[Model Output Discussion - Colder but how cold and for how long?]

5 minutes ago, makelikeaturkey said:

How do the models 'learn' from their mistakes? 

Sorry if this has been asked and answered before - I've been struggling to keep up at points over the last few days!

Hi Malarkey, unfortunately they don’t at the moment. this is something I mentioned recently about due to the change in the weather we can’t rely on algorithms from whether patterns from the past as much because the weather now seems to behave in a different way and pressure. Pens have changed a lot with more stubborn high-pressure systems to our south and Southwest, and sometimes stubborn, low pressure systems, even as well, I’ve noticed but much has been controlled by high-pressure the Azores, high Iberian, high Bartlett high, and Euro slug
there was a fantastic poster recently called midnight, ice age. 

He has a vast amount of experience with working in the industry, and he said the problem is that the software needs updating. Here was his post..

 

“You can have all the modern technology you want, you are trying to predict the future, that's simply impossible on a scale you're asking for  

I am afraid that we are now getting to the point that it is now the errors and assumptions in the software are causing the greater errors.

Until you get those sorted out you will just magnify the errors, and end up with even more chaos.

Spoken after 40 years in the software industry.

MIA

[Model Output Discussion - Colder but how cold and for how long?]

3 minutes ago, andy989 said:

I’d like to hope people won’t be lambasted in the future for simply talking about the possibility that the “nailed on 2010 repeat teleconnection porn fest” won’t happen. Even the more “respected” members jumped on people who discussed the possibilities that it won’t come off. Before I myself get jumped on, yes there is a cold spell happening right now, 2010 it is not, this is a post to say let people discuss all possibilities because all possibilities are always on the table. 

Hi Andy , I can see your point. I don’t jump on anyone personally all I tried to do was when all this started to be flagged up as a possibility in the early days, I just said to air with caution. For some of the members that have been around long time like myself, we’ve seen it happen all too often before things get watered down longevity decreased. Then, as this was mentioned to be more of a certainty and nailed on because supposedly all of the ensembles showed this myself and others were just saying be cautious as it can all change as it was quite a long way out as this has happened, multiple times before. I certainly wouldn’t lambast you or anyone as to be honest, nobody knows and that is the fact the super computers don’t know half the time, especially as the weather is even more unpredictable now than ever. I just like to try and save this appointment for a lot of people as it’s horrible feeling when everything is pointing towards cold. Weather only need to be let down so I always try and air on the side of caution until much nearer the timeframe. They certainly nothing wrong with discussing for some people like to discuss stuff that’s so far out which is fine as create discussion and I guess what the threads for models discussion, but a lot of people including myself have the child within us of cold and snow excitement. I always just felt it was better to wait until it seems much more of a very high probability within a few days as opposed to a week away or 10 days and More where so much can change has has been proven so many times.
nothing gets nailed on in my opinion, especially a week out, despite what all the ensembles would say, and then GFS came along to show either a breakdown or mild weather, more unsettled which seemed to pick up on that. unfortunately, it seems to me that the only thing that forecasters some super computers can nail on is wet and windy weather when we are zonal. 
winter weather is so fickle and difficult to protect as the weather literally can come from any of the four directions north south east or west, especially at this time of year, although rare from the east but can happen. that’s the issue with being an island next to a giant ocean.. 

so all I say to save disappointment is when weather charts start showing cold, 12 days out 10 days out just be cautious. Nothing is nailed on even at a week out. 
 

[South West and Central Southern England Discussion - Dec 2023 onwards]

Funny how they have said it’s going to get much colder from Sunday and yet today and this week has gone. Seems like it’s been much colder than what is coming

 

 

[Model Output Discussion - Colder but how cold and for how long?]

6 minutes ago, MJB said:

You clearly have a lot to learn, people chase snow, they don't come on here looking for a frost lol

Most are looking beyond this week because there are no snow depths to talk about, no chat of warnings etc . Searching the next signs of a block.

Will it be cold yes, will there be many Ice days ? I doubt it, will most of the population see -10 at night ? I doubt it. 

Reading the MET update - Towards the end of the week or more likely over next weekend, the start of a transition back to less cold conditions is likely as Atlantic systems start to arrive from the west brining spells of rain 

So I doubt very much models are overdoing zonal flow....................as much as it pains us.

I do agree with you. It does to me look like zonal and possible dry are times of the Bartlett comes far enough up to influence otherwise it will be back to Wet n Windy not what we needs with Thames and many other rivers levels so very high due to recent high rainfall totals 

[Model Output Discussion - Colder but how cold and for how long?]

17 minutes ago, MKN said:

Hopefully this week has acted as a learning curve to alot of a I assume fairly new followers of the various model outputs. In the UK away from Scotland, charts showing cold and snow at 120-200 odd hours away are just pure fantasy. Unless the charts which are showing cold and snow are within +48 to +72 they are far from reliable. It's one area that even after all the years of model development they simply are not that good at getting right. I Mentioned the same at the start of the week incase people assume I'm just saying it now. Anyway fingers crossed for one or two surprises and if not this week there's still a good 6 weeks left for another chance to come along.

You’re not the only one ,myself and many others have always expressed a little caution so far away. Just the nature of our location, and also the very fickle nature of the weather now more so than ever. 

[Model Output Discussion - Colder but how cold and for how long?]

15 minutes ago, Kentspur said:

GFS spotted it again the trend finder model

Echoed my thoughts, as when all the ensembles recently were pointing to cold prolonged period of weather, the GFS was the first one to start indicating a breakdown and getting earlier and earlier on in the period this has done the same thing on several occasions has been the first one to model or breakdown, and then the others start falling in line with it

[Model Output Discussion - Colder but how cold and for how long?]

24 minutes ago, booferking said:

It seems just above zonal wind speed is enough to set the zonal express train kicking in happened last cold spell to back in the start of December everything looked rosy but quickly turned on its head like a flick of a switch.

Ps Maybe it will all switch back tomorrow and we see these wedges that's been talked about yea never know  

Unfortunately, with the chance, this happens all too often again, it always gets water down and longevity shortened

[South West and Central Southern England Discussion - Dec 2023 onwards]

2 hours ago, Dorsetbred said:

Met Office update today

Good luck to you all down on the south coast. You’ve had some good luck recently with snow, the other day, Salisbury, Downton and many other areas.
looks like you could all get hit quite nicely again down there so good luck

[Model Output Discussion - Colder but how cold and for how long?]

1 minute ago, Severe Siberian icy blast said:

But are you even taking the time to talk about a chart so far away?! 

Surely the focus has to be on the reliable and looking for hints of change or not in the early stages. 

Posting a ppn chart 9 days away is as much use as trying to cook without a cooker. 

Excellent post it seems this happens far too. Often the charts are just so unreliable that far out. Can’t even be an indication as they are so topsy-turvy even in shorter time frames of 5 to 6 days can be so inconsistent, let alone nine days or above completely agree with you.

[Model Output Discussion - Colder but how cold and for how long?]

24 minutes ago, Midlands Ice Age said:

Derecho...

30 years of modelling, programming , analysis, technology design and implementation has left its scars on me. 

Twenty years ago we needed more machine power.

Now, unless we sort out the assumptions  and get them analysed correctly into mathematics  we will only magnify the errors as we push ever deeper into being able to look at ever more local detail.

Tamara is showing where some of the current assumptions are wrong, or at least give a totally different view. .

People on here showed how the stratosphere could affect things greatly. (10 years ago) This was incorporated into the models (though I believe is now possibly over responsive (see below)).

The only people to gain by increasing the computing power right now are 1) computer salesmen, and 2) system design specialists who have no further ideas as to how to improve the current models.

Incorrect assumptions ALWAYS increase volatility in output. Increased  computing power will only serve to show up any design errors.

I appreciate that increasing the spread by the changing the input criteria, is one method of reducing any errors. But the initial 'errors' are still present, and will only serve to produce more chaos.

I am talking about from where the models are standing today (which is much better than they were 5-10 years ago).  Spending hundreds of millions on more computing power right now will not give the same rewards as ensuring some of the 'unknowns' are fully researched and included  as regards the accuracy of the models..

1)     Extra notes.... Take today and this graphic taken from the above (to be saved)....

  

The whole of even  the western hemisphere has been impacted by the the slope and shape of the high forecast to develop  over  the Bering Ocean.  Changing its inclination or intensity will dramatically change the forecast for the whole Northern Hemisphere via mechanisms previously discussed on here. Both the ECM and GFS have shown this happening over the last week. Both had to withdraw from their positions . Something is not 'spot on'.

Will they this time? Can you tell me? I assume that this situations has come about as to their incorrectly handling the totality of the 'telecomms' signals. 

2) As to whether the models are accurate from a scientific pov.    I cannot make any comments  (you are correct).  However they have now released documentation on the 'CC' models, which were based upon forecast models  (I  am told),  and for the previous release of them there were still around 25 assumptions which were not yet proven. If they were missing in the 'CC' models then I assume they were not in the forecast models.  - In the CC models they substituted parameters. 

Sorry to mention the XX word banned on here  - MODS.

So IMO opinion,   (FWIW)  there is still quite a way to go before we can say that all the assumptions (and therefore the absolutely correct science) built into the models, means that models are correct.

Until that happens  putting in more computing power will increase chaos (not decrease). that is what my 30 odd years of experience has taught me.

MIA

Fantastic post highlighting improvements that can be made with modelling charts. Absolute pleasure reading your posts backed up by all that experience.  Thank you for your input

[Model Output Discussion - Colder but how cold and for how long?]

9 minutes ago, Midlands Ice Age said:

I am afraid that we are now getting to the point that it is now the errors and assumptions in the software are causing the greater errors.

Until you get those sorted out you will just magnify the errors, and end up with even more chaos.

Spoken after 40 years in the software industry.

MIA

 

Brilliant post, and thank you for confirming all your experience in the industry as that is gold to me and the fact you have acknowledged about the changes and also what needs to be done to prevent this happening so much i.e. software issue. 

[Model Output Discussion - Colder but how cold and for how long?]

7 minutes ago, Frank Trough said:

it may just be my perception, but is this lobe of high pressure starting become more of a feature? 

i guess maybe just GFS. Would like to see that undercut

 

edit - beaten to to it! 

You are not wrong at all this is become a much more common feature on the pressure charts. I’ve discussed this recently and another couple of others. have also agreed and a couple of posters I’ve been really informative of putting links in with more detailed information and pattern changes that have occured with the subtropical area moving slightly further north, and also the increasing intensity and size in these high-pressure systems

[Model Output Discussion - Colder but how cold and for how long?]

Thank you Blast,

 finally found the MetOffice charts

[Model Output Discussion - Colder but how cold and for how long?]

3 minutes ago, BLAST FROM THE PAST said:

You looking at the same UKMO I am?  

T120

BFTP

Hi Blast,

Can you post a link for me please to that graphical chart UKMOffice chart. Thanks 

[Model Output Discussion - Colder but how cold and for how long?]

45 minutes ago, northwestsnow said:

Yep totally run of the mill January weather..

5 days of temps struggling to get above zero and 5 nights of -3 /-6 

UKMO looks cold throughout

Hi Northwesr snow ; 

are these charts you posted from  UKMO ( Ukmetoffice) ?

if so can you post a link to these please. I asked recently someone kindly responded but gave me a link to fax charts from the metoffice. 
 

Can you give me a link please to the other charts ( the graphical ones you have posted from the metoffice- UKMO) 

 

thanks 

[Model Output Discussion - Colder but how cold and for how long?]

21 minutes ago, Malarky said:

PLEASE bring it home UKMO. Bank x100.

Could you post a link for me please where i can view these met office charts. Thanks 

[Model Output Discussion - Colder but how cold and for how long?]

Just now, MidnightSnow said:

With the colder temps filtering further south on this run... I'm wondering if this will have any impact on the midweek low? i.e. Does it increase the likelihood of the low heading into France, or does it make no difference?

I think there’s going to be rain and sleet, looking at the charts for much of southern England, but most probably snow from Wales all the way across Middle England over to East Anglia
I won’t be surprised if anything below that has settling snow

I think more likely to be sleet and cold rain across the south

[Model Output Discussion - Colder but how cold and for how long?]

1 minute ago, Ali1977 said:

Way colder than the apps are saying, I’ve day there for most

You’ve just hit on a really good point though Ali,

I noticed quite often the apps undercook the temperatures in summer by 3 to 4° at the beginning of the week when a hot spell has been forecast for the week they raise it by a degree each day and in the winter they seem to put milder temperatures on the apps compared to the actual temperature. I don’t know why that is? Any idea? 

[Model Output Discussion - Colder but how cold and for how long?]

Have heard a lot of talk about charts bringing in colder weather on Sunday but blimey down where I am. It’s been bitterly cold here already all week although today is a fraction milder if you can call it that but it really has been freezing cold!  So are we talking even colder air coming down from the north from Sunday onwards? 

[Model Output Discussion - Colder but how cold and for how long?]

Just now, Nick2373 said:

Often cloudy across clearing central and southern areas at first. Thereafter, turning colder from the North, with brisk northerly winds likely developing widely across of the UK, bringing a risk of snow showers, most frequent across the north. Temperatures remaining cold, and a marked wind chill especially in the north. There is risk of unsettled weather pushing in from the south through this period, which could lead to a band of snow and sleet where it meets the colder air across the country. Confidence is low with regards the timing of the arrival of any such disturbance, but there is an increasing risk of something potentially disruptive at some point in this period. Widespread frosts continue to be a feature by night, with a risk of ice in places.

Updated: 14:00 (UTC) on Wed 10 Jan 2024

Thanks for posting this Nick, but we should all be careful as they said the confidence is low.

it’s like looking at charts in FI 

It’s still a long way off and the confidence is low, would be great if it comes off, just have to be a bit cautious really. 

[Model Output Discussion - Colder but how cold and for how long?]

8 minutes ago, Nick2373 said:

Wow just read it. 

Hi Nick

Can you post a link please to it? Thanks 

[Model Output Discussion - Into 2024]

Just now, January Snowstorm said:

And that's the headline really, the block  all but gone by Day 11!! All decent cold spells must maintain heights either over Greenland or Scandi/Iceland... still it's day 11 so all can change

Will probably just become back to normal mobility with a bias towards Western based driven low pressures from the Atlantic

[Model Output Discussion - Into 2024]

4 minutes ago, Richard Fisher said:

Why is the border between snow and rain always the M4? Is this by chance or is there a geographical explanation why this happens a lot?

this is so true, even the weather people local presenters would often say M4 corridor, and quite often it would be snow to the north of this and rain in the south quite often

[Model Output Discussion - Into 2024]

22 minutes ago, GSP said:

What has been startling is the loss of oomph from the cold from the north.

It did seem nailed on there would be an artic sweep well through the country, now at best we are relying on developments to the south to dictate what happens.

This is a good point about what you said about it being nailed on. There were a lot of comments about ensembles all pointing into the same direction that it was nailed on despite but GFS had wobbled and gone the other way, but turns out it looks like GFS could be correct.

I think what needs to be addressed, is that when all the ensembles are nailed on what are driving factors for such dramatic changes to something that was to a dead due to the ensembles.

I always say when it’s so far out, there’s always gonna be an inconsistency and chance that it won’t happen despite what ensembles say as the weather seems to be so fickle now super computers and symbols. Everything just seems to struggle much more now than ever.

not a criticism of anyone or computers it’s just the fact that it’s extremely difficult and I don’t know if there is anything or will be anything around that will make it more accurate, especially at this time of the year.

when we get locked into zonal patterns, they do seem to be easier to predict with low pressures just rolling in from the Atlantic non-stop at times.

and in summer high-pressure generally seem to be easier to forecast.

I think at this time of year, with so many different weather types possible from all directions, it just makes it so much more complex with all the other added factors

 

[Model Output Discussion - Into 2024]

6 minutes ago, Mr snowman 2018 said:

It’s not negative at all I’ve been following charts for 15 years and I know where this is going we all get our hopes up every single year every time we see decent outputs but it gets watered down and eventually we get nothing we haven’t had anything significant UK wide since  2018 and I can’t see that changing either 

Completely agree with you I’ve stated that several times before that everything seems to get watered down more often than not.  But being constructive moving forwards, I would really like to see the amount of times these charts have gone wrong fed into a system That could potentially improve accuracy or reduce the amount of times this happens. I don’t just mean because people like myself want to see some cold and snow. I mean from an interested point of perspective whether it’s rain heat thunderstorms snow if there could be a way to identify when something was forecast did not happen what was the outcome and keep feeding those patterns in to system so for long-term benefits I wonder if then the chances and ratios could be increased for accuracy no matter what the weather is
. 

I know it’s becoming extremely difficult to forecast the weather as with Arctic high-pressure systems in a mess and on a day-to-day basis seems to be difficult to forecast what’s going on up there also combined with a warmer sea temperature than normal and also the fact of subtropical, high-pressure belt slightly further north and also expanding and strengthening Hadley Cell. 
weather pattern seem to be slightly more clearcut years ago, whereas now everything does seem to be in much more of a mess, so if they could start feeding in to a system all the predictions all of the outcomes to then improve pressure charts, moving forward.

I would also like to see if there’s a possible way to track underwater volcanic activity and also the continuous streaming of underwater vents. these release fast amounts of heat and they have to also play apart in Ocean temperatures so I wonder if there’s a way to integrate this information as the sea temperatures also affect the weather conditions and weather systems. 

 

 Just an idea.