Model Output Discussion - 30th November onwards

[chionomaniac]

Eastward bias my friend-  .....

 

So before any one else thinks I make it up- whos got the link to the NCEP model bias page-

 

S

The biggest change is with the GEM ensembles though, Steve.

Was just about to post that ed (distracted by athletico beating barca) - thats the naefs swinging back and forth (point taken steve and i'll look out for that on the 12z. in future when block building to our east though it isnt something ive noticed in previous winters).Soon be able to assess ecm full ens. They have been fairly consistent on that scandi height rise and judging by the dutch graphs, they look like continuing thus.

Will need to wait until they are in day 10 range because at day 10 the 12Z Euro ensembles have consolidated the Scandi positive anomaly compared to the 00Z.

 

Oh and Steve - I don't necessarily think that this is a case of eastward bias - rather just 'routine' unpredicatability when the outputs are searching for the pattern.

 

And to show the ECM day 10 anomalies

 

00Z

 

 

12Z

 

 

oppsite direction to the NAEFS!!!

 

Edited December 1, 2013 by chionomaniac

[NorthNorfolkWeather]

.....until we get nearer to the time and it downgrades by 70% or more. 

 

The reliability of the models seems to be getting worse the past couple of years? Why is this, it's getting so disheartening to watch. Even when all the models agree at times, the whole lot collapses and dashes our hopes to the floor again.

Perhaps it's the fact we are all over them like a rash as soon as they come out.  Would be interesting to see if your comment is reflected in the verification stats.

 

By the way, I'm not disagreeing with you, I think they are worse, I think part of the the reason is that the models were developed during the years of raging zonality, every year that the zonality doesn't happen throws a spanner in the works.  It will lead to better models eventually, but you are bound to have a dip in performance if the pattern of a complete hemisphere changes.

[Ali1977]

Looks like a stronger northerly is about to happen on this run

[SMU]

The biggest change is with the GEM ensembles though, Steve.

Will need to wait until they are in day 10 range because at day 10 the 12Z Euro ensembles have consolidated the Scandi positive anomaly compared to the 12Z.

 

Oh and Steve - I don't necessarily think that this is a case of eastward bias - rather just 'routine' unpredicatability when the outputs are searching for the pattern.

 

And to show the ECM day 10 anomalies

 

00Z

 

00zecmwfens500mbHeightAnomalyNH240.gif

 

12Z

 

12zecmwfens500mbHeightAnomalyNH240.gif

 

oppsite direction to the NAEFS!!!

 

LOL- I will take the ECM!!

 

 

 

So 18z- who can spot the trough at T90....

 

http://modeles.meteociel.fr/modeles/gfs/runs/2013120118/gfsnh-0-90.png?18

 

S

[Jason M]

Cant complain when this is shown 4 days away

http://www.meteociel.fr/modeles/gfse_cartes.php?&ech=96&mode=2

 

Already better than the 12z.

Edited December 1, 2013 by Jason M

[Mucka]

Looks like a stronger northerly is about to happen on this run

 

Yup, hope it is trend to move things back west but probably just natural run to run  variation

 

[bobbydog]

Eastward bias my friend-  .....

 

So before any one else thinks I make it up- whos got the link to the NCEP model bias page-

 

S

 

steve, i respect your knowledge of the models and their various bias or nuances but bearing in mind these are used by meteorological organisations around the world, surely they must be neglible enough to avoid wild inaccuracies? if not, why have the producers of these models not addressed these issues?

[EML Network]

It's looking a bit breezy up North on Thursday !!

 

Edited December 1, 2013 by EML Network

[SMU]

steve, i respect your knowledge of the models and their various bias or nuances but bearing in mind these are used by meteorological organisations around the world, surely they must be neglible enough to avoid wild inaccuracies? if not, why have the producers of these models not addressed these issues?

 

If they didn't make a difference they wouldn't have made a whole page about it!!!

 

I don't know why they haven't rectified it- however the bias DOES make a big difference in amplified situations especially for the NW part of Europe-

 

Perhaps NCEP don't care about the UK, or possibly rather than fix it they just buy the ECMWF package for around 50k.... Its cheaper.

 

I hope they finish with it soon- I will buy it off ebay

 

S

Edited December 1, 2013 by Steve Murr

[damianslaw]

2 key factors which will determine whether we see a cold pattern sticking through to the middle of the month or a milder pattern are -

 

1. The position of the Polar Vortex (not strength), there are indications the PV will positioned in a westerly position over N Canada which in turn increases the likelihood of heights building over scandinavia.

 

2. The position of the jetstream, this is forecast to cut due south of the UK during the second week of December which would increase the chance of a split flow, energy off the eastern seaboard being forced SE into the mid altantic which again increases the likelihood of heights building to our NE.

 

There are increasing signs a continental feed establishing itself during the second week of December, yes NW parts may well see a milder SW feed for a time, but the chance of a full on atlantic onslaught look increasingly slim.

 

ECM output is certainly picking up these signals. Whilst GFS will probably continue to toy with various scenarios showing the atlantic in a different guise.

[Vorticity0123]

More easterly options showing on tonights ECM ensembles for Holland:

 

http://www.weerplaza.nl/15daagseverwachting/?r=midden&type=eps_pluim

 

Some very low dew points shown there down to -15c, difficult though to tell whether this is an inversion type scenario, earlier the ECM operational run is at the top end of the main cluster.

 

The wind direction shows quite a scatter but again more east/ne/se wind flows than this mornings run. I think we have to sit tight for a few days and see how the models develop that troughing in the Atlantic because this is quite crucial in helping the high ridge ne.

 

In terms of the UK in terms of how long any cold lasts that depends on what sort of flow is picked up, theres quite a difference at this time of year between a southerly and se flow from that high, it could remain cold at the surface well past the point that the models lift out the cold uppers.

 

Note that the amount of easterly solutions consists only 5 of the 50 lines around monday. Most solutions are going for a south-southwesterly around that timeframe. Therafter, the scatter is so large that, in my opinion, no real conclusions can be drawn about that timeframe.

 

EPS "pluim" of the wind direction forecast of Holland.

 

The flow is indeed forecast to veer more to the south around Dec 10, but keep in mind that the average always go to south in the long term. This is because, for example, two EPS members with wind directions of 350 and 010 degrees, respectively (both almost fully northerly), give an average in the graph of 180 degrees (which is a southerly!) This indicates that the average will always converge to the middle of the graph, which is in this case 180 degrees.

 

However, when one looks at the temperature "pluim" itself, one can clearly see that there are quite a large amount of members (around 40%, decreasing slowly with increasing timeframe) which sustain the cold up to tuesday. This shows that there are indeed some cold solutions, but I can't the link with the southeasterly winds.

 

EPS "pluim" of the temperatures in Holland.

 

Perhaps the cold solutions might have something to do with cold uppers being drawn into Holland from the southeast?

For analysis, below is the upper temps chart of GFS 12z for 11 December (it showed a very cold solution for Holland at T+192, temperatures only reaching about 2 degrees at 12Z))

 

 

A quick analysis shows that there are indeed quite some cold uppers to the southeast of Holland, but this is way too far to actually reach Holland. Therefore, this is highly unlikely the right explanation.

 

The only explanation I could come up with is that this might have something to do with the so called "radiation cold" (when the outgoing radiation is larger than the incoming radiation). Does anybody have another possible answer?

 

Sources:

http://www.weerplaza.nl/15daagseverwachting/?type=eps_pluim

http://www.wetterzentrale.de/topkarten/fsavneur.html

Edited December 1, 2013 by Vorticity0123

[OldMetMan]

I have watched the models flip-flop over the forecast N blast at the end of the week, for the last few days, and with the latest run, they seem to have opted for this northerly merely giving us a glancing blow. Something doesn't look right to me, so let's look at the current position:

Strong upper ridge still over us - yes. But what's happening further afield? A vigorous upper low heading into the Pacific NW with very cold Arctic air spreading SW across the W USA. All this suggests a likelihood of amplification of the flow, as indeed the NWS guidance in the States has been talking about for some days.

Strong troughing then over the W US, adding amplitude to the Canadian vortex, helping our upper high to retrogress into mid Atlantic - down comes the upper trough at our longitude along with the Arctic flow.

But instead, the models push the HP back over us again - peculiar to say the least, not impossible, but to me, it just looks all wrong. I think the crucial development point will be at around T120, in terms of how the cyclogenesis actually takes place S of Greenland. Given the strong baroclinic zone that seems likely to develop at that point in that area, it seems quite likely that a vigorous LP will peel off south-eastwards and fully engage the cold air, and pull quite an intense cold pool towards the UK.

As I said the other day, developments over, and to the east of the Canada/US east coast will clearly have a big impact on this development, and to what follows from there, but I don't see the models have got a proper handle of all this, as of yet. Signs of more pronounced upper flow amplification around the hemisphere have been there for some days.

An interesting twist has come with the 12Z GFS suggesting HP blocking starting over Scandinavia, but that is very long term. It still needs to get a grip on more immediate developments.

It will be interesting to see how the Met O FAX handle it later. I suspect they will go for the glancing blow, but we'll see.

[bobbydog]

If they didn't make a difference they wouldn't have made a whole page about it!!!

 

I don't know why they haven't rectified it- however the bias DOES make a big difference in amplified situations especially for the NW part of Europe-

 

Perhaps NCEP don't care about the UK, or possibly rather than fix it they just buy the ECMWF package for around 50k.... Its cheaper.

 

I hope they finish with it soon- I will buy it off ebay

 

S

 

thanks for the reply steve, though i was talking in general about all the models, not just the NCEP.

maybe as you say, the bias is towards the region it (each model) mainly serves, therefore it doesn't matter to them if it's way out on the other side of the world!

[Interitus]

The NAEFS doesn't feature on the NCEP bias page because it combines US, Canadian and Mexican ensemble forecasts which are already bias corrected.

[forecaster]

thanks for the reply steve, though i was talking in general about all the models, not just the NCEP.maybe as you say, the bias is towards the region it (each model) mainly serves, therefore it doesn't matter to them if it's way out on the other side of the world!

Not easy to fix these apparent biases. Models: take equations and integrate forwards in time. They don't see things the same way we do.

The NAEFS doesn't feature on the NCEP bias page because it combines US, Canadian and Mexican ensemble forecasts which are already bias corrected.

Not the same thing as fixing subjective biases. Bias correction is the most bog standard statistical technique around, it has little impact on what people here complain about.

[swfc]

18z looks very flat at 168 hrs with much less ridging than earlier.looks more likely a wsw than any build of pressure towards scan!!!

Edited December 1, 2013 by swfc

[NorthNorfolkWeather]

I have watched the models flip-flop over the forecast N blast at the end of the week, for the last few days, and with the latest run, they seem to have opted for this northerly merely giving us a glancing blow. Something doesn't look right to me, so let's look at the current position:

hemi1_00011213.png

Strong upper ridge still over us - yes. But what's happening further afield? A vigorous upper low heading into the Pacific NW with very cold Arctic air spreading SW across the W USA. All this suggests a likelihood of amplification of the flow, as indeed the NWS guidance in the States has been talking about for some days.

Strong troughing then over the W US, adding amplitude to the Canadian vortex, helping our upper high to retrogress into mid Atlantic - down comes the upper trough at our longitude along with the Arctic flow.

But instead, the models push the HP back over us again - peculiar to say the least, not impossible, but to me, it just looks all wrong. I think the crucial development point will be at around T120, in terms of how the cyclogenesis actually takes place S of Greenland. Given the strong baroclinic zone that seems likely to develop at that point in that area, it seems quite likely that a vigorous LP will peel off south-eastwards and fully engage the cold air, and pull quite an intense cold pool towards the UK.

As I said the other day, developments over, and to the east of the Canada/US east coast will clearly have a big impact on this development, and to what follows from there, but I don't see the models have got a proper handle of all this, as of yet. Signs of more pronounced upper flow amplification around the hemisphere have been there for some days.

An interesting twist has come with the 12Z GFS suggesting HP blocking starting over Scandinavia, but that is very long term. It still needs to get a grip on more immediate developments.

It will be interesting to see how the Met O FAX handle it later. I suspect they will go for the glancing blow, but we'll see.

Thanks OMM,

 

Nicely explained and put in terms that most of us can understand.  I agree with your comments that "it just looks all wrong".  I cannot believe the "glancing Northerly",  I think the charts shown yesterday will be closer to the mark than those produced today.

 

Why? I can't tell you, but, like you, looking at the charts there does seem to be a tendency of the models to shove everything off East (Steve Murr's Eastward Bias)

[Mucka]

GFS 18 has gone off on one, who would of thought?

 

 

Rather it show one of the flattest and mildest solutions for this range than one of the most amplified and promising if it is going to be wrong.

[SnowObsessor90]

Did people say there was no sign of Zonal? I really hope this run isn't the first  sign of the mist beyond next weekend beginning to clear....

 

Okay maybe not but what GFS has gone onto show really isn't much better.

Edited December 1, 2013 by SnowObsessor90

[Frosty.]

Bin the 18z it looks all wrong in low resolution, the gfs is always slower at picking up blocking signals. Cold and blocked is the trend tonight once the heavy snow clears.

[Ali1977]

Heights about to

Build towards scandy I think at 204

[Cheshire Freeze]

Once again, an easterly cometh

 

 

I think people are going to have to accept a few days of southerlys to enable the potential easterly.

Yep, the high builds once again

 

Edited December 1, 2013 by CreweCold

[Cheshire Freeze]

WAA right into the Arctic

 

[phil nw.]

It does look like a brief Arctic spell.

The T96/120 fax's

 

 

much in line with the raw output by the looks.

A cold,frosty weekend to follow under the high then it all depends on where the block goes from there.

Disagreement even in the ens tonight on next week-a lot to decided yet.

 

[swilliam]

18z looks very flat at 168 hrs with much less ridging than earlier.looks more likely a wsw than any build of pressure towards scan!!!

Well it gets there in the end with a Russian super block - lots of variations on offer I suspect at this range.

 

Plenty going on the other side too.