep all to do with how fast you drain energy on Greenland tip if you go back through the runs you will see that is the critical part for runs that make it and runs that dont.

[Met4Cast]

Just now, Banbury said:

Can someone tell me 'if' ICON is good why do people say " the big 3" 

Big three = ECM, UKMO, GFS

They’re the top performing models globally. ICON is generally quite poor 

[kold weather]

4 minutes ago, Banbury said:

Can someone tell me 'if' ICON is so  good why do people say " the big 3" 

If I remember correctly that the icon was originally the DWD. It normally ran close to the ECM in terms of pattern.

It's specialty tended to be continental flows and high ressure cells in the euro area. Almost like an anti GEM which tends to be too progressive. 

It's had big upgrades since then and is generally regarded as a decent enough model, got some features not even the ECM/GFS has yet.

It's probably around similar level to GEM, which itself has been outperforming at least the two versions of GFS last month.

Edited January 19, 2019 by kold weather

[legritter]

16 minutes ago, snowangel32 said:

If the ECM, ICON, switch back i will go skinny dipping in lake Windermere lol. It's the other models as well though they took a big step towards the Icon.

Better go skinny dipping quick  as the  water will be   frozen over soon , charts from Ecm  are on a wobble again , but still we have a cold signal ,fingers crossed ,cheers gang , back to bed now to dream ,cheers .

[Jason M]

1 minute ago, Banbury said:

Can someone tell me 'if' ICON is so  good why do people say " the big 3" 

Because it’s not very good. That said you simply can’t ignore any output at days 4 through 6. 

 

[The Eagle]

Just now, Banbury said:

Can someone tell me 'if' ICON is so  good why do people say " the big 3" 

I think because it is only live little over a year. Some stuck in their ways? It's just as fallible as any of "the big 3" but it's a really good model that needs more attention than it gets here.  

 

[Ben Sainsbury]

As many have said already it is a knife-edge scenario! The low coming out of the US is EXTREMELY hard to model, and the likeliness of any models nailing its track and intensity down at this time is still unlikely. Until we know how it acts in the next 48 hours is crucial to what we see. If the low takes an unexpected track we could still see large changes at such a small range.

I hardly see people comment about the GFS P when it does verify better than the standard GFS. Don’t forget GFS P also follows the ICON & ECM solution.

Edited January 19, 2019 by Ben Sainsbury

[Man With Beard]

Oh

So we now have the rather awkward situation, at just T72, of the ECM op being at odds with... just about every other run, including its own ensembles. 

The ECM should be the leader here, but it is so far out on its own, it seems hard to believe. 

Help! 

edit - OK, just seen the FV3, so it is not completely on its own. 

Edited January 19, 2019 by Man With Beard

[The Eagle]

1 minute ago, Jason M said:

Because it’s not very good. 

 

Why is it not very good?

 

[Met4Cast]

We’re never going to hear the end of it from ICON lovers of by pure luck it has managed to call this one

[kold weather]

Just now, Man With Beard said:

Oh

So we now have the rather awkward situation, at just T72, of the ECM op being at odds with... just about every other run, including its own ensembles. 

The ECM should be the leader here, but it is so far out on its own, it seems hard to believe. 

Help! 

Not sure that is true, GFS ensembles are about 40-60 in favour of cold, para GFS is similar as is icon.

Looks almost 50-50 balance overall to me.

[seabreeze86]

Don't panic everyone the CMA still going for it, seriously though this could go either way if the UKMO looked like the ECM they I'd be very pessimistic, massive 12zs later

[northwestsnow]

Not much joy on EC det this morning with any snow tues, and it looks a bit scarce away from northern hills generally..

Wouldnt want to be writing Exeters outlook today, we wait for the ens..

Edited January 19, 2019 by northwestsnow

[The Eagle]

1 minute ago, Daniel Smith said:

We’re never going to hear the end of it from ICON lovers of by pure luck it has managed to call this one

And the summer heatwave, and the two separate easterlies last Spring. All luck?

[kold weather]

3 minutes ago, Daniel Smith said:

We’re never going to hear the end of it from ICON lovers of by pure luck it has managed to call this one

In all fairness it nailed both easterlys last year as well, better than any other model.

As I said it's actually got some good features that n it's programming and that is cutting edge so I'd not say it's poor, it's at worst a middle level model and one worth watching.

Edited January 19, 2019 by kold weather

[Met4Cast]

1 minute ago, seabreeze86 said:

Don't panic everyone the CMA still going for it, seriously though this could go either way if the UKMO looked like the ECM they I'd be very pessimistic, massive 12zs later

Ironically the CMA has been the most consistent out of all the models..

It’s going to be a very tense 6z GFS this morning

[mountain shadow]

3 minutes ago, northwestsnow said:

Not much joy on EC det this morning with any snow, and it looks a bit scarce away from northern hills..

Wouldnt want to be writing Exeters outlook today, we wait for the ens..

What  actual use are the ens in an all or nothing scenario? 

The ECM is going down a route that only had 8% cluster support yesterday. 

Edited January 19, 2019 by mountain shadow

[Rayth]

Just now, seabreeze86 said:

Don't panic everyone the CMA still going for it, seriously though this could go either way if the UKMO looked like the ECM they I'd be very pessimistic, massive 12zs later

Is this the Chinese model? It’s been steadfast 

@northwestsnow indeed , given the positive update overnight  

[Lloyds32]

Looks like to me the ECM has the pressure sunk over the U.K. Looks like my feeling was right. I have noticed the change since yesterday mornings ECM. Cold uppers pushed away at 144 and gone at 168.

 

[kold weather]

Just now, mountain shadow said:

But actual use are the ens in an all or nothing scenario? 

The ECM is going down a route that only had 8% cluster support yesterday. 

GFS support was also only at 8-10% yesterday as well, now it'sabout 60-40 in favour of cold. Range of options gave opened up.

Regardless of what the 06z shows, someone PLEASE save that 00z GFS one, it THAT GFS run for good or I'll!