Jump to content
Sign in to follow this  
Paul

Model output discussion - into 2018

Recommended Posts

GFS control and mean are a lot better than the op. The ENS will be interesting.

 

gens-0-1-192.png

gens-21-1-192.png

Share this post


Link to post
Share on other sites
8 minutes ago, Joshua said:

Worth looking at this fantastically in-depth forecast from Gavin Partridge. 

If the GFS fails I will put it in the shredder, stick the pieces back together one by one and then claim I have a more accurate GFS model than before. 

It's standards continue to slip* away in winter forecasting terms. 

*Wash is probably a better word to put in there, no ice with a GFS slice!

Turned it off when I saw he was using the CFS.

Share this post


Link to post
Share on other sites

HMMM - a very interesting GEFS set so far, lots of undercutting going on prolonging at the very least surface cold.

Share this post


Link to post
Share on other sites
8 minutes ago, Joshua said:

Worth looking at this fantastically in-depth forecast from Gavin Partridge. 

If the GFS fails I will put it in the shredder, stick the pieces back together one by one and then claim I have a more accurate GFS model than before. 

It's standards continue to slip* away in winter forecasting terms. 

*Wash is probably a better word to put in there, no ice with a GFS slice!

So the US taxpayers finance 6 global weather models: gfs, cfs , coamps, NAVGEM, nasa/Geos and fim and all of them are crap 😎, what a waste of money 

Share this post


Link to post
Share on other sites

Nope, even the ENS can't seem to lower heights in the Med. Without that, you can forget it.

Looks like GFS doesn't deal with a split jet very well. Seems to always want to push the energy north.

Share this post


Link to post
Share on other sites

Interesting, both the Opp and control at the milder end of the spectrum, lots of colder options.:)

 

London and Manchester there.

graphe_ens3.gif

graphe_ens3-1.gif

Share this post


Link to post
Share on other sites

12z GEFS set certainly showing plenty of interest in less energy going over the top of the block. All options still wide open.

Share this post


Link to post
Share on other sites
4 minutes ago, ArHu3 said:

So the US taxpayers finance 6 global weather models: gfs, cfs , coamps, NAVGEM, nasa/Geos and fim and all of them are crap 😎, what a waste of money 

whats the best one to use please.

Share this post


Link to post
Share on other sites
Just now, offerman said:

whats the best one to use please.

ECMWF 😁

Share this post


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

The 00z NCMRWF was model filth :rofl:

in all seriousness didn’t someone post the GEM was doing joint 3rd with the GFS recently or am I utterly model fatigued 

its not in my capability to decipher that chart 

 

Current verification for the Northern Hemisphere 12z runs. Over the 30 day average -

Day 5:

cor_day5_HGT_P500_G2NHX.png

1. ECMWF

2. UKMO

3. GFS

4. GEM

5. JMA

 

Day 6:

cor_day6_HGT_P500_G2NHX.png

1. ECMWF

2. UKMO

3. GFS

4. GEM

5. JMA

 

Day 10:

1. ECMWF

2. GFS

3. GEM

 

That's for 500mb heights. If we look a bit closer to the surface:

Day 5:

cor_day5_PMSL_MSL_G2NHX.png

1. ECMWF

2. UKMO

3. GEM

4. GFS

5. JMA

 

Rather than clogging up the thread any further, this is the exact same at Day 6 and Day 10, with the GEM marginally ahead of the GFS over the 30 day average.

Incidentally, I know I switched from heights to pressure for the surface charts (for reasons which are hopefully obvious in terms of what we are interested in at the surface level), however we see the exact same results for heights at 1000mb.

A few things for everyone to take away:

1) Given the above, you would probably say that GFS and GEM are generally on a par over the last 30 days, the GFS perhaps just about edging it in terms of average score across disciplines

2) GEM is worse than GFS for 850mb temperature skill (not shown above)

3) You can probably throw all of the above away when dealing with the current situation because the dynamics of the atmosphere are vastly different to anything we have seen these models try to handle over the past 30/31 days, and so this is probably not the best measure of which model has the best handle on our unfolding situation (which I'm sure anyone who has read through all of this post to this point will be really pleased to hear).

So that's the 12z stats, a few things to note for the 0z's:

1) GFS generally remains ahead of the GEM over the last 30/31 days

2) ECMWF 0z has performed really rather significantly better than the ECMWF 12z in recent weeks (by a correlation score different of around 0.4-0.5 - which at the day 10 range is probably quite noteworthy)

3) And the final nail in the coffin for anyone wanting to blindly follow the GEM believing it to be better than the other models out there....at 0z, it has performed worse on SLP than the NAVGEM. So, all hail the NAVGEM. (admittedly, the NAVGEM only has 29 days of data available for some reason vs 31 for the rest).

 

 

 

Share this post


Link to post
Share on other sites
9 minutes ago, ArHu3 said:

So the US taxpayers finance 6 global weather models: gfs, cfs , coamps, NAVGEM, nasa/Geos and fim and all of them are crap 😎, what a waste of money 

I wasn't going to put it like that, but let's just say: there's plenty of room for improvement? 😬

Share this post


Link to post
Share on other sites
9 minutes ago, s4lancia said:

12z GEFS set certainly showing plenty of interest in less energy going over the top of the block. All options still wide open.

That's not what Warren said. Very confusing in here!

Edit. Just checked the ens for myself and yes plenty of interest.

Edited by Johnp

Share this post


Link to post
Share on other sites
36 minutes ago, nick sussex said:

NOAA's last update re the MJO came down in favour of the ECM forecasts not the GEFS because they have better verification and seeing as the MJO is playing an instrumental role in the outlook then I'd be wary of what the GFS does.

We are due another update later this evening from the Global Tropics Hazards outlook so it will be interesting to see what they have to say then.

I would be wary of putting to much faith in the MJO. Not everyone agrees that the MJO is playing an instrumental role in the outlook at all. There are plenty more factors at play.

Disregarding the gem ( I think I have made my feelings clear on that model) just when I start to be persuaded that this could possible be a noteworthy cold spell along comes another set of runs and my worst fears of poor amplification and the cold disappearing to the south are re-affirmed. This could still turn into a blink and you  will miss it affair unfortunately. By now  you would have thought the emphasis of topic would be distribution of snow showers and how cold will it get, not will we get an east or north easterly or not. bah humbug.

Edited by comet

Share this post


Link to post
Share on other sites

I know the gem isn’t particularly trusted on here but it’s value imo is that occasionally it picks a winner. It picked this Scandi ridge  first and it might just be right on this occasion. 

About forty five mins till we know where ec op is heading ...... initial ridge may be a bit flat (ironically advecting lower uppers) but the second coming via the stalling trough could be the interesting part. 

Share this post


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

I know the gem isn’t particularly trusted on here but it’s value imo is that occasionally it picks a winner. It picked this Scandi ridge  first and it might just be right on this occasion. 

About forty five mins till we know where ec op is heading ...... initial ridge may be a bit flat (ironically advecting lower uppers) but the second coming via the stalling trough could be the interesting part. 

NOAA uses the gem ensemble too for their forecasts, they wouldn't do it if it were a rubbish model 

 

http://www.emc.ncep.noaa.gov/gmb/ens/NAEFS.html

Share this post


Link to post
Share on other sites

Can we keep it to the models please, and those responding to the off topic stuff are just making any issues that much worse, so please don't. 

Share this post


Link to post
Share on other sites
1 minute ago, bluearmy said:

I know the gem isn’t particularly trusted on here but it’s value imo is that occasionally it picks a winner. It picked this Scandi ridge  first and it might just be right on this occasion. 

About forty five mins till we know where ec op is heading ...... initial ridge may be a bit flat (ironically advecting lower uppers) but the second coming via the stalling trough could be the interesting part. 

All models have slightly different algorithms otherwise they would all churn out basically the same, equally I'm sure inputs for each are not the same hence the need to take each with a pinch of salt and see how each handles different situations. Gem does has its moments and if the scandi ridge does come off then your right it pick it first.

I'm not sure why some get so het up about individual models if you don't trust it don't use it

Share this post


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

I would be wary of putting to much faith in the MJO. Not everyone agrees that the MJO is playing an instrumental role in the outlook at all. There are plenty more factors at play.

Disregarding the gem ( I think I have made my feelings clear on that model) just when I start to be persuaded that this could possible be a noteworthy cold spell along comes another set of runs and my worst fears of poor amplification and the cold disappearing to the south are re-affirmed. This could still turn into a blink and you  will miss it affair unfortunately. By now  you would have thought the emphasis of topic would be distribution of snow showers and how cold will it get, not will we get an east or north easterly or not. bah humbug.

Normally with Scandi highs the problem is the how the cold pool advects west but in this instance any initial cold is coming from the north/ne.

The pattern isn't too far south but too far north, if the high stays further north at the start of the evolution then that suggests high pressure in the Med is coming into play which isn't what we want to see.

The key is to drive low heights as far east and se into Europe to block off that ridge from the Med.

Low heights over the north Med mean the block can't sink, the issue we've had is a sag of the high on the eastern flank.

If you get to the point of the trough disrupting upstream then you'd want the high to then be forced further nw to increase the forcing on the Atlantic.

I can understand your frustration but this model drama seems par for the course with anything that involves high pressure to the ne.

Share this post


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

I know the gem isn’t particularly trusted on here but it’s value imo is that occasionally it picks a winner. It picked this Scandi ridge  first and it might just be right on this occasion. 

About forty five mins till we know where ec op is heading ...... initial ridge may be a bit flat (ironically advecting lower uppers) but the second coming via the stalling trough could be the interesting part. 

Maybe of interest to folk here - stats show that GEM is verifying 3rd/4th and often swaps places with GFS. :D

Share this post


Link to post
Share on other sites

Gem has this nailed why can't it  ... its been very  consistent run to run ..:D

IMG_0316.PNG

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...