Jump to content
Sign in to follow this  
Paul

Model output discussion - Winter has arrived and the hunt for cold is on

Recommended Posts

3 minutes ago, Mike Poole said:

Nice try.

The last ten data points have nothing to do with the weather that is going on now, nor does  it depend the implosion of the election day snow storm for that matter, they merely rate to the performance of the the model on the previous days.  Here's the chart you posted, verification stats at T144:

image.thumb.jpg.95eda4b32884d47603e26cd8d8d68840.jpg

It s pointing to days before now, today, not the period were talking about.  And if you think that UKM score of 0.911 and GFS at 0.917 are statistically different you want to brush up on your maths, mate.

edit. Just like the election. FAKE NEWS. I'm so sick of it...

You are missing the point and I am probably not going to make myself clearer on paper so best to leave it. Cheers.

Share this post


Link to post
Share on other sites
39 minutes ago, Mike Poole said:

That is complete and utter nonsense, a few third significant figure points on a statistic at T144, you clearly don't understand the concept of statistical significance, and further you seem to keen to cherry pick evidence that supports an incorrect view. To clarify, as I follow these statistics regularly and UKMO on average verifies better than GFS at all of its range to T144.

Actually how are those verifications worked out? Does a computer calculate it got that  ridge right size, that low pressure the right shape, that high pressure in the right position, that low pressure the right intensity.

Share this post


Link to post
Share on other sites
5 minutes ago, Weather-history said:

Actually how are those verifications worked out? Does a computer calculate it got that  ridge right size, that low pressure the right shape, that high pressure in the right position, that low pressure the right intensity.

The ones I quote are 500mb height anomalies, computed across the whole northern hemisphere, so maybe iffy for UK, that is for sure.  Detailed methodology is here:

WWW.EMC.NCEP.NOAA.GOV

As far as I know this is as good as it gets for model verification...

Share this post


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

The ones I quote are 500mb height anomalies, computed across the whole northern hemisphere, so maybe iffy for UK, that is for sure.  Detailed methodology is here:

WWW.EMC.NCEP.NOAA.GOV

As far as I know this is as good as it gets for model verification...

Thanks Mike, so OK to analyse at a macro level but not when it comes to the a small area such as the UK. I thought that.

Share this post


Link to post
Share on other sites
Just now, Weather-history said:

Thanks Mike, so OK to analyse at a macro level but not when it comes to the a small area such as the UK. I thought that.

Well some may want to, but it isn't that, there just isn't verification statistics on local scales, the information isn't there, I can remember. @Steve Murr has suggested that more local verification stats used to be available, but aren't now. 

Share this post


Link to post
Share on other sites

Just a quickie from moi...Looking at the GEFS 12z mean it's a very unsettled (disturbed) outlook with temperatures predominantly below average across northern uk, cold enough at times for snow across northern hills / mountains but temps closer to average further south.  

Share this post


Link to post
Share on other sites
1 hour ago, Mike Poole said:

Nice try.

The last ten data points have nothing to do with the weather that is going on now, nor does  it depend the implosion of the election day snow storm for that matter, they merely rate to the performance of the the model on the previous days.  Here's the chart you posted, verification stats at T144:

image.thumb.jpg.95eda4b32884d47603e26cd8d8d68840.jpg

It s pointing to days before now, today, not the period were talking about.  And if you think that UKM score of 0.911 and GFS at 0.917 are statistically different you want to brush up on your maths, mate.

edit. Just like the election. FAKE NEWS. I'm so sick of it...

Can we keep politics out of it please, had enough of to last me a lifetime!:bad:

Anyway, maybe the model output that we are getting is fake too, and we are in fact just about to enter a New Ice Age next week as predicted by the Daily Express.😱

:oldlaugh:

Edited by snowray

Share this post


Link to post
Share on other sites

What on earth happened here? Really throwing out the toys because of the models verification? I honestly thought that judging on here that Kevin the carrot got sold out 😂 

I don't get all of the negative vibes tonight perhaps a sight improvement in the middle term especially up north 😉 

I think we need to be mature it's only the weather for crying out loud 

Share this post


Link to post
Share on other sites
20 minutes ago, Weather-history said:

Thanks Mike, so OK to analyse at a macro level but not when it comes to the a small area such as the UK. I thought that.

You can get a pictorial view of bias and extrapolate from that. Here is the D5 UKMO raw bias:

2016687348_Screenshot2019-12-08at20_11_18.thumb.png.5eced105b67c93e083d1cb5e5fff8ffd.png

We can see most of the NH has little bias, there are certain hot spots where we can see errors in verification, the UK being a high-level example. Therefore, say we have a NH verification of 0.91 at D5, we can be pretty sure that for the UK that is worse, maybe significantly. The above shows the UKMO was wrong with the heights around the Atlantic/UK based on the last 5-runs on the 0z. We can see the bias at T96:

1378820527_Screenshot2019-12-08at20_21_33.thumb.png.bba1fe4df87105dee792acbdf32fd526.png

The same area of bias but over the extra 48h that bias grows exponentially as highlighted in the first chart. This possibly explains why the last 5-days of verification (previous post) on the 0z and 12z UKMO forecasts have performed badly compared to normal, dragged down by amongst other areas its failure to handle the Atlantic/UK corridor (all models struggled here to varying degrees).

This is an example of that error via charts(UKMO):

6-day f/c: 1953021899_UN144-21(1).thumb.gif.790038127b507dd8dd8bac19b021af1b.gifCurrent 24h f/c:UN24-21.thumb.gif.b0538df4e85178f42950dc764ae57478.gif

The first forecast was for Dec 9 from 6-days ago, and the second is the forecast for that date today. We can see the UKMO was poor. For fairness the ECM:

ECH1-144.thumb.gif.2f00a5780232d9e85cc967538cc5e507.gif

We can see the UKMO and ECM were similar in the long-wave NH profile but UKMO failed over the UK/Atlantic. In this case my feeling is that UKMO loves to undercut and that is why sometimes it gets it right during those times when it is a possibility, but we forget the many times it undercuts when it is wrong, as per this example!

Charts for bias: https://www.wpc.ncep.noaa.gov/mdlbias/

Share this post


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

Plus its the NH..... 

UKMO will forever outperform the GFS in Europe > Another example today with ECM / UKMO identical at 120 V the random GFS.

Anyone putting GFS above UKMO for Europe has been wearing horses blinkers for years.


Where does that leave the ECM for us here in terms of performance? Somewhere between GFS and UKMO?

Share this post


Link to post
Share on other sites
3 minutes ago, MP-R said:


Where does that leave the ECM for us here in terms of performance? Somewhere between GFS and UKMO?

Up to D6 ECM was the best model until recently I'd say, in the last year or so it's been UKMO. Obviously D7-D8 it's the ECM, that's my experience anyway.

Share this post


Link to post
Share on other sites
41 minutes ago, Skullzrulerz said:

What on earth happened here? Really throwing out the toys because of the models verification? I honestly thought that judging on here that Kevin the carrot got sold out 😂 

I don't get all of the negative vibes tonight perhaps a sight improvement in the middle term especially up north 😉 

I think we need to be mature it's only the weather for crying out loud 

Good point but more importantly will it snow in Sheffield this coming week !!!😜😜🤑

Share this post


Link to post
Share on other sites

Let's get off of the verification subject and get back to model output discussion.

Thanks.

Share this post


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

Would that be this winter, only 8 days old, with another 82 days to go?  Which has already delivered a lovely cold start to December with widespread frosts for many, and an Atlantic storm now in the western approaches?  Some have been reporting hail and thunder in the south west today as well.  Goodness me, hasn't this been a promising start to the season?  No reason to lose heart so soon, my friend, just because we're not all buried in snow - yet.  Every year I see the best winter weather conditions develop almost overnight from nothing in the 5-7 day period so we should all keep looking because the next great blizzard may only be just around the corner!

Thank you. Safe to say the models are making a hash of anything past the 14th. Some building heights, some not. What appears to be a weakening PV over greenland. It's still game on. Plenty to keep positive about. 

Share this post


Link to post
Share on other sites

Amazing how the 18z GFS at192 hrs sends the jet towards the canaries and the UK stays on the mild side has the euro high makes its move!!!😂😂😂you coudnt make it up.

Share this post


Link to post
Share on other sites

Ok i have just put my 50p in the meter 

the fact is that no model will be accurate at five days,they might be near but not spot on,from five days> the goal post's get smaller and smaller as in terms of accuracy,so people righting off the rest of Dec and for the rest of winter needs to calm down because it's only 8th DEC,most cold spells happen after Christmas with the exception of 2009/10

right....back to the models 

the latest from the eps/gefs 500mb and 850's at day ten are in harmony with trough extension in the Atlantic exiting NE seaboard with heights eastern Europe/Western Russia joining forces with the Arctic high,gefs to a lesser extent

ecmwf-ens_z500a_nhem_11.thumb.png.bc1e9dfc08cd3c70fd6380221e0a502c.pngecmwf-ens_T850a_nhem_11.thumb.png.834faa9ae0650f4702c5357a3466ce72.png

gfs-ens_z500a_nhem_41.thumb.png.2b85297d8b59a9df89c8a840a3a93a1e.pnggfs-ens_T850a_nhem_41.thumb.png.103451d10bfe59c6caa6558373424f9a.png

EDH101-240.thumb.gif.5fe0633f38f56f9ac26ada4b0eba1c12.gifgensnh-21-5-240.thumb.png.aa87abbf19f3c2b5bd7b229f8f88e084.png

the cpc in the same room

814day_03.thumb.gif.04814c3845580e138b1069a39ab4ff00.gif

so an unsettled week coming up thanks to a strong jet extension firing lows at us but maybe signs of settling down later but that's not for me to judge right now

what will happen after next week?

we will never know.

 

Edited by Allseasons-si

Share this post


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

Amazing how the 18z GFS at192 hrs sends the jet towards the canaries and the UK stays on the mild side has the euro high makes its move!!!😂😂😂you coudnt make it up.

If ever i have seen a solar min vortex and jetstream signature that is it, just get the feeling this is gonna keep happening, seems like we will continue the hard luck stories unless we get a cross polar flow, only thing that will drive that is an SSW, forget that until mid Jan at the very earliest - gutted - so near yet so far.

Share this post


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

Amazing how the 18z GFS at192 hrs sends the jet towards the canaries and the UK stays on the mild side has the euro high makes its move!!!😂😂😂you coudnt make it up.

Horror show!😝

gfseu-0-210.png

gfseu-1-210.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...