There's 3 problems with the models as I see it:

1. Inevitably some processes aren't modelled perfectly. This aspect is being improved, eg with GFSP as compared to GFS.


2. The models represent the atmosphere as a grid in three dimensions. That means that there will inevitably be small-scale features falling between the grid points that aren't picked up, and such features can develop to become significant. This is gradually being improved as more powerful computers make it possible to have more grid points, closer together.


3. Linked to #2 is the fact that the atmosphere is a chaotic system, which means that any minor errors are going to tend to snowball as time goes by. You might need to be able to individually model every molecule in the atmosphere to avoid that, so however good the models become we are told that we are never likely to get perfect accuracy beyond about 5-7 days.

Taking all that into account, I think the models do remarkably well. I'm old enough to remember the days before computer models were used for forecasts, when 48 hours was about the limit for anything fairly reliable. The models do even better if you don't focus too much on the operational runs, but concentrate on what the ensembles indicate. Apart from the milder "bump" that has recently appeared in the 850 mb temperatures for next Friday/Saturday, the GFS and ECM ensembles have been remarkably consistent for a good many days now in showing a mean 850 mb temp of about -5 persisting for most of the next couple of weeks.

Terrific post JHall,  and I agree. 

About 10 years back, I treated myself to an expensive but snazzy looking traditional wall barometer. Must admit I rarely look at it, but can you imagine what an important piece of equipment this must have been to the generations before ours before the onset of TV/Radio forecasts and later, NWP outputs? And I wonder, was the 'humble' barometer any less accurate back then than even the hi-res model outputs are for today for any one location? 

 

Yet more interest in FI. Before that it looks like Tuesday will be a bust for lowland areas. All in all very disappointing from this Bullseye Winter

Hang on a minute , wasn't this forecast as a backloaded winter.?

All the action was meant to be after any SSW 

And the reason it’s called Fantasy Island!  How many times do we hear that and those other famous words ‘more runs needed’?  Yet some people take it as read!  It’s the MOD thread. Discuss the models but don’t take it to heart. 

I have personally found this thread more interesting than the actual weather this winter, even if it’s been a case of ‘look at what you could have won’.  It just goes to show how very close we sometimes come to getting the weather we crave.  

 I think many, including me, will find it hard to disagree with this Caz 

My eyesight seems to be going really downhill this last year, and as a result, I often read 'more runs needed' as 'more nuns needed'. Or perhaps this could be just my brain just trying to tell me something 

Snow for many on that run 

Excellent run - the chase straight away begins 

Another cold NWly approaches

 

Hang on a minute , wasn't this forecast as a backloaded winter.?

All the action was meant to be after any SSW 

Yes, most forecasts went for a colder second half of winter and I suspect that is almost certain (not just because the bar is not exactly set high!) Having said that the forecasts at the beginning did not/could not have taken the SSW into account. 

Things still look cold and not just in FI. The problem is it may be cold but possibly not “cold enough” during any precipitation. I think hindsight may be needed for this one as well as a bit of nowcasting.

I would expect quite a bit of cold rain, some wintry or snow showers and some severe frosts in parts of the north especially. Frustratingly what we may end up with is a persistent cold outlook (end of the week blip excepted) without anything very snowy or prolonged deep cold. I do however think now that we are in a generally cold regime there will be more opportunities (as well as more disappointments). 

Snow for many on that run 

The U.K. view is impressive: -7c 850s and 500-1,000 hPa thickness around 520 dam. Definitely cold enough for snow to low levels even in the south

you wont have to wait that long Peter about 6c  drop behind that cold front

 

The U.K. view is impressive: -7c 850s and 500-1,000 hPa thickness around 520 dam. Definitely cold enough for snow to low levels even in the south

That is a very tight and potentially volatile looking thermal gradient between NFL and GL at 240:

http://www.wetterzentrale.de/maps/ECMOPEU00_240_2.png

you wont have to wait that long Peter about 6c  drop behind that cold front

 

Problem is that even if the pattern proves to be right , it will be heavily modified come the time ,with any significant snow for the usual spots in a cold north westerly. They are rubbish for most of low land southern and Eastern England

indeed Zub but I do think we may see a little of the white stuff down here from that, anyway set looks cold at the surface 

http://www.meteociel.fr/modeles/gefs_table.php?x=0&y=0&lat=51.5085&lon=-0.12574&run=12&ext=fr&mode=7&sort=0

 

Problem is that even if the pattern proves to be right , it will be heavily modified come the time ,with any significant snow for the usual spots in a cold north westerly. They are rubbish for most of low land southern and Eastern England

That is a very tight and potentially volatile looking thermal gradient between NFL and GL at 240:

http://www.wetterzentrale.de/maps/ECMOPEU00_240_2.png

 

 

 

The extreme cold coming off the eastern seaboard will fire up the atlantic and probably flatten the pattern

 

Hang on a minute , wasn't this forecast as a backloaded winter.?

All the action was meant to be after any SSW 

Yes it was, glad someone remembers.  Nothing has changed including the need to cling to fantasy charts out in FI.  Get 8 hours of one page post, then a little change and all hell is breaking loose.  It's still only January, loads to play for.  

Get your heating fixed Geoff hope you have been cold today without it in a nice way

 

The extreme cold coming off the eastern seaboard will fire up the atlantic and probably flatten the pattern

That is a very tight and potentially volatile looking thermal gradient between NFL and GL at 240:

http://www.wetterzentrale.de/maps/ECMOPEU00_240_2.png

 

 

That's the 0Z run. The 12Z isn't that different, and looks like this. I like the suggestion of WAA up the west coast of Greenland.

 

Problem is that even if the pattern proves to be right , it will be heavily modified come the time ,with any significant snow for the usual spots in a cold north westerly. They are rubbish for most of low land southern and Eastern England

It looks like a very marked cold front rather than just shower activity, so I don't think what you say necessarily applies in this case.

Looks better here uppers about-7

 

That's the 0Z run. The 12Z isn't that different, and looks like this. I like the suggestion of WAA up the west coast of Greenland.

it is but zub is right on his comment, white stuff on back edge

http://www.meteociel.fr/modeles/gefs_table.php?x=0&y=0&lat=51.5085&lon=-0.12574&run=12&ext=fr&mode=3&sort=0

 

It looks like a very marked cold front rather than just shower activity, so I don't think what you say necessarily applies in this case.

Looks better here uppers about-7

 

I was following up the point the OP made about the sharp temperature gradient just off the coast of Newfoundland, when I said it wasn't that different to the 0Z, rather than considering our own neck of the woods.

 

The extreme cold coming off the eastern seaboard will fire up the atlantic and probably flatten the pattern

Yes, the ECM is the sort of sequence that downgrades both in upper temps and synoptics nearer the time, with the Azores High ending up playing much more of a role than progged at this stage.

I'm afraid after the Northern blocking failure of the NWP we are now left with the crumbs of this.