Jump to content
Sign in to follow this  
Paul

Model output discussion - late November

Recommended Posts

19 minutes ago, Ed Stone said:

It reminds me of the days when I'd keep buying newspapers, until I found one with a forecast for snow in it...:oldgrin:

I still buy em all pete just to read those great headlines!!!! Vortex destruction to bring a 6 month maunder minimum style winter this year!!! You can't beat it.... They must be getting there facts from me! 🤣 Anywho.... The 6z ensembles show some rather spiffing members.... Hey its Friday.... I'm happy 😎

graphe3_1000_250_94___.png

Share this post


Link to post
Share on other sites
1 hour ago, Steve Murr said:

I cannot believe the total memory loss of the same old people here, thats the same ones posting the same replies as they do every year - defending the GFS when its going to be wrong.

When you have UKMO aligned to ECM & ECM mean at 120 they are NEVER trumped by an out of kilter GFS especially the 06z which is like comparing a pedigree with a stray.

So for all the neutrals who wonder if its just opinion or is it sound education over a few years lets take yesterday as a prime example of how bad the 06Z GFS is - 

UKMO 144 yest V 120 Today

B5153E79-F386-4A94-83D2-7178CCF0300A.thumb.jpeg.56d5f6a62ca521a53b211bb96025d2c5.jpeg6B525265-FABE-486A-B2C7-E12BFD6ECAB3.thumb.jpeg.05b8b45ae61e0f0de19ee54ecb80d3b3.jpeg

That is a superb match across day 6 down to Day 5. Metronomic consistency with minimal adjustments all around the Globe.

Now look at the runt of the litter the GFS. Same timeline.

6F54335A-95E0-4E68-BE5E-AFD258481719.thumb.jpeg.4c8cd2d0b3dce2a6712f2deac53938e8.jpeg

8A5D94D1-6B0D-4290-9BBF-FE963766558D.thumb.jpeg.18e5dc3a0be635442e434dd94de26c4c.jpeg

Yesterday at 144 No Greenland High, No decent arc of Cold coming through Scandi

UK in southerlies -- Look how much this model has changed in that short time span...

* So if yesterday was awful at 144 with  no clue of the pattern then the probability is todays 144>168 is the same bin fodder

How anyone who indicates they follow the models can put any credibility on such a pants model (06z) - means to me they are deliberate in their attempts to derail the thread.

Back to the original well proven point 

* If the 06z goes off on a tangent through 144 then scrap it *

Steve, as I said before, with all due respect it is not the same GFS though. The new version hasn't yet had a winter run as the op, so making the same assumptions based on years old observations is rather unscientific don't you think?

Share this post


Link to post
Share on other sites

Well I'm someone who usually champions the gfs cause but not so sure of late. I think the gfs upgrade is not performing so well tbh. Saying that I think every model fails at times and I don't think you could bet your mortgage on the others either!!

So it's jury out until the 12z's

Right then back to arguing until then lol

Share this post


Link to post
Share on other sites

The GFS06Z run could well be more on the money but it carries on the theme of low pressure near the UK, threat of some colder air heading southwards and some height rises to the south of Greenland. As per usual with the weather, if the short term is incorrect, the rest of the output will be and that is the same with the euro models. 

It's very easy to get sucked into each run and expect it to play out like that, there will still be some variation. Hopefully the variation on the Euro models will be more of a northerly instead of a ENEly as it means the colder air gets down quicker and its brighter instead of cold and wet weather. 

Share this post


Link to post
Share on other sites

I did say yesterday that it would be clever to take a watching brief for a few runs ........currently I estimate wintry nirvana to be around 20% and I don’t believe that the current modelling offers too much help re direction of travel .......

I wouldn’t want to bet my house that days 8/12 won’t see a return of mobility but suspect if we don’t see poss snowy cold setting in then a MLB may be the interim result which of course, in early December, could be seasonal re surface conditions.....

sundays 00z runs will hopefully resolve next weekend and the first few days dec re general feel of things .......

Share this post


Link to post
Share on other sites
5 minutes ago, bluearmy said:

We did watch the gfsv3 in parallel last winter ......it generally had the same failings as its predecessor (just as ec does with over amplification ) 

one thing I haven’t mentioned is the jet continues to correct south as the runs tick closer .....

You're right, but my observation still stands. Not quite enough data yet for us to make the same assumptions, and in fact we shouldn't assume at all unless it's backed up with something verifiable ;)

Share this post


Link to post
Share on other sites
16 minutes ago, Seasonality said:

Steve, as I said before, with all due respect it is not the same GFS though. The new version hasn't yet had a winter run as the op, so making the same assumptions based on years old observations is rather unscientific don't you think?

What i would say though is we have had these so called improvements to the GFS many a time but nothing ever seems to change, the same old failings emerge when it moves from parallel to op run, just at different times thats all. a leopard never changes its spots perhaps?

Share this post


Link to post
Share on other sites
9 minutes ago, Seasonality said:

You're right, but my observation still stands. Not quite enough data yet for us to make the same assumptions, and in fact we shouldn't assume at all unless it's backed up with something verifiable ;)

I'm no expert on NWP but I do have experience in high end data processing. First off, I have to take Steve's point, that based on his experience and indeed as demonstrated in previous post there is more variability in output on that run. So why is that? My top two guesses; is that they may run these on older hardware (night shift run) with more limited input to cope with less processing ability, or they may do that and use it to trial certain features or tweaks as a pre release or control run overall in their mind. There are many possibilities and we have little to no facts on the matter. In more general terms, why  anyone feels the necessity to get so hung up on this is totally beyond me. It is an absolute total waste of time. 

Share this post


Link to post
Share on other sites

Just a few points I would like to make.

We are approaching the FIRST week of winter not the LAST week of winter although reading through some of the unnecessary panicky posts you would be forgiven for thinking the latter! 

@Steve Murr has highlighted the known issues which I also mentioned in one of my previous posts (with the same Feb 2009 example) 2 key points - Underestimating strength of blocking higher pressure and overdeveloping lower pressure which we know from the model runs over the years is a GFS specialty (although it is the new FV3 probably still improvements / ironing out needed) Although I would add that other models are also likely to suffer from this too even though GFS is usually the front runner at doing so. That being said I still believe that these current heights that are being shown to develop around Greenland will be brief (possibly 3/4 days) and it may be a secondary go at this and also restrengthening of the higher pressure out east (especially toward Siberia) that may have a better and longer lasting chance of the UK benefiting from a cold POV, If I had to pick a time for monitoring it looks to POSSIBLY coincide with the beginning of the warming in the strat (first week of December) although plenty of uncertainty with this, then it will possibly become a question of how quick the feedback from the warming which may disrupt any likely pattern although we are a long way off knowing details so worth keeping an eye on the strat thread. 

Also something that becomes tedious quite quickly is the overuse of the word ZONAL so here is a chart that illustrates it (purely for example purposes not saying it is likely or unlikely to happen) gensnh-17-1-384.thumb.png.5c43166f39a9bd6899acf2594e7d1130.png

MJO may be heading into COD but still uncertainty with that so I will look to do an update as things progress.

Share this post


Link to post
Share on other sites
24 minutes ago, feb1991blizzard said:

What i would say though is we have had these so called improvements to the GFS many a time but nothing ever seems to change, the same old failings emerge when it moves from parallel to op run, just at different times thats all. a leopard never changes its spots perhaps?

I guess that depends on whether you believe the deficiency that GFS has v ECM/UKMO is down to the model (i.e. the physics included) or in the initial conditions.  I understand that ECM and UKMO have a more sophisticated algorithm for accurately calculating the T0 state of the atmosphere than GFS, which would explain why the new model seems to have similar failings.

Share this post


Link to post
Share on other sites

Today's 06z gefs are more progressive compared to the 00z gefs and eps at day 10.The ht anomalies towards S,Greenland are much reduced and the Atlantic high is toppling quicker heralding the return of westerlies soon after.

I know i have seen a lot of GFS op runs become less progressive as we get closer to 0 hrs and not just the 06z run.In fact i showed this in a post i did yesterday on comparing the last  3 x12z runs.

Still it's best never to completely disregard any run but a look at where it sits within the ens spread usually gives idea of it's probability.

 

Share this post


Link to post
Share on other sites
10 minutes ago, Paul said:

It's none of the above. It's simply differences between input data and the amount and type of data available. It's run on the same super computers regardless of run and every run uses the same physics.

There's some evidence to suggest the 00z and 12z runs are stronger, and based on the presentation linked below, there's also less data available to the 06z (the presentation suggests ~30%)  than all of the other runs. 

https://ams.confex.com/ams/27WAF23NWP/webprogram/Handout/Paper273676/GFS4cycle_2015_NWP_Chicago_15A.pdf

Praise the lord!Thankyou for clearing that up.so going threw the next 7 days the rain will be main player and yet again causing flooding. Anything in regard to a rise in heights would suit me over the UK going in to December 

 

Edited by swfc

Share this post


Link to post
Share on other sites

Through this evening and tonight the trough to our south west will become more negatively tilted under pressure from the west. This results in the associated fronts, and heavy rain bands, swinging north west within the circulation and effecting most of England, Wales and N. Ireland by morning

gfs-deterministic-natl_wide-z300_speed-4488800.thumb.png.bb76d57b7a070f23356fd65e38f2ce86.pngindex.thumb.png.1b01ac764598ff2228f8b251904a2509.png

PPVA89.thumb.gif.cb67ee23d0fa1d757487628dcf5ce149.gifPPVE89.thumb.gif.a6c8635f2eba2470647427436299bfc2.gif

meanreflec_d02_10.thumb.png.29d7a57650f9178266caf9974dc9c4b9.pngmeanreflec_d02_13.thumb.png.6c7061d5b16ba106ee6ea6deb9e5fd24.pngmeanreflec_d02_16.thumb.png.faa2ffd88a86fdb3a4f6a537aeaaf097.pngmeanreflec_d02_20.thumb.png.ade62657677db85cceb36bd71dfb4214.pngmeanreflec_d02_24.thumb.png.3c46935c5b93c5cd63feb42ddfcebf25.png

Share this post


Link to post
Share on other sites

It still looks extremely cold in both GFS/ECM even the 6z was cold so below zero degrees at daytime temperatures for whole western europe and NAO is still negative for the whole period so no possibilty for any  milder temperatures than normal

Share this post


Link to post
Share on other sites
5 minutes ago, Faronstream said:

It still looks extremely cold in both GFS/ECM even the 6z was cold so below zero degrees at daytime temperatures for whole western europe and NAO is still negative for the whole period so no possibilty for any  milder temperatures than normal

I think you may be getting a tad confused with ground temperatures and 850 heights temperatures there

Share this post


Link to post
Share on other sites

I don't know about historically, but the GFS 6z is currently verifying better than the 0z, 12z, 18z:

T120

cor_day5_HGT_P500_G2NHX.png

T144

cor_day6_HGT_P500_G2NHX.png

The differences are almost negligible though. The issue is not so much the difference between individual runs, but that the GFS performs significantly worse than the ECM, UKMO, and even the GEM. So for all the expensive upgrades, they still can't seem to catch up.

cor_day5_HGT_P500_G2NHX.png

 

Edited by Yarmy

Share this post


Link to post
Share on other sites
3 minutes ago, StretfordEnd1996 said:

I think you may be getting a tad confused with ground temperatures and 850 heights temperatures there

We already know that the GFS 06z run is an EXTREME outliner, it won't verify

Share this post


Link to post
Share on other sites
2 minutes ago, Faronstream said:

We already know that the GFS 06z run is an EXTREME outliner, it won't verify

Is It?

graphe3_1000_244_23___.gifgraphe4_1000_244_23___.gifgraphe6_1000_244_23___.gif

Share this post


Link to post
Share on other sites

Icon at 120, not great but better than the last GFS with the block 

261B97FA-42BC-46B1-B5ED-09CB16DFF936.png

91F4F148-096D-493B-A318-CE8B3E9FB957.png

Share this post


Link to post
Share on other sites
5 minutes ago, ICE COLD said:

Icon 12z looks good at T120 . 8E71154F-E3A4-4FE1-9A60-980146F21296.thumb.png.f8fb0ba019efdb3424c929a625578132.png

Not so good at 144, more gfs 06z than ECM....This is the ICon though, big UKMO coming up 

C3F8EB66-FB5F-49DA-9892-A9518DF40170.png

Edited by Ali1977

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.
Sign in to follow this  

×
×
  • Create New...