Mike Poole

Forgetting the nonsense in the recent posts about Liam Dutton - which now seem to have been removed - edit! (who I don't rate after his provably wrong comments about the GFS 6 and 18 runs), I think we will see a move to the ECM tonight with a quicker move to Greenland around T192 bringing in snow opportunities for significant parts of the country thereafter. That's my punt anyway!

Just comparing ECM 12z and GFS 12z, there looks to be a significant difference at around T192 where the ECM has pulled the High through towards Greenland, GFS much slower: As others have alluded to, and we saw from the later charts, this means in the case of ECM lower pressure and disturbances thereafter leading to potentially significant snow. As ECM is the best performing model, I wonder if this slight potential change of direction might be repeated in future runs. Maybe something to look out for on the pub run!

You are so right - from a psychology point of view I reckon the winter models chase hits the same highs and lows in the brain as gambling - particularly something like roulette where you can go on big wining and losing steaks with rapid changes of fortunes. I should know - I do both!

ECM T240, snakes alive! Seriously good run this.

Fair cop, too big fingers too small iPad! T192 great though, back to direct hit territory

Hmmm, spicy

ECM at 168 looks good

One question on this, I'm still trying to try to understand the reasons why this block sets up where it does - is the Hannah Attard plot that you quote not just itself an model output from a single GFS run? So how does it give predictability for future runs, or are you saying that as this is at 100hPa the model will have a better handle on this at 7 days than it does for the surface weather?

Let's say the chance of getting a massive high-latitude block like currently modeled in a year are 20%, 1 in 5 (only for illustration), this is probably a requirement for all three things MIA quoted: beast from the east, channel low and polar vortex low. This means if you take into account the correlation and not count this probability three times, you'd need to reduce the odds of the combination by about a factor 5*5 = 25.

OK, sorry, my bad, you have noted the issue - easy to read this thread too fast when the run is rolling - however, in my opinion this correction would no where near account for the correlation as all three things are much more likely in the modeled synoptics, which you can only take credit for once, and therefore you need to correct for a triple counting error of a significant part of the probability which would be greater than a factor 5, in my opinion.

This is not appropriate and misleading. These are not independent variables, and will show some significant correlation - based in particular on the underlying synoptic pattern of high latitude blocking, so it is NOT appropriate to multiply them together. The true odds of these things happening are much shorter than the 2250/1 you quote.

I laid a square egg when viewing the pub run! Really stung too. Surely this run can't come off, it's bonkers! I think the more significant development was the ECM getting on the same page. Interested in the width of the cold pool at T192, much wider (more room for error) on ECM than GFS:

GFS mayhem at T216,, just WOW

ECM T192 aligning towards direct hit

As always, you've got to read the wording very carefully with these updates, but for me this, and the 6-15 day update, looks like an upgrade for snow potential in the south of the UK.

My worries after last night's pub run pale into insignificance if this one verifies. If this run came off I'm worried I might have kittens. Even though I'm male and not a cat. It's concerning because I'm not sure I'm competent to look after them. GFS at T198

Retrogression as early as T144 on this one, wow

Take that back about too far north, sorry, looking good at T132

Yes, that's the worry with this one, it does look very north at T102, offsets the ECM anyway!

18z rolling now - what do we want, I mean a repeat of last night's pub run would be unlikely, I'd suggest, trousers still damp from that one!! I'd take a run that doesn't follow the GFS 0z, or the ECM 12z model output. Anything else is fine.

Hi Big Snow, The high pressure is first building towards the UK and then north as a result of the first SSW, only then does it look like heading to Greenland probably forced by the second warming. Sooner it gets it's skates on the better, for this upcoming cold spell to really count.

Very happy with the ECM ensemble mean But we've seen two op runs that highlight the risks that we face, the GFS 0z taking the High too far north, and the ECM 12z leaving the block too close to the UK so the cold air goes south. And, of course, all the other runs that say Game on! Should we be worried? Not today, let's see what tomorrow's runs bring.

We'll I don't but it was the first chart from a major model that looks like a wobble, but hey ho, the T240 looks just fine, high looks like it's off to Greenland, maybe a delay to get the cold in though. But different, to the other output today, I think.

Is this a bit of a worry at T216?

With the ECM at T168, it's coming, just not as fast