The Weather Outlook

Remove ads from site

Quantum
  • Quantum
  • Advanced Member Topic Starter
15 January 2025 18:09:41
We all use the term semi satirically but I just can't shake the notion there may be something real in it. Why is the 18Z GFS run always better for coldies (and is the 0Z run notably worse than the 6Z and 12Z for that matter?). I should say that I don't think finding Nhem correlation charts can dispel this one way or the other; what we need is winter correlation charts for a specific area including the UK and north of it. Is the 18Z more likely to predict high latitude blocking for some reason? And if it is, what possible explanation could there be for it? Fewer non satellite observations? But if that's true what about the 6Z? And why doesn't this just mean worse performance rather than some kind of systematic bias towards blocking.

So I want to know: Is this systematic bias real or just in our minds, and if it is, why?

25/26 (850hpa temp) 11 days snow/sleet falling

18/11 (-4) 19/11 (-6) 20/11 (-6) 01/01 (-7) 04/01 (-10) 10/01 (-7) 11/01 (-3) 30/01 (-1) 13/02 (-6) 15/02 (-4) 18/02 (-6)

24/25 10d

18/11 (-6) 19/11 (-6) 23/11 (-2) 22/12 (-5) 04/01 (-5) 05/01 (0)14/02 (0) 15/02 (0)12/03 (-6) 13/03 (-6)

23/24 8d

29/11 (-6) 30/11 (-6) 02/12 (-5) 03/12 (-5) 04/12 (-3) 16/01 (-3) 18/01 (-8)08/02 (-5)

22/23 7d

18/12 (-1)06/03 (-6) 08/03 (-8) 09/03 (-6) 10/03 (-8) 11/03 (-5) 14/03 (-6)

21/22 12d

Retron
15 January 2025 18:26:19
I've never subscribed to the theory, and TBH though the 6z and 18z *are* a little less reliable (due to the lower amount of surface observation data entered into 6z and 18z vs 0z and 12z runs), I've never thought it enough of an effect to be noticed (unlike at Christmas due to the marked decrease in aircraft data).

It's interesting, though, that the Christmas data loss tends to manifest as increased blocking and a weaker jet in our part of the world. Perhaps the mild data loss of the 6z and 18z runs has a similar effect?

(ECM data input stats are here:

https://charts.ecmwf.int/catalogue/packages/monitoring/products/dcover?Flag=all&base_time=202501150600&obs=synop-ship 

...and FWIW ECM has the highest data input of the models, the others run before all the data is in. It's the reason why ECM comes out that much later than the other models!)

Leysdown, north Kent
Gandalf The White
15 January 2025 18:28:08

I've never subscribed to the theory, and TBH though the 6z and 18z *are* a little less reliable (due to the lower amount of surface observation data entered into 6z and 18z vs 0z and 12z runs), I've never thought it enough of an effect to be noticed (unlike at Christmas due to the marked decrease in aircraft data).

It's interesting, though, that the Christmas data loss tends to manifest as increased blocking and a weaker jet in our part of the world. Perhaps the mild data loss of the 6z and 18z runs has a similar effect?

Originally Posted by: Retron 

Brian posted the stats a few weeks ago and there was a negligible difference between any of the four runs, to my surprise.

Location: South Cambridgeshire

130 metres ASL

52.0N 0.1E



Brian Gaze
15 January 2025 18:34:17
I don't subscribe to the theory. However, it would be interesting to run some programmatic tests over a (arbitrary) 30 day period. I guess you could log 850hPa and 2m temps for various locations and then do some basic analysis on them. TBH, on TWO you could simply do it be picking X number of 16 day forecast locations and then pulling out the applicable variables.
Brian Gaze

Berkhamsted

TWO Buzz - get the latest news and views 

"I'm not socialist, I know that. I don't believe in sharing my money." - Gary Numan

Quantum
  • Quantum
  • Advanced Member Topic Starter
15 January 2025 18:37:34

I don't subscribe to the theory. However, it would be interesting to run some programmatic tests over a (arbitrary) 30 day period. I guess you could log 850hPa and 2m temps for various locations and then do some basic analysis on them. TBH, on TWO you could simply do it be picking X number of 16 day forecast locations and then pulling out the applicable variables.

Originally Posted by: Brian Gaze 

I didn't used to, but I just can't shake it these days. It happens so much. There are often kernels of truth in mythos.

I can well believe the effect doesn't show up generally; but for the specific case of HL blocking near the UK, Idk.

25/26 (850hpa temp) 11 days snow/sleet falling

18/11 (-4) 19/11 (-6) 20/11 (-6) 01/01 (-7) 04/01 (-10) 10/01 (-7) 11/01 (-3) 30/01 (-1) 13/02 (-6) 15/02 (-4) 18/02 (-6)

24/25 10d

18/11 (-6) 19/11 (-6) 23/11 (-2) 22/12 (-5) 04/01 (-5) 05/01 (0)14/02 (0) 15/02 (0)12/03 (-6) 13/03 (-6)

23/24 8d

29/11 (-6) 30/11 (-6) 02/12 (-5) 03/12 (-5) 04/12 (-3) 16/01 (-3) 18/01 (-8)08/02 (-5)

22/23 7d

18/12 (-1)06/03 (-6) 08/03 (-8) 09/03 (-6) 10/03 (-8) 11/03 (-5) 14/03 (-6)

21/22 12d

Essan
15 January 2025 20:25:34
For the past 20 or more years I have always ignored the 06z and 18z runs.   

It's nothing new

Andy

Evesham, Worcs, Albion - 35m asl

Weather & Earth Science News 

Look in the doubt we've wallowed, look at the leaders we've followed, look at the lies we've swallowed, and I don't want to hear no more

tierradelfuego
15 January 2025 21:40:33

For the past 20 or more years I have always ignored the 06z and 18z runs.   

It's nothing new

Originally Posted by: Essan 

I would class the 6z as the Hangover run, not the 0z as per thread title, never awake for the 0z after a night out, or I wasn't 20 years ago anyway hence the term I used earlier as I'm sure others must have... so I agree in terms of runs I take less interest in, 6z and 18z... again whether that is true or not, it just seems to have been something many say or do.

Bucklebury

West Berkshire Downs AONB

135m ASL

VP2 with daytime FARS

Rainfall collector separated at ground level

Anemometer separated above roof level

WeatherLink Live (Byles Green Crew )

idj20
16 January 2025 10:56:24
I'm sure I've read somewhere that a "wild card" algorithm is usually thrown into the 18z run which result in tasty looking op outputs. The 0z and 12z runs tend to be more "sensible" while the 06z seem to produce the worst looking outputs.  With all that to mind, I usually take the 0z and 12z more seriously when it comes to my own medium range model watching. But at the end of the day it's all pretty academic anyway. 
Home location: Folkestone Harbour.
Retron
16 January 2025 11:02:21

I'm sure I've read somewhere that a "wild card" algorithm is usually thrown into the 18z run which result in tasty looking op outputs. The 0z and 12z runs tend to be more "sensible" while the 06z seem to produce the worst looking outputs.  With all that to mind, I usually take the 0z and 12z more seriously when it comes to my own medium range model watching. But at the end of the day it's all pretty academic anyway. 

Originally Posted by: idj20 

Wherever you read that - don't bother going back, as it's rubbish. The algorithm is the same for all the GFS runs, the only differentiator is the amount of data fed into them. As mentioned earlier, the 6z and 18z have fewer surface obs compared with the 0z and 12z runs.

Leysdown, north Kent
Chunky Pea
19 January 2025 06:41:27

...and FWIW ECM has the highest data input of the models, the others run before all the data is in. It's the reason why ECM comes out that much later than the other models!)

Originally Posted by: Retron 

I'm not sure if I imagined it, but I do recall talk that the GFS was to incorporate data supplied by the ECM many years ago. Did this ever come to pass?

I look at both the GFS and the ECM, and I don't find the ECM (op) to be any better than the GFS after about 144hrs. If anything, I find the GFS to be the model that picks up on longer term trends quicker. That said, I do buy into the 18z 'pub run' theory. Very often, it'll run with charts like something out of the 1960s!, and more often than not, the 00z will slap down those childess notions of the 18z with a more sensible and mature approach. 😅

Patrick,

East Galway, Ireland.

Remove ads from site