Model Discussion Thread - 15th Jan

[LomondSnowstorm]

There has to be a reason for the GFS op's stubborness, but I'm just a bit mystified regarding why it has this signal when most other models and most background signals don't favour it at all. It's still an outlier http://cdn.nwstatic.co.uk/ensimages/ens.20120116/00/t850Fife.png

The GFS ensembles are good, not quite as good as last night's but if you recall the 00Zs haven't been anything like as promising as the 12Zs in the last 2 days.

The ECM doesn't look as good either though, which is a major worry http://www.meteociel.fr/modeles/ecmwf/runs/2012011600/ECM1-192.GIF?16-12

I'll stick to my view that the models haven't yet managed to find the right solution and things will resolve themselves favourably for blocking, because that's what the background signals appear to show, but to be honest the output this morning is miles apart from the ensembles of last night.

[nick sussex]

Apparently NOAA in the USA were sending out some recon missions from Hawaii yesterday to gather extra data from the Pacific as they think there were too many starting errors in the model output.

Given the chaos this morning obviously this never happened! As we can see the models all disagree, although the ECM has been consistent with its trend from yesterday, the UKMO has joined in the uncertainty by suggesting a northerly after 144hrs, the GFS continues to churn out nonsense not supported by its ensembles , the GEM thinks of supporting the GFS for a while but realizes thats a bad idea and changes course, the NOGAPS looks a bit of everything with shortwave heading south and then a slack easterly flow and finally the BOM at least keeps high pressure to the east and troughing over the UK later and still makes more sense than the GFS.

Given the teleconnections and where the major support is in the ensembles the following IMO can get themselves to the nearest landfill site.

GFS

GEM

I've left the BOM in as if its going to be a stalemate type solution upto 240hrs then thats the more likely way it would happen, generally the models most likely to verify are the ones who keep a strong block to the east and displace the Azores high further west, whether that brings fun and games we'll have to wait and see as this is dependent on what happens to the troughing over Europe.

I was hoping we'd have a resolution this morning but the models just don't want to know and are determined to drag this on and on and on........

[Cecil Hogmanay Melchett]

Apparently NOAA in the USA were sending out some recon missions from Hawaii yesterday to gather extra data from the Pacific as they think there were too many starting errors in the model output.

Given the chaos this morning obviously this never happened! As we can see the models all disagree, although the ECM has been consistent with its trend from yesterday, the UKMO has joined in the uncertainty by suggesting a northerly after 144hrs, the GFS continues to churn out nonsense not supported by its ensembles , the GEM thinks of supporting the GFS for a while but realizes thats a bad idea and changes course, the NOGAPS looks a bit of everything with shortwave heading south and then a slack easterly flow and finally the BOM at least keeps high pressure to the east and troughing over the UK later and still makes more sense than the GFS.

Given the teleconnections and where the major support is in the ensembles the following IMO can get themselves to the nearest landfill site.

GFS

GEM

I've left the BOM in as if its going to be a stalemate type solution upto 240hrs then thats the more likely way it would happen, generally the models most likely to verify are the ones who keep a strong block to the east and displace the Azores high further west, whether that brings fun and games we'll have to wait and see as this is dependent on what happens to the troughing over Europe.

I was hoping we'd have a resolution this morning but the models just don't want to know and are determined to drag this on and on and on........

Personally nick i think the ecm loses it between 144h and 168h.

The evolution up to 144h looks ok and we can see heights stretchinh across NW russia Scandy towards Iceland then all of a sudden it is blown away.

Im not sure it will happen like that.

[The Northern Ramper]

Apparently NOAA in the USA were sending out some recon missions from Hawaii yesterday to gather extra data from the Pacific as they think there were too many starting errors in the model output.

Given the chaos this morning obviously this never happened! As we can see the models all disagree, although the ECM has been consistent with its trend from yesterday, the UKMO has joined in the uncertainty by suggesting a northerly after 144hrs, the GFS continues to churn out nonsense not supported by its ensembles , the GEM thinks of supporting the GFS for a while but realizes thats a bad idea and changes course, the NOGAPS looks a bit of everything with shortwave heading south and then a slack easterly flow and finally the BOM at least keeps high pressure to the east and troughing over the UK later and still makes more sense than the GFS.

Given the teleconnections and where the major support is in the ensembles the following IMO can get themselves to the nearest landfill site.

GFS

GEM

I've left the BOM in as if its going to be a stalemate type solution upto 240hrs then thats the more likely way it would happen, generally the models most likely to verify are the ones who keep a strong block to the east and displace the Azores high further west, whether that brings fun and games we'll have to wait and see as this is dependent on what happens to the troughing over Europe.

I was hoping we'd have a resolution this morning but the models just don't want to know and are determined to drag this on and on and on........

With the info GP posted last night i am very confident the Gfs is miles out.

[cold snap]

yes nick i was convinced to,infact i done indepth work and was so sure but this morning is shortwave madness.i know we have had this coversation before,but you can acount for shortwaves they just throw a spanner in the works.one thing im sure about this pattern is very complex.more changes today one would guess.edge of seat stuff

[nick sussex]

Personally nick i think the ecm loses it between 144h and 168h.

The evolution up to 144h looks ok and we can see heights stretchinh across NW russia Scandy towards Iceland then all of a sudden it is blown away.

Im not sure it will happen like that.

Who knows in the current Twilight Zone called the model output!

The trend of the ECM remains the same from yesterday,as the PV becomes elongated its likely to eject a mini PV segment towards the UK and its how far east this gets thats uncertain, also how much if any trough disruption occurs.

But if you look at its NH pattern its no different from yesterday evening, troughing near or over the UK, block to the east, Azores high displaced to the west at 240hrs.

All will be revealed at some time in the future, just don't ask me when!

[Jayces]

A question sorry about the models themselves in general, sorry if this is OT for this particular thread.

When people make comments about a particular model going off on one with regards to a certain outcome at a certain time frame and then querying why it would do that or making statements like it's not possible, what is that based on?

I mean surely the models computer modelling software has been written by weather experts who know what will happen under certain circumstances based on various weather factors. Does this not mean that every single model run no matter how bizarre should in theory at least be completely plausible?

Otherwise there must be some serious data errors with a model if it could model a forecast that just can't physically ever happen weather wise?

[nick sussex]

A question sorry about the models themselves in general, sorry if this is OT for this particular thread.

When people make comments about a particular model going off on one with regards to a certain outcome at a certain time frame and then querying why it would do that or making statements like it's not possible, what is that based on?

I mean surely the models computer modelling software has been written by weather experts who know what will happen under certain circumstances based on various weather factors. Does this not mean that every single model run no matter how bizarre should in theory at least be completely plausible?

Otherwise there must be some serious data errors with a model if it could model a forecast that just can't physically ever happen weather wise?

Of course generally the models show solutions that can realistically occur, although you do have problems with the drop down in resolution on the GFS at 192hrs.

Can the GFS verify yes but not with the current teleconnections, if you're building a case for it, theres not much to go on, its ensembles are not really supporting it and the two top models whilst disagreeing at 144hrs in terms of detail agree with the displacement of the Azores high. In fact thats one of the few aspects of the modelling thats more certain than anything else.

if the Azores high displaces into the Atlantic then no the GFS cannot verify, its a synoptic impossibility! as the Azores high ridging north there and retrogressing and block to the east would end up with troughing over Europe.

Given the set up the Azores high displacement is crucial for people wanting to see cold, a flat high is no good, hence its the amplitude of that high that will determine what happens in Europe, the more this displaces and ridges north the more digging of that trough will occur in Europe.

Edited January 16, 2012 by nick sussex

[ANYWEATHER]

What I do see from the models in the next ten days is a continuation of sorts of what we have had for the majority of Winter so far...! Yes, there will be some cold shots at times but milder interludes as well. A mobile weather set-up looks to be the way things are to be, some folks will see some snow, but for the Majority of the Uks population ,deep Winter conditions will "not prevail for the forseeable future. Thats my take on it this morning.

[The Northern Ramper]

Worth noting that the strong message coming out of all NWP output at the moment is for an opening out of the contours (mean deceleration of the westerly winds) across the North Atlantic as the low heights to the north lift out and pull back west, associated with the upstream wavelengths being seriously altered by that Aleutian ridge. The Azores ridge is going to get a serious yank back west allowing the longwave trough and attendant shortwaves to cut across on a NW-SE axis. If you follow the longwave pattern and the ensconsed 850 values, the only way for this to go is underneath the block.

GFS ensemble mean package suggests a strong building ridge over the top of this block by day 10 and a split flow developing over Europe by day 12 (easterly potential high). Of note, the 00z parallel GEFS came up with a strongly negative NAO and the 12z GEFS ensemble mean has popped in a strong ridge developing south of Greenland at the very end (day 15), but a strong signal at that range.

The interesting part of this post for me is the azores being pulled west which the gfs doesnt follow but others do. A less amplified jet, azores west and undercutting lows suggests cold from the east to me.

[chionomaniac]

This GFS ensemble anomaly chart suggests a westwards shift of the Azores high with the PV shifting to Alaskan territory. Svalbard pressure rises too. Not quite an undercutting scenario...yet.

[Gizu007]

A question sorry about the models themselves in general, sorry if this is OT for this particular thread.

When people make comments about a particular model going off on one with regards to a certain outcome at a certain time frame and then querying why it would do that or making statements like it's not possible, what is that based on?

I mean surely the models computer modelling software has been written by weather experts who know what will happen under certain circumstances based on various weather factors. Does this not mean that every single model run no matter how bizarre should in theory at least be completely plausible?

Otherwise there must be some serious data errors with a model if it could model a forecast that just can't physically ever happen weather wise?

There must be serious error in weather forecast models. Today is a good example for that. Current GFS run 00z shows at 0600 minimum temperature of the place I live -2 deg. GFS T800 is 0 deg. Yesterday Met Office stated that today at 0600 will be -1 or -2 deg. Real temperature at 0600 was -5 deg. Met office update data to -5 at around 0300-0400. So if there is error of -3,-4 deg in T+6 so I don't know what sort of percentage confidence we can have in T+144 and more.

[Frosty.]

In my opinion the Ukmo 00z wins 1st prize for wintry potential going into next weekend, closely followed by the Gem and Ecm with the Gfs lagging behind in 4th place, the Gfs shows a milder weekend with temps between 8-10c but the other models keep it colder with wintry showers and overnight frosts. I think going forward, the Gem 00z is the best as it has more of a nw/se angled jet, the ecm is a bit flatter but not mild, there is plenty of rather cold weather in the outlook. The Gfs 00z looks a bit too progressive to me in that it pushes the ridge eastwards more than the other models and then allows milder air into the uk next weekend although it does gradually turn a little colder into the following week and then shows an arctic blast at the end of next week.

So this week starts cold with frosty nights, briefly turning milder on wednesday and then turning colder from the northwest with increasing risk of wintry showers, initially in nw britain and then gradually spreading to other areas.

Edited January 16, 2012 by Frosty039

[Paul]

Temperature differences on the ground are caused by resolution - GFS has a resolutiuon of 0.5 degrees so each grid square of it's data is relatively large (50km approx), which is why it'll not pick out localised low (and high) temperatures at times. All weather models are the same in that respect - the metoffice as you noted showed a similar temperature so their model was looking at similar things, the adjustment was due to them using current conditions to update forecasts at the closest range if there are any major differences.

This sort of variation on the ground is because of topography, towns, effects of the sea etc - once you go up into the atmosphere there isn't any of that to cause the same sort of issue, so the resolution as it is doesn't hurt the forecast of pressure systems, upper air temps etc.

[full_frontal_occlusion]

...........

I mean surely the models computer modelling software has been written by weather experts who know what will happen under certain circumstances based on various weather factors. Does this not mean that every single model run no matter how bizarre should in theory at least be completely plausible?

Otherwise there must be some serious data errors with a model if it could model a forecast that just can't physically ever happen weather wise?

Models are purely mathematical, constructed to obey the laws of physics. However they are an approximation - albeit one which is continuosly improved.

They are not perfect and do not contain every single physical-variable possible nor is their resolution capable of crunching to a scale below which error is eliminated. That feat is a very, very, long way into the future.

As each computed time step is based on the output of the previous computation, the errors will accumulate getting larger at each successive step. Hence why FI is taken with a very large pinch of salt.

The equations used are simultaneous and there are several solutions, hence the need for the ensembles which introduce a random 'seeding' in an attempt to understand the 'spread' of potential solutions which gives a reasonable indication that any given output has the highest probability of success.

All this to say some runs will be more successful than others and hence the observation that sometimes the models 'go off on one'.

Hope this goes some way to answering your question.

ffO.

Edited January 16, 2012 by full_frontal_occlusion

[Duane S.]

Good grief!!!

For the first time since this debacle began, and after looking at all options on the table last night, I went to bed 90% sure we would see a definitive swing to cold from the 00z model suite. Wrong.

The NOGAPS 18z looked on the money (I know, shoot me for saying that) and it sort of looks OK again this morning. The Azores does not seem to be responding to upstream signals, I am still convinced (errr 'ish') in the next 24-48 hours we are going to see much better ridging appearing in the 5 day timeframe, WAA in a SE-NW direction into and around southern Greenland, not creating some monster G/L high pressure but a ridge brought about by increased amplification and leading to low pressure settling around Mid Western Europe and a much colder scenario developing. But then again...

[Isolated Frost]

It will be nip'n'tuck when it comes to SES on Friday to Monday.. the ECM 0z output gives 168 (Monday 0z next week) as the best chance with the atlantic stuck out in Ireland and a slight n/ne flow, very temporarily though. The 12z output most likely to be colder with more SES chances.

[john mac]

06z coming out now and looking better early on t68 high further south and west

[Harsh Climate]

The GFS 00z was one of the biggest mild outliars you will ever see!

[JonMillar]

Model comparison spreadsheets for the last few runs, showing their evolution and attempts to reconcile. Hope they are of interest.

2012-01-20 500hpa + SLP GFS-ECM Comparison.xls

2012-01-21 500hpa + SLP GFS-ECM Comparison.xls

2012-01-22 500hpa + SLP GFS-ECM Comparison.xls

[john mac]

Snow showers in the west and possibly E,EA &SE by friday luchtime on the 6z

[Harsh Climate]

Snow showers in the west and possibly E,EA &SE by friday luchtime on the 6z

I think It would be just wintry showers for low ground to be honest.

Edited January 16, 2012 by Harsh Climate

[TEITS]

Very frustrating at the moment considering the UKMO, GFS, ECM all disagree with each other at +144. I feel plenty more changes are likely to happen based on the last few days.

[KTtom]

Snow showers in the west and possibly E,EA &SE by friday luchtime on the 6z

North of Wales and only over higher ground, temps far to high going by 06z anywhere else. (5 - 7c)

[TEITS]

The difference between the 0Z, 06Z GFS at +138 is laughable with regards to the HP to the NE.

http://www.wetterzentrale.de/pics/Rtavn1381.png