Model Output Discussion October 2012

[The Eagle]

Am I alone in thinking that any northern blocking, whilst it's good to see, might be just a bit too early or is it ok at this stage?

Obviously it would be better in Dec-Feb for example, but is it a good thing to potentially have it coming up now, or a waste?

Was no harm late October and through November in 2009/10 winters. Many areas saw ice and snow at the time very early.

Edited October 16, 2012 by The Eagle

[Severe Siberian icy blast]

No not on your own.

We don't get many stabs at these Synoptics and to get them in October is a waste as is a southerly plume this late.

The weather is not without a sense of humour it seems.

Well to be fair we have had these Synoptics since April , northern blocking has dominated us for the last 6mths now , long may it continue, I understand what you mean though, but if the warm pole can prevent the vortex from forming now and in November then we really may hit the jackpot this winter, for the vortex to form then the poles need rapid cooling over the next 6wks , and as of yet the vortex is showing no sings of that, intact quite the opposite.

[Nick F]

Am I alone in thinking that any northern blocking, whilst it's good to see, might be just a bit too early or is it ok at this stage?

Obviously it would be better in Dec-Feb for example, but is it a good thing to potentially have it coming up now, or a waste?

You are not alone, Octobers of 2002 and 2003 had some quite strong Nern blocking in place bringing those months below par CET average wise. But the following winters were little to write home about.

Although Nern blocking may hang into late Autumn, there's no guarantee it will persist through winter.

Quite alot of uncertainty on where the block will end up in the meantime, though there's still a signal for a mean trough to the SW into the medium term.

Edited October 16, 2012 by Nick F

[SMU]

I will try & do a longer post later-

However to have these synoptics now especially into november is a good thing- as I said though more explanation later-

Remember people the 06z is the worst run of the day so unless it bears some resemblence to the overnight outputs then its bin fodder..

S

[Liam J]

I will try & do a longer post later-

However to have these synoptics now especially into november is a good thing- as I said though more explanation later-

Remember people the 06z is the worst run of the day so unless it bears some resemblence to the overnight outputs then its bin fodder..

S

If the 6z is the worse run of the day then why should we give it any credit at all?

Edited October 16, 2012 by Liam J

[SMU]

NO.

in a word- nor the 18z.

S

[Laurence Hill]

Steve

Why are the 6z and 18z unreliable?

[Liam J]

NO.

in a word- nor the 18z.

S

Then if the 6z run continues the synoptics on from the previous, how would your perspective change? ie - for the bin if it doesn't show resemblance to the 00z.

Edited October 16, 2012 by Liam J

[A.J]

So let me see if I've understood correctly what's been previously posted.....If the 06z runs mirrors the 00z (ie showing synoptic patterns that gets cold lovers interested) then we should pay close attention to it....If, however, the 06z differs from the 00z run then we should ignore it because it's an 'unreliable' run.......hmm....interesting

[swfc]

So let me see if I've understood correctly what's been previously posted.....If the 06z runs mirrors the 00z (ie showing synoptic patterns that gets cold lovers interested) then we should pay close attention to it....If, however, the 06z differs from the 00z run then we should ignore it because it's an 'unreliable' run.......hmm....interesting

And now im really confused!!!!

[BLAST FROM THE PAST]

Am I alone in thinking that any northern blocking, whilst it's good to see, might be just a bit too early or is it ok at this stage?

Obviously it would be better in Dec-Feb for example, but is it a good thing to potentially have it coming up now, or a waste?

Am I alone in thinking that any northern blocking, whilst it's good to see, might be just a bit too early or is it ok at this stage?

Obviously it would be better in Dec-Feb for example, but is it a good thing to potentially have it coming up now, or a waste?

Wasted synoptics don't exist. December is likely to be as it is whether this block forms as per GFS or not. BFTP

[Summer Sun]

Next week continues to look much more settled than this week

Pressure slowly begins to rise to the east

By Wednesday this shifts west to move over the UK

By Thursday it shifts a bit further west

By Friday its acting as a blocking system sending the cold air into eastern Europe whilst we stay under high pressure

Into the weekend and the following week high pressure stays over the UK

This is turning out to be a much more settled run the 00z could have just been a tease, lets not forget GFS was showing a more settled outlook for next week over the weekend and earlier this week

Edited October 16, 2012 by Summer Sun

[swfc]

Oh well it was nice while it lasted i guess!!!!!

[AWD]

Some very cold upper air temps for the time of year as the HP slowly drifts NW.

http://www.meteociel.fr/modeles/gfse_cartes.php?&ech=252&mode=1

[bluearmy]

i notice references to northern blocking and atlantic chart thumbnails being posted. whilst its good to see blocking in our part of the NH, at this time of year the hemispheric profile is whats important re the upcoming winter. whilst the greenland area might look blocked, a developing raging large polar vortex elsewhere in the NH would not be great news. surely much better to post the NH view ?

[The watcher]

When Northern blocking does occur (as is showing) it tends to bounce around quite a bit, back and forth between Scandi and Greenland. This is what happened in 2010 and for pretty much a whole month, winds dominated between East and North. Even if the HP remains in situ over the UK (if it's mnot a cloudy high), temperatures will tumble by night and it will render us colder than average.

Edited October 16, 2012 by The watcher

[BLAST FROM THE PAST]

Why look beyond 24th October? Even that's really stretching it a bit with model reliability.

BFTP

[chionomaniac]

Time to put the model accuracy debate to bed once and for all.

I am afraid that what Steve says regarding binning outputs should be ignored. The difference between each GFS run is miniscule in the terms of input and therefore does not make a significant impact compared to the other respective runs. The ensembles also run for each run to try and counteract any minor discrepencies and these should be referred to rather than 'binning' runs.

Steve you run the risk of putting bias over objectitivity if you continue to peddle the line of 'binning' runs if they don't show what you would like. And that would be an utter shame because the amount of information and quality of your posts can be second to none in the area that you operate.

So what are the statistics showing currently?

Well here they are:

These clearly show that the ECM is the first model to be believed currently in accuracy rates. This is followed by the GFS 12Z, then the GFS 6Z, then the GFS 18Z then in last place the 00Z.

Now these statistics change from time to time, but the point remains that we do not know which run is the most accurate until after the event. And because of this we simply cannot discount any run - surely the most valid run is the one with the most up to date data fed into it - ie the latest.

Links to statistics are below:

http://www.emc.ncep....gmb/STATS_vsdb/

http://www.emc.ncep....TS/STATS4x.html

Any future references to 'binning' runs may be binned themselves as this is not borne out by the statistics.

What we should still be concentrating on is the disrupted polar vortex this month and how this does open up the possibilities of polar outbreaks.

I like to consider that the zonal train is very much like a travelator and that this has very much been replaced with escalators. These are moving warm up to the poles and cold air down from the poles. i hope that we can get on the down escalator this winter.

And remember SM's quote from yesterday evening.

Those looking for mild will look at the UK, but those looking for cold will look at the pole.

Edited October 16, 2012 by chionomaniac

[BLAST FROM THE PAST]

C

That's all very well but how on Earth is there such wild model swings from run to run? If the data is nigh on the same and miniscule differences then why such huge differences...even at relatively short/near timescales? Your point is taken and indeed my suggestion of even looking/comparing beyond say 24th is likely to be fruitless.

Then there is JH POV that compare like for like eg 12 z with 12z etc etc.

BFTP

[SMU]

Hi Chiono-

Binning runs is relevent when looking at the 06z Past 144- Its a bit 'frustrating' to comment & say ignore my post-

I have watched every run of the GFS every winter for the last 8 years,- & the 06z is cr*p.

More recently over the last 4 years time & time again the morning 06z has created the train wreck of emotion on here in relashon to pending cold spells, & to a certain extent the 18z as well-

The balloon data is fed into the 00z & 12z & not 06z + 18z..- therefor the upper air profiles & info is poor-

I think we can agree to disagree as generally the GFS isnt that good anyway,- However you will have to take my word that the 06z is useless.....- I could go back to the archives of cold spells & show that its a nonsense run- but not today...

Anyway just going on a few questions I think the models overall have it right- by around day 8-10

Greenland High pressure with a deep scandi trough & CAA moving south-

I have been going on about this for quite some time, but we have the milder interlude that interupts proceedings-

im 95% sure northern & north west europe gets an early dose of winter, the key is whether we can get enough low pressure over central europe to force the cold air west-

Point in case- look at the 06z drop the greenland high over the UK & force it out of the pole- this ISNT my expected outcome, I expect more undercutting & the higher pressure to stay closer to greenland/ svalbard...

regards

S

Edited October 16, 2012 by Steve Murr

[A.J]

C

That's all very well but how on Earth is there such wild model swings from run to run? If the data is nigh on the same and miniscule differences then why such huge differences...even at relatively short/near timescales? Your point is taken and indeed my suggestion of even looking/comparing beyond say 24th is likely to be fruitless.

Then there is JH POV that compare like for like eg 12 z with 12z etc etc.

BFTP

The fairly straightforward explanation is that weather modelling uses vast amounts of computer processing power with incredibly complex & intricate mathematical/algorithmic calculations meaning that even a miniscule change in just one parameter at T+6 can have a huge knock on effect deeper into the calculations for subsequent time frames. Until the technologies of Quantum computing becomes advanced enough to be able to ditch parallel processing (that is currently used), the present infrastructure for weather modelling is about as good as its going to get

Edited October 16, 2012 by ajpoolshark

[BLAST FROM THE PAST]

Sorry AJ....at 7-10 days yes but not 5 days. I take your point and I understand that as I say re looking beyond [at a stretch] 24th Oct. But humans can judge the kind of set up 5 days ahead, even 7 the wild swings are ridiculous sometimes.

Anyway, backend of Oct for cold from N/NE looks on cue. I can't post images on this computer at work so if someone can post the jetstream at t51.....that is incredible.

BFTP

Edited October 16, 2012 by BLAST FROM THE PAST

[Radiating Dendrite]

Sorry AJ....at 7-10 days yes but not 5 days. I take your point and I understand that as I say re looking beyond [at a stretch] 24th Oct. But humans can judge the kind of set up 5 days ahead, even 7 the wild swings are ridiculous sometimes.

Anyway, backend of Oct for cold from N/NE looks on cue. I can't post images on this computer at work so if someone can post the jetstream at t51.....that is incredible.

BFTP

Here you go!

I have never seen it like that before!

[lorenzo]

Statistics are just like mini-skirts, they give you good ideas but hide the most important thing.

Think in this case the most important thing is the NH view.

MJO has developed from it's update on the 15th with some difference between GFS and ECM. GFS firing it eastward, phase 8 18-22nd October before re-entering phase 1.

Composites are here if you want to browse, a plus 8 day lag on impacting the NAO. NAO+ regimes tend to be preceded by phase 3-4 of the MJO ( Cassou 2008).

NAO projections here and you can see the GFS weighting the negative vs ECM.

Edited October 16, 2012 by lorenzo

[Glacier Point]

If I could just chuck this grenade in and run...

This is just my own personal view, but I do regard the 06z GFS and 18z GFS runs as having something in them that makes them less reliable as a forecasting tool, mainly when they seem to develop mad ideas in the day 5 and 6 timeframe which sends them off in different directions, particularly around the polar field. The ensemble suite should be less variable but even here there are notable differences in GEFS ensemble mean from 00z and 06z runs which cannot alone be explained by variability in data.