Oh come on GFSP... give me a break:

See the massive change at just 5 days range between the 00z (left) and 06z (right)...?!

There's then too much development to our west for a cold outcome within the next 10 days though - the UK storm isn't able to dig south and it's a pretty tragic outcome for England and Wales. Becomes a snowfest for Scotland by day 8.

But really. My sanity is being tested.

I think it has lost the plot but nonethess rather entertaining! 

I'm not an expert at all but unless we (Darlington) get something from the North East or a true Easterley the most we can expect are squally sleety showers.

It takes an age for the cold upper air to make it all the way to the SE though - snow-fest for days in the North though.

It just needs to move 200 miles further SE and we're all happy.

Indications of a strospheric warming now showing in the models

http://users.met.fu-berlin.de/~Aktuell/strat-www/wdiag/ts.php?plot=temps&alert=1&lng=eng

Stonking thriller of a run from 6Z GFSP - amazingly snowy for the north.

Total chaos in model land 

Scotland and parts of Northern England would be under feet of snow.   It might be a bit parky at times down south, of a little wet.

GFS and GFSP could hardly be more different by 10 days out. The colours are virtually negatives!

http://www.wetterzentrale.de/pics/Rtavn2401.gif - Greenland low, Azores high

http://www.wetterzentrale.de/pics/Rpgfs2401.gif - Greenland high, Azores low.

Nobody has a scooby.

All options are still available. It's interesting though to see the possibility of an Easterly appearing in the output of late.  

What's totally bonkers is how a slight change in the angle of the jet in 4 days time causes the U.S. storm to dramatically disrupt with a lot of energy heading under the Greenland High, which then results in a very robust block that just goes on and on with no sign of abating even at day 16.

If anything too much energy is under the block on this run 

In fact I think it would also wreak havoc on the strat. vortex.

Total madness really. The run is pretty much in a territory unexplored by recent GEFS and ECM ensemble output, with the only run coming close being yesterday's 12z JMA op run.

As hard as I look, I can't see a trigger in the 500mb charts for this dramatic change of output so an extremely suspicious run, only worth this many words because of the ridiculous uncertainty levels at play right now.

The GFSP isn't *terrible*. It keeps heights high over Greenland for the hi-res portion of the run at least.

I think the very best I can hope for here (and this would be a near-miracle) is some back edge snow settling, followed by a freeze as HP builds over the UK.

All rather feeble.

(One year I'll get the 'gridlocked UK' festive super-snowstorm. One year...)

Well the models are farcical of late, I think I will stick to the ensemble means beyond Christmas - much more stability and sanity there. If you ignore the wildly swinging op runs there's good consistency in the means at day 9 although certainty is clearly low ATM.

NAEFS 

ECM

GEM

You call those horrendous charts 'sanity'???

I have literally said WTF to my PC.

Where did that come from?

Greenland High back in the mixer.... i need a lie down.

Next weekends LP is being modeled further north and east by GFS 06h and a lot further north by GFSP.The result is lower intensity and less cold event.Still time to change,but trend is towards less severe event.

Post depression GFSP provides a genuine looking GH -if it ever happens like that.!

I would be very wary about using an ensemble 500hPa/SLP mean chart as they will simply average out all the scenarios. If many options exist in the output it will come to some form of halfway house as shown. I guess if they are used with the "spread" charts to show confidence levels they may give more of a helpful indication.

I'm afraid the ensemble means are misleading at the moment, because the individual members have a fairly even spread across a huge range of outcomes. The overall mean usually tends to have the appearance of a zonal setup in these situations, giving a false impression of low uncertainty and inter-run consistency.

I think this is due to the fact that runs with LP to the north tend to have deeper storms with negative 500 mb height departures (from a 'neutral' height field, about 550dam I suppose) far larger than the positive height departures achieved on runs with high pressure in that area, due to the effect of colder air (lowers heights). The situation is reversed at lower latitudes, where the low heights with low pressure are countered or overridden by the high values in high pressure.

This is just a bit of theoretical thinking on my part based, so If anyone knows of a reason why this is incorrect, please don't hesitate to correct me 

3 or 4 bitter Easterlies in the mix as well...    all to play for.

Added:  I'll take perturbation 1 please....     snow heaven 

They're not to be taken literally of course especially with high divergence but they do show the balance of probability of a particular weather type, and as you said to be used in conjunction with the spreads and other data available.