Model output discussion - is the beast awakening?

[Methuselah]

From a quick perusal of the GEFS/GFS ens, I'll be taking the 06Z Operational with a few thousand lorry-loads of finest rock salt!

[kold weather]

14 minutes ago, nick sussex said:

I’d have been less concerned if the GFS was being the grinch . Not sure anyone knows how this is going to turn out . It’s quite bizarre because normally when the op starts losing interest the ensembles start moving quickly . The ECM has had now 4 poorish runs so you would have expected the ensembles to be fleeing the sinking ship by now . When I looked at today’s ECM op I was quite surprised to see its ensembles .

I'm really torn about this Nick.

1: The ECM does have a bias towards taking too long to shift these sorts of upper lows out of the way, indeed the 00z never exactly succeds as it takes so long it ends up getting absorbed by the LP to the west. Thats nearing worst case scenario levels. Not quite, but getting close.

2: However the OP runs of most of the models have made a definate nod of the hat towards the ECM between 96-120hrs. On its own as long as it stops there, its no major problem. BUT we've seen countless times that once the models start making a nod, it rarely ends wih just one run and normally becomes a trend that extends and strengthens within in the suite.

The other worry is the models have done something like this earlier in the winter and advertised a cold spell, but in the end what happened.

A limpet low, just like the ECM is showing...

[Froze were the Days]

First look at the models (ECM & GFS 0z/06z) and oh dear!...ECM op still not having it! and that's a warning sign (4 op runs in a row) and GFS slowly trending away from any prolonged cold spell though okay towards central and northern areas for a snowy spell (quelle surprise).

Down where I live...'Goodnight Vienna' but that's been the tale of winter 2020/21.

[Anthony Burden]

As a very well respected person commented on a different weather site,percentage wise follow 

the majority of ensembles not the operation if tracks away from the majority.Not 100% but more

favourable.

[WeatherWatchmanG]

Metcheck has snow down for next weekend in London but the BBC has it as dry I know what I'd rather have it down as but is there a reason the two forecasts are different from the models I assume it's automatic and two different models being used don't agree on the location of the high along with the cold air too. Metcheck shows -5c bbc shows +6c.

[ALL ABOARD]

I didn't think so much of it at first but surely the issue of flight data is playing a massive part. 

I know it's difficult to forecast snow but the met have been way off on timings, amounts, how far north how far south...

New warnings gone out for Wales today as its already snowing heavily, same for the updates ones for Tuesday. 

Maybe the models are struggling majorly because of the lack of flights? 

[Pennine Ten Foot Drifts]

So, not wanting to be Mr Grumpy, but having been on here model watching for 16 (!) years:

FI = anything beyond T180, and what Operational models are showing is the slight favourite amongst a wide range of possible outcomes. no point viewing these for any local detail, or really even with any great confidence.

T144 - T180 = most likely broad synoptic pattern i.e. likely to be warmer / colder / wetter / drier than average, but in terms of local detail, barely more reliable than FI.

T72 - T144 = broad synoptic pattern likely to be accurate (but not 'nailed on'!). Detail starting to be defined, but for a small area such as the UK, in a predominantly maritime influenced region, confidence in accuracy of detail is still not high. And it mustn't be forgotten that what are actually relatively minor changes to the broader pattern can still have significant changes for UK localities, so even at this range, for 'my location' confidence in just how much warmer / colder / wetter / drier than average is not going to be very high at all.

T36 - T72 = confidence in broad synoptic pattern is high, confidence in local forecast is moderate to high dependent on the volatility of the forecast pattern i.e. a forecast large high pressure can give pretty high confidence of at least dry weather locally (although, for example, cloud amounts even at this range cannot be forecast with absolute certainty), whereas what we're looking at now i.e. highly volatile synoptics, means even at this range confidence of local weather conditions is no better than moderate (and in this case probably lower than that).

T12 - T36 = broad synoptic pattern highly likely to have been resolved, fairly high confidence in local detail, but again, as for T36 - 72, dependent on the type of synoptics being forecast, changes can happen.

T0 - T12 = radar watching!

So, for the upcoming 10 days, trying to predict where snow may fall next weekend is at this juncture pretty futile, (and even now, it's quite possible the synoptics will change such that nowhere outside the usual northern upland suspects see any at all). I'd say, assuming the currently modelled synoptics play out broadly as forecast, the soonest we'll be able to confidently start anticipating widespread lowland snow will be Wednesday.

[Frosty.]

Oh my..I’ve been drinking..yeah I’m a souse..weekends only!..I had something too say..hiccup..but I’ve forgotten what it was..?...anyway..perhaps these ECM 0z ens mean charts speak for themselves?...hiccup..hurrah!

Edited January 31, 2021 by Jon Snow

[northwestsnow]

Just now, WeatherWatchmanG said:

Metcheck has snow down for next weekend in London but the BBC has it as dry I know what I'd rather have it down as but is there a reason the two forecasts are different from the models I assume it's automatic and two different models being used don't agree on the location of the high along with the cold air too. Metcheck shows -5c bbc shows +6c.

Beeb will use EC det.

I doubt very much they are using UKMO model if they are projecting 6 degrees.

[jacke]

3 hours ago, That ECM said:

@nick sussex do one of your drawings please for newbies as this must be frustrating to view. However a good learning setup to illustrate why this feature is crucial. Im down the beach on my phone so can’t. 
 

lots of seagulls down here but there is always

I trust my gulls as much as the charts at the mo!

[Chasbrown]

1 hour ago, Lampostwatcher said:

It's only a short drive up north

During lock down?! Doh! 

[MJB]

That's some chart from The Control

[MJB]

I'd settle for that I think lol

[WeatherWatchmanG]

As long as it's clean of snow down south on the between 5th and 9th, these models are looking favourable for my needs. I do wonder if anyone has collected data with the models when one week, two weeks and so on ahead how accurate they get them during different seasons.

[turkishfella]

Is it me, am I not seeing something others are seeing or the models are not very nice anymore? They were bringing very cold air from the east to whole UK on Friday’s output but now nothing major for south

[Wivenswold]

12 minutes ago, Pennine Ten Foot Drifts said:

So, not wanting to be Mr Grumpy, but having been on here model watching for 16 (!) years:

FI = anything beyond T180, and what Operational models are showing is the slight favourite amongst a wide range of possible outcomes. no point viewing these for any local detail, or really even with any great confidence.

T144 - T180 = most likely broad synoptic pattern i.e. likely to be warmer / colder / wetter / drier than average, but in terms of local detail, barely more reliable than FI.

T72 - T144 = broad synoptic pattern likely to be accurate (but not 'nailed on'!). Detail starting to be defined, but for a small area such as the UK, in a predominantly maritime influenced region, confidence in accuracy of detail is still not high. And it mustn't be forgotten that what are actually relatively minor changes to the broader pattern can still have significant changes for UK localities, so even at this range, for 'my location' confidence in just how much warmer / colder / wetter / drier than average is not going to be very high at all.

T36 - T72 = confidence in broad synoptic pattern is high, confidence in local forecast is moderate to high dependent on the volatility of the forecast pattern i.e. a forecast large high pressure can give pretty high confidence of at least dry weather locally (although, for example, cloud amounts even at this range cannot be forecast with absolute certainty), whereas what we're looking at now i.e. highly volatile synoptics, means even at this range confidence of local weather conditions is no better than moderate (and in this case probably lower than that).

T12 - T36 = broad synoptic pattern highly likely to have been resolved, fairly high confidence in local detail, but again, as for T36 - 72, dependent on the type of synoptics being forecast, changes can happen.

T0 - T12 = radar watching!

So, for the upcoming 10 days, trying to predict where snow may fall next weekend is at this juncture pretty futile, (and even now, it's quite possible the synoptics will change such that nowhere outside the usual northern upland suspects see any at all). I'd say, assuming the currently modelled synoptics play out broadly as forecast, the soonest we'll be able to confidently start anticipating widespread lowland snow will be Wednesday.

Could we pin this to the top of every page on the MOD thread and leave it there forever? It's so much easier to refer those hyperventilating over permutation 34 at day 9 to this invaluable advice rather than explain how it all works.....again.

[WeatherWatchmanG]

2 minutes ago, turkishfella said:

Is it me, am I not seeing something others are seeing or the models are not very nice anymore? They were bringing very cold air from the east to whole UK on Friday’s output but now nothing major for south

nice? I see that as a nice outcome if nothing major for the south. As long as models a week out are reliable.

[Gustywind]

Was it Jan 2013 when the UKMO was rock solid for a wedge & subsequent snow event, while the ECM refused to come on board with the rest? Think it went down to literally T+72 before it finally caved in.

Just my opinion but the ECM seems to have been all over the place recently, so not as concerned it’s right this time as I would be normally. 
 

If UKMO & ECM end up meeting in the middle is that pretty much the GFS solution?

[jacke]

53 minutes ago, IDO said:

Yes, we have gone from the models relatively persuasive solid gold cold-synoptic to a watered down version and now a cold spell dependent on micro features behaving! 

That seems the modus operandi of models; post-d10 we get clean, almost perfect cold charts. Then at around d7-8 we get macro changes that downgrade the potential, then at d5 we get the spoilers that can scupper the whole cold spell. It is a masochistic pastime for sure!

I DO get your para above framed! That is the essence of computer programming. Limited data being built on as the time gets nearer. Surely for chart watchers the fun is guessing what the new data will influence? Don't panic going on a full count later, at this point confident in colder input by next weekend for much of the east.!

[JamesL]

And while we all ponder on where we will be in 7 days, I write this with heavy settling snow falling in S Wales 

a nice surprise of which there will be a lot of over the next 5-7 days countrywide! 
 

Bring on the heavy hitters of the 12z’s! 

[turkishfella]

5 minutes ago, WeatherWatchmanG said:

nice? I see that as a nice outcome if nothing major for the south. As long as models a week out are reliable.

Snow is what I want, that’s why i said the models are not nice anymore!

[snowy36]

1 minute ago, JamesL said:

And while we all ponder on where we will be in 7 days, I write this with heavy settling snow falling in S Wales 

a nice surprise of which there will be a lot of over the next 5-7 days countrywide! 
 

Bring on the heavy hitters of the 12z’s! 

The computers are struggling 7 hours away never mind 7 days.

[Cold Winter Night]

1 hour ago, Vikos said:

ECM d10 vs d3

    

This is an excellent illustration of the poor modelling by the EC Operational lately.

The Eastern Hemisphere and the Arctic were modelled completely different on day 10 from what it makes of it on day 3.
Not even that far from being the exact opposite, I'd say.
That does not inspire much confidence in what it is showing today at those FI timeframes.

Below I've added the GFS at 240h for that same 0z on January 24th. Much better match!

The past few days EC has been consistent only at finding all the things that can go wrong, but different things every time.
It's been at the very poorest end of its ensembles for several days now, which is an odd thing. After day 5/6 it diverges from the ensemble mean, which shows sensitivity to starting conditions.

Really folks, watch FI charts (120h and later) just for entertainment, not for reliable information.

[Cloud 10]

There are lots of descriptive  words for the model output at the moment but "boring" certainly isn't one of them.

GFS brings blizzard conditions to parts of the Midlands and Northern England next weekend,with over half a metre of snow in quite a few favoured locations by day 10.

 

[Snowman.]

15 minutes ago, Gustywind said:

Was it Jan 2013 when the UKMO was rock solid for a wedge & subsequent snow event, while the ECM refused to come on board with the rest? Think it went down to literally T+72 before it finally caved in.

Just my opinion but the ECM seems to have been all over the place recently, so not as concerned it’s right this time as I would be normally. 
 

If UKMO & ECM end up meeting in the middle is that pretty much the GFS solution?

It was at T72 when the GFS and to a lesser extent the ECM were wrong and the UKMO beat them all. 
 

After the UKMO killed the Dec 2012 easterly that year as it never brought into it and trumped the ECM again in Jan is why I personally along with many others view it in such high regard. It’s the decider model. 
 

I expect 12z downgrades like the 6z but the ECM to have something to show. We may get a delayed Easterly but I wouldn’t be surprised if it’s a lower thickness one.