Jump to content
Sign in to follow this  
phil nw.

Model Output Discussions 06z 04/11/16

Recommended Posts

35 minutes ago, Summer Sun said:

@Gibby posted them more or less daily

You can see his model analysis below but he doesn't post the verification stats now from what I can see

http://www.norton-radstockweather.co.uk/latest-model-analysis

Interesting to note that he disagrees with most here on the interpretation of the 06Z GFS output - to quote him:

Quote

By Day 10 the HIgh has declined and has become replaced by Low pressure having developed and moved in from the West. The weather would turn unsettled and wet for many as a result.

As if to emphasise his convictions, in his summary paragraph, he says the following:

Quote

Longer term there looks to be a slow trend towards the chance of a spell of more Atlantic based weather returning late next week which would mean spells of rain and showers for many. It still looks very unlikely that there will be any wintry type weather falling from the skies over the next few weeks with just the traditional seasonal offerings of frost and fog being the only wintry type weather likely in the upcoming 2 week period.

This suggests that we not only have to contend with wildly differing model outputs, but also wildly differing interpretations of the same (GFS 06Z at T+240) output!

Edited by chrisbell-nottheweatherman

Share this post


Link to post
Share on other sites
6 minutes ago, chrisbell-nottheweatherman said:

Interesting to note that he disagrees with most here on the interpretation of the 06Z GFS output - to quote him:

 

Just to quickly point out to all

Gibby thoughts are on the 00z charts, unfortunately, the charts he posts auto update and thus can look out of place especially given the wild swings we are seeing at present.

His thought's are updated once per day now around 9am

Share this post


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

I would normally say that the 06z is an outlier (it is); and its the 06z, so pretty rubbish after D6, but there is so little instruction as to what happens with the N Pacific Ridge from D6 and the knock on effect to the downstream Atlantic Ridge that it is difficult to rule it out. It could well be the Eureka moment that the models start veering towards, or not?

As always, more runs needed and we require these to be pumped out inter and cross model before we can tell our families and friends to baton down the hatches:

graphe6_1000_228_1___.gif

The 06Z is the same computer and same software as the 0Z. The only difference being the initial conditions are updated since the 0Z was executed. 

Verification stats have shown time and again that there is no long term trend which show 06Z better or worse than 0Z etc. At times all 4 runs have been shown to be best performing. 

Share this post


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

Just to quickly point out to all

Gibby thoughts are on the 00z charts, unfortunately, the charts he posts auto update and thus can look out of place especially given the wild swings we are seeing at present.

His thought's are updated once per day now around 9am

How come the T+240 output I see on his site is a 23rd November 06Z 500hPa chart?  Has he mixed analysis of the 00Z with images of the 06Z for GFS?

Edited by chrisbell-nottheweatherman

Share this post


Link to post
Share on other sites
Just now, chrisbell-nottheweatherman said:

How come the T+240 outout I see on his site is a 23rd November 06Z 500hPa chart?  Has he mixed analysis of the 00Z with images of the 06Z for GFS?

He uses wetterzentral charts which auto update unless you save them and upload

The text analysis is based on what he saw on the 00z outputs ignore the charts now until tomorrow morning

Share this post


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

The 06Z is the same computer and same software as the 0Z. The only difference being the initial conditions are updated since the 0Z was executed. 

Verification stats have shown time and again that there is no long term trend which show 06Z better or worse than 0Z etc. At times all 4 runs have been shown to be best performing. 

The 06z has less data than any other run, between 10-30% so I am not sure where that info comes from. Here is the scientific paper on this (2015):

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

Share this post


Link to post
Share on other sites
40 minutes ago, IDO said:

I would normally say that the 06z is an outlier (it is); and its the 06z, so pretty rubbish after D6, but there is so little instruction as to what happens with the N Pacific Ridge from D6 and the knock on effect to the downstream Atlantic Ridge that it is difficult to rule it out. It could well be the Eureka moment that the models start veering towards, or not?

As always, more runs needed and we require these to be pumped out inter and cross model before we can tell our families and friends to baton down the hatches:

 

To be honest IDO I'm not quite sure to which ridge you refer. Is it the GOA one which seems to retrogress?

gefs_z500a_nh_61.png

Edited by knocker

Share this post


Link to post
Share on other sites

Hi ^^^^^

Its more to do with the HP cell off the west coast of the conus that has been subdued of late. It ebbs and flows but from T90 to T384 it is active and provides some amplification upstream and with the propensity of the Azores High to ridge, this may be influential within the confines of a weak PV

gfsnh-0-90.pnggfsnh-0-384.png

The PNA index has been positive for sometime but is now forecast downwards possibly negative (>N Pacific blocking):  pna.sprd2.gif

This may be useful in the Atlantic sector for a sympathetic ridge and "potential" again the watchword for UK cold.

Edited by IDO

Share this post


Link to post
Share on other sites

I think we are n the same one. And talking of teleconnectione the latest EPS mean for the AO is just negative until the 8th and the GEFS around -1.

Share this post


Link to post
Share on other sites
30 minutes ago, IDO said:

The 06z has less data than any other run, between 10-30% so I am not sure where that info comes from. Here is the scientific paper on this (2015):

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

I've just had a quick scan through that PDF and despite the number of abbreviations after the authors name, no way could that ever be described as a "Scientific Paper". At best it's a series of slides, probably missing most of the commentary someone would use if they were actually doing a presentation.

I'm not qualified to say if the author is right or wrong, but to me it seems to portray quite a 'mixed message' despite his claim of less data on certain runs. In fact, even he says

"Value of the GFS 06Z and 18Z cycles: a) provide updated initial and
boundary conditions for down-stream jobs; b) provide users with updated
and more accurate guidance than the earlier 00Z and 12Z cycles for short
to medium-range weather forecast."

The verification stats he presents at Day 5 are 0z - 0.877 , 6z- 0.871, 12z - 0.876, 18z - 0.872. Hardly huge differences there, and the GFS went through a major upgrade last year as far as I remember, so those stats may be out of date anyway. Maybe using data that is 6 hrs 'newer' makes more difference than the amount of data?

He also says,

"18Z and 00Z have the same amount of conventional data, but 18Z
has worse forecast skills than 00Z. On the other hand, 12Z has
less conventional data than 00Z, but 12Z and 00Z have similar
forecast skills."

How does anyone explain that if it's all to do with the amount of input data? He doesn't.

No offense to the Author, but I'd not be placing  any weight on his apparent findings since they seems to raise more questions than they provide answers.

Share this post


Link to post
Share on other sites

Back to model discussion, GFS 12Z is running out up to t120 so far and is fairly similar to the 6Z

6Z gfsnh-0-126.png?6   12Z gfsnh-0-120.png?12

The main changes of Interest that I can see are slightly lower heights over Med, low over Greenland stronger, but slightly more WAA going up west coast of Greenland.

Share this post


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

 

The verification stats he presents at Day 5 are 0z - 0.877 , 6z- 0.871, 12z - 0.876, 18z - 0.872. Hardly huge differences there, and the GFS went through a major upgrade last year as far as I remember, so those stats may be out of date anyway. Maybe using data that is 6 hrs 'newer' makes more difference than the amount of data?

H

Hi

I did say the 06z is OK before D6 as the data is more relevant, but after that the lack of data will make it perform poorer. Of course over a period of a year where you have weather patterns that are much easier to predict the margins of error will be less and drive down the average, but in situations of high volatility like the current post D6 NH profile any poor data in = poor data out exponentially. 

Anyway the 12z is coming out and we will see if the 06z was onto something.

Share this post


Link to post
Share on other sites

Similar but looks like heights from the Atlantic up towards Labrador may not be as strong, next few frames are key. 

Looks like a Northerly is still on the cards !!

Edited by Ali1977

Share this post


Link to post
Share on other sites
1 minute ago, Ali1977 said:

Similar but looks like heights from the Atlantic up towards Labrador may not be as strong, next few frames are key. 

It`s on a Knife edge!!

Share this post


Link to post
Share on other sites

12z looks to be heading the same way sypnotically as the 6z at t144, very similar charts comparably...

Surface pressure over Greenland a notable increase.

Share this post


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

Hi

I did say the 06z is OK before D6 as the data is more relevant, but after that the lack of data will make it perform poorer. Of course over a period of a year where you have weather patterns that are much easier to predict the margins of error will be less and drive down the average, but in situations of high volatility like the current post D6 NH profile any poor data in = poor data out exponentially. 

Anyway the 12z is coming out and we will see if the 06z was onto something.

I don't necessarily disagree, but if that is the case then how do you explain the second quote I posted from the paper?

"18Z and 00Z have the same amount of conventional data, but 18Z
has worse forecast skills than 00Z. On the other hand, 12Z has
less conventional data than 00Z, but 12Z and 00Z have similar
forecast skills."

That's within day 5 too, where you'd think it wouldn't make as much difference.

Share this post


Link to post
Share on other sites

GFS pretty much going for a repeat at t168...

06Z gfsnh-0-174.png?6   12Z gfsnh-0-168.png?12

Maybe not quite as good.

 

Share this post


Link to post
Share on other sites

it doesnt look likely to happen on this run at 168h a spoiler shortwave south of greenland looks like it will force its way through still it can all change on the next run

Share this post


Link to post
Share on other sites

Looks to me, based on those charts, as if much hinges on T+150 and that that shortwave development (or not). As others say, thereafter the changes are notable from what we've seen on the 06z.

 

Share this post


Link to post
Share on other sites

168 isn't as good as the 06z but that's still not a bad chart at all. Can't post pictures as I'm on my phone but generally the theme is the same as the 6z up to a reliable timeframe, we just need some luck

Share this post


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

it doesnt look likely to happen on this run at 168h a spoiler shortwave south of greenland looks like it will force its way through still it can all change on the next run

It's not awful though at t192...

gfsnh-0-192.png?12 gfsnh-1-192.png?12

Just not the nirvana that the 6z was, but it was unlikely to be anyway.

Share this post


Link to post
Share on other sites

Synoptically it is close enough to the 06z to consider we will see a sharpening/amplification of the pattern in the Atlantic sector as this is also played out by UKMO though it is more progressive.

UN144-21.GIF?23-17

So some kind of Northerly outbreak seems quite plausible but how potent and how long lived should it verify is another question.

I would expect to see the same upstream improvements in tonight's ECM or at latest tomorrow morning though that doesn't mean it will show a screaming Northerly, just that it will more amplified 144/168+ than recent efforts.

Edited by Mucka

Share this post


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

A solid day so far with the date 30th still the hinge day-

all afternoon models so far are similar to the earlier GFS solution -

12z GFS 7/10

12z UKMO 8/10

12z GEM up to 144 8.5/10

S

Steve

Could you post please?, I'm on mobile and cant access.

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...