Model Output Discussion - Into Winter

[northwestsnow]

00z day 8

06z

[Derecho]

37 minutes ago, Mike Poole said:

I think you might be trying to have your cake and eat it here?  You’ve argued the ensembles are the right way to go week 2 and beyond, but with that comes the spread.  The BOM chart shown by @Downburst says it’s an ensemble of 33 members and they are all amplified.  So there is a difference - I’ll be much happier when the amplification into phase 7 is confirmed by other models, the BOM being the go to model all of a sudden has a whiff of cherry picking to me.

EC and GFS bias corrected forecasts all go for a weak phase 7 (within the COD)

We did have a decent phase 1 though in the later stages of November which corresponds to the following:

 

Kinda looks like what we've had in the last week or so. However if we go through an amplified phase 4-6 and then a weak phase 7, the composites suggest...

Amplified phase 4

 Amplified phase 5 - yuk

 Amplied phase 6 - nah

 Weaker phase 7 - still mild

So I don't really hold much hope with the MJO composites here and presuming there is a lag time that would take us from now to late December with mild weather... if the MJO was an exact science.

 

Edited December 6, 2023 by Derecho

[Met4Cast]

51 minutes ago, Downburst said:

The South Korean weather model has a 7 day warranty apparently, has ECM worried with their 5 day warranty.

What does this mean? I know what a warranty is obviously but not sure how that applies to weather models!

[northwestsnow]

Day 10 

6z top ,0z bottom.

A chilly 6z due to placement of the anticyclone..

[bluearmy]

2 minutes ago, Met4Cast said:

What does this mean? I know what a warranty is obviously but not sure how that applies to weather models!

He’s making a joke ….

[Met4Cast]

Just now, bluearmy said:

He’s making a joke ….

Evidently I need more coffee 

[jmp223]

The amplitude of the MJO on the most recent plot is noticeably downgraded, could anyone explain what that would entail for NW Europe? A lesser chance of high latitude blocking I'd assume?

[Johnp]

4 minutes ago, Met4Cast said:

What does this mean? I know what a warranty is obviously but not sure how that applies to weather models!

South Korea's main car manufacturer is Kia and one of their main selling points when they first appeared on the scene was a 7 year warranty...

 

[bluearmy]

48 minutes ago, Mike Poole said:

I think you might be trying to have your cake and eat it here?  You’ve argued the ensembles are the right way to go week 2 and beyond, but with that comes the spread.  The BOM chart shown by @Downburst says it’s an ensemble of 33 members and they are all amplified.  So there is a difference - I’ll be much happier when the amplification into phase 7 is confirmed by other models, the BOM being the go to model all of a sudden has a whiff of cherry picking to me.

That’s fair on the basis that I like cake 

my main point on the 46 spread is that it has 101 members 

 

[Mike Poole]

1 minute ago, bluearmy said:

That’s fair on the basis that I like cake 

my main point on the 46 spread is that it has 101 members 

 

I’m still not entirely clear why this matters, as I see it, the BOM spread with 33 members is nowhere near the spread of the ECM, even accounting for the fact that the latter has about 3 times the number of members.  

[Johnp]

The Theta-E charts would indicate this high on GFS 6z would be a frosty one, so at least some seasonal, sunny weather would be nice if we can't have snow on the run up to Xmas.

[Mark wheeler]

1 hour ago, Derecho said:

Hahaha that's a shame! It already looks very different to the UKMO at T168....

Is it better then the NAVGEM though? 

Never

[minus10]

This is what we want to see...evolving further north as time goes on..

...unfortunately it is outside the guarantee period...

Edited December 6, 2023 by minus10

[Cambrian]

The early hours of last Saturday:

The night was magical.

Owls-a-hooting under an upended plough 

in the hour after midnight.

The tile of each roof glistening 

in the crispest light of the moon.

The frost a sparkling enchantment, 

sharing edges with the silence 

of the hard, white ground 

on the first full night of winter.

 

And it was -6 degrees and I needed my jacket. Needs too such a night to free the spirit of a season. 

 

Early hours this Saturday:

Forget it, I’m not going out in that.

[StretfordEnd1996]

25 minutes ago, Johnp said:

The Theta-E charts would indicate this high on GFS 6z would be a frosty one, so at least some seasonal, sunny weather would be nice if we can't have snow on the run up to Xmas.

Yes both immediately before the high is over the top of us (where it forms something of a weak Atlantic ridge very briefly) and during it being over the top of us - 

 

Daytime temps also look on the cool side under this higher pressure

I'd therefore say at this stage, in terms of surface temps there appears little in the way of mild from around D7 onwards for the most part, into the following week.

[Downburst]

41 minutes ago, Met4Cast said:

What does this mean? I know what a warranty is obviously but not sure how that applies to weather models!

Sorry, a joke on the Korean "models" (car models) having 7 year warranties, which beat the likes of European cars (re ECM) with 3 to 5 year warranties and extrapolating to the NWP

[Frostbite80]

CFS sniffing out a SSW?

[Met4Cast]

9 minutes ago, Downburst said:

Sorry, a joke on the Korean "models" (car models) having 7 year warranties, which beat the likes of European cars (re ECM) with 3 to 5 year warranties and extrapolating to the NWP

Thanks, not a car person so it went straight over my head

[Purga]

Suffice it to say that FI GFS offers a significant difference 0z to 06z with the later offering a glimmer of hope for a longer term colder evolution.

[Blessed Weather]

Plenty of discussions around the accuracy of models in forecasting the MJO. Here's another paper (in addition to the one posted on the previous page by Downburst) which identifies a known weakness of underestimating the amplitude of the MJO as it progresses across the Maritime Continent and Western Pacific:

"One of the greatest challenges in current dynamical forecast systems is the fast damping of the MJO signal as the forecast lead time increases, which results in a rapid decrease of forecast skill. This systematic damping of the MJO convection signal is particularly apparent when the MJO starts over the Indian Ocean and is expected to propagate through the Maritime Continent and move further into the western Pacific. The frequency of MJO events not crossing the Maritime Continent in forecast models is more than twice as large as it is in observations, known as the Maritime Continent prediction barrier."

Does this help explain the slump of amplitude in many of the RMM forecasts we are currently seeing? Here's the ECM Ensemble Extended Range marked as illustration:

In the same paper the Average forecast errors by model (unfortunately no BOMM inclusion):

Source: Deep learning for bias correction of MJO prediction

I've put both papers into the Netweather Research Library where there's lots of other MJO papers.

[Derecho]

1 minute ago, Blessed Weather said:

Plenty of discussions around the accuracy of models in forecasting the MJO. Here's another paper (in addition to the one posted on the previous page by Downburst) which identifies a known weakness of underestimating the amplitude of the MJO as it progresses across the Maritime Continent and Western Pacific:

"One of the greatest challenges in current dynamical forecast systems is the fast damping of the MJO signal as the forecast lead time increases, which results in a rapid decrease of forecast skill. This systematic damping of the MJO convection signal is particularly apparent when the MJO starts over the Indian Ocean and is expected to propagate through the Maritime Continent and move further into the western Pacific. The frequency of MJO events not crossing the Maritime Continent in forecast models is more than twice as large as it is in observations, known as the Maritime Continent prediction barrier."

Does this help explain the slump of amplitude in many of the RMM forecasts we are currently seeing? Here's the ECM Ensemble Extended Range marked as illustration:

In the same paper the Average forecast errors by model (unfortunately no BOMM inclusion):

Source: Deep learning for bias correction of MJO prediction

I've put both papers into the Netweather Research Library where there's lots of other MJO papers.

Yup I've heard about this, the charts I posted have a bias correction applied to them but obviously in the long term that correction becomes more difficult as the uncertainty increases also. 

[Big Gally]

[Cambrian]

An interesting indication of the direction of travel might be gleaned by comparing yesterday’s 0z EPS for day 15 with today’s 0z day 15 chart. 

Both are extraordinary charts, yesterday’s 0z for its emphatic blocking, strong heights over both the Barents / Kara Sea area and the Atlantic, leaving the question begging - where’s the weather going to come from?

Today’s day 15 chart, for the winter solstice, looks like a much more definite nod toward the European trough, and markedly stronger heights over the Atlantic, the contour lines indicating a trough to our northeast over the Norwegian Sea. The trough is linking back through the pole to the core PV that by then has migrated back west towards Alaska, and extends southeast to the east of the UK and Ireland, down into Southern Europe…

…hinting at a sharp northwest-southeast alignment to the jet stream, diving southeast into Europe, the UK and Ireland very clearly on the western flank, the cold side of the trough. Notice too the loss of the contoured trough to the south of Greenland, now much further east and the beginning of a return of lower heights around the back of the pole to Siberia. 

In this growing momentum, there’s plenty of scope for a notably seasonal change to the weather in the approach to Christmas. 

[ScottSnow]

CFS on wetterzentrale leading to Christmas.

I don’t place much faith in it but it did ‘seem’ to pick out the late Nov cold well. 
 

Edited December 6, 2023 by ScottSnow

[Catacol]

3 hours ago, Downburst said:

 

As we all know the MJO is ticking away all the time since the ice age ended and there is a danger it will be taken too seriously for cause and affect when needed for a cause. I think the main issue I have is this BOM chart, seems always to be in insane amplification, like 100% more. So I'll be sticking with ECMWF and I certainly don't buy the argument that Australians have the MJO nailed in their computer models and the rest don't, hard to believe that given the age we are in and funding ECMWF and GFS get etc. 

 

Dec 5th ECMWF

Dec 4 BOM

A few thoughts in relation to this given the publication of a peer-reviewed paper on MJO modelling skill back in 2018. A few years old now, but likely still relevant:

 

1. Initial starting phase is important. This quote:

We have seen this ourselves now. The forecast for the MJO while it was sitting in real time in 2/3 was low amplitude. This is now being revised up as we get close to entering phase 4.

2. An overall assessment of skill is reflected in this summary table. BOM comes out 2nd - ECMWF is top. 

3. BOM has a significantly larger ensemble size and reforecast period. I would argue that this is because it matters to the Australian Met that they get Pacific conditions as accurate as possible. Locality does matter.

4. Further snippets on skill include:

and this substantive table for starting phase of the forecast.

This table for lead times given amplitude.

My take away would be that ECM, BOM and UKMO are ahead of the pack.

5. And from the paper's conclusions some relevant statements:

 

Edited December 6, 2023 by Catacol