Mike Poole

Hey, this is beyond satire now GFS! T210: Earth shattering run...

Pub run T168, just awesome: Snowfest: Just need to clear that low at T72! Stella run

GFS pub run special, NH vortex sliced like butter, T120: I repeat, T120!

GFS 18z T72 compared with 12z T78. Look at the two lows highlighted, the southern one (black) is deeper, the northern one (red) shallower, this should weaken the system over the UK:

Well just as the pub run starts trickling, a word on the background signals, AAM is forecast by CFS to stay just above average for the foreseeable. The MJO looks about to go haywire in phases 6 and 7 according to GEFS - one member has gone off the chart, I hope it will be OK... But we know the driver for amplification is there, we are seeing it all over the modelling, aren’t we....while we are producing pages and pages of discussion about the behaviour of one small low near Ireland. The perils of being a weather enthusiast in a small island off the Atlantic.

ICON 18z T90 v 12z T96: The low west of Ireland is shallower on this run and ‘closer’ in a sense I’m not sure what I mean (!) to the one to the SSW, so hopefully it won’t go the way of the 12z run.

My concern with the ensembles (all of them, but particularly GEM/CMC) is that the current dilemma seems to be troubling the highest resolution op runs at short range, so the lower res ensembles could be collectively leading us up the garden path. Anyway, raise a glass to you all for the pub run

Maybe, but I dont rate the GEM ensembles at all, they can be totally different to the GEFS which causes a problem for the NAEFS as it uses both. The GEM op has performed very well since its upgrade a year or two ago. And has consistently predicted cold for UK - a decent pattern, apart from, obviously, the awful run this evening. But I’d forgive it one run if the trend was right.

Well on the clusters at that key T120-T168 time window, there are two, one where the low sticks around in some form in the vicinity of the UK, and one where it doesn’t. 33-18 in favour of the low sticking: This is so important for the UK, it isn’t really worth looking any further until this is resolved.

Must be true then Edit, but seriously it probably is coming from the east or northeast, it is just for a small island like ours the details do matter and that is why the conversation we’ve just had over the last 10 pages or so matters.

It isn’t as clear cut as that, depends what is causing the snow, if it is frontal snow, then yes that might be true near the coast, but if it is convective snow off an easterly, you’d ideally want the sea as warm as possible and the upper (T850) temperature as cold as possible to generate convection.

Certainly is Nick! I was not expecting that the run would suddenly dump the connection with the arctic high at T168 and build heights over Greenland and even west of it, instead. So it looks like a completely new evolution, and may therefore fall by the wayside in due course, but it is interesting. At least ECM is consistent in being inconsistent!

ECM T192, and now it has gone so far from what we’ve been seeing from the other models, need to check the T850s too: Marginal for the south only, with plenty waiting in the wings!

Weird, I tend to use my brain!

ECM T168...be honest, who saw that!

It probably will now, as it shouldn’t phase with the upstream one.

Absolutely, touch and go! T144 and it is still not certain, looks messy though:

ECM T120, I think we might be OK just, the weak ridge should push the low away slowly now. Won’t know for sure til the T144...on it goes.

ECM T96 compared to UKMO: Edit, what we need to see on the next one is a gap opening to join the high, its on UKMO already, here illustrated where it isn’t on the ECM 0z T120: (12 hr time difference).

GFS // T120: This looks to be clearing the low without any issues, and good alignment for advection of cold air from the NE with arctic high and atlantic high engaging nicely. Good run this will be, I think.

Yes, I’ve noticed that the means have tended to do this - suggests to me that the higher resolution of the ops runs is going to be needed to resolve this properly at such a short range. I would view all the op runs as a super-ensemble here and for the moment the GEM/ECM 0z solution seems to be a bit of an outlier, but that could still change over the next day or so.

It is vile, worse than 0z ECM, here T138: Thing is, that envelope of uncertainty I spoke of earlier obviously does include this evolution, we just hope it doesn’t happen. This may be the one bad model this set of runs, and the ECM may show better, just the way the cookie crumbles.

GFS T198, what a time to get the link with the arctic high, excellent alignment and pressure not too high over UK. Cold in place and on tap to reload. What’s not to like?

UKMO, GFS and ICON at T144: I don’t see this as models flipping between runs, it is more there is an envelope of uncertainty that we have in our minds, and what changes things is when a model shows something outside of it - like the ECM 0z, for me. All runs so far on the 12z suite are within the uncertainty envelope I have in my mind, so that is all good. The UKMO is not as good as this morning but it is still fine, the GFS looks good and so does the ICON. On to the GEM, GFS // and ECM.

Your question was originally about the GFS // i think? Well the same T0 data initialisation issues apply, no change there. But the vertical resolution increase would likely give it an advantage during a SSW as that’s where the important things are occurring. So, at the moment, I would definitely rate the // over the GFS op run. But come spring and summer, I expect it will be equally good/bad as the op as the strat doesn’t impact our weather then.