Model Output Discussion:

[NorthantsSnow]

has anyone thought gfs was right all along in showing the easterly first if it comes off

No because it was NOT right about how we got there.

[russh]

i am not sure if this is the right place to put this if it is wrong please move it

i am sending a link from paul Hudsons blog from tonight

he is talking about the Ensembles and he is saying "the most likely outcome is conditions to turn colder again next week as air spreads in from the near continent"

http://www.bbc.co.uk/blogs/paulhudson/

[KTtom]

Here you go.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz6.html

Bear in mind the GFS verification stats won't include the easterly fiasco yet, expect its stats to nosedive in a few days, the UKMO could well be pushing the ECM very close in the top model stakes soon.The ECM over the last week has still been fine although all models have dipped slightly.

Cheers Nick. Whilst even a numpty like me can understand the graph, can anyone explain how they are scored?

[shiver]

No because it was NOT right about how we got there.

lol doesnt matter how we get there as long as we get there easterly is an easterly

[chionomaniac]

There was talk a while back regarding whether it was the data fed into the GFS rather than the actual model itself that was the problem, I'm trying to remember exactly here perhaps someone else remembers an experiment they did with feeding GFS data into the ECM or perhaps the other way round.

It was something along those lines Nick, but I too can't remember what the result was. I have also searched for the original thread on Eastern but with no luck.

[Frosty.]

The differences between the gfs & ecm 12z @ T+120 hours are enormous, ecm bringing a colder flow in so quickly where the gfs has a gale force S'ly, some serious backtracking required by the gfs to get even close to that scenario but the true cold is out at T+192 onwards so a looooong way to go yet.

[NorthantsSnow]

lol doesnt matter how we get there as long as we get there easterly is an easterly

Oh I quite agree with you on that score. I want the easterly to occur as much as anyone. However, if it happens it would be wrong to say that the GFS had it nailed all along, as that is simply not true. The devil is in the detail and that's where the GFS falls down every time. It may have just about got the start point and end point right, but the bit in between has been a real struggle for the model.

Off topic but poor Vancouver! No snow so far in 2010 and none likely for February, with temps only dipping below freezing once. Perhaps Scotland would have been a more suitable home for this year's Winter Olympics.

[ribster]

OMG +168 is frankly stunning.

Sit back, have a cup of tea and enjoy +192/+216.

Indeed, and oh my - look at the date, seems to me nothing has changed, it's not getting further into FI, it's getting closer, after a wobble at the weekend. Nothing set in stone obviously, but all the finger wagging and climbing on to high horses was slightly premature I think...

[johnholmes]

I thought someone had posted that the ECM had performed worse out of the 'big 4' over six consecutive days?

There was a link, but I can't find it?

it has not been at its best but its not really been outperformed over the past 3-10 days, the link to see the performance figures for the 6 day checks NOAA run is below

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz6.html

[Ian Fergusson]

Oh I quite agree with you on that score. I want the easterly to occur as much as anyone. However, if it happens it would be wrong to say that the GFS had it nailed all along, as that is simply not true. The devil is in the detail and that's where the GFS falls down every time. It may have just about got the start point and end point right, but the bit in between has been a real struggle for the model.

Off topic but poor Vancouver! No snow so far in 2010 and none likely for February, with temps only dipping below freezing once. Perhaps Scotland would have been a more suitable home for this year's Winter Olympics.

In fairness, the UKMO 10-15 day briefings have consistently called for the easterly influence to establish sometime around start of next week (i.e., briefings over the past two or more days) and not based on GFS output but instead largely on the EC ENS (refer to previous posts containing abstracts of the briefings).

An equally consistent - and crucial - component of the latest briefings has been the almost universally low signal for heavy PPN into the same period; thus the emphasis on a colder spell does not necessarily equate to an especially snowy one..... thus a provisional note of caution worth injecting, in case of myopia developing about the 'white stuff'.

In any case, it's all light years away in terms of the detail, and as many can clearly attest over the past 48hrs, much can - and inevitably does - change!

Edited February 1, 2010 by Ian Fergusson

[Bobby]

JMA is tasty at +144

[nick sussex]

Cheers Nick. Whilst even a numpty like me can understand the graph, can anyone explain how they are scored?

Yes they're using a correlational package so 1 is the top score, logically i think they map the forecast 500 chart over the actual recorded chart and then the computer does the rest by working out the error rate in terms of how far the model was out, I can't really think of how else they would do it, perhaps I'll have a look at the NOAA site to see what statistical package they're using.

[johnholmes]

Nothing set in stone obviously, but all the finger wagging and climbing on to high horses was slightly premature I think...

it is usually the case-but it provides lots of excitement for those wishing for the hourly roller coaster rides.

[Cecil Hogmanay Melchett]

Good news from NOAA they've ditched the GFS 12hrs operational run because its an outlier and like the ECM and UKMO aswell as the GEM for events over there which is important regarding the Atlantic trough.

THE 12Z/01 DETERMINISTIC GFS WAS AN OUTLIER TO ITS ENSEMBLE

MEAN...AS WELL AS THE NEW GEM GLOBAL...UKMET...AND ECMWF...WITH

THE SUPPRESSED TRACK OF THE CYCLONE ALONG THE EAST COAST DAY 5.

And more good news.

BUT WITH ITS STRONG VERIFICATION THIS

SEASON...STUCK CLOSEST TO THE DETERMINISTIC ECMWF FOR THE DETAILS

OF THE PRELIMINARY UPDATE PACKAGE.

The GFS fall from grace continues!

Nick i nearly always finding myself agreeing with your posts but im not entriely sure i agree about GFS fall

from grace again comment.Cast your mind back a few days,ukmo and eventually ecm BOTH went for complete Atlantic

domination,did you see the ecm ens suite yesterday?they have flipped around terribly over recent days.GFS on the other

hand was first to pick up on a slider and now the ruors are going with the energy going SE rather than NE.

Granted gfs now doesnt know what to do but i personally think all the models have struggled badly over the last 72 hours

[The PIT]

Seems all the models are dipping down at the moment. Is this the first time they've really had to deal with major blocking for a sustained period of time ????

[smich]

What a day! You couldn't make it up could you?!

I was so puzzled yesterday and posted as such. All the signals were there -AO, -NAO, SSW, MJO, all favourable, but the models were showing mild? Many thanks to Stephen Prudence for the reply about "West based" -NAO.

And that is still a big risk IMO, the great charts we're likely to see in the next day or so will evolve further. We need to keep a keen eye out for the trends.

Great debate earlier about the ensembles; having seen the GFS mean doggedly stick to -5 to -10 850hpa temps for days on end, then to suddenly show +5......it was commented that maybe the ensembles were useful to illustrate certainty in the runs ie. great scatter = great uncertainty. But now I'm not so sure!

One things for sure. Given the "I told you so" posts on both sides over the last few days, I've learn't one thing:

Humble Pie is a dish best served cold!

smich

[bluearmy]

Nick i nearly always finding myself agreeing with your posts but im not entriely sure i agree about GFS fall

from grace again comment.Cast your mind back a few days,ukmo and eventually ecm BOTH went for complete Atlantic

domination,did you see the ecm ens suite yesterday?they have flipped around terribly over recent days.GFS on the other

hand was first to pick up on a slider and now the ruors are going with the energy going SE rather than NE.

Granted gfs now doesnt know what to do but i personally think all the models have struggled badly over the last 72 hours

although we are paying for it, we dont get to see the ecm ensembles sypnotically HD. however, the meto 15 dayer, derived from these ens, has never ditched the idea of it turning colder next week so i assume the ecm ens didnt flip mild in the longer term. maybe thats the advantage of there being 49 of them.

http://www.knmi.nl/exp/pluim/Data/PLUIM_06260_NWT.png

se flow by sunday looking solid for holland - the uk by monday ??

[johnholmes]

The differences between the gfs & ecm 12z @ T+120 hours are enormous, ecm bringing a colder flow in so quickly where the gfs has a gale force S'ly, some serious backtracking required by the gfs to get even close to that scenario but the true cold is out at T+192 onwards so a looooong way to go yet.

Frosty-it must be my eyesight?

I've just drawn the 12z GFS, ECMWF and Met 1025/1030mb isobar and its almost coincident

The centre of the surface high with all 3 not very far apart as are the centres of low pressure in the Atlantic?

Nor, on a closer look is the even more important upper pattern greatly different.

It is after all at T+120-the 3 of them appear to me to be singing from the same hymn sheet. IF they are near enough correct then it could be time at T+96 to feel that the long term prediction-a week or more ago, of some kind of block with a flow from east rather than west would develop?

Edited February 1, 2010 by johnholmes

[andymusic]

Oh I quite agree with you on that score. I want the easterly to occur as much as anyone. However, if it happens it would be wrong to say that the GFS had it nailed all along, as that is simply not true. The devil is in the detail and that's where the GFS falls down every time. It may have just about got the start point and end point right, but the bit in between has been a real struggle for the model.

Off topic but poor Vancouver! No snow so far in 2010 and none likely for February, with temps only dipping below freezing once. Perhaps Scotland would have been a more suitable home for this year's Winter Olympics.

not as bad as that - whistler has snow for the alpine events (but the company that owns the resort is being made bankrupt - lifts might suddenly dissappear i suppose) - and cypress creek - snowboard + freestyle - they're shifting tonnes and tonnes of snow from further up a mountain to where it's needed! - crazy - canada usually has tonnes of the stuff and mega low temps - everything is all about face this winter and we're about to go back into the freezer ourself - be nice to have a foot or two of snow outside while we watch the winter olympics - make us feel like we're really there (well maybe not if they don't have any of the white stuff)

Can't wait to see what the GFS 18z will serve up for supper tonight

Edited February 1, 2010 by andymusic

[Blizzard_of_Oz]

I'm preparing myself just now lol, its always good to get your well balanced thoughts on the runs, I'm afraid I'm sitting on the fence also. Wow its amazing how things have changed but I thought something must be wrong because of a certain poster on the boards from the B.B.C. kinda like reading in between the lines of his posts.

Ian Fergusson:

In fairness, the UKMO 10-15 day briefings have consistently called for the easterly influence to establish sometime around start of next week (i.e., briefings over the past two or more days) and not based on GFS output but instead largely on the EC ENS (refer to previous posts containing abstracts of the briefings)

Thanks for that Ian while others on here yesterday evening were saying Winter was over some people never learn about our British weather do they lol.

However I'm not 1 of those people who would go on and say its nailed until the fat lady is clearering her throat as some were saying last night. Im glad the forum is back in the agreement now and the models are starting to settle down its amazing how much human influence can change things isnt it?

[johnholmes]

Yes they're using a correlational package so 1 is the top score, logically i think they map the forecast 500 chart over the actual recorded chart and then the computer does the rest by working out the error rate in terms of how far the model was out, I can't really think of how else they would do it, perhaps I'll have a look at the NOAA site to see what statistical package they're using.

this link takes you into their various correlation checks

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

[Cloud 10]

I think bashing only the gfs is a bit unfair on this occasion as the rest of the models have struggled over the last few days,with more swings than tarzan on steroids.

PLUIM_06260_NWT.png (PNG Image, 640x880 pixels)

A very good run from the ecm tonight,although no suprise to see it on the colder side of the ensembles later on!

A good trend today from the models regarding some cold air spreading from the east in the medium term which fits in with the forecast negative AO/NAO combo.

[nick sussex]

Nick i nearly always finding myself agreeing with your posts but im not entriely sure i agree about GFS fall

from grace again comment.Cast your mind back a few days,ukmo and eventually ecm BOTH went for complete Atlantic

domination,did you see the ecm ens suite yesterday?they have flipped around terribly over recent days.GFS on the other

hand was first to pick up on a slider and now the ruors are going with the energy going SE rather than NE.

Granted gfs now doesnt know what to do but i personally think all the models have struggled badly over the last 72 hours

The problem with the gfs is its extreme and hangs on to the wrong solution for too long, the ecm makes mistakes but corrects a wrong solution very quickly this is why it verifies ahead of the gfs. It doesn't matter that the models might show an easterly now and the gfs originally went for this because it got the route to the possible easterly totally wrong and didn't want to know about the renewed Atlantic energy and then when it did realize it went overboard. The initial GFS energy ejection se'wards was from a low in the Atlantic and not the culmination of the USA shortwave deepening and basically drawing the life out of the low near the UK which as you can see fills as the main upper trough forms from the deepening shortwave, its as the low fills that energy in a shortwave is sent se'wards by the ECM and UKMO.

To be honest I'm still not convinced of the model output because its been so volatile over the last few days however given the NOAA comments and general background teleconnections the ECM and UKMO look plausible upto 120 hrs but I still would not like to call things after that point because the models might still retrogress the pattern further, what I would say though is the further west and south the upper trough is the better the chance for colder weather, given the signals if this is the case even the UK couldn't make Little Chef food out of Dover Sole, Scallops and a fine Chablis!

this link takes you into their various correlation checks

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

Great John, many thanks, hope you had a good skiing trip.

Edited February 1, 2010 by nick sussex

[Frosty.]

Frosty-it must be my eyesight?

I've just drawn the 12z GFS, ECMWF and Met 1025/1030mb isobar and its almost coincident

The centre of the surface high with all 3 not very far apart as are the centres of low pressure in the Atlantic?

Nor, on a closer look is the even more important upper pattern greatly different.

It is after all at T+120-the 3 of them appear to me to be singing from the same hymn sheet. IF they are near enough correct then it could be time at T+96 to feel that the long term prediction-a week or more ago, of some kind of block with a flow from east rather than west would develop?

No it's ok John I think I might need glasses,

Hope you had a good skiing holiday.

A fairly significant move towards the next cold spell hopefully evolving.

[Jason T]

Well the ECM being consistent with a cold spell building increases the odds significantly for me.

GFS is really having a go, but is still on the super-sized-storm drink.

UKMO is going somwhere more interesting for once.

GEM looks decent at 120 hours and then goes for a cold SE'rly draw as a ridge comes up from the south to further pump up the high.

This is basically that thing where an easterly is hinted at, dropped and then springs up again a bit closer, this time with stronger agreement. Its gone from GFS ramping it to ECM. Will ECM be the winner and get GFS plus UKMO fully behind it for tomorrows +120 charts?

Only time will tell, but it would be wierd to see it dropped again only to spring up at 96 hours so hopefully consistency will be of the order for once