Model output discussion - winter proper underway

[northwestsnow]

27 minutes ago, FiftyShadesofWine said:

Is that a 3 wave pattern at t120, troughs and ridges looks bizarre

Really like the look of that...

[Man With Beard]

Very interesting seeing how the ECM ensembles have developed over the last 5 days. The ridge has moved NW a little on each one.

        

 

[northwestsnow]

Looking increasingly possible its going to get colder for christmas itself IMHO..

[shotski]

2 minutes ago, northwestsnow said:

Looking increasingly possible its going to get colder for christmas itself IMHO..

It certainly looks more interesting than it did at the start of the week but I can’t see anything extreme or deep cold on the cards ? A sort of repeat of what we had last week with an Atlantic Ridge ? Or am I missing something ? 

[Blessed Weather]

Interesting discussions earlier about which model has been performing best at the moment in our neck of the woods. It does seem that GFS has been doing well lately. So while it's quiet, here's some NWP model comparisons.

For some reason the European plots wouldn't work, but here's the latest (Aug 2017) RMS error rates (lower is better) for the 500hPa NH forecasts up to 240 hours. ECM clearly performing best throughout; with barely a hair's breadth between the other models at 120 hours; but at 240 hours GFS moving into 2nd place. http://apps.ecmwf.int/wmolcdnv/scores/mean/500_z

Next, courtesy of the Canadian Met, the RMS error rates for the 500hPa forecasts at 120 hours each month since Nov 2011 to Oct 2017. I'm assuming this covers forecasts for the USA and Canada. Looking just at 2017, ECM again clearly ahead, although 2nd place switching back and forth between UKMet and GFS. https://weather.gc.ca/verification/monthly_ts_e.html

[Chesil View]

52 minutes ago, Man With Beard said:

The ECM - not long ago it was the one model that overdid amplification between D7 and D10 - but since its upgrade last winter, I've felt it has actually slightly underdone amplification. I wonder, then, if this backtrack tonight is just the beginning of a move to another potent northerly.

I thought the very same thing MWB its almost as if they wrote an algorithm to counteract its overamplification tendencies and now it underamplifies instead.

[shotski]

3 minutes ago, FiftyShadesofWine said:

Prob get told off again but here you go...ext eps...

 

 

 

Looks pretty mobile/typical set up to my un trained eye ? 

[Summer Sun]

UKMO extended still has enough cloud under the high to produce some rain at times

[bobbydog]

Just now, Summer Sun said:

UKMO extended still has enough cloud under the high to produce some rain at times

thats good news SS. i'm glad you told us that....

[nick sussex]

I think its best to wait a few more runs in terms of the ensembles.

At the moment we've seen an increase in amplitude in the ops. Is that the end of that trend or will it continue? Its always risky drawing too many conclusions when you're moving from one pattern to another.

So we saw the positive PNA, then that relaxed and now we're seeing a bit more amplitude developing.

 

 

 

 

[Blessed Weather]

4 minutes ago, ShortWaveHell said:

That’s me told ?

Not at all SWH and it certainly wasn't my intention to challenge anything you said. Just interest on my part about how the models are performing. It's frustrating not to be able to get hold of stats for Europe. I agree that GFS appears to be doing somewhat better than ECM lately - but of course final verification of current forecasts catching our interest will be the proof of the pudding!!

[Methuselah]

Right then, with Nick's immortal words of advice (more runs needed!) still ringing in my ears, I'll bugger off and listen to the radio...

Tomorrow's the day!

[bobbydog]

1 minute ago, Ed Stone said:

Right then, with Nick's immortal words of advice (more runs needed!) still ringing in my ears, I'll bugger off and listen to the radio...

Tomorrow's the day!

just a shame the england cricket team didn't take nick's advice...

Edited December 14, 2017 by bobbydog

[feb1991blizzard]

I also think its important to not get wires crossed here, IMO people thinking that bout of amplification inside 10d range will lead to a white xmas or indeed any chance of a potent outbreak in the xmas week will be wide of the mark, however, it could be the first stage towards getting an Easterly in for the first week of Jan, all I am looking for over the next week to week and a half is that a reasonable number of ensembles start bringing the scandi high nearer timeframe wise and ultimately (hopefully) start seeing flatliner runs sub -10c in quantity, this could actually depend on the trough and frigid cold dropping nicely into Europe first.

[Daniel*]

5 minutes ago, FiftyShadesofWine said:

Not really as again it doesn't distinguish the key area around Greenland, which GFS totally out performs ECM, all day long.

Steady on no? GFS has got this renewed amplification signal seemingly right possibly fitting with MJO progression. ECM years ago I’m sure would have been on it first, ‘chucking’ ridges to our north. ECM is good at fine tuning GFS is pretty good in the broad pattern sense of things in a little area off Greenland I suspect not. 

[phil nw.]

The ext eps shown earlier pretty much mirrors the favoured Gefs 500 pattern

A westerly flow for the UK with the outpourings from the Canadian trough extending east into the Atlantic.Just signs of ridging north from height anomalies over E.Europe,nothing dramatic yet.It all looks pretty average with probably the best chance of some colder shots is from the nw depending on the amount of amplification in the Atlantic jet.

Edited December 14, 2017 by phil nw.

[bluearmy]

maybe something, maybe nothing

the cluster from the 00z eps at T204 which best represents where we find ourselves at T192 12 hours later becomes the smallest at 16% post day 10. It's the one which keeps the upper ridge amplified across the uk and eventually establishes further ne over w scandi by day 15. 

probably nothing and looking at the eps mean/anomoly, I doubt it's grown too much on the 12z suite 

[Singularity]

2 hours ago, nick sussex said:

Oh dear  whats happened to the ECM and UKMO.

Both have performed poorly for several days and are now having to climbdown, very unusual to see this given its normally the GFS which has a flat bias.

Looks like NOAA were right to berate recent ECM op runs.

The once star of the model stage turning into Norma Desmond!

Can't speak for UKMO as its MJO output has been missing from these plots for some time now, but the ECMF prediction has increased in amplitude with each of 5 new days in a row now, taking it ever closer to the enthusiastic GEFS. This isn't the first time this season the model has been correcting toward GEFS either, so this is likely a factor in those periods of strangely flat output relative to the other models (or even in general, in this most recent case) that have been observed. 

With the Alaskan ridge it really is critical to get it sufficiently poleward enough to flip the switch to easterly potential from relentless westerly dross.

Not sure why GEFS have so many members (13 or 14 out of 20; one of those only just) going in the zonal direction given the MJO projections. Could it be a resolution thing... or a sign that even with such MJO amplitude, victory is far from assured...?

[lorenzo]

21st to 22nd of December remains an absolute moving feast of anything and everything as far as modelling is concerned. Sorry for plugging in strat charts on NWP thread but entirely relevant and imprints on model volatility imo. Dr Butler showing interest in flux events 'if verified' that's good enough for me given the body of work delivered.

In 48 hours things go from one extreme to another..

And a helluva bunched up vortex 

Neat cold pool building out east...

And just for fun the wacky analogs go for a prime year as top analog two days in a row...one for the history lovers / statisticians.. am sure many of the other cited years were horrendous!, nevertheless - adds to the really enjoyable viewing this season..

[damianslaw]

Short one from me - as won't be in the thread again until later Monday - the watch word - amplification, ECM trending towards a more amplified flow as we approach christmas, GFS is holding firm.

I expect come 4 days time, both will be resolute with this theme for the last week of the month, azores high forced NW with a trough swinging down NW-SE through UK just 23/24 Dec weekend, possible snow in the north, heights building to the NE, then the azores high sucked back NE elongating and ever stretching towards scandi, setting up a cold frosty high. Could be wrong but that's what I'm expecting to see. So a few mild quite lacklustre days from 18-22 Dec then a change. 

Edited December 14, 2017 by damianslaw

[Daniel*]

32 minutes ago, FiftyShadesofWine said:

Gfs models Greenland far better than any other through the years, you have only to look at the shortwave ECM debacle of 2014 and others to see it, ECM over amplifies or flattens around Greenland, it's poor in that specific area, always has been.

That region of the world all models are terrible at it.

The number of Greenland highs we’ve seen churned out on both GFS and ECM to a lesser extent? Getting a Greenland high in winter time is very rare, I’m always extremely weary unless within 120. We do not need a blockbuster of a Greenland high sometimes the most potent cold spells have been brought by a slither of +ve heights to north.

[I remember Atlantic 252]

if it's mild, let's have it dry! chart dosen't stink! should feel decent, if it's wet and windy at 5C or so that stinks

Edited December 14, 2017 by I remember Atlantic 252

[feb1991blizzard]

Edited December 14, 2017 by feb1991blizzard

[tight isobar]

A very tropical maritime looking gfs 18z into next week!!!

The signal needs to keep repeating' for decent evolve from around 21-23dec!!!

Or the panic will rise to the surface in here!

Edited December 14, 2017 by tight isobar

[Man With Beard]

4 minutes ago, feb1991blizzard said:

Yes was just scanning those. The 850s are average or above average until the end of January.

Probably doesn't mean much.