General Model Output Discussion

[nick sussex]

I find it so amusing, not being such a poster on here but now rather a watcher, at how peoples bias towards which model is showing what they want rather than going with what is being shown.

As for instance, i remember a lot of people on here slagging the ECM off just before xmas because it was not being consistent. The GFS was consistently showing much colder weather whereas the ECM was all over the place. People slagging this model off because of inconsistencies. Now the ECM is showing what people want, every1 is now slagging off the GFS which showed and predicted the cold spell correctly to begin with.

Not having a go, just was thinking at the bias towards the model which looks the best.

Im sure if the GFS was showing the cold weather extending and the ecm was showing it ending, every1 would be moaning about how inconsistent the model is and that every run that doesnt show what people want should be binned.

Sorry but the verification stats don't back your argument, the ecm has had one wobble on one single operational run in the last 3 weeks, the gfs has been all over the place and did not predict the cold spell, it doesn't matter what the ecm shows at 168hrs and beyond as all output is subject to variability at that timeframe including the gfs. The gfs has constantly gone AWOL within 144hrs something the ecm rarely does. In terms of people picking a model because it shows them what they want to see, well that might happen sometimes but i think if you look at the regular posters we've all been objective about the output and have viewed this whilst also keeping in mind the background teleconnections. Also bearing in mind an obvious inherent bias of the gfs to remove European blocking prematurely and overdo cyclogenesis. This isn't hopecasting, it's not just looking at operational runs and accepting them as fact it's taking into account alot more than that.

[weather eater]

I find it so amusing, not being such a poster on here but now rather a watcher, at how peoples bias towards which model is showing what they want rather than going with what is being shown.

As for instance, i remember a lot of people on here slagging the ECM off just before xmas because it was not being consistent. The GFS was consistently showing much colder weather whereas the ECM was all over the place. People slagging this model off because of inconsistencies. Now the ECM is showing what people want, every1 is now slagging off the GFS which showed and predicted the cold spell correctly to begin with.

Not having a go, just was thinking at the bias towards the model which looks the best.

Im sure if the GFS was showing the cold weather extending and the ecm was showing it ending, every1 would be moaning about how inconsistent the model is and that every run that doesnt show what people want should be binned.

Not sure who you are referring to as I see no model bias, just balanced judgements on a broad range of model outputs. Nobody's going to start taking any models outputs seriously if the model in question is jumping from one evolution to the other as is the case with the GFS. Yes the 12z shows a swift breakdown, however the 06z was a frigid run throughout. Just as before Christmas the GFS was showing consistency now its the ECM, I think that a fair way to view model output, especially given that its the GFS that has fallen into line behind the ECM over the last few days. If over the next few days the GFS becomes consistent and the ECM starts wavering then the GFS will be more believable even if it does show a return to milder conditions.

[Cloud 10]

ECM+168 GFS mean+168

Good agreement between this evening's ecm and gfs mean for day 7 with low pressure going under the block prolonging the cold theme.

The gfs mean chart from this time last week has verified very well indeed with the pressure patterns we have today.

[Mark Bayley]

I find it so amusing, not being such a poster on here but now rather a watcher, at how peoples bias towards which model is showing what they want rather than going with what is being shown.

As for instance, i remember a lot of people on here slagging the ECM off just before xmas because it was not being consistent. The GFS was consistently showing much colder weather whereas the ECM was all over the place. People slagging this model off because of inconsistencies. Now the ECM is showing what people want, every1 is now slagging off the GFS which showed and predicted the cold spell correctly to begin with.

Not having a go, just was thinking at the bias towards the model which looks the best.

Im sure if the GFS was showing the cold weather extending and the ecm was showing it ending, every1 would be moaning about how inconsistent the model is and that every run that doesnt show what people want should be binned.

People back the ECM because it is the most consistent model. It has had its wobbles this winter, in fact the GFS parallel was doing very well. Overall though the ECM has been and still is the best performing model. The NOAA discussions have constantly slated there own model going with the ECM and its ensembles more times than not. The ECM is not bulletproof however it handles these situations far better than the GFS which wants to bring back mild weather when ever it can and over blow Atlantic lows. The GFS could be correct however it is pretty much on its own with little support.

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

Model comparisons show the GFS is currently the worse performing model with the ECM at the top

[FetchCB]

Looking at the Ensembles it looks if the Op run has overdone the depth of the pressure( as the GFS is prone to doing) .It is on its own and I supsect that is why it goes off on a tangent like it does in blasting away the European Block

[Long haul to mild]

I find it so amusing, not being such a poster on here but now rather a watcher, at how peoples bias towards which model is showing what they want rather than going with what is being shown.

As for instance, i remember a lot of people on here slagging the ECM off just before xmas because it was not being consistent. The GFS was consistently showing much colder weather whereas the ECM was all over the place. People slagging this model off because of inconsistencies. Now the ECM is showing what people want, every1 is now slagging off the GFS which showed and predicted the cold spell correctly to begin with.

Not having a go, just was thinking at the bias towards the model which looks the best.

Im sure if the GFS was showing the cold weather extending and the ecm was showing it ending, every1 would be moaning about how inconsistent the model is and that every run that doesnt show what people want should be binned.

I agree with that some may only want to see cold weather so they will cherry pick the model that show this best set up. Back before I learnt more about the models I used to pick out 36 hour Northerly Topplers at T+240 and treat them as gospel!

However, the GFS has performed pretty poorly over the last few weeks; it picked out a phantom easterly (albeit a slack one) for the S in mid Dec as close as T+48 that never came to fruition and that the ECM never bought into. And earlier this week for several runs, the GFS brought High pressure across all but the far SE for the coming weekend before gradually backtracking to the ECM / UKMO over the course of the next few days!

Nick Sussex has also often mentioned that the NOAA have binned the output and they must have good reason for doing that.

Nowadays I try to follow the ECM or UKMO (or even the GEM) as I think they are likely to be closer to the truth (even if I don't like what I see).........

Edited January 7, 2010 by Chalk Serpent

[nick sussex]

Not so sure sizzling Heat

I do not tend for bias, in fact, i tend for mild but like to see snow for the benefit of everyone else.

The GFS has really not performed well and even the NOAA (The people whose model it is)have stated how badly it has performed. The GFS has always been a bit over-progressive with breakdowns,

I wonder if someone could post the verification charts up to see how well GFS has done ??

These are the verification stats for the ecm against just the gfs at day 5. Both ecm runs beat all the gfs output.

http://www.emc.ncep.noaa.gov/gmb/STATS/html4x/acz5.html

In terms of the 00hrs for day 6, the ecm again beats all the models, the gfs scraped into joint 2nd place but i think it's main problems are it's other runs, the 00hrs hasn't been so awful.

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

So the ecm without a doubt is the best global model.

Edited January 7, 2010 by nick sussex

[Mondy]

I haven't read the last 10 pages or previous model threads due to various reasons, but is there good agreement on a breakdown? I'm personally missing the Atlantic weather!

[snowwonder]

Mods feel free to delete if needed, but as a watcher on here for some years I find some posts very, very confusing. I get fed up reading one post saying this then the next saying complete oposite. I know its opinions and its good that some differ but sometimes it makes it so hard for someone like me ond prob others who cant read charts etc very well at all who come in here for information and interest in the weather. I do filter, where poss, who I read in depth though and I am sure these experienced posters know who they are, but some of the stuff that goes on here in time will prob push those particular posters away.

This is a fantastic website and some serious time must be given by a lot of people to keep this up, I just wish there was a little more respect sometimes, from some, towards these people.

Rant over.

I hope in time I can get a grip on models so I can contribute a bit more to discussion.

[AtlanticFlamethrower]

12z ECM looking much more like a Reload rather than a Slow Thaw. Iceland-Scandi high might not have to go to Greenland, it could elongate above the British Isles pulling down an Arctic high, then retrogress NE to become a Scandi and fire over a viciously cold NEly.

Or the high could be progressively moved towards Greenland as ECM gets a better grip on what is happening.

I think 12z ECM is a big boost to hopes/fears of Reload, with perhaps only a minor thaw or none at all in between.

[Frosty.]

Ignoring the gfs 12z which is all over the place in FI from run to run, 3 totally different fi evolutions so far today with another to follow in a few hours. The ukmo 12z looks very cold out to T+144 with a roar ESE'ly wind off a frozen near continent although it does show a low to the southwest which may inject some milder air into the far sw of england later next week or maybe not if the ecm 12z verifies, it shows the cold continuing right out to T+240 with another reload possible beyond that, not so sure how much snow there will be beyond say monday next week but probably some flurries in the east mainly and frosty nights for all with temps by day no more than a degree or two above freezing in most areas for a few hours late am / early pm.

[AtlanticFlamethrower]

12z ECM looking much more like a Reload rather than a Slow Thaw. Iceland-Scandi high might not have to go to Greenland, it could elongate above the British Isles pulling down an Arctic high, then retrogress NE to become a Scandi and fire over a viciously cold NEly.

Or the high could be progressively moved towards Greenland as ECM gets a better grip on what is happening.

I think 12z ECM is a big boost to hopes/fears of Reload, with perhaps only a minor thaw or none at all in between.

[Rollo]

Sorry but the verification stats don't back your argument, the ecm has had one wobble on one single operational run in the last 3 weeks, the gfs has been all over the place and did not predict the cold spell, it doesn't matter what the ecm shows at 168hrs and beyond as all output is subject to variability at that timeframe including the gfs. The gfs has constantly gone AWOL within 144hrs something the ecm rarely does. In terms of people picking a model because it shows them what they want to see, well that might happen sometimes but i think if you look at the regular posters we've all been objective about the output and have viewed this whilst also keeping in mind the background teleconnections. Also bearing in mind an obvious inherent bias of the gfs to remove European blocking prematurely and overdo cyclogenesis. This isn't hopecasting, it's not just looking at operational runs and accepting them as fact it's taking into account alot more than that.

As an avid watcher of both of GFS and ECM Nicks statement just backs up my thoughts,of course at some point the GFS will get it right but when is the question. I suppose perhaps we should take into account that GFS goes out to T384 while the ECM only to 240.

[johnholmes]

A little more Info on this :

The 850mb level of the atmosphere is about 5,000 ft. above sea level. Roughly, freezing temps at 850mb can produce snow. However, if temps are warmer between the surface and 850mb, precipitation could still be rain. Sounding data from the models is then useful to determine the temperature at each level of the atmosphere, which helps determine "thicknesses." Again, roughly speaking, a thickness of 5400 or below will probably provide snow, while thicknesses above 5400 probably won't.

apart from saving frayed nerves on the forum the 1000-500mb thickness is almost immaterial in a frontal set up such as MAY develop over/after the weekend.

Nick F explained the element of 850mb and below well just above.

Like I posted earlier the scenario which MAY be played out into next week is the worst possible in terms of ANY forecaster, with or without model help, being very sure more than 3-6 hours ahead what is going to happen.

Accept that and lots of frayed nerves, tempers and such will be spared-trust me I've been there many times and no what its like-a nightmare.

I find it so amusing, not being such a poster on here but now rather a watcher, at how peoples bias towards which model is showing what they want rather than going with what is being shown.

As for instance, i remember a lot of people on here slagging the ECM off just before xmas because it was not being consistent. The GFS was consistently showing much colder weather whereas the ECM was all over the place. People slagging this model off because of inconsistencies. Now the ECM is showing what people want, every1 is now slagging off the GFS which showed and predicted the cold spell correctly to begin with.

Not having a go, just was thinking at the bias towards the model which looks the best.

Im sure if the GFS was showing the cold weather extending and the ecm was showing it ending, every1 would be moaning about how inconsistent the model is and that every run that doesnt show what people want should be binned.

S I don't see it as you having a go-just another person trying to inject some realism into model watching rather than model preference with what anyone may want to see.

Believe me its VERY difficult to keep a personal preference in professional check-so I can understand why posters tend to 'see' what they want to see in any model.

Its not realistic though-objectivity will yield much more consistent results

Not so sure sizzling Heat

I do not tend for bias, in fact, i tend for mild but like to see snow for the benefit of everyone else.

The GFS has really not performed well and even the NOAA (The people whose model it is)have stated how badly it has performed. The GFS has always been a bit over-progressive with breakdowns,

I wonder if someone could post the verification charts up to see how well GFS has done ??

go to the NOAA web site and find their verification tables for 5 and 6 days and for as far back as you like, well this decade anyway.

ECM, short and long term almost always outperforms the others, UK and GFS vie with one another.

Ask our Storm chase team which model is often used by them and the local forecast services?

best I stop quoting or I'll be even more unpopular than now!

Mods feel free to delete if needed, but as a watcher on here for some years I find some posts very, very confusing. I get fed up reading one post saying this then the next saying complete oposite. I know its opinions and its good that some differ but sometimes it makes it so hard for someone like me ond prob others who cant read charts etc very well at all who come in here for information and interest in the weather. I do filter, where poss, who I read in depth though and I am sure these experienced posters know who they are, but some of the stuff that goes on here in time will prob push those particular posters away.

This is a fantastic website and some serious time must be given by a lot of people to keep this up, I just wish there was a little more respect sometimes, from some, towards these people.

Rant over.

I hope in time I can get a grip on models so I can contribute a bit more to discussion.

I suggest you read the forecaster blogs they may help you sort the wheat from the chaff or even the technical discussion.

perhaps try the blogs first, we try to write them such that they interest all but do explain in basic terms, wherever possible, any technicalities and attempt to produce as unbiased a forecast as we can.

honest!

I Promise

last one from me this evening.

Edited January 7, 2010 by johnholmes

[Thundery wintry showers]

Re. SizzlingHeat's post, I think it was perhaps too quick to tar many of us with one brush. Quite often you do see many people cherry-picking one particular model because it suits their own preferences, be they cold or mild. But in this case many people are suspicious of the GFS because it has performed so much worse than the ECM recently. In essence, for GFS to be right and ECM to be wrong, it would require the model accuracy trends of the last three weeks to be binned.

I do remember the time around 10 December when the ECMWF churned out a few runs which overdid the northern arm of the jet and restricted our cold spell to a brief 2 day NE'ly at around T+144 to T+168 followed by high pressure, and this shows that the ECM gets it wrong sometimes. But, over the last month or so, it has got it wrong far less often than the GFS and to a lesser extent UKMO.

[HighPressure]

I don't post much in this thread as I am no expert with models but have read it avidly over the past few years. It appears to me that the GFS never feels comfortable outside a predominately Atlantic theme, be it summer or winter it nearly always seems to want the status quo back sooner rather than later? Whereas the ECM seems to cope better in the opposite situation? If GFS keep calling for a breakdown at +144 it will eventually be right, I do wonder whether the GFS has ever coped with such an intrenched block successfully?

I have a bias for ECM outside Atlantic influence and GFS when we are in it.

Just my thoughts sorry if its a bit of topic.

[phil nw.]

Looking at the nearterm there`s little doubt that we continue with the v.cold temps.and an increasing N.Easterly wind.

http://www.meteociel...h=60&carte=2000

It will feel bitter across the snowfields as renewed cold uppers move across from the East over the weekend.

http://www.meteociel...h=60&carte=1007

Increasingly snowy ,especially across the South and Central UK.

http://www.meteociel...h=72&carte=1594

Of course Percip. amounts are not certain at this stage but in this setup,strong cold flow over a warmer n.sea a further covering of snow is likely in the areas mentioned.

The Scandi block gradually orientates a little NW-SE by T144hrs. with the flow turning towards S.Easterly and falling lighter.

This is where GFS is keen to push a low close to the South West against the High pushing milder air into the South.

http://www.wetterzen...s/Rtavn1441.png

UKMO and ECM pushing the low under the block and extending short waves into Europe maintaining the cold east flow across the UK..

http://www.wetterzen...cs/Recm1441.gif

http://www.wetterzen...cs/Rukm1441.gif

Interestingly GFS Ens. at T144hrs shows the Op. run an outlier for lower air pressure which indicate to me that it brings the low too close to the UK. at this point.

http://charts.netwea...arwickshire.png

Edited January 7, 2010 by phil n.warks.

[mrmonopoly]

I have been reading with a lot of intrest and all compeling arguments.

What i would like to know if possible lets just say the adlantic does win, does that mean back to normal for the rest of the season as its only early feb still a good 6weeks left of winter yet.

What do you think the chances of another similar setup occuring in feb because thats forecast a below average CET also.

Personally being in wales I hate the Adlantic weather, damp cold misserable rain constantly. Least with snow and ice it feels fresh, looks nice, and its something away from the norm. Hell the adlantic destroys our summers so its like 340 days of the same thing differant day.

[sebastiaan1973]

temperatures for De Bilt based 12 hr run ECM

8: -3/-8

9: -1/-4

10: 0/-1

11: -2/-4

12: -3/-5

13: -4/-9

14: -5/-9

15: -1/-7

16: +1/-1

17: -3/-4

[Luke Best]

Hi

I understand that the 'front' denoted in the attached map may actually be a dry line rather than a front. Given that the temperature of the air masses on either side of a dry line is usually fairly similar (and it is the undercutting of the moister air which causes convection), can anybody let me know what effect the dry line might have on my area (Bournemouth) in this set-up. I'd be interested in understanding more about this as the comments I've heard thus far have suggested a warm front that will bring slightly milder air bringing precipitation of a more transient nature to the extreme South Coast. Thanks in advance for your thoughts on this.

Luke.

[TEITS]

12z ECM looking much more like a Reload rather than a Slow Thaw.

I agree and in some respects I actually find the model output scary as I shall explain.

Probably the most severe cold weather synoptics to hit the UK is when the PV sits underneath a HP system. This is why Jan 87 remains one of the most severe cold spells albeit a short one.

Now obviously this isn't be shown on the models at the moment. However I do think this could be a possibility and it wouldn't require much change in the synoptics to bring that cold pool to our shores.

http://www.wetterzentrale.de/pics/Rjma1921.gif

The only way the UK can see upper temps nudging -20C is via this synoptic evolution. Lets hope for our sakes it doesn't happen because it will make our current cold spell seem tame. Certainly one to watch in future model output as i've been saying for a few days now.

[nick sussex]

An interesting difference of opinion tonight on the ECM De Bilt ensembles for the Netherlands with the control run looking like bringing in a very cold easterly later on with the operational run less cold but still cold in any other winter.

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

[Polar Continental]

Evening everyone,

Well there is at least some agreement from the models for the weekend, The GFS has slowly come in line to what has been projected for some time from the ECM and the UKMO for that matter.

A tight squeeze of bitter easterlies from southern Scotland down to the south coast. Most of the heaviest drifting snow will be over inland areas of eastern England, though quite moderate falls are likely to be driven well inland to the midlands and central southern England, I wouldn’t rule out light some snow over western areas also.

This set-up has similarity to January 12th 1987, though the 850’s are not as cold unfortunately, so the snow blitz which occurred then, doesn’t look like happening this time, also the surface temperatures won’t be quite has low. more in the region of –1c to 0c, rather than the ridiculously lows of –8c during 1987 storm.

Beyond that the ECM holds out the cold spell to the end of its run, similar to the 0z GEM. Though I doubt we will see the extremely low temperatures, which many areas have seen recently, as we begin to run out of very cold upper’s.

Paul

[PubliusEnigma]

Hi

I understand that the 'front' denoted in the attached map may actually be a dry line rather than a front. Given that the temperature of the air masses on either side of a dry line is usually fairly similar (and it is the undercutting of the moister air which causes convection), can anybody let me know what effect the dry line might have on my area (Bournemouth) in this set-up. I'd be interested in understanding more about this as the comments I've heard thus far have suggested a warm front that will bring slightly milder air bringing precipitation of a more transient nature to the extreme South Coast. Thanks in advance for your thoughts on this.

Luke.

Hi Luke Id say its not so much a dry line as an upper level warm front. meaning that the upper section of the warm front is well..warm but surface temps are unaffected to a greater

extent. So if the ground level is 0oC or better -1oC or lower snow or possible freezing rain will fall, if the surface levels are at 1oC or more then mixing of the air occurs and

temps rise a little through the heights to you would see rain or wet snow. You will only be able to tell on the day. If you are at -1oC ( quite possible considering current temps) you are like to see snow especially if the wind direction there is ENE, ESE would be off the solent, and that could spell disaster with the temps a degree up. An easterly would be..well nail biting. Good luck

[nick sussex]

Broad agreement from the models for the general set-up for the middle of next week.

However the progression from the GFS is most unlikely; the Low to the SW would simply not be able to make that movement NEwards with HP where it is.

I think it more likely that aproaching systems will be intially deflected away from the British Isles, perhaps skirting the far South of Ireland and the far SW. Then it's a matter of where the next pulse of jet energy goes, if it goes into the Northern arm then I would expect a gradual retreat East of the block as pressure rises from the South, if the energy goes South then it will just reinforce the block and so the cold will go on.

That's a good point, the gfs did something even more bizarre than usual! what do you think of those ECM ensembles, could the control run be doing what Dave mentioned regarding the PV dropping down to our east and much colder air being advected westwards, if the PV splits this must be a possibility.