Model output discussion - 5th December 2016 - Into Winter we go

[Seasonality]

@Steve Murr haven't quoted your post to avoid excessive page scrolling. You've obviously put a lot of time and thought into this. You seem to be suggesting zonal winds are in themselves a sort of magic bullet for forecasting winter weather. Can it really be as simple as this?

[Catacol]

4 minutes ago, tcc said:

Interesting stuff Steve and thanks for your explanation as to your "mood" yesterday!

Question I have, and this is to anyone really, is why has it all gone Pete Tong? Why are we not on the brink of a decent cold spell?

If it's been answered elsewhere then forgive me :o)

Steve has provided plenty of detail on his opinion of this above. If you want another angle on it I think people got very excited towards the end of November and saw too much cake and honey on the horizon. There were plenty of voices who did not join in with the "early cold" rallying cry, but when the forecast of a front loaded winter came out I think a number of people thought it was nearly nailed on.

The alternative view to Steve's above would be that a stratospheric / tropospheric disconnect is quite possible, and that the tropospheric vortex continues to be disrupted. The ENSO disconnect in terms of the GWO phase and forecasted MJO progression would see amplification as returning. I for one never really saw any proper winter until mid December onwards, though the crazily mild air of this week was not really expected - I had thought the mid lat block would hold on longer.... but that wasn't to be as it sank SE under falling AAM conditions. However we are on the cusp of the next phase as mid month approaches, and we will see whether pacific forcing combined with a rather perturbed vortex at low levels can produce the goods.

No guarantee EVER of snow,  but the idea that winter is pretty much over now until mid January because the models have taken a turn for the worse, and the winds at 10hpa are forecast to increase to 40m/s + is a brave call. Other signals suggest a different outcome.

As ever - only time will tell. The only thing that is certain is that the Snow-watch will continue on here!

[SMU]

21 minutes ago, Seasonality said:

@Steve Murr haven't quoted your post to avoid excessive page scrolling. You've obviously put a lot of time and thought into this. You seem to be suggesting zonal winds are in themselves a sort of magic bullet for forecasting winter weather. Can it really be as simple as this?

Not totally however what it dows is give us good context of where the starting point is ( good / bad or ugly ) & what way the trending is going ....

[Seasonality]

Just now, Catacol said:

Steve has provided plenty of detail on his opinion of this above. If you want another angle on it I think people got very excited towards the end of November and saw too much cake and honey on the horizon. There were plenty of voices who did not join in with the "early cold" rallying cry, but when the forecast of a front loaded winter came out I think a number of people thought it was nearly nailed on.

The alternative view to Steve's above would be that a stratospheric / tropospheric disconnect is quite possible, and that the tropospheric vortex continues to be disrupted. The ENSO disconnect in terms of the GWO phase and forecasted MJO progression would see amplification as returning. I for one never really saw any proper winter until mid December onwards, though the crazily mild air of this week was not really expected - I had thought the mid lat block would hold on longer.... but that wasn't to be as it sank SE under falling AAM conditions. However we are on the cusp of the next phase as mid month approaches, and we will see whether pacific forcing combined with a rather perturbed vortex at low levels can produce the goods.

No guarantee EVER of snow,  but the idea that winter is pretty much over now until mid January because the models have taken a turn for the worse, and the winds at 10hpa are forecast to increase to 40m/s + is a brave call. Other signals suggest a different outcome.

As ever - only time will tell. The only thing that is certain is that the Snow-watch will continue on here!

Thanks @Catacol. 

@Steve Murr the sentence in bold is what I was referring to in my earlier post.

[SMU]

Just to add - I dont want to add any specifics to any forecast for late Dec / Jan because thats not what its really about- suffice to say with a low amplitude wave forecast & combined with the fact to deliver snow a scandi high needs a higher peak ( circa 75 maybe 80N ) then the favoured route for any snow would be ditch the scandi high for an atlantic high - or hold both (rare) as the atlantic high can have a lower peak amplitude but still deliver cold air to the Uk that can produce snow -

So in a nutshell - No GH or artic high or svalbard high to effect the UK ( HLB ) 

we need the MLB to locate favourably ...

[SMU]

18 minutes ago, Seasonality said:

Thanks @Catacol. 

@Steve Murr the sentence in bold is what I was referring to in my earlier post.

You can never say never- however Its of no surprise that all the peak days that I have seen in the ERA data are strongly positive AO months & just as important all the date record months ive seen were strongly negative AO months-

& whilst there isnt a 1:1 correlation for AO mode > UK CET - the higher up you go in anomaly the closer the match-

45 -50 M/S would be my max ceiling - past that we are in Dec 15 territory!

S

[bluearmy]

re the MJO - it was always my understanding that predictions which were low amp (ie. within the circle of doom or just outside) were unlikely to affect our background pattern. I cant see any decent amplitude to current model output

[Singularity]

Catacol makes very important points (as usual). When it comes down to it the 10hPa zonal wind is only a reliable predictor when the stratospheric vortex is coherent right down to the troposphere.

When you have a disconnect and mechanisms for disrupting the lower levels still further (bearing in mind GFS neg AAM bias causing it to suppress some important forcing), the 10 hPa situation is off little consequence.

Our troubles next week are down to movement of the lower vortex independent of the upper, a catalyst for which is tropical forcing, which itself is subject to change (hence vortex position over N America could still shift notably). 

Late Nov, those low down troubles weren't apparent due to overlooked Indian Ocean tropical interference, so with the 10 hPa vortex unlikely to propagate down for many weeks, a cold Dec was odds on.

Below average overall is still achievable in my eyes, mind.

[Singularity]

15 minutes ago, bluearmy said:

re the MJO - it was always my understanding that predictions which were low amp (ie. within the circle of doom or just outside) were unlikely to affect our background pattern. I cant see any decent amplitude to current model output

If I understand it right - the plots are less useful than usual when you have dual centres of convection. It might even be that the two signals 'average out' with the current plotted proximity to phase 2/3 over 5/6 reflecting the dominance of the former (which is expected to give way to the latter... some day!).

IR satellite shows plenty of convection ongoing.

[Mucka]

For me it really doesn't matter what we put most stock in because every teleconnection is interdependent on another and there is no way to know which is the chicken and which is the egg - that is to say we have no start point where the atmosphere was set and all the starting data correct and that is even assuming we have a handle on all the variables that affect our weather (which we don't)

This is why it is impossible for even super computers to model the weather with any degree of worthwhile accuracy beyond a few days or in a predictable pattern beyond say 10 days.

So in this sense the more deep our understanding and the more stock we put in teleconnections the further out in time we are forecasting because we are stretching out causality and thus more prone to errors our forecasts become - not because those thoughts are misplaced and teleconnections/interactions unscientific, but because they are interdependent and one error in one calculation (which becomes more inevitable the more teleconnections we introduce - the further out in time we go) means several errors in the next.

This compounding of errors works in the  same way as the model output. It only requires a tiny miscalculation by 72 hours to have a huge error by 192 hours.

In that sense less is more in forecasting accuracy, the shorter range we deal with the less teleconnections we need consider and the less any errors will be in our calculations but that doesn't mean we should give up trying to forecast at longer ranges and just deal with the output out to 5 days. - unless we believe we should have give up trying to fly or get to the moon.

We know there is good scientific theory behind many of these long range background signals but none is a silver bullet to long-range forecasting because we cannot know which will be the dominant driver from whatever starting condition we choose, unless we only deal with shorter time frames, because we cannot know exactly how we arrived at that starting condition - we are using a static snapshot to describe something in a constant state of flux - our data is wrong and outdated before we start.  

This is easily proved as well. Simply run the models backward with our starting data be it longrange models or shorter range and we will quickly see they can no more tell us the history of our weather than predict the future.

So the brave (the frontiersmen and women of meteorology) will continue to break new ground and beat a path but also be wrong - often. The medium range forecasts will follow behind them, a little clearer and safer with a path ahead, but unsure of their destination. And the shortrange forecasters will follow once the signposts are up and the streets paved.

That may be a little unfair on shortrange forecasters because responsibility also increases as we get to shorter ranges and the devil is in the detail as they say. I guess the moral of the story is, if you want to be a weather forecaster in the public domain, wear a bullet proof vest.

 

Edited December 7, 2016 by Mucka

[Johnp]

Steve  - those zonal wind forecasts are generated from the GFS. So surely these forecasts are just as volatile as the 500hpa charts etc?

 

[bluearmy]

3 minutes ago, Johnp said:

Steve  - those zonal wind forecasts are generated from the GFS. So surely these forecasts are just as volatile as the 500hpa charts etc?

 

Modelling high up is far more reliable - it can go wrong as per the gefs late November but generally it's much better than in the trop (I think Andrej showed the op forecast to be twice as reliable at day 10)

extrapolate that for the ens ......

 

[Catacol]

20 minutes ago, bluearmy said:

re the MJO - it was always my understanding that predictions which were low amp (ie. within the circle of doom or just outside) were unlikely to affect our background pattern. I cant see any decent amplitude to current model output

True - current forecast is low amplitude. But when one considers how radically the MJO forecast has changed in the last few days / week then this is not a concern at present. I think the trend is key, and if I understand those who have a much better grip than I do on the projected view of the MJO going forward I think they are seeing increased amplification on the horizon. @Singularity may well be correct above in the "average out" effect. Wind speeds in the troposphere are not a concern at present. We'd all prefer to see the vortex shattered from top to bottom... and who knows: maybe Cohen's theory of snow cover feedbacks will help overcome the WQBO factor and give us a SSW somewhere in the heart of winter... but for now I'm not too concerned with conditions at 10hpa. We have a trop vortex prediction that remains favourable for ridging with vortex centres in the Canadian and Siberian sectors, a continued lack of any dynamism to atlantic weather systems and pacific teleconnections suggesting amplification.

If these carry any weight then expect to see the mid lat heights next week move to higher latitudes as we approach the final third of the month. Not much point saying more than this yet because it goes without saying that we need the block orientated correctly to get a feed suitable for snow. That's the luck factor that is always required for snow in our part of the world... If we don't get the pacific amplification then chilly, dry and frosty under a mid lat high is as good as it will get this side of Christmas. Eyes down on those MJO forecasts and overall GWO phase.

[Snowy L]

19 minutes ago, bluearmy said:

Modelling high up is far more reliable - it can go wrong as per the gefs late November but generally it's much better than in the trop (I think Andrej showed the op forecast to be twice as reliable at day 10)

extrapolate that for the ens ......

 

Yeah this myth has to go.I think we're stuck in 2012/13 where everything went right in the strat forecasts. In the following 3 winters and the beginning of this one I've lost count of how many times a forecast warming at 240 hours has completely disappeared the following day.

[bluearmy]

23 minutes ago, Snowy L said:

Yeah this myth has to go.I think we're stuck in 2012/13 where everything went right in the strat forecasts. In the following 3 winters and the beginning of this one I've lost count of how many times a forecast warming at 240 hours has completely disappeared the following day.

Can I advise you to look at heights rather than temps then. You will find them far more reliable and the Berlin ones (ex ecm op) are usually pretty good. 

I have screenshots saved and whenever I check to see how accurate they are I am nearly always impressed

Edited December 7, 2016 by bluearmy

[Gael_Force]

30 minutes ago, Snowy L said:

Yeah this myth has to go.I think we're stuck in 2012/13 where everything went right in the strat forecasts. In the following 3 winters and the beginning of this one I've lost count of how many times a forecast warming at 240 hours has completely disappeared the following day.

There's also a myth that it's all about temperature ... it's not. The geopotential heights at the different levels are more useful for shorter term patterns, to gauge whether there is a disconnect and whether the top down/bottom up is more influential.

[booferking]

No wonder the jet is firing up with these temps predicted over the states...

[SMU]

17 minutes ago, Gael_Force said:

There's also a myth that it's all about temperature ... it's not. The geopotential heights at the different levels are more useful for shorter term patterns, to gauge whether there is a disconnect and whether the top down/bottom up is more influential.

& the wave heights in terms of lattitude-

Edited December 7, 2016 by Steve Murr

[Timmytour]

41 minutes ago, Snowy L said:

Yeah this myth has to go.I think we're stuck in 2012/13 where everything went right in the strat forecasts. In the following 3 winters and the beginning of this one I've lost count of how many times a forecast warming at 240 hours has completely disappeared the following day.

I'd gladly take another 2012/13.  My memory is that it began after THAT Ecm and then a huge let down after so much was "promised" by the models.  Lots of toys were thrown out of the pram I seem to recall with winter being written off by a good few. And then began a series of cold spells that gave me some great snow memories at least!  

[Seasonality]

While the 12z rolls out, Happy Christmas from the CFS!! 

 

 

[Methuselah]

1 minute ago, Seasonality said:

While the 12z rolls out, Happy Christmas from the CFS!! 

 

 

The CFS comes to the rescue, just like The Lone Ranger...Hi Ho, Silver!

[Weathizard]

There's definitely some interest at the end of the GFS ensembles, I know it's difficult to get excited about but it's worth watching, even at this range. 

 

London ENS after 18th/20th are intriguing, at this range it's just a trend but good to see amongst the ensembles, you can't write December off just yet as some appear to be doing.

 

 

[Frosty.]

2 minutes ago, Ed Stone said:

The CFS comes to the rescue, just like The Lone Ranger...Hi Ho, Silver!

Not forgetting P11 & P16 GEFS 6z

[Wayneywoo]

6 minutes ago, Weathizard said:

There's definitely some interest at the end of the GFS ensembles, I know it's difficult to get excited about but it's worth watching, even at this range. 

 

London ENS after 18th/20th are intriguing, at this range it's just a trend but good to see amongst the ensembles, you can't write December off just yet as some appear to be doing.

 

 

The Op is also at the higher end of the Ensembles from the 12th to 21st!

Edited December 7, 2016 by Wayneywoo

[Tim Bland]

2 minutes ago, Wayneywoo said:

The Op is also at the higher end of the Ensembles from the 12th to 21st!

Usually I think there is more confidence in the OP if it is followed by the control and in this run it is very closely followed so perhaps a likely evolution.  CFS run for Xmas is not impossible and shows how we can still get cold & snow in the UK with a slight Atlantic Ridge and a strong PV over Greenland, this is our best bet for snow this side of New year IMO.