Winter Model Discussion 18Z 3/2/13 onwards.

[cheeky_monkey]

i think what many forget when commenting on models is that they model the global weather pattern and not a single region..often people only look at the European region and dont look at the rest of the northern hemisphere to see how what is going on outside of Europe that may effect the pattern a few days down the line...then they get upset and frustrated and slate the models for being inaccurate when things dont quite pan out as hoped.

[SomeLikeItHot]

My point was more that GFS isnt "Cannon fodder" or "Dogs biscuits" as its often refered to in here. I agree that the no 1 model in our region over a long period is the ECM but its not by a massive difference and GFS is still in the top 3 and shouldnt be dismissed. Its still better than the JMA / NOGAPS / BOM etc. the 12Z and the 00Z are often the more accurate of the 4 and I personlay dont take much notice of the 06z and 18z.

Yes I agree. Personally I think the 6z and 18z should be regarded more as ensemble members rather than individual runs.

[weather eater]

Lots of posts about Stats over the last couple of days, we know the ECM is the best verifying model but that does not and will not guarantee it being right on this occasion. My own view is, that it is the ECM that models easterlies the poorest; I cannot remember it modelling a decent easterly at the 144hr mark, so it’s not a case of trusting the GFS it’s a case of not trusting the ECM. Yes the ECM is consistently modelling an easterly but that’s as far as its consistency goes, a small correction east and it’s not going to be of use, we need to lose the PV over Greenland and see the high retrogress west to give us the best shot of something memorable, we have some tentative signs of that but that’s all they are at the moment. I’d call it 50/50 at the moment for an easterly.

[IDO]

Hence why i come to the conclusion that the 2 are likely to meet in the middle somwhere and we may see a slack easterly flow with tough disrution over us and marginal snow.

A cold few days and then at the weekend we see the pivot point. "In the Middle somewhere" is the most likely option within the UK; a battleground scenario. The cold fails to go far enough west and the UK will be on the borderline of the cold.

BritishWeatherServices have been very good this Winter with their calls. At the moment (low confidence) they suggest the ECM option as about 10% chance

:

At this time of year, battleground scenarios may be our best chance for large accumulations of snow, somewhere.

Edited February 4, 2013 by i'm dreaming of...

[Frosty.]

BritishWeatherServices have been very good this Winter with their calls.

At the moment they suggest the ECM option as about 10% chance:

never heard of them, once again you are drinking from an empty glass, as usual.

Edited February 4, 2013 by Frosty039

[SnowBallz]

For me, the inconsistency of the GFS is what give it plausability (assuming there to be such a word)

There are so many factors that come into play when it comes to delivering to the UK a cold spell, so many of them which will always be on a knife-edge, that one almost expects and understands the variety of solutions put forward in so many runs and therefore goes for the trend rather than the results of any one individual run.

I really want the ECM outcome, I really do. But I'm actually suspicious of how rock-solid it's been at delivering the goods time after time. In the end, it's not consistency that matters but ending up getting it right, whatever solution you eventually decide upon.

I don't dismiss the idea that the ECM could prove to have been right all along. But I do doubt it.

And, believe it or not, you do make a critical observation. Indeed, a mathematical model which seems to show incredible consistency isn't necessarily to be trusted prima facie; it may, at a deterministic level, have a very strong error, which is recursive enough to deliver consistency.

To give you a good idea, a good example of such error would be if, for instance, there was a data blind spot. You can run an algorithm to normalise the data set, thus 'smoothing over' the lack of true values. However, if the output of the normalised data is inherent flawed - and especially if the data has a heavy weighting - then you'll tend to see recursion, which can appear to be consistency, which may belie accuracy.

It's easy enough to spot though, and - in principle - you'd run a data governance parameter, in concurrence, which will - at the end of the run - give you a score as to how 'true' or reliable the initial parameters were. It would also spit out a report which would identify data blind spots or smoothing issues, and then a forecaster can detemine how influential those variables would be in the overall model.

So, in summary: consistency isn't always what it appears to be - accuracy.

[london-snow]

For me, the inconsistency of the GFS is what give it plausability (assuming there to be such a word)

There are so many factors that come into play when it comes to delivering to the UK a cold spell, so many of them which will always be on a knife-edge, that one almost expects and understands the variety of solutions put forward in so many runs and therefore goes for the trend rather than the results of any one individual run.

I really want the ECM outcome, I really do. But I'm actually suspicious of how rock-solid it's been at delivering the goods time after time. In the end, it's not consistency that matters but ending up getting it right, whatever solution you eventually decide upon.

I don't dismiss the idea that the ECM could prove to have been right all along. But I do doubt it.

I really don't understand your post here and disagree on a few points ~

You don't buy the ecm idea even though its acting rock solid ~

Why not?

Until it either falls back into line with the GFS or to a extent the ukmo then you have got to believe it because it is showing consistancy.

The other point & major factor is that the gfs does not handle winter patterns that come from a e~ne well & is well known for throwing the pattern to far east at the start which slowly backs west day by day & run by run. Have we seen the gfs do this thus far & the answer is yes.

Imo its very easy to trust the ecm idea when its acting rock solid & its very easy not to trust the gfs when we know how it behaves when patterns come from the east.

Let's see how we get on this afternoon but i am backing the ecm all the way.

[Captain Shortwave]

And, believe it or not, you do make a critical observation. Indeed, a mathematical model which seems to show incredible consistency isn't necessarily to be trusted prima facie; it may, at a deterministic level, have a very strong error, which is recursive enough to deliver consistency.

To give you a good idea, a good example of such error would be if, for instance, there was a data blind spot. You can run an algorithm to normalise the data set, thus 'smoothing over' the lack of true values. However, if the output of the normalised data is inherent flawed - and especially if the data has a heavy weighting - then you'll tend to see recursion, which can appear to be consistency, which may belie accuracy.

It's easy enough to spot though, and - in principle - you'd run a data governance parameter, in concurrence, which will - at the end of the run - give you a score as to how 'true' or reliable the initial parameters were. It would also spit out a report which would identify data blind spots or smoothing issues, and then a forecaster can detemine how influential those variables would be in the overall model.

So, in summary: consistency isn't always what it appears to be - accuracy.

Pretty much what I think. Simply put the ECM has spotted something. It might be correct or incorrect. But the fact it's still holding firm with other models going towards it must be a good sign that the ECM op has picked up on the correct signal and pretty much ran it from when it appeared at T240, which if it proved correct is a major victory for the ecm op, not a fail for the gfs as frankly until recently the ECM op was on it's own.

@timmytour, you were using the reverse argument for the last cold spell. Political spin doesn't work on a weather forum

Edited February 4, 2013 by Captain shortwave

[Tim Bland]

A cold few days and then at the weekend we see the pivot point. "In the Middle somewhere" is the most likely option within the UK; a battleground scenario. The cold fails to go far enough west and the UK will be on the borderline of the cold.

BritishWeatherServices have been very good this Winter with their calls. At the moment (low confidence) they suggest the ECM option as about 10% chance

:

At this time of year, battleground scenarios may be our best chance for large accumulations of snow, somewhere.

Very much agree with what they are saying... its not a glass half full, its just realistic. Its also what the Metoffice believe will occour (see their long range forecast). As you say battleground scenarios are the UKs best source for snow. 1964 is the perfect example. We V rarely will get heavy widespread snow without a risk of an atlantic influence. 2 weeks ago being a prime example, it wasnt deep cold with frozen lakes but we had widespread and fairly deep snow. IMO best possible scenario is undercuting lows or stalling atlantic bringing deep snow. Maybe its becuase Im a snow fan and not a deep cold fan... Cold just means higher heating bills and more uncomfortable... Snow means put the skis on and head to the hills or build a snowman with the kids,

[Tamara]

It isn't about consistency per se - but it is about consistency when that consistency clearly shows a trend being carried to a certain point and almost insiduously gains support along the way - up to a certain point that only time can reveal when that is. There is a big difference in implication here between this and simply consistency on its own that is quickly found wanting if it is founded on a completely wrong idea. It doesn't always happen by any means, but it is quite conceivable for a persistent apparent 'non consensus' solution to be the right one too, and on that basis it is not dismissable until proven to be dismissable.

So in terms of the ECM and a trough followed by an easterly, it is everything about the amplification and number of degrees that any trough non disrupts east (warmer) or disrupts south east, or, better still south at the sharpest angle (colder). But it isn't black and white, on or off by any means where there is at least a little scope either way to play with. One only has to look at the margins between the current UKMO and ECM to illustrate this. The UKMO on its latest ouput could still disrupt the trough far enough west to allow the cold to back west to some degree afterwards. On that basis one can logically argue quite easily that the ECM has *some* margin to correct itself eastwards and not drastically alter the outcome other than make the area of marginality shift east accordingly and mean that any subsequent easterly (if it happens) does not advect quite as far west.

We saw all this happen in January, so we know, or should know that it isn't cut and dried/black or white .

Edited February 4, 2013 by Tamara Road

[MR EXTREMES]

gfs and ukmo look similar.

[Captain Shortwave]

T90, miles better than the 6z, high pressure already building over scandinavia

http://modeles.meteo...gfs-0-90.png?12

well it looked better for a while

Edited February 4, 2013 by Captain shortwave

[surrey slush]

So, in summary: consistency isn't always what it appears to be - accuracy

.

Indeed it's not, two independent measures, but consistency taken with the verification results, do give some credence to the ECM Atm, Whereas GFS flapping + the lower verification tends to undermine it. I think we can relegate Timmy's ramblings to the 'noise' pile, along with seagulls and grannie's joints.

[cold snap]

looks like the ukmo is going with the gfs

ecm is surely wrong

[Tamara]

looks like the ukmo is going with the gfs

Yes UKMO is flatter at t96 than at the same time this morning

Holds up a bit better though at t120 - flatter than ECM though

Edited February 4, 2013 by Tamara Road

[Tim Bland]

I realy hope the ECM sticks to its guns toinght or its going to be very quiet in here tommorow with very little to debait. As i said earlier though I expect the GFS to become more amplified and the ECM less so until we reach a middle ground.. Perhaps UKMO is the model to watch and is about to show us what the middle ground is?

[SMU]

This threads a joke sometimes. Just like the GFS.

UKMO 120.

Fine

[No Balls Like Snow Balls]

looks like the ukmo is going with the gfs

ecm is surely wrong

Excellent analysis

[Tim Bland]

Steve, Can you clarify; do you believe that the ECM has it spot on and is unlikely to back tract at all in the next couple of days? Im genuinely interested (and sure others are) as you are a very knowledgable and respected forum member

Thanks

Tim

[Jan 87]

This threads a joke sometimes. Just like the GFS.

UKMO 120.

Fine

http://www.meteociel...20-21.GIF?04-17

Is it? Atlantic on the door, could go either way but too close for comfort!

[winterof79]

This threads a joke sometimes. Just like the GFS.

UKMO 120.

Fine

http://www.meteociel...20-21.GIF?04-17

Quite a lot of over analising going on.Down she goes http://www.meteociel.fr/ukmo/runs/2013020412/UN144-21.GIF?04-17

[Ice BIast]

This threads a joke sometimes. Just like the GFS.

UKMO 120.

Fine

http://www.meteociel...20-21.GIF?04-17

I must say that I was struggling to see what was bad about it at T96, the position of the atlantic low is the important part and it is fine.

Edited February 4, 2013 by lce Blast

[Tamara]

Perhaps UKMO is the model to watch and is about to show us what the middle ground is?

On the evidence so far this evening I would agree with that part at least. t144 on UKMO to come yet

Edit : Looks to have the trough slightly further east of this morning - through the centre of the UK and further east for sure than ECM

Edited February 4, 2013 by Tamara Road

[SMU]

UKMO 144-

Low sliding South- one of the most easterly solutions seen V the ECM.

S

[Snowman.]

UKMO is low res after T72 or whatever.

If it was higher you would be able to see the differences between the UKMO and GFS.

OFC the T144 chart is fine people really need to take a chill pill and stop calling out the ECM saying its wrong before it even happens.

Edited February 4, 2013 by SN0WM4N