Winter Model Discussion 18Z 5/2/13 onwards.

[Coast]

the ecm backtracks ,and the gfs after faffing about for days comes on board and shows even greater cold coming!!

But surely a model can't 'backtrack' 'upgrade', 'downgrade', or be judged good, bad or indifferent about a future weather event until after that event has passed? Using the football team analogy your Man U. Man C, Chelsea etc are of course in a premier league, but until Saturday evening how can you say that they will win the match they are due to play on Saturday afternoon - no matter how good they have been in the past? If we could, we would all be millionaires.

I get the feeling that some pick a future weather event they favour and then go looking for a model that fits that scenario and call it good?

[weather eater]

Oh dear! Why do so many folks still persist in 'supporting' models, as though they were football teams?

They can be fickle as well altering support, often local factors plays a large part, for one or two anyway.

In truth it’s hard to fault the ECM this time round it looks a good bet now, and the GFS has indeed played catch up, but that’s not always the case we have seen plenty of times when things have been the other way round and in January both the GFS and the ECM played catch up with the UKMO, in December the GFS never really went for an easterly in the way that the ECM did. Even this time round we are possibly (only time will tell) going to see a backtrack by the ECM on the duration of an easterly, it is starting to look a very short affair indeed, although it seems pretty obvious to me that there is much to resolve before drawing conclusions, I do have to wonder if now that the GFS has caught up with the ECM it now goes on to handle the further evolution better, will it get the same kudos.

[SMU]

But surely a model can't 'backtrack' 'upgrade', 'downgrade', or be judged good, bad or indifferent about a future weather event until after that event has passed? Using the football team analogy your Man U. Man C, Chelsea etc are of course in a premier league, but until Saturday evening how can you say that they will win the match they are due to play on Saturday afternoon - no matter how good they have been in the past? If we could, we would all be millionaires.

I get the feeling that some pick a future weather event they favour and then go looking for a model that fits that scenario and call it good?

Im afraid you are in denial. The mere fact that theres an event progged for the weekend eliminates the GFS - if we would have used this model until todaythe event would have been heading NE to lapland....

[Methuselah]

I get the feeling that some pick a future weather event they favour and then go looking for a model that fits that scenario and call it good?

It might be more than a 'feeling', Coast?

[cornish snow]

What i dont understand is why certain people even look at the GFS runs if they are so poor.

My advice is DONT. and then you wont get all worked up.It's not healthy you know.

[bluearmy]

As usual, things getting a bit silly now. Why are we debating specifics at range (and also backwards).

Ecm op was the only forecast tool that progged an undercut at fi range. There is nothing else to debate. Next time a model goes out on a limb it could be gfs and it could be right. However, in the current pattern where it seems undercutting could be a recurring theme, is gfs likely to be the most trustworthy tool? As it happens, I don't think next week looks easy to call as there look to be too many bundles of energy and blocks fighting for supremacy.

[lord stratus]

But surely a model can't 'backtrack' 'upgrade', 'downgrade', or be judged good, bad or indifferent about a future weather event until after that event has passed? Using the football team analogy your Man U. Man C, Chelsea etc are of course in a premier league, but until Saturday evening how can you say that they will win the match they are due to play on Saturday afternoon - no matter how good they have been in the past? If we could, we would all be millionaires.

I get the feeling that some pick a future weather event they favour and then go looking for a model that fits that scenario and call it good?

You are correct, for example if you take yesterday's output from ECM and ukmo you would expect to see a snowy Sunday for much of the UK. If we have this snow event on sunday then you can call the ECM rock solid. Ukmo tends to play it safe though so could not be given the same amount of praise!

[Gibby]

Looks like I'm missing out again then not seen a covering this winter

What i dont understand is why certain people even look at the GFS runs if they are so poor.

My advice is DONT. and then you wont get all worked up.It's not healthy you know.

FWIW I could perhaps be of as good authority as anyone to comment on how I feel the models perform and my findings are as follows.

GFS is generally fairly good particularly when it comes to short ranging of Atlantic Westerly regimes for say 4-5 days head. Thereafter, I find it falls away somewhat in reliability with a tendency to exaggerate depth of Low pressures and over progressiveness. It handles blocking situations to the NE less well, especially in Winter. The fact that it runs four times per day can sway people's views that it often is better than it actually is as it has two additional sets of data fed in that the General Public can see at 06z and 18z.

UKMO is usually the form horse that I put my short term faith in. It is nearly always spot on within the 72hour timeframe with the usual less reliability factors coming into play thereafter. The 144hr chart is often somewhat suspect and can swing as frequently as GFS at that range. I do feel it's strengths are best displayed in cold blocking situations such as the period approaching and once inside t96hrs it rarely sways.

ECM is the best model for longer term trends and through my experience over the years I find it rather more reliable in the t144 to t240 hr range than GFS and having found a trend generally sticks to it. One of the notable exceptions was of course the raging Easterly it showed in December but that really was a rare debacle from an otherwise well respected model by me. It though like UKMO handles blocking situations relatively well and certainly better than GFS.

Of the others I find GEM generally relatively good for verification which falls off quite quickly later in its ranges. The other models I do not profess to look at much but fom what I have read here and elsewhere they don't appear to be of the standard of the above in our little corner of the World.

On a final note remember models are just Mathematical calculations based upon a million of starting values which compute to what we see. They all get it wrong at times and this is where the faxes are a valuable human tool in fine lining the shorter term outlooks.

Edited February 6, 2013 by Gibby

[trickydicky]

But surely a model can't 'backtrack' 'upgrade', 'downgrade', or be judged good, bad or indifferent about a future weather event until after that event has passed? Using the football team analogy your Man U. Man C, Chelsea etc are of course in a premier league, but until Saturday evening how can you say that they will win the match they are due to play on Saturday afternoon - no matter how good they have been in the past? If we could, we would all be millionaires.

I get the feeling that some pick a future weather event they favour and then go looking for a model that fits that scenario and call it good?

Because this weekend there is going to be some sort of slider low that will bring to somewhere a snowfall event. Up until yesterday this 'slider' was shown on the GFS for days to be 'brushing' over the top of Scotland. So, therefore, for a good few days, the GFS was wrong while the ECM and UKMO which were showing the slider were correct. Why is this so hard to understand for a couple of folk on here?

[Yarmy]

And Bermuda would be under -10C uppers!

Could happen

http://upload.wikimedia.org/wikipedia/en/d/dc/SnowballSimulations.jpg

Bermuda has shifted somewhat since then though, and the entire NH landmass is missing, so perhaps this is unlikely to pan out synoptically.

[A.J]

Im afraid you are in denial. The mere fact that theres an event progged for the weekend eliminates the GFS - if we would have used this model until todaythe event would have been heading NE to lapland....

Fair enough Steve, and indeed everyone is perfectly entitled to air their views on the merits and fails of the weather models of their choosing....but you've made your personal point of view not only obvious but made it ad nauseum. This thread is for 'across the board' model output discussion, there are more appropriate threads on this forum for discussing the merits of weather models

For all members, can we please stick to actually discussing what each relevant model run is showing, this is what the thread is for...cheers!

[Ryan Dutch Weatherman]

Ok letse see CMA 00z at 240 Hours en see GFS chart 384 hours.

Its the same almost.

The lock down en the alliantie with the H [ Greenland/Atlantik/Russian Federation]

I think it will be good times with snowfallls.

[Methuselah]

Because this weekend there is going to be some sort of slider low that will bring to somewhere a snowfall event. Up until yesterday this 'slider' was shown on the GFS for days to be 'brushing' over the top of Scotland. So, therefore, for a good few days, the GFS was wrong while the ECM and UKMO which were showing the slider were correct. Why is this so hard to understand for a couple of folk on here?

Slating the GFS off - just because it doesn't always show Snowmageddon - is a far cry from suggesting that it ain't always right...

What do the official verification stats say?

[MR EXTREMES]

Look i saw alle the models 00z .

I could be wrong , i admit there is lot of uncertain.

But wen i have my life or money on it i would choose this in the weather roulette models.

This one is gone be worked out i think.

Its just like januari little bit.

Thats my true opinion .

I could be wrong , but thats the way forward , i believe CMA 00z at 216 hours.

End maybe some experts can for you people explain wat this chart means , becouse i cannot express my self good in your language in weather jargon.

its a very valid point u have put across and january had a very good coldspell here on the southcoast and it was the same sort of setup with the slider lows.

its not 100% but the likely hood of another coldspell is there for the taking get rid of some of these shortwaves and its game on but we see flip flops yesterday even from the ecm and very much anything beyond t144 is fi right now even sooner to be honest.

[trickydicky]

Slating the GFS off - just because it doesn't always show Snowmageddon - is a far cry from suggesting that it ain't always right...

What do the official verification stats say?

I wasn't slating it off because it wasn't showing snowmageddon, none of the models have really have they. What I was saying is that the GFS was showing for this weekend something that we now know isn't going to happen. What it was showing, and what is going to happen is irrelevant. I couldn't care less which of the models get it right.

[mountain shadow]

The answer to the GFS, is to ignore it unless there is agreement with the ECM and UKMET.

Once their is agreement then look at the various GFS charts on Netweather to determine whether you will see snow or not.

[Tamara]

ECM postage stamps show that there are a fair number of members that still have the sliding low sharper than this mornings operational, with less WAA ahead of it. The evolution on the back of the low looks better on many of them as well. In that sense I think the operational might be warmer this morning than the reality. Time will tell of course

http://www.ecmwf.int...0600!!chart.gif

Edited February 6, 2013 by Tamara Road

[AWD]

ECM postage stamps show that there are a fair number of members that still have the sliding low sharper than this mornings operational, with less WAA ahead of it. The evolution on the back of the low looks better on many of them as well. In that sense I think the operational might be warmer this morning than the reality. Time will tell of course

http://www.ecmwf.int...0600!!chart.gif

It is in the later stages of the run;

[bluearmy]

ECM postage stamps show that there are a fair number of members that still have the sliding low sharper than this mornings operational, with less WAA ahead of it. The evolution on the back of the low looks better on many of them as well. In that sense I think the operational might be warmer this morning than the reality. Time will tell of course

http://www.ecmwf.int...0600!!chart.gif

tamara, looing at the graph posted by AWD, you can see the spread on sunday is large and therefore a raft of solutions is possible re the trough track (and indeed if it will be a single circulation). i would be waiting for the T95 FAX to see what exteter think, irrespective of the 12z runs.

[TEITS]

Looking at the ensembles its becoming pretty obvious to me that a shift E of the LP is going to occur like the UKMO/ECM. The 06Z was a colder run for my location which is why the precip didn't even reach the E. However you look below and it was colder than the mean.

http://modeles.meteociel.fr/modeles/gens/graphe_ens3.php?x=302&ext=1&y=78&run=6&runpara=0

So contrary to the Met O forecast I feel W areas are more likely to see rain. I feel the risk of snow is going to be mainly focused on central, maybe E areas.

[snowmadchrisuk]

Looking at the ensembles its becoming pretty obvious to me that a shift E of the LP is going to occur like the UKMO/ECM. The 06Z was a colder run for my location which is why the precip didn't even reach the E. However you look below and it was colder than the mean.

http://modeles.meteo...run=6&runpara=0

So contrary to the Met O forecast I feel W areas are more likely to see rain. I feel the risk of snow is going to be mainly focused on central, maybe E areas.

I agree , If you take my area in Somerset , all the Ensembles bar 1 have shot up to near 0deg at 850 level. The Operational was virtually an outlier for Sunday ,

[That ECM]

I agree , If you take my area in Somerset , all the Ensembles bar 1 have shot up to near 0deg at 850 level. The Operational was virtually an outlier for Sunday ,

If you are going by the latest info then I am not sure central ares we see much. This looking out to next Monday. Cold rain would be likely in most areas. It can change tho. Lets hope so.

[bluearmy]

my anecdotal view of the 06z suite is that it tends to be a little eastward biased with a slightly stronger jet. others will have a different opinion.

[Tamara]

tamara, looing at the graph posted by AWD, you can see the spread on sunday is large and therefore a raft of solutions is possible re the trough track (and indeed if it will be a single circulation). i would be waiting for the T95 FAX to see what exteter think, irrespective of the 12z runs.

Yes I can see that clearly from the graph . I guess for my own purposes I like to get an insight into the actual synoptics that the individual members show. I like to be able to visualise the maps against the lines on the graph, rather than just the ensemble graphs on their own which mean less to me. I find it helpful in trying to see and decide in my simple little way what may happen

Its interesting that the GFS operational, after so many recent days, has a more westward suggestion than the European models for Sunday/Monday! But it does seem that the consensus is slightly more east than the Euros suggested yesterday - which is not the most ideal for many. But we will see.

Edited February 6, 2013 by Tamara Road

[beijingboy]

"GFS - didnt even look."

I think this is my first - and possibly my last, once it's removed - post under new membership; currently living in Beijing, although I've been following this (model discussion) forum for approaching a decade.

Steve, you are a VERY respected member on here. And I, myself, have a lot of time for your (along with others') analysis. There's no question you can interpret the info available to us, and you have a very friendly posting tone. But I am GMT+8 - so when the GFS rolls out at (approx) 3.30am in the UK, the first thing I do is look to see who is up for it.

Given you were logged-on/monitoring at the time, I can't imagine why you would say you didn't even look at it - surely the temptation would be too much (and is it not the only run filtering out at that time?) Forgive me, I genuinely mean no offence, but I know a shortwave from a shortcrust, and 'tis the former that might well scupper any (immediate) chances of a prolonged cold spell. The GFS run has progged this, and makes for very uncomfortable viewing for snow/cold lovers. Might as well tell it as it is, despite our yearnings for a sustained E'ly feed, or N/E'ly with some significant heights over Greenland (my personal preference).

As is the case every single winter, people almost invariably ignore (or try to ignore) a bad run. I'm sure there are others on this forum who just want an honest appraisal, and particularly those who can't yet read charts and are trying to learn. To an extent, I am one of them. I really do just appreciate an honest appraisal of what's being shown, beyond my own limited understanding.

OK, I'm done now. My apologies if I've upset/offended anyone; that was not my intention.

I would say I'll get my coat. But perhaps I'll only need a cardigan.