Model Output Discussion: 12Z (02/12/12) And On...

[Cloud 10]

Purely because in about 10 years of model watching its never been taken seriously by most of us!

In fairness to us, the verification stats have never justified it being a big player until rather more recently...

Have a read here SK.

http://forum.netweat...-supercomputer/

[fergieweather]

The ukmo model goes further behind close doors? Can't remember

UKMO-GM prog never gets cited in operational forecasts beyond days 4-5. Ditto anythIng churned out of GFS. Only EC and MOGREPS used in any deterministic sense for MR stuff beyond day 6. The raw UKMO-GM as sent to public websites often bears little resemblance to the modified version we use in any case: hence UKMO forecast story may quite often not align with what you see on web output (ditto re NAE).

[snowking]

Have a read here SK.

http://forum.netweat...-supercomputer/

That clears up the reasoning for the vast recent improvements then

It was only as recent as last winter that we were discussing on here whether the big 3 should become the big 4 to include the GEM, with the JMA nowhere in sight

Clearly a very good update by the JMA

SK

[If Wishes made Weather]

Currently it is hard to see which, if any of the possibilities beyond circa +96 is going to prove more correct. The greater concern that I have as a snow fan is that, so far no control or operational run has shown temperatures to be consistantly in the range for snow. Rather, they remain somewhat boarderline with possible snow events, rain events and back to snow etc. Ian's comment though suggests now that this position is possibly one that is transitory into 'proper' cold later once a Scandinavian high pressure cell builds. Note though (if I am correct) that the teleconnective indicators and strat charts suggest a greater anomoly somewhat to the West of Scandi.

[swfc]

After the near suicide posts a week ago when the"big freeze"failed maybe its best for some folk to stick to the fax charts which are even changing at short notice!.Run after run will vary as is the nature of the models and i think to be expected.

[Stuie]

How do we know that a model is 'performing badly' if things haven't happened yet?

Very true. 1 way of benchmarking I suppose you could monitor precipitation forecast. Lets see is the GFS is correct on Thursday (when it comes). With it only being 4 days away that seems a reasonable time frame to get it right or wrong. (Good/Bad)

[fergieweather]

PS as of 2 yrs ago, Ops Centre no longer bother providing us with their T+144 GM prognosis, which says everything about their view on how fruitless it is examining that sort of stuff beyond that timescale (as rightly suggested above by another poster). Sticking with EC & MOGREPS is the way of things into timeline beyond that.

[JCP]

He is right, there will be no cold spell later on if that 12z UKMO comes off, its NOT good at all...

How do you know?

I can easily see how that could progress to a cold evolution within two days.

[Jonan92]

GFS's Run of Control showing pretty much the same

[bluearmy]

On the subject of gefs, as we enter low res, the spread is much lower than we've seen for a long time. We maybe approaching some confidence in the probable solution.

[mcweather]

He is right, there will be no cold spell later on if that 12z UKMO comes off, its NOT good at all...

I'm afraid that's complete tosh Kold.

I presume you wouldn't see a cold spell coming from this chart either then

http://www.wetterzentrale.de/pics/archive/ra/1947/Rrea00119470118.gif

[Bobby]

Well with all the talk of snow and cold, there could be a lot of rain over the next week or so, GFS rain accumulations

(hopefully some of that will be snow)

[bobbydog]

can someone define a "poor run"?

[IDO]

The short ensembles have wide scatter on Thursday:

http://modeles.meteociel.fr/modeles/gens/graphe_ens3.php?x=306&ext=1&y=141&run=12&ville=Londres&runpara=0

From -7 to +3c on the 850s. That is where FI starts.

The GEFS mean:

It has the tropical low mixing in with the low over the UK. With heights to our NW being progged:

However the control follows the Op with the tropical low:

Its a waiting game.

[fergieweather]

Very true. 1 way of benchmarking I suppose you could monitor precipitation forecast. Lets see is the GFS is correct on Thursday (when it comes). With it only being 4 days away that seems a reasonable time frame to get it right or wrong. (Good/Bad)

Actually I think PPN is about the WORST benchmark at that sort of range (indeed, even at T+48!!). More valid parameters would be MSLP, 500hPa geopotential & 850hPa temps.

[nick sussex]

The GEM shows just how volatile outputs can be when you view its 12hrs at 144hrs and compare it to the 00hrs at 156hrs.

The difference is laughable, regarding the UKMO if you look at the NOGAPS that shows how that could develop if it went further as they both look very similar at the 144hrs timeframe.

The GFS is really fine in the higher resolution although it blows up a few shortwaves to the north.

We wait for the ECM to see what that comes up with.In the shorter term theres still some uncertainty with the shortwave at 96hrs so I wouldn't be surprised to see the ECM have a different take on this.

[chionomaniac]

Actually I think PPN is about the WORST benchmark at that sort of range (indeed, even at T+48!!). More valid parameters would be MSLP, 500hPa geopotential & 850hPa temps.

Yes, I think a lot would agree.

I would always use the 500 hpa ensemble mean as any benchmark. The less spread the better and work back from there.

[Neilsouth]

He is right, there will be no cold spell later on if that 12z UKMO comes off, its NOT good at all...

UKMO has been the worst model so far this last month so I think we need to put it to the back burner for the time being.

[Buriedundersnow]

Well with all the talk of snow and cold, there could be a lot of rain over the next week or so, GFS rain accumulations

(hopefully some of that will be snow)

hopefully it will all be snow the totals look better towards the top of the map towards me in scotland that would do me fine if it was all snow

[chionomaniac]

UKMO has been the worst model so far this last month so I think we need to put it to the back burner for the time being.

Another of those statements not backed up by fact.

Actually apologies it is slightly under the GFS last month when looking at the big three. ECM the leader by a way.

but as ever short term past performance is no indicator of future performance - the more reliable long term performance is bettter here.

Edited December 2, 2012 by chionomaniac

[weather eater]

Firstly the UKMO, one it’s not terrible and two its verification is not the best at 144hrs plus it keeps changing from run to run.

The GFS is showing remarkable consistency not just out to 120hrs but also in the 120 – 180hr mark, yes there are detail differences but in the broad evolution very similar and it’s been that way now for eight consecutive runs. I suspect that it is over deepening and being over progressive with next weekend’s LPs but that’s to be expected of the GFS

Finally lots to resolve post next weekend but check out the 06z chart and the 12z charts for the 17^th I call that pretty consistent.

[bluearmy]

The short ensembles have wide scatter on Thursday:

http://modeles.meteociel.fr/modeles/gens/graphe_ens3.php?x=306&ext=1&y=141&run=12&ville=Londres&runpara=0

From -7 to +3c on the 850s. That is where FI starts.

The GEFS mean:

It has the tropical low mixing in with the low over the UK. With heights to our NW being progged:

However the control follows the Op with the tropical low:

Its a waiting game.

In the shorter term, the detail is very difficult to in down due to the number of systems and their interactions. Its feasible that the mid range to proper fi period (6 to 8 days) might be easier to predict if we have a solid upstream solution.

Edited December 2, 2012 by bluearmy

[fergieweather]

UKMO has been the worst model so far this last month so I think we need to put it to the back burner for the time being.

Ah: but that risks leaving faith only in the output solutions that favour any personal biases (e.g. wanting cold)... and that's poor science!! Objective and impartial comparison the only valid route, I hope (on this thread at least!!).... we see too much guff about "poor runs"; "bin this or that run"; etc etc. The weather will be what it will be.

[swfc]

Ah: but that risks leaving faith only in the output solutions that favour any personal biases (e.g. wanting cold)... and that's poor science!! Objective and impartial comparison the only valid route, I hope (on this thread at least!!).... we see too much guff about "poor runs"; "bin this or that run"; etc etc. The weather will be what it will be.

Yes but will it snow in sheffield Ian??? . Sorry.Id agree with you there regarding selective model picking.Any thoughts from you on the next 10-15 days outlook wise?Thanks in advance

[If Wishes made Weather]

I've been through the contol perbuatations for the GEFS +12z and there is only one that I can see of offering severe winter weather (i.e. -8 805 temps). The remainder all flirt in the 0 to -4 range most of the time indicating continuation of cool weather (similar to today). This will bring the CET down for December (assuming notihng above average thereafter), but presumably won't satify those who would see a 30cm+ dentritic snow, often. A blacked Atlantic may be showing, but our position on the boundary to continental weather remains.