Mid-Winter: Forecast and Model Discussion commentary

[WYorksWeather]

 Mike Poole I think there's also different types of easterly as well which people consider within the broad idea. You have a light near-continental easterly, say from Germany or southern Scandinavia. That is the least cold option. You then have a long-draw easterly from eastern Europe - Poland, Ukraine, western Russia, etc. That is generally much colder. The final option and by far the rarest is more of a long-fetch easterly to north-easterly from the heart of Scandinavia or even Siberia. All three of those are technically an easterly, but they range from fairly underwhelming to absolutely freezing cold BFTE-type scenarios. And then of course you have the surface conditions to consider - are the 850hPa cold enough, dew points, etc., all that much closer to the time.

I would say that overall we're in a position where there is a 50-50 chance at days 8-10 of a pattern that is broadly conducive to a cold easterly. If that then verified, we would have to work out the detail of whether the coldest air is likely to make it to the UK.

[knocker]

 Mike Poole I am well aware that the contours are not an anomaly as I mentioned in my first post. I have drawn up a few charts in my time. 

Edited January 29 by knocker

[blizzard81]

We also have to bear in mind that this is a highly unusual situation for the models to digest and process. Normal accuracy rates at day's 6/ 7 go out of the window imo in such circumstances. 

[Mike Poole]

2 hours ago, Met4Cast said:

Because it’s a known model bias with both the GFS & ECM, although less so with the ECM at least out to day 10,  no idea what the biases are beyond that as it’s only recently been extended.

So many really good thought provoking posts in here tonight .

I’ve a few posts in mind on various topics as a result, in this one, I want to discuss model biases in terms of an amplified or otherwise pattern.  A few years ago I would have said that the GFS has a moderate bias in the direction of a flatter zonal pattern on the 10 day timescale, and the ECM had a moderate bias towards over-amplification on a similar timescale.

So if there was a model standoff between the two of them, a decent bet was to go for a 67% ECM 33% GFS blend as a likely middle of the road solution.  The GFS had gained something of a reputation for having an advantage for Greenland blocking, and I think this was partly because if the GFS was going for it, it was going for it against its own bias, and therefore leant some credibility to the scenario happening. 

Unfortunately, the model biases are no longer so clear cut, to my mind.  In particular, the GFS - since its FV3 ‘upgrade’ a few years ago - seems to now have a bias just towards being wrong, but not in any particular way!  And the ECM’s bias is also less predictable.  Many times recently the GFS has gone for solutions that prove to have been over-amplified.  So it is more difficult to use known biases to pick the correct solution.

Edited January 29 by Mike Poole

[WYorksWeather]

 Mike Poole We also have other models in the mix as well - the ECM AI being a key one, plus also GEM seems to have improved in recent years and possibly surpassed the GFS for accuracy.

[Mike Poole]

 knocker I know you know that - and I’m not suggesting otherwise.  But I was also replying to @Jacob who had asked for more information in response to your post. 

[Wedgesmakesledges]

Im keeping an eye on Day 7 > Each run of the models are sharpening up the ridges & troughs - so much so that the chart below could see the low becoming a cut off trigger low. 

We would need a more vertical profile west of Scandi for that to happen.

[blizzard81]

Such is my total lack of faith in the JMA, I'm actually ecstatic that it is showing this in the extended.

[Mike Poole]

 blizzard81 I’d take the Korean model at T234 though:

Let’s see what the pub run can come up with!

[blizzard81]

 Mike Poole Lol....Indeed Mike

[Mcconnor8]

 Mike Poole We're overdue a GFS pub run special, if not tonight then maybe tomorrow we will get one

[WYorksWeather]

 Mcconnor8 Still way too early to tell about the second push of heights, but so far it looks poor at day 5 compared to the 12z. The Atlantic is much closer, the whole pattern is shifted further south and east. Not great.

Still chance for the second push to redeem the run, of course. Arguably that is the real period of interest anyway.

[Severe Siberian icy blast]

 Mcconnor8 yes i agree, i want a good old fashioned drunk gfs 18z, dare to dream? I think so! 

[blizzard81]

 WYorksWeather Don't ever underestimate the pubrun's propensity to pull a rabbit out of the hat just when a couple of frames earlier it seemed like all was lost

Edited January 29 by blizzard81

[Scandinavian High.]

The Atlantic as completely been  shut down doesn’t look like anything is coming through the massive wall of high-pressure to the north east and east of UK ..

[Welsh_Weather]

 Met4Cast I really don’t understand the term ‘model bias’. It’s bounded around in here way too often. There are tendencies for models to overdo certain scenarios but they are still only picking up on what is fed into them and how they deal with it based on history. 
 

if we are dealing with biases then the background drivers, i.e. the MJO, which you used last year to predict a very cold period was completely overridden by a strong PV, was completely wrong. There is very little to suggest that the PV won’t override anything this year and the MJO backs that up. However, what’s to to say that the PV won’t keep shunting west to allow the block to build? 
 

I’m not here to swipe at you because I feel you a very good amateur forecaster, but you do keep putting down models and posters if they don’t agree with your theories, which aren’t always right. 

[WYorksWeather]

 blizzard81 Atlantic is blocked off but there's no clear route to get the proper cold in. At no stage do we ever get the cold 850s in, and by day 10 it's like night and day compared to the 12z.

Certainly not going to throw in the towel on the basis of one pub run! But this run does illustrate the danger if a Scandinavian High attempt fails - you end up with a fairly non-descript outcome that fails to achieve anything notably cold and whittles away valuable time. It would be relatively dry and seasonal of course, so could be a lot worse.

 

 

 

[nick sussex]

The GFS highlights the issue with this set up. The main block is centred too far east and the UK is at the flabby weaker end with no low pressure over northern Italy .

It’s a struggle to get the deeper cold west . 

At the moment a convective easterly looks unlikely and so it’s going to need to be trying to get enough cold in case the Atlantic moves in and you have a more battleground scenario .

 

[feb1991blizzard]

 Scandinavian High. Its not as good as 12z but your point still holds true, which is why i can't see why people were writing off an SSW, mind you looking at that high on the 12z, i just wish we could have had it in Oct & Nov, the vortex would never have got going at all and we'd have been looking at another 47 or 63.

[WYorksWeather]

As a consolation prize, the pub run has now found a decent polar maritime attack at day 11. Only transient, but might deliver something for the north perhaps?

[Paul]

Locking this one ready for a fresh thread - to start the thread I'm going to split out all of the posts from today..

Thread now up: