Model Output Discussion 13/11/13 >

[Timbo]

Yeah, don't recall any outages on previous upgrades and presumably they've been running the new version in parallel for a while anyway.

 

They may be already be using the new data in there current model, and that could be why there outlook is a lot different to the other models!!

[shedhead]

That's unless we just end up with a 'cloudy High' God forbid!

 

 

Cool dank days and mild nights

Indeed, always a danger if the main core of the HP sets up as shown above, but tbf that chart is for the middle of next week. I was talking about frosts later this week, which still look likely for many away from exposed eastern coasts, but by next week cloud may well start to predominate IF the high eases W. 

Edited November 18, 2013 by shedhead

[MR EXTREMES]

well ive been looking at the models and I really strongly disagree with the ukmo and frosty I think todays models run is a lot better than the last few days and that's just not one model but all.

[Frosty.]

wow I just come on here think I have a look at the models thinking more doom and gloom wow what a shock fantastic the ecm heights bloody hell now that's a block if that goes in the right place 09/10 rerun im not counting my chickens though im gonna wait a couple of days to see if this firms up but very nice indeed gfs 6z nice ecm nice ukmo getting there but I wait to see if the ukmo goes with the others if not then I think it will be the model closest to the right solution.

I agree the ecm is fantastic, and I have said as much earlier but it has zero support which is disappointing..however, it's still 13 days until the meteorological winter begins and the models are constantly churning out different ideas. Edited November 18, 2013 by Frosty.

[SMU]

there is NO point, ZERO, nada, using the varification statistics for the NH to disprove the GFS is at these times cannon fodder- because it is- especially the 06z & 18z.

 

Why are people using the NH when the UK is a tiny fraction of that statistic.

 

Sadly i cannot get the regional data anymore, however whent he page was available it divided the NH into around 10 sections, Europe, pole, atlantic, asia, pacific & so on

 

The GFS 06z SMOOTHED its weaknesses out in other areas not associated with the UK because they were probably easier to model, however when you take the 3 areas most important for the UK in terms of cold & HLB's, atlantic, europe & Pole the GFS 06z was MILES behind on varification stats.

 

Im sick & tied of writing this every year.

 

the 00z & 12z faired better but he model bias of the GFS always showed through in at least one of the 3 focus areas above ^^^

Sadly there is no regional statistics available for ECM + UKMO. - but just on the last 3 years performance they are FAR superior, again the GFS will catch them up on NH because some areas are easy to model so the model accuracy is skewwed better by those easy areas.

 

can we put this to bed now.

Edited November 18, 2013 by Steve Murr

[Mucka]

there is NO point, ZERO, nada, using the varification statistics for the NH to disprove the GFS is at these times cannon fodder- because it is- especially the 06z & 18z.

 

Why are people using the NH when the UK is a tiny fraction of that statistic.

 

Sadly i cannot get the regional data anymore, however whent he page was available it divided the NH into around 10 sections, Europe, pole, atlantic, asia, pacific & so on

 

The GFS 06z SMOOTHED its weaknesses out in other areas not associated with the UK because they were probably easier to model, however when you take the 3 areas most important for the UK in terms of cold & HLB's, atlantic, europe & Pole the GFS 06z was MILES behind on varification stats.

 

Im sick & tied of writing this every year.

 

the 00z & 12z faired better but he model bias of the GFS always showed through in at least one of the 3 focus areas above ^^^

Sadly there is no regional statistics available for ECM + UKMO. - but just on the last 3 years performance they are FAR superior, again the GFS will catch them up on NH because some areas are easy to model so the model accuracy is skewwed better by those easy areas.

 

can we put this to bed now.

Not really. 

 

I happen to agree with you regarding GFS bias although I don't think it is cannon fodder, I have seen it beat down ECM in similar situations before even if ECM tends to be better in blocking situations overall. I'm not sure but GFS may have been more accurate regarding the current cold snap with its flatter pattern.

 

The problem is that calling that bias is based purely on observation (subjective to some degree) and while you say GFS was miles behind in the important zones with verification stats, you also say there are no such statistics for the Euros. (I assume you were saying the 06z was miles behind the 00z and 12z - all GFS) so there is no way to prove statistically that ECM or UKMO are much better than GFS 00z/12z in forecasting our weather though we can probably assume they are better than the 06z.

 

Still, at least there is/was statistical proof that the GFS 06z was not a run to put our faith in.

 

 

Edited November 18, 2013 by Mucka

[mountain shadow]

Agreed Steve.

After circa 8 years on here and using our own experiences and knowledge of GFS over that timeframe, it is undeniable fact this model is inferior to UKMET and ECM in UK weather modelling, particularly in potential winter blocking scenarios.

Anyone saying otherwise has not been around the past ten years.

[Snowy L]

Both Hirlam and Bracknell FAX charts supporting ECMWF, where they both link up our high and the developing one over Greenland, also JMA obviosuly did more than support the ECM yesterday with that 12z run. Not sure where the idea of no support for ECM has come from? UKMO is almost there, once again GFS is the one firmly on its own.

Edited November 18, 2013 by Snowy L

[snowmadchrisuk]

Models showing different solutions regarding 98L in the Atlantic , With Gfs even wanting to turn it back West. I imagine how this system behaves will make a difference to the setup we end up with in week 2 ..

 

[shedhead]

Steve...without question you thoughts are well documented, however can I just make this one simple point to you. 

there is NO point, ZERO, nada, using the varification statistics for the NH to disprove the GFS is at these times cannon fodder- because it is- especially the 06z & 18z.

 

Why are people using the NH when the UK is a tiny fraction of that statistic.

 

Sadly i cannot get the regional data anymore, however whent he page was available it divided the NH into around 10 sections, Europe, pole, atlantic, asia, pacific & so on

 

The GFS 06z SMOOTHED its weaknesses out in other areas not associated with the UK because they were probably easier to model, however when you take the 3 areas most important for the UK in terms of cold & HLB's, atlantic, europe & Pole the GFS 06z was MILES behind on varification stats.

 

Im sick & tied of writing this every year.

 

the 00z & 12z faired better but he model bias of the GFS always showed through in at least one of the 3 focus areas above ^^^

Sadly there is no regional statistics available for ECM + UKMO. - but just on the last 3 years performance they are FAR superior, again the GFS will catch them up on NH because some areas are easy to model so the model accuracy is skewwed better by those easy areas.

 

can we put this to bed now.

 

Not everyone is going to read every post you make and fully understand them, especially the rather technical ones.  Therefore it can be confusing to see you lambasting the GFS one moment, then as with this morning 'at least appearing' to support what it shows a short time later....including stuff like 'smelling the coffee' won't always resonate with everyone.

[Barry12]

Both Hirlam and Bracknell FAX charts supporting ECMWF, where they both link up our high and the developing one over Greenland, also JMA obviosuly did more than support the ECM yesterday with that 12z run. Not sure where the idea of no support for ECM has come from? UKMO is almost there, once again GFS is the one firmly on its own.

 

Don't forget the BOM!

 

[SomeLikeItHot]

Ian has tweeted the met still going for a return to average temps

 

they (met office) still go with return to average temps into 10-15d period albeit uncertain timings

 

So MOGREPS must still be going for a return to something more normal

 

Its worth noting that as much as I like ensemble models, that at this longer range they do tend towards the normal the  further out you go. Eg check the Met office update from 16th of Feb this year (First post in the Met office 16 to 30 day outlook Forum).

 

Its states regarding March  this year  "towards the middle of March temperatures may recover close to the seasonal norm.". If anything today's update is more pesimistic about that chances then it was then today states "there are signs of a slight recovery nearer to normal by the middle of December."

 

So not trying to come down on the met but you've got to read it in the context of increasing uncertainty always pulling us towards normal.

 

Edit: Pessimistic is with regard cold being good. From Gavin's point of view I should have said optimistic!

Edited November 18, 2013 by SomeLikeItHot

[Tamara]

The first amplified pattern of the winter in our part of the NH has already drawn a lot of scrutiny due to the uncertainties of the split flow around the atlantic High pressure system and hence the inflection and track of the mid week feature. We see it has taken till at least this morning to get some kind of agreement on this alone - with the GFS, arguably, slowest to get the final energy percentages correct. That said, all the models have struggled with this and there is some credence to the fact that from a longer distance the ECM and occasionally the UKMO slightly over amplified the pattern and the position of the High for the midweek/latter part of the week is slightly further south than suggested some 6 to 8 days since

 

Uncertainty remains into the weekend and beyond as to how sharply aligned the next buckle in the jet stream is and whether we see the atlantic High retrogress itself NW wards or whether there is more energy over the top of the High and it is pushed it closer to our Island (the UKMO currently suggests the latter f.e).

 

The better evolutions which the ECM is currently hinting at, would allow a deeper burst of cold air than the current week offers to come southwards from the pole on the eastern flank of the retrogressed High and not only sustain the cold(er) spell but also increase its wintry potential for the UK.

 

We know that the current polar stratospheric profile is favourable towards a colder pattern as wave breaking activity punches into its lower levels in such a way to disrupt the polar vortex to some extent. The modelling of this as often is the case by the tropospheric models, has come up with rather different ideas about the transport and movement of these parcels of disturbed cold vortex energy and hence that further muddies the picture in terms of amouts of jet energies that will determine the amplification of the jet stream and hence the eventual position and latitude of the High. No surprise that this provides the different model solutions we find as discussed above.

 

A lot of this will seem obvious to many seasoned posters, but there may be quite a few who read these pages and don't understand what is going on, and thus a simple overview is helpful. And I only try and do simple anyway!

 

The fate of the pattern, as such, rests on how these wave breaking sequences play out and sustain themselves. The core of the polar vortex at the highest upper levels is cold and stable - and as we move into actual winter proper, then much as Chiono said I think yesterday, the current disturbances at the troposphere/lower stratosphere level (if sustainable) are going to have a much harder task at disrupting the vortex as a whole without a more significant warming of some kind that penetrates deeper into its core and effectively splits it. At least for a while before it reforms - as inevitably for sure it will.

 

I think in my humble little opinion that the ECM model is mirroring the stratospheric developments the best of the main models at the moment, but we have to bear in mind that some of the most promising stratospheric forecasts at the moment are just that. Forecasts. We depend on these to verfiy in order to not only sustain the upcoming cold spell for as long as possible, but also to provide potential for any future reload. If there is to be one of course. But we also have to take into acount the fact that the ECM model has been prone to over amplifying the pattern further out - and this point has been well made earlier on the thread

 

My guess would be that the latest known MOGREPS signal is a correct one for when the current vortex disruption pauses (at least) and the pattern inevitably flattens out again as a consequence with the atlantic returning at least for a while. But in this sense, the coming days are very important in terms of how forthcoming forecasts quantify vortex disruption and also future angular momentum (measuring jet energy and pressure system strength in very simplistic terms) as these (and obviously more besides) will determine the longer term outlook as winter proper starts.

 

As stated, the atmosphere is going to have to work hard against what looks to be a stronger and less yielding vortex than last year when we consider that a complete destruction of the vortex is much less likely once we get past the festive season for technical reasons best left in the, um, technical thread(s). All a long way ahead obviously, and beyond the remit of this thread, but it nevertheless puts into context the importance of how long the current colder patterns can sustain, and perhaps more importantly whether it can and will recyle itself.

Edited November 18, 2013 by Tamara תָּמָר

[Frosty.]

Mogreps has delivered a punch to the stomach as far as the further outlook is concerned, temps returning nearer to average according to Met but in the shorter term, becoming settled in the north with fog and frost, unsettled and rather cold to the southeast with wintry showers and night frosts but eventually the high to the north drifts south so it's like flipping the coin, the north turning unsettled and atlantic driven whereas the south then becomes settled with fog and frost but temps returning to normal towards mid dec.

Edited November 18, 2013 by Frosty.

[nick sussex]

I see the GFS is causing yet more drama in here! oh well nice to know in an ever changing world some things remain constant! The ECM refuses to back down on its trend and it would have to be wrong at T96hrs as that's when we see model divergence between it and more especially the GFS.

 

Generally speaking for newbies all models have bad runs and the odd rogue output but its exceptionally rare for the ECM to continue on a wrong path for days, in the time I've been viewing it in here I can only remember two times when it was wrong at T96hrs and that's going back along way! The GFS I'm afraid has been wrong countless times at that timeframe and will not budge for run after run.

 

If you're looking upstream then its unlikely the high will sink or that milder conditions will have a chance especially as the pattern in the USA is expected to set up east Pacific ridge, amplified eastern NOAM troughing

 

Some of the complications we're seeing is how exactly that amplified wave interacts with low pressure already in the Atlantic, as this happens the models have to decide where the energy goes ne or se, generally though as a more amplified low heads east it helps to sharpen the pattern ahead of it, its very unlikely you'll see a spill east of energy at worst more a ne so with high pressure near the UK this generally would help this ridge ne, if the upstream pattern is sufficiently amplified then the energy should spill se under the high.

 

We have seen in the past the models bias with energy,they generally always prefer to send this ne until nearer the time. There is no middle ground in these types of situations, you either get energy over the top and high pressure forced further south or energy underneath and the block reinforced.

 

Because the ECM has stuck to its key pre 120hrs trend and done so for several days then I would be shocked if it hadn't called this correctly, especially as the foundation is laid as early as T96hrs.

 

We'll see this evening, hopefully!!!

[radiohead]

there is NO point, ZERO, nada, using the varification statistics for the NH to disprove the GFS is at these times cannon fodder- because it is- especially the 06z & 18z.

 

Why are people using the NH when the UK is a tiny fraction of that statistic.

 

Sadly i cannot get the regional data anymore, however whent he page was available it divided the NH into around 10 sections, Europe, pole, atlantic, asia, pacific & so on

 

The GFS 06z SMOOTHED its weaknesses out in other areas not associated with the UK because they were probably easier to model, however when you take the 3 areas most important for the UK in terms of cold & HLB's, atlantic, europe & Pole the GFS 06z was MILES behind on varification stats.

 

Im sick & tied of writing this every year.

 

the 00z & 12z faired better but he model bias of the GFS always showed through in at least one of the 3 focus areas above ^^^

Sadly there is no regional statistics available for ECM + UKMO. - but just on the last 3 years performance they are FAR superior, again the GFS will catch them up on NH because some areas are easy to model so the model accuracy is skewwed better by those easy areas.

 

can we put this to bed now.

 

This topic often crops up in US weather forums too, often around hurricane season. There is a thread on AmericanWx.com where someone who works for the NCEP wrote that the only difference between the the 0Z and 06Z runs is that the 06Z gets assimilated with the latest observations since 0Z. If the 06Z ran without these new observations, then the output would be indentical to the 0Z run. It's still the same computers doing the same calculations with the same formulas afterall. The only differences in the number of oberservations between the runs is the aircraft observations, which themselves are only a small part of the total observations that go into the model. On average the 18Z has the highest number of aircraft observations but that doesn't seem to lead to the 18Z being significantly more accurate than the others.

 

It's a shame we don't have regional verification stats. Though I think in this case it is changes on the wider NH scale which lead to cold getting to the UK in the first place.

[MPG]

Its quite basic Shedhead, the model seems to take time or lack behind the UKMET and ECM. Steve has made that very clear and there is no confusion. So 'smelling the coffee' is an easily understandable term to describe that the GFS is on catch up. 

 

Edited November 18, 2013 by MPG
last sentence removed.

[Team Jo]

Hello! Could we please avoid the point scoring and nit picking and digging of one another on this glorious Monday?

 

Thank you Kindly....

[terrier]

Mogreps has delivered a punch to the stomach as far as the further outlook is concerned, temps returning nearer to average according to Met but in the shorter term, becoming settled in the north with fog and frost, unsettled and rather cold to the southeast with wintry showers and night frosts but eventually the high to the north drifts south so it's like flipping the coin, the north turning unsettled and atlantic driven whereas the south then becomes settled with fog and frost but temps returning to normal towards mid dec.

Very true frosty a real kick to the stomach.But hopefully jan will bring us better news. Fingers crossed

[Tim Bland]

Whilst I would love the ECM to be correct here, Im afraid with the Metoffice (and all the info at their disposal) and MOGREPS not backing it, leaves me thinking it will backtrak in the next 24hrs. Hope it proves me wrong though!

[fergieweather]

Mogreps has delivered a punch to the stomach as far as the further outlook is concerned, temps returning nearer to average according to Met but in the shorter term, becoming settled in the north with fog and frost, unsettled and rather cold to the southeast with wintry showers and night frosts but eventually the high to the north drifts south so it's like flipping the coin, the north turning unsettled and atlantic driven whereas the south then becomes settled with fog and frost but temps returning to normal towards mid dec.

To clarify: MOGREPS merely veers the E/SE faster than EC. There's no suggestion of 'mild', merely return to average. Moreover, UKMO analysis is focused on next 10d (to 15 at a push). The story into December is another matter.

[MR EXTREMES]

I would not read to much into the ukmo update seriously although a very good model im not so sure its a punch in the stomach until the issues are resovled.

[MPG]

Whilst I would love the ECM to be correct here, Im afraid with the Metoffice (and all the info at their disposal) and MOGREPS not backing it, leaves me thinking it will backtrak in the next 24hrs. Hope it proves me wrong though!

EC ensembles, 36 out of 51 go for pressure rises to the NW by next weekend.

[Tim Bland]

To clarify: MOGREPS merely veers the E/SE faster than EC. There's no suggestion of 'mild', merely return to average. Moreover, UKMO analysis is focused on next 10d (to 15 at a push). The story into December is another matter.

 Stope teasing us Ian  Whats the story into December !?!?!?

[Frosty.]

To clarify: MOGREPS merely veers the E/SE faster than EC. There's no suggestion of 'mild', merely return to average. Moreover, UKMO analysis is focused on next 10d (to 15 at a push). The story into December is another matter.

I guess that average will feel relatively mild, the point I'm trying to make is the met office in nov 2010 had an emphatically high confidence of repeated cold blasts from the north and east but this time there is nothing remotely like that indicated, I'm just keeping it real rather than giving false hope. :-)