Model output discussion - Closing in on Christmas

[Tim Bland]

IF we got to this position early Jan I’d be hoping there is no SSW to shake things up...I’d rather stick! 

[weirpig]

18 minutes ago, mushymanrob said:

For the GFS 06z @ t240 onwards to be right, then every anomaly chart would need to be wrong.

Personally, i cant see that myself.

Yep Mushy    not infalliable    but close      i will continue to ignore GFS  for the mo

[Zak M]

Ladies and gentleman, let me introduce you the NASA model. This shows frequent snow showers coming from the North Sea on Christmas Day, along with decent, cold uppers.

   
 

Oh, if it verified!

[offerman]

4 hours ago, bluearmy said:

Beware ecm op runs bearing n Atlantic amplified solutions without cross model op support !

of course it has gefs support so let’s just forget about the Gsf and gem ops this morning ........

I'll support Geoff go for it son. Forget the GoodForSodalls charts

[phil nw.]

The overnight ecm clusters  for Christmas -day 9-all go for a colder nw/n flow.It's just the angle of flow to be pinned down as we wait for the degree of ridging to be decided.

 

Looking just beyond Christmas -the 28th-we can see real divergence with 5 clusters.Does the Greenland trough flatten the high and return us to Atlantic westerlies or does the high hold and maybe ridge ne pushing low heights south over Europe?

An interesting period coming post Christmas.Today's London ens.temp.graph 

continues to show surface temps. staying below average although as expected quite a wide spread in members from the holiday period with the uncertainty on which way the pattern will develop.

[Ali1977]

I think we can say with some confidence that Xmas day will be cold, with Frost for many in the U.K.
Any precip falling could be wintery in Scotland, Northern Hills with 850 temps at -5/-6 and anything colder could mean snow showers possible anywhere. (These temps aren’t guaranteed yet of course) 

it now depends on how watered down the Northerly becomes, longevity may be the main factor and not necessarily the cold for Xmas Day itself.

Edited December 16, 2020 by Ali1977

[Chesil View]

39 minutes ago, mushymanrob said:

For the GFS 06z @ t240 onwards to be right, then every anomaly chart would need to be wrong.

Personally, i cant see that myself.

Interestingly though  Mushy. the GFS 06Z  is almost a carbon copy out 384 of the GFS 0Z run from a couple of days back that went to become a Cobra run and that run mimicked very much the 8-14 day  profile on those upper air charts that you and JH like to post with big atlantic/greeny ridge and low pressure to the east.

[Zak M]

Just went through a few of the GFS ensembles. P10 is the best I have found so far.

      

[Kentspur]

14 minutes ago, Zak M said:

Just went through a few of the GFS ensembles. P10 is the best I have found so far.

      

Still manages rain over Kent not good if its the best we have

[chris55]

1 hour ago, Met4Cast said:

Are we really getting excited over a toppler northerly? Deary me it has been a poor couple of years hasn't it 

There's a huge amount of spread within the EPS as we approach the Christmas period, though there is a clear signal for reducing temperatures as we approach Christmas day itself, although I would expect the "coldest" period to be just after Christmas. Snow risk looks rather muted at the moment. 

In the context of a classic zonal flow I may agree, but with the consistent signals of northern blocking, the weak SPV and TPV and the generally blocked December we are having, I think the potential is there for something a little more significant to develop. 
 

Still a long long way to go, but a classic two day toppler sandwiched in a Zonal flow this is not! 
 

 

Edited December 16, 2020 by chris55

[Kentspur]

GFS 6z ensembles still showing a great deal of uncertainty, the ops not a cold outlier at all and if anything is a mild outlier nearer the end of Dec

Edited December 16, 2020 by Kentspur

[chris55]

That’s a lovely ‘Christmas dip’ on GFS ensembles, perfect timing . Mild hump also flattening, and pretty good agreement for the dip. Overall a good set. 
 

 

[Ali1977]

2 minutes ago, chris55 said:

That’s a lovely ‘Christmas dip’ on GFS ensembles, perfect timing . Mild hump also flattening, and pretty good agreement for the dip. Overall a good set. 
 

 

That mean is dragged way up by a few really mild outliers, the majority of runs are colder around Xmas than the mean. 

[Weather vane]

Although I am new to posting here, I have been following this amazing forum for quite a few years now. One thing that has constantly piqued my interest is that when it comes to major negative comments about how a particular model is performing, it seems like >90% of these comments are directed at the GFS (very negative comments of ECM charts seem way less frequent), and almost always when the chart doesn't show what is being hoped for. I have to admit I find this complete dismissal bias somewhat surprising since GFS charts are posted all the time, and used as evidence to support an interpretation.

I have no vested interest whatsoever in any model; just trying to find my feet with regards to how they all compare, which ones perform better in certain scenarios, but less well in other scenarios. In the critical 7-10 day forecasting timescale that so many of us focus on, how often is the GFS model completely wrong or very inconsistent on a day-to-day basis, and how does this compare with ECM - is it that much better ? I imagine this is not easy to define empirically, though there must be studies addressing it.?

Not trying to upset the apple cart here, just trying to wade my way through learning about models and chart interpretation.

 

[Frosty.]

 

 

 

That’s a cracking day 9 ECM 0z ensemble mean for crimbo day!...fingers ...all this cold chasing is quite exhausting but I think it’s going to bear fruit during the second half of next week...hopefully!  

 

Edited December 16, 2020 by Jon Snow

[SLEETY]

Is  it time to order the sledges for the kids, and adults! yet? normally when you do that the models flip to mild and never revert back, seen that happen also many times before. lol. 

Time to hide behind the sofa soon, is coming up again.. 

Edited December 16, 2020 by SLEETY

[chapmanslade]

30 minutes ago, Weather vane said:

Although I am new to posting here, I have been following this amazing forum for quite a few years now. One thing that has constantly piqued my interest is that when it comes to major negative comments about how a particular model is performing, it seems like >90% of these comments are directed at the GFS (very negative comments of ECM charts seem way less frequent), and almost always when the chart doesn't show what is being hoped for. I have to admit I find this complete dismissal bias somewhat surprising since GFS charts are posted all the time, and used as evidence to support an interpretation.

I have no vested interest whatsoever in any model; just trying to find my feet with regards to how they all compare, which ones perform better in certain scenarios, but less well in other scenarios. In the critical 7-10 day forecasting timescale that so many of us focus on, how often is the GFS model completely wrong or very inconsistent on a day-to-day basis, and how does this compare with ECM - is it that much better ? I imagine this is not easy to define empirically, though there must be studies addressing it.?

Not trying to upset the apple cart here, just trying to wade my way through learning about models and chart interpretation.

 

You have to be very careful how you interpret comments on this thread. The 'GFS is rubbish' can actually mean, 'the model is poor and rarely verifies', 'the model is not showing what I want', 'the model is showing something different to the others and isn't what I want to believe'. As you can see these are three very different things. 

Verification stats are available for all the models and none are perfect. The stats are not that different for all models in the grand scheme of things but all will get it 'wrong' on occasions or be slow to pick up on changes.

So - read with caution !

probably the biggest problem is that the GFS goes out a long way into the future with 'apparently' the same resolution - as in when you look at the chart it looks just the same at 06z as 240z. I have often thought the charts should blur out as you get further into the future to an ensemble average to make this obvious.

[Mike Poole]

36 minutes ago, Weather vane said:

Although I am new to posting here, I have been following this amazing forum for quite a few years now. One thing that has constantly piqued my interest is that when it comes to major negative comments about how a particular model is performing, it seems like >90% of these comments are directed at the GFS (very negative comments of ECM charts seem way less frequent), and almost always when the chart doesn't show what is being hoped for. I have to admit I find this complete dismissal bias somewhat surprising since GFS charts are posted all the time, and used as evidence to support an interpretation.

I have no vested interest whatsoever in any model; just trying to find my feet with regards to how they all compare, which ones perform better in certain scenarios, but less well in other scenarios. In the critical 7-10 day forecasting timescale that so many of us focus on, how often is the GFS model completely wrong or very inconsistent on a day-to-day basis, and how does this compare with ECM - is it that much better ? I imagine this is not easy to define empirically, though there must be studies addressing it.?

Not trying to upset the apple cart here, just trying to wade my way through learning about models and chart interpretation.

 

I think there are 3 reasons for this.  First, the evidence, from the verification stats the GFS is not as good a model as ECM:

At day 5 it is behind the ECM, UKMO and just behind the GEM (CMC on chart).

Secondly, it runs 4 times a day so gets more airtime.  Thirdly, all its output is freely available, whereas only limited output is available free from some of the other models.  So you see more output from GFS on here for those reasons.  

 

[snowlessayr]

34 minutes ago, Weather vane said:

Although I am new to posting here, I have been following this amazing forum for quite a few years now. One thing that has constantly piqued my interest is that when it comes to major negative comments about how a particular model is performing, it seems like >90% of these comments are directed at the GFS (very negative comments of ECM charts seem way less frequent), and almost always when the chart doesn't show what is being hoped for. I have to admit I find this complete dismissal bias somewhat surprising since GFS charts are posted all the time, and used as evidence to support an interpretation.

I have no vested interest whatsoever in any model; just trying to find my feet with regards to how they all compare, which ones perform better in certain scenarios, but less well in other scenarios. In the critical 7-10 day forecasting timescale that so many of us focus on, how often is the GFS model completely wrong or very inconsistent on a day-to-day basis, and how does this compare with ECM - is it that much better ? I imagine this is not easy to define empirically, though there must be studies addressing it.?

Not trying to upset the apple cart here, just trying to wade my way through learning about models and chart interpretation.

 

Well in verification stats, the GFS is usually always below the ECM. Only in very rare instances does it have a period where it verifies better. The UKMO is nearly always the best over the previous 2. I think the main reason that the GFS gets a lot of flack is that it does have some more erratic tendencies shall we say. What I mean by that is that it tends to flip a lot more than the ECM and the UKMO. All models have there issues, they are never going to be perfect as the weather will do what it wants to at the end of the day. Just always have a huge caveat for single GFS runs or any single model run. Always better to look for trends rather than individual runs.

Most in here Including myself can get too excited with 1 run, especially when looking for winter Narnia charts

 

[Methuselah]

The 12Z it's a'coming, it's coming round the bend; I ain't seen a snowflake, since I don't know when...?

[TSNWK]

36 minutes ago, Jon Snow said:

 

 

 

That’s a cracking day 9 ECM 0z ensemble mean for crimbo day!...fingers ...all this cold chasing is quite exhausting but I think it’s going to bear fruit during the second half of next week...hopefully!  

 

I agree on exhausting..;)

Own up? Who's gone to bed after the 18z and then find themselves up earlier than normal to catch up on any overnight comments from the 0z

I also enjoy  the 12z when there is interest sitting down for an hour with  a cuppa, viewing the models and laughing and eye rolling at the comments here.

And with that time to boil the kettle and buckle up for the 12s

 

 

 

[snowlessayr]

As much as I would love a winter wonderland white Christmas, I would more than happily take a frosty seasonal Christmas. I think most of us would after this year, nice to have something to look forward to. Hopefully we have some afternoon upgrades, and some fabulous posters recently.  Very helpful for people who like myself are still learning some of the ropes

[Nick F]

Afternoon all, quick round up of the morning output.

Good support from the 00z EPS clusters for a cold N or NW flow on Christmas Day as per 00z EC deterministic run.

00z EC det looking cold for Christmas lunch

GEFS support for flow N or NW on Xmas day despite the operational run-to-run inconsistencies - however, it does flatten out the upper flow from the NW/W thereafter ... so uncertainty whether the cold has staying power

Although yesterday's run, being the strat doesn't chop and change as much as the trop modelling, but EC shows at the 150 hPa level above the top of the troposphere, ridging appears over Greenland too, as waves develop at the base of the SPV, which may help in the troposphere.

 

[DavidS]

6 minutes ago, General Cluster said:

The 12Z it's a'coming, it's coming round the bend; I ain't seen a snowflake, since I don't know when...?

Gotta love a Johnny Cash reference 

[WxHerts]

10 minutes ago, Mike Poole said:

I think there are 3 reasons for this.  First, the evidence, from the verification stats the GFS is not as good a model as ECM:

At day 5 it is behind the ECM, UKMO and just behind the GEM (CMC on chart).

Secondly, it runs 4 times a day so gets more airtime.  Thirdly, all its output is freely available, whereas only limited output is available free from some of the other models.  So you see more output from GFS on here for those reasons.  

 

B..but I though ICON was best? Great analysis, I would add a fourth point in that ECM's solutions, even if very wrong, often simply appear less outrageous when put side by side next to GFS - as in I'd say it's much rarer for ECM to churn out those random, 50mb drop in 24h type bowling ball lows range that GFS is infamous for. Oh and speaking of ICON, why hasn't it rolled out yet?