Model Output Discussion

[MR EXTREMES]

you say that lastnight about 2am i watch bbc24 weather and he was talking about maybe turning sleety by sunday night.

looks like weather online update are picking up on a prolonged cold outlook.

Looks like the cold spell may be with us for quite a while with frigid temperatures later next week.

[weather eater]

First off all its good to see the ECM move away from its position of yesterday, however I would caution against believing that anything is set in stone, because although the ECM has re positioned, both the NOGAPS and GEM models are not playing ball this morning, so some questions still remain. My hat is still tipped to the parallel this morning as it has maintained its consistency. No model however has lost credibility, the sheer dynamics of climate modelling, means that errors will creep into model runs the further they get from the input data; it will be after the weekend that we see which one of the models has handled this best.

[TEITS]

Not sure why this had been the case, though I think some have already touched on a possible reason, possibly because the ECM has been more finely tuned to cope with a + AO, where I suspect the GFS has been closer tuned to work better with –ve AO.

So maybe the reason the ECM as a consequence has struggled with the developments over the last few days.

Sounds plausible Paul.

Im sure I once read that historical weather patterns are programmed into these models. This could be why the models struggle when our weather patterns change from the norm. This is certainly true of the AO/NAO because they are dropping off the cliff.

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.sprd2.gif

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif

These couldn't be more opposite to what we have become used to in recent years which is probably why IB suggested a +AO/NAO in his Dec forecast.

I believe this model uncertainity will definately continue which is why im just going to enjoy next week and focus on +0 to +72.

[John Snow]

First off all its good to see the ECM move away from its position of yesterday, however I would caution against believing that anything is set in stone, because although the ECM has re positioned, both the NOGAPS and GEM models are not playing ball this morning, so some questions still remain. My hat is still tipped to the parallel this morning as it has maintained its consistency. No model however has lost credibility, the sheer dynamics of climate modelling, means that errors will creep into model runs the further they get from the input data; it will be after the weekend that we see which one of the models has handled this best.

There can't be any doubt now though that the cold spell will happen, it is now about how much snow and how long the cold spell is going to last

The snow early on in the week is pretty meaningless, dusting at best for those who get it but the real deal is the event on Thursday that is forecast with the shortwave, that would dwarf the levels of Feb 2009 if everything falls into place so all there is left to play for is what happens Thursday, the rest before that is set in stone now

[Methuselah]

looks like weather online update are picking up on a prolonged cold outlook.

Looks like the cold spell may be with us for quite a while with frigid temperatures later next week.

'Morning badboy...

Please can you put these posts in the 'Cold Spell' discussion - it's not really much about models??

[andy_leics22]

NOGAPS is a shocker compared to the other models

Pleased with this mornings runs though, ECM now upgraded from yesterday and GFS is not backing down from what it's said for the past few days.

[Shiny_Bottom_1]

Certainly worth having a look at the NW Model Comparison Charts for the 00z runs of the three main models -> http://www.netweather.tv/index.cgi?action=modelcomp;sess=

Up to T+120

Sea Level Pressures are pretty similar

H500 charts are similar*

Northern Hemisphere H500 charts are similar*

*It is worth noting though, that whilst the pattern is similar, there are some very very minor differences. GFS and ECM sit on either side of the fence, with the UKMO somewhere in the middle. All are good, but worth perhaps looking at.

[Neilsouth]

This is just awesome though:

http://www.wzkarten.de/pics/Rmgfs1592.gif

[Guest North Sea Snow Convection]

Well I am glad that I am spared the ECM shooting squad after my comment about ignoring the 12z yesterday evening

The cold spell as others have said looks to have good model agreement now - still the details for next week are not wholly certain but it does look like some wintry showers for some eastern and south eastern areas, especially, during the early part of the week and then the chance of even colder air still, modelled to come south westwards later in the week to bring snow to quite a few places perhaps.

The presence of low pressure over southern europe is very nice to see and has been a missing ingredient at times in terms of underpinning blocking and helping to advect cold air westwards rather than southwards.

Looking pretty good

[Altostratus]

Sounds plausible Paul.

Im sure I once read that historical weather patterns are programmed into these models. This could be why the models struggle when our weather patterns change from the norm. This is certainly true of the AO/NAO because they are dropping off the cliff.

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.sprd2.gif

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif

These couldn't be more opposite to what we have become used to in recent years which is probably why IB suggested a +AO/NAO in his Dec forecast.

I believe this model uncertainity will definately continue which is why im just going to enjoy next week and focus on +0 to +72.

They will almost certainly analyse historic data and check the performance of the models via simulations. If anyone knows more about this it would be interesting to hear. The only way I can see historic patterns being programmed into the models would be via the use of correction factors to overcome any known flaws in the modelling. I could see this being particularly useful for areas with minimal observation and subsequent data input. Based on the model simulation analysis I suppose it would be possible to produce calibration curves as well. I would expect some large error margins however, as no two patterns are ever identical.

Regarding the AO and NAO forecasts, it is my understanding that these are based on the model outputs from the GFS. They are purely reactive. They are not based on an independent model or forecast, and as such, provide no additional support.

[Turnedoutniceagain]

This is just awesome though:

http://www.wzkarten.de/pics/Rmgfs1592.gif

Nice eye-candy, but 6 days away so don't dust off the sledge yet.

I'll rest easier when NOGAPS gets into line though !!

[phil nw.]

Hi Phil, everyone,

Yes it's very good so see the ECM joining in the fun at last.

It does now look as though we could be in for a significant cold spell.

Have been watching the very much-expected drama from the models over the last couple of days, though one normally expects the GFS to have the mood swings, but this time to my surprise it has been the ECM.

Not sure why this had been the case, though I think some have already touched on a possible reason, possibly because the ECM has been more finely tuned to cope with a + AO, where I suspect the GFS has been closer tuned to work better with –ve AO.

So maybe the reason the ECM as a consequence has struggled with the developments over the last few days.

Hi Paul,

Yes the ECM threw a spanner into the works last night.

Although it`s with the others again in the short term it`s interesting to see the differences in the latest 8-10 mean ht.comp.

http://www.meteo.psu.edu/~gadomski/ECMWF_0z/hgtcomp.html

ECM still seems keen on sinking the block into the mid atlantic later on,whereas GFS keeps it further north,although both move it further west in time.

A very different pattern though Paul,from what we are used to and i think we are in for some twists and turns yet.

[Southender]

Nice!

Many cm's of snow throughout next week... (Maybe!)

Edited December 11, 2009 by LeighShrimper

[weather eater]

There can't be any doubt now though that the cold spell will happen, it is now about how much snow and how long the cold spell is going to last

The snow early on in the week is pretty meaningless, dusting at best for those who get it but the real deal is the event on Thursday that is forecast with the shortwave, that would dwarf the levels of Feb 2009 if everything falls into place so all there is left to play for is what happens Thursday, the rest before that is set in stone now

No, there is no doubt that we will see a cold spell, indeed judging by the frost here this morning it’s already started. However what still remains up in the air is how that cold spell will develop, you have to remember that the models once they get past a the first couple of days are making increasingly large guesses as to future synoptic patterns, that’s why an easterly with bucket loads of snow is not set in stone, the models will continue to change and evolve. TEITS has just put up a post saying that he will just focus on the next 72hrs, which is wise. I would say as well, just because a model run is unlikely to play out as programmed, this does not automatically mean a downgrade it could well be the opposite. The GEM and NOGAPS are considered minor models but I find it’s wise to look at the whole picture and not get caught up in the moment, not easy given some of the runs we have seen.

[Geordiesnow]

I still think we need to wait until around this afternoon/Saturday morning time before we can say this is nailed. I'm not too fussed about the breakdown at this stage either, aslong we get some easterly blast, then that what all that matters.

Once again, just because we got some model agreement does not mean it will happen as it is showing but it does look more likely an easterly with a convective flavour might occur.

[TEITS]

Still some slight differences even at +60 with the 06Z likely to pull in colder E,lys for early next week compared to the other models.

Now at first glance you would say the other models are right. However for the period between +0 & +60 I find the GFS more reliable than even the Fax charts. This was the case with the precip charts during the cold spell in Feb.

[J07]

Agreed J07 - folks do your research before hammering the ECM model -

25 November 2009 ECMWF has pioneered a system to predict forecast confidence. This system, operational at ECMWF since 1992, is the Ensemble Prediction System (EPS). A new leaflet on EPS has been published on the ECMWF web.

Amazing that the mods see fit to delete my post (which you quoted), but not yours?

Hello mods...I am merely pointing out something quite useful and *relevant* to "model output"!

[Yeti]

Nice!

Many cm's of snow throughout next week... (Maybe!)

That just says "Schnee im Bergland" (snow in the mountains). Things really aren't looking that snowy at all I'm afraid from the 00z runs, except maybe on the coasts. Before next Thursday the pressure is either too high or the uppers aren't cold enough to produce either good convection or anything other than wintry showers (again near the coasts due to a weak flow). Thursday itself (144) looks a lot better:

http://www.wetterzentrale.de/pics/Rtavn1441.png

However, by Friday the flow is much weaker again, even though the uppers are cold, especially over Europe:

http://www.wetterzentrale.de/pics/Rtavn1681.png

http://www.wetterzentrale.de/pics/Rtavn1802.png

Obviously still a long way out at that stage, but recent runs haven't been much better either in that respect. Not that we have much to complain about - there should be some really cold days coming up...

Edited December 11, 2009 by Yeti

[Southender]

That just says "Schnee im Bergland" (snow in the mountains). Things really aren't looking that snowy at all I'm afraid from the 00z runs, except maybe on the coasts. Before next Thursday the pressure is either too high or the uppers aren't cold enough to produce either good convection or anything other than wintry showers (again near the coasts due to a weak flow). Thursday itself (144) looks a lot better:

http://www.wetterzentrale.de/pics/Rtavn1441.png

However, by Friday the flow is much weaker again, even though the uppers are cold, especially over Europe:

http://www.wetterzentrale.de/pics/Rtavn1681.png

http://www.wetterzentrale.de/pics/Rtavn1802.png

Yes, sorry got a little carried away there!

Snow in the mountains of London!? Where would you find those!

Still, fingers crossed we all get a nice covering next week. Bring on the Shortwave!

[Glacier Point]

Models wobbling around in the last 24hrs, most probably related to a slight westward shift in organised tropical convection in the Pacific.

However, the Global Wind Oscillation, which factors both tropical and extratropical wind signals continues to show a phase 4 projection. This is strongly in line with the general longwave pattern and ensemble mean guidance from GFS and GEM for their forecast ranges out to day 15.

Day 11-15 guidance:

http://raleighwx.easternuswx.com/models/cmcensemble/00zCMCENS11-15day500mbHeightAnomalyNH.gif

http://raleighwx.easternuswx.com/models/gfsensemble/00zENS11-15day500mbHeightAnomalyNH.gif

The block over Iceland / southern Greenland is forecast to retrogress towards Canada in the timeframe, allowing the jetstream to shift back slightly north from an extremely depressed latitude days 6-10.

What's very interesting with the evolution we are seeing is the really cold air is held back over Russia and Scandinavia for the time being which is reasonable considering we will evolve a blocking structure more towards Iceland rather than Scandinavia which would bring about the more rapid advection. However, what this does give us is a large puddle of -5 to -10C 850 hPa air covering much of the NE Atlantic sector.

I do not expect the long wave pattern to alter much, indeed we could well see the Greenland block beef up again if tropical convection peps up again east of the Dateline. So we are likely to keep a depressed jet and for the UK to remain in a mean position on the poleward side up to and beyond the Christmas period.

If we continue to see this kind of evolution in the longwave pattern, and global angular momentum increases in early January, we should see a very negative NAO pattern setting up.

[John Snow]

That just says "Schnee im Bergland" (snow in the mountains). Things really aren't looking that snowy at all I'm afraid from the 00z runs, except maybe on the coasts. Before next Thursday the pressure is either too high or the uppers aren't cold enough to produce either good convection or anything other than wintry showers (again near the coasts due to a weak flow). Thursday itself (144) looks a lot better:

http://www.wetterzen...s/Rtavn1441.png

However, by Friday the flow is much weaker again, even though the uppers are cold, especially over Europe:

http://www.wetterzen...s/Rtavn1681.png

http://www.wetterzen...s/Rtavn1802.png

Thursday especially with the "lake effect" should be enough to produce a foot or so of snow to Eastern counties from the Wash northwards to about Fife unless there is some backtracking by the GFS in next couple of days

[i luv snow]

That just says "Schnee im Bergland" (snow in the mountains). Things really aren't looking that snowy at all I'm afraid from the 00z runs, except maybe on the coasts. Before next Thursday the pressure is either too high or the uppers aren't cold enough to produce either good convection or anything other than wintry showers (again near the coasts due to a weak flow). Thursday itself (144) looks a lot better:

http://www.wetterzen...s/Rtavn1441.png

However, by Friday the flow is much weaker again, even though the uppers are cold, especially over Europe:

http://www.wetterzen...s/Rtavn1681.png

http://www.wetterzen...s/Rtavn1802.png

Obviously still a long way out at that stage, but recent runs haven't been much better either in that respect. Not that we have much to complain about - there should be some really cold days coming up...

Yes but we have been here before...get the cold (which is nailed on) and wait for the ppn details later. Unfortunately people always seem to froget this...snow events tend to be out of the blue anyway, they just happen due the unpredictability of snow itself!

Edited December 11, 2009 by i luv snow

[Yeti]

Still some slight differences even at +60 with the 06Z likely to pull in colder E,lys for early next week compared to the other models.

Now at first glance you would say the other models are right. However for the period between +0 & +60 I find the GFS more reliable than even the Fax charts. This was the case with the precip charts during the cold spell in Feb.

Yep, GFS proved pretty awesome last year when it came to precipitation, especially wrt the track and northern extent of fronts.

[Polar Continental]

Sounds plausible Paul.

Im sure I once read that historical weather patterns are programmed into these models. This could be why the models struggle when our weather patterns change from the norm. This is certainly true of the AO/NAO because they are dropping off the cliff.

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.sprd2.gif

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif

These couldn't be more opposite to what we have become used to in recent years which is probably why IB suggested a +AO/NAO in his Dec forecast.

I believe this model uncertainity will definately continue which is why im just going to enjoy next week and focus on +0 to +72.

Hi Paul,

Yes the ECM threw a spanner into the works last night.

Although it`s with the others again in the short term it`s interesting to see the differences in the latest 8-10 mean ht.comp.

http://www.meteo.psu.edu/~gadomski/ECMWF_0z/hgtcomp.html

ECM still seems keen on sinking the block into the mid atlantic later on,whereas GFS keeps it further north,although both move it further west in time.

A very different pattern though Paul,from what we are used to and i think we are in for some twists and turns yet.

Hi Dave, Phil,

Yes, very interesting as to why the ECM is continually sinking the high. Maybe a flaw has been recognised, as its latest output this morning is much more in line with the other models, it could be that and there has been more manual tweaks in to the latest output.

Maybe Steve M has been at the controls, lol.

[weather eater]

Still some slight differences even at +60 with the 06Z likely to pull in colder E,lys for early next week compared to the other models.

Now at first glance you would say the other models are right. However for the period between +0 & +60 I find the GFS more reliable than even the Fax charts. This was the case with the precip charts during the cold spell in Feb.

Demonstrates the fact that the synoptics are constantly evolving, and nothing should be taken for granted.