Model Output Discussion; into 2015

[Barry12]

The ECM is far better than the GFS, the stats prove this. It's ironic that the same members that go on about how poor the ECM is past day 7, post day 16 GFS charts to back up there views! ECM day 7-10 charts are far better than GFS 16 day charts that's for sure.

[johnholmes]

so another question Nick

How much worse, in terms of getting the correct 'feel' for the surface weather at 240h do you feel GFS is over ECMWF?

[snow mad]

The ECM is far better than the GFS, the stats prove this. It's ironic that the same members that go on about how poor the ECM is past day 7, post day 16 GFS charts to back up there views! ECM day 7-10 charts are far better than GFS 16 day charts that's for sure.

i think it's the manner of its failing or bias as it gets a lot of people excited as it tends to be bias towards building heights in the desired places time and again only to jump ship last minute now.

[Captain Shortwave]

Well going by the current output, it would suggest that the milder interludes will tend to dominate over the cooler spells for the first week at least of January. There is the potential for some stormy weather at times in the north, but further south it looks like being damp and windy rather than the wet and stormy weather of last year. 

Very little to suggest any change from the GFS ens with a flat westerly pattern with weak ridging and troughing allowing drier and cooler weather to develop more widely for brief periods.

Here are some mean charts

No real signal yet for change, these of course can change as we approach, but low heights look very much stuck to our north/north west.

We continue to wait for change. Lets hope it comes soon and that the ECM ens and in house metoffice products hint for change in the near future.

 

@Barry95 - At day 10 all the models tend to not verify to well, though the ECM is probably the best. Beyond day 6 we should be trying to look at ensembles and clusters within that to guide our thoughts. But of course it is fun to follow the operationals as they are by far the most accessible tools we have.

Edit - ECM ens not really seeing any meaningful cool down in the first half of week 2, though day 11-15 could be different.

Edited January 1, 2015 by Captain shortwave

[bluearmy]

Nick, did you, like me, assume that the spread to our west on the 00z ECM quite showed the trough getting further south? Seems the opposite was possibly true?

[Jason M]

The problem with ECM is that we are not allowed to see the ensembles, which means the opp gets over analysed. If we could see the level of data we get with GFS it wouldn't be an issue.

 

Anyway, not sure there is any real disagreement at present. Its all just a variation on a theme i.e westerly / zonal with zero chance of any meaningful HLB.

[fergieweather]

The ECM is far better than the GFS, the stats prove this. It's ironic that the same members that go on about how poor the ECM is past day 7, post day 16 GFS charts to back up there views! ECM day 7-10 charts are far better than GFS 16 day charts that's for sure.

http://apps.ecmwf.int/wmolcdnv/scores/mean/500_z

The WMO dataset we follow.

Edited January 1, 2015 by fergieweather

[Barry12]

i think it's the manner of its failing or bias as it gets a lot of people excited as it tends to be bias towards building heights in the desired places time and again only to jump ship last minute now.

 

GFS is completely over progressive with the Atlantic though, each model has there faults.

[nick sussex]

so another question Nick

How much worse, in terms of getting the correct 'feel' for the surface weather at 240h do you feel GFS is over ECMWF?

:smiliz19: I think we'd better draw a line under this John! I didn't realize my one line criticism of the ECM much earlier in the thread  tonight would lead to so much drama.

[shaky]

Any updates fergie on saturdays prospects?

[nick sussex]

Nick, did you, like me, assume that the spread to our west on the 00z ECM quite showed the trough getting further south? Seems the opposite was possibly true?

Yes I was in the same boat I think. It's disappointing that the zonality looks more the milder type but the PV blow up might hopefully be shorter in duration. That rounded PV blob  is certainly something we want to see gone ASAP!

[fergieweather]

Any updates fergie on saturdays prospects?

Still low confidence on exact track and resultant PPN phasing but more southerly track currently garners higher confidence (=likely reducing snow issues). Thus best estimate is some snow primarily 300m+, but perhaps transiently to low levels,roughly N Wales-N/Central England. Risk of this being disruptive currently assessed as low, but remaining under close scrutiny.

[ANYWEATHER]

Yes I was in the same boat I think. It's disappointing that the zonality looks more the milder type but the PV blow up might hopefully be shorter in duration. That rounded PV blob  is certainly something we want to see gone ASAP!

Nick ,Please take a break from looking at the models so in detail, Anyway go ,outside and have a breath of model weather

[shaky]

Still low confidence on exact track and resultant PPN phasing but more southerly track currently garners higher confidence (=likely reducing snow issues). Thus best estimate is some snow primarily 300m+, but perhaps transiently to low levels,roughly N Wales-N/Central England. Risk of this being disruptive currently assessed as low, but remaining under close scrutiny.

thanks ian!! At least theres a chance of upgrades then mate!! It all looks extremely similar to last friday but just further south! !! The euro4 model has it south aswell and with -4 uppers across central areas to boot!!

[legritter]

Well i must say tonights ecm run does not fill me with cold weather expectations .even looking at gfs both models gives me a fear of panick   and IFF the met office were to suddenly show us all their data out to the end of January i think i would dive behind the settee and stay there till well april .But we are talking Weather ,give it a couple of days and we could be looking at some good eye Candy .i,v probably got well over 300 books on meteorology a thousand Roy Met Society journals and bought many old historical weather diarys ,surprises turn up all the time ,and i have met well respected scientists who have stated that Meteorology is so hard to predict out past 10 days yet alone 15/30 days .Well things will be kicking off big time across the pond so i surely expect a stormy spell after about the 8th of january ,ok Ecm looks very bog standard mild tonight but things could change quickly on future runs .but its great to debate which charts are more accurate and in advance thanks to anyone who puts in the work to find these things out .perhaps if the budget for collecting weather data was increased that would help ,but it would have to be one hell of an increase .anyhow anyone lucky enough to see the holy grail sat /sun ,all the best   

[Nick F]

Trying to find some straws to clutch in a rather bleak outlook for any pattern change from the Atlantic zonal train. I have found a few, some of the model plots today for the MJO point it heading towards phase 7- GFS op and GEFS for example:

 

 

MOGREPS - ensembles heading in a similar direction:

 

Without the SSW, getting into 7/8 at a decent amplitude could be another way of forcing a pattern change for the better ... though other models seem keen to drop the MJO back into the COD before reaching the Dateline or anywhere near the favourable phases.

 

Neutral to weak El Nino, as we appear to be in atm, can favour blocking to develop later in winter - more particularly February. 2004-05 winter seems to be a good analog to use, as that had a weak Nino with December and January that winter generally mild and zonal apart from a brief northerly at the end of Dec '04, then the pattern suddenly flipped with -AO appearing in late February'05 with Greenland blocking appearing.

 

QBO in easterly/-ve phase, which tends to favour a weaker PV, with potential for more strat warmings for the rest of the winter, that may work down.

 

But, any pattern change from these drivers seem unlikely to take effect until at least late Jan ... unless something unforseen happens.

Edited January 1, 2015 by Nick F

[knocker]

Tonights extended anomalies.

NOAA 8-14 has the Hudson Bay trough and another just north of Scotland with the HP west of Iberia. Weak east Pacific and Alaskan ridge and also weak HP eastern seaboard giving a mainly cyclonic Atlantic zonal flow. The GEFs has the LP a tad further north and gives more play to the HP in the western Atlantic. The extended ECM EPS certainly makes more of the heights to the west and also has an Alaskan trough. There would appear to be a lessening of the more intense cyclonic activity.

Edited January 1, 2015 by knocker

[SMU]

Guys.....

Your discussion on model stats is overall POINTLESS.........

Why?

You are arguing over results that correspond to the northern hemisphere which, relating to the UK is a waste of time.

Its such as expanse of space that each models accuracy in the important areas like greenland are not quantified & therefor we are none the wiser on these key specifics.

It was now probably 3 or more years ago that the model performance site removed the regional results which showed quantifiable results for areas that are key to us.

It was the GFS only that had error scores shown in all NH regions & there was 2 key points that will remain in my mind

* GFS day 6 scores dropped off a cliff over the atlantic & moreso the pole versus the overall mean NH average

* 18z & 06z scores were consistently lower than 00z & 12z scores.

In terms of the ECM its still clearly the market leader in the statistical results @ NH level however since its upgrade 2 years ago where prior to this it pretty much pwned everything it has become very volatile post day 8 for our all important forecast areas ( pole & atlantic ) & now its sadly as bad as the GFS for anomaly charts, usually at the other end of the scale for over promoting blocking.....

I suppose the net result of this is that it keeps us on our toes......

s

[Cloud 10]

 

 

I'd be happy though to conduct a study into the emotional impact of mild weather on snow lovers! :smiliz19:

 

That would be a fascinating study,but who would be first on the couch. :w00t:

 

 

Regarding the model verification,the ECM is still the best across the Northern Hemisphere,but it does have a rather nasty habit of over-amplifying high pressure in the Atlantic/Southern Greenland area which means downgraded northerlies come verification day,which is why it gets a bit of a bashing!

 

The GFS is pretty much the opposite and routinely flattens any ridging,particularly to our Northeast.

 

The GEM has a major issue with its 850 hpa temperature predictions,sometimes at very short range so i would think that model will be getting some more tweaking very shortly.

[SMU]

Just seen fergies link -

Nice, pity no gfs or ECM data for pole or atlantic specifically....

[Cloud 10]

Just seen fergies link -

Nice, pity no gfs or ECM data for pole or atlantic specifically....

 

There is an option for North Pole in the drop down boxes at the top Steve.

 

 

Meanwhile on the pub run,that cold pool to the East gets closer and closer on seemingly every run.

 

Edited January 1, 2015 by Cloud 10

[john mac]

Could someone explain to me why Saturdays low will have less risk of snow the further south the track is ? Thanks

[Jason M]

Spot the difference between GFS P and GFS opp :-)

http://www.meteociel.fr/modeles/gfse_cartes.php?&ech=192&mode=0&runpara=1

http://www.meteociel.fr/modeles/gfse_cartes.php?&ech=192&mode=0

Remarkable really at 192 hours out. Ugly charts though!! Even Nick won't find a straw to clutch here!

[Jason M]

Could someone explain to me why Saturdays low will have less risk of snow the further south the track is ? Thanks

Sure thing!!

Its simply too warm further south. To be honest its pretty marginal further north anyway, which is why no one has got particularly excited. The system is developing in warm air and running into not particularly cold air. End result = Rainfest.

See 850s below. Well above zero in the south. #

http://www.meteociel.fr/modeles/gfse_cartes.php?ech=48&code=0&mode=1&runpara=1

[Captain Shortwave]

Well at least we get something later on

At 14 days out, there is little hope as of yet for this to trend to something significant. The day 8 charts from the op and parallel are really bad for those who like cold with a massive black/purple blob over Greenland and a strong west to south westerly flow.

Edited January 1, 2015 by Captain shortwave