Jump to content
Sign in to follow this  
Paul

Model output discussion - heading into 2018

Paul

Please only post model discussion in here. 
For more general chat and banter, moans and ramps, please head to the banter thread.
For general weather chat, please go to the regional threads.

Notice a problem with a post? Please hit the report button.
Thank you!

View the latest forecast models in the Netweather DataCentre

Message added by Paul

Recommended Posts

The models continue to show an exciting prospect.

The meridional / cold zonal will have ongoing snow potential over the next week. What will particularly wet the appetites of cold lovers, however, are the repeatedly clear hints of something much more exciting evolving in 10 days time. We're seeing this now across models and within runs and ensembles. Whereas the 0z operational had little support in FI, this is not true of the 6z which shows some exceedingly cold members.

I continue to be very encouraged. Yes yes yes ... it's FI. But I like the way this is evolving. It looks to me to be plausible. It fits the patterns seen so far this winter and there is good support within and outwith models. Keep hoping oh ye cold lovers.

5a4639d34270c_ScreenShot2017-12-29at12_45_31.thumb.png.679ac39d0e855371648701f0c16fef8a.png

5a4639ddae618_ScreenShot2017-12-29at12_45_47.thumb.png.e0443836b2b0b43c3759a8f58befcbaa.png

5a4639e7dbff7_ScreenShot2017-12-29at12_46_13.thumb.png.c558d919a5d1f2a5e4e0e4bb240f4747.png

Share this post


Link to post
Share on other sites
6 minutes ago, Timmytour said:

I think the reason for Nick's apprehension is that this time last year on too many occasions we WERE jumping up and down looking at charts even better than that at T180+

And we remember how it turned out

We didn't start the winter like this, though. I've seen two big snow events this side of New Year and I can't remember that happening for many many years. I'm not naive enough to think one colder month makes the other two cold, and there have been some infamous examples of that, but I do tend to think colder winters show their hand and start as they mean to go on.

A touch anthropomorphic but not without meteorological backing. The signs looks good to me.

Edited by West is Best

Share this post


Link to post
Share on other sites

npsh500_192.thumb.png.5513ef5df0d38cf299d0d638750c5daf.pngnpsh500_192.thumb.png.5513ef5df0d38cf299d0d638750c5daf.png

Edited by legritter
my post is further down the page .

Share this post


Link to post
Share on other sites
32 minutes ago, TEITS said:

Nothing to do with one upmanship. I just disagree with how Teleconnections are used on here.

Personally I feel Teleconnections are useful for seasonal forecasting. They are useless when I see posts trying to predict an E,ly for the UK via a Scandi HP. You simply cannot put that amount of detail for such a tiny Island like the UK. The same rule applies if you're trying to predict a N,ly or SW,ly.

Let me put it another way. Why is it at +384 the GFS runs can vary so wildly? The simple answer is despite these supercomputers working out a massive amount of equations, the runs end up different due to slight changes much earlier which only increase as the run progresses. The chaos theory will always be troublesome for long range forecasting. So if a supercomputer struggles with these slight changes then how on earth can teleconnections compensate for the chaos theory. The simple answer it can't!

So sorry but I would rather use seaweed than rely on teleconnections for the next few weeks. if someone made a post in the Autumn about the possible trends for the winter then I would take more note. So nothing to do with one upmanship. I am more interested in accuracy than who can make the most technical post by using as many abbreviations as possible!

Finally didn't you say in a recent post that "winter might be over"?

Good post. A cold forecast was made only to be replaced by a much milder zonal pattern but just like the NWP models, forecasts are changing all the time.  Trying to put to much detail into a long range forecast will probably lead to failure. Rather than trying to pin point where highs and low pressure areas for our little Island will be surely it would be better to give a broader brush idea of what to expect ie colder, milder, wetter, dryer, etc.

Their is amplification to be had in the northern hemisphere weather pattern over the next few weeks which may be conducive for us but as always the models are struggling with the amount and location of such.

Share this post


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

But height rises to the north east are a long way from a surface high cell plonked over Scandinavia. If care is not taken this will take on the same mythical properties of the now famed 'Greenie High'. :shok:

Bang on, at the moment we are just in NW/ SE territory......well not quite there yet but you get my drift 

Share this post


Link to post
Share on other sites

Look at these SLP ensembles for Iceland. Anyone dare make a forecast!

prmslReyjavic.png

Share this post


Link to post
Share on other sites

After New year there look to be a couple of days for snow chances via the last several runs off the GFS op. Around the 1st/2nd further north with a weak incursion of PM air, before a more significant chance around the 5th with a low/slider again bringing TM air against PM air. After that it does look as though Atlantic ridging and height rises over Scandinavia are trending more so than lower pressures, but as has been said the uncertainty and unpredictability of this winter's synoptics and variations has been high. From my point of view it is an extremely interesting time and one loaded with potential for snow and cold.

Share this post


Link to post
Share on other sites

*Teleconnections gate*

maybe you’re both right let me try and explain. How far out is the short term 2,3 or 4 weeks? And when does the long term start 2,3,4,5 or 6 weeks? There must be a overlap or grey area where one starts and one finishes so you can both be right. 

Ok I will get my coat 🧥 

Share this post


Link to post
Share on other sites

Considering what Teits was posting re uncertainty in forcasting in the next 14 days this seems appropriate. 

Fergie

‘Yes, interesting turnaround from ECMWF monthly model yesterday (frm previous +ve to now -ve UK temp anomalies through fair stretch of Jan wks 1-3). There's been similar to-&-fro in recent GloSea runs & given all manner of conflicting signals/output, extended range confidence low!’

 

Share this post


Link to post
Share on other sites
Guest

THE GFS LONGER TERM OUTPUT HAS ACTUALLY BEEN REMARKABLY CONSISTENT!!!!

As I just said in my very long post late last night (on page 47) I wasn't intending to post again until early next week but there have been various comments in relation to some of the longer term output which I strongly feel are unjustified and very wide of the mark. I will endeavour to address these now with the appropriate evidence.

We know that the GFS goes far deeper into FI territory than the other main models and certainly compared to the other two of the "big three" UKMO (day 6) and the ECM (day 10). This extended period exposes the GFS to far more variability and inaccuracies as it explores all the possible evolutions. When there are one or more broader scale pattern changes in the offing, the ensemble suite might well fall into two (or more) clusters. If these possible changes continue to be indicated with several distinct destinations, the following few operational and control runs may typically swing from one solution to the other. It's this variability which causes many model watchers to say that it's pointless to look beyond week 1 and to conclude "that the GFS is totally unreliable" or "it must be on some strong medication" or "the pub run special is completely out of kilter with all the earlier output" or comments to that effect. The criticism is usually far stronger when the most recent run moves away from a cold evolution to something far less palatable for the many "coldies" on this thread. I like the use of humour on this thread but there are too many times when the GFS is ridiculed unfairly which exaggerates its indifferent reputation.

When I sometimes look into the longer term output, I always do so cautiously and emphasise the likely unreliability. I feel that the problem is that some posters (and probably some readers too) use this extended period in the wrong way. Yes, it's great to post some day 15 "eye candy" charts but it's wrong to view these in the same way as one would, for example, a day 5 chart (I'm not just talking about "reliability" here). It's also wrong to dismiss the longer term charts as "utterly meaningless" or that "at the moment FI begins as early as day 4". Just because there are considerable uncertainties in the very short term does not automatically mean that one should conclude that "any longer term output does not have a hope in hell of verifying". What the extended charts should be used for is to see the variety of different outcomes that are being suggested. Then we can study the evolution to each of those scenarios. We can look for the timing of a change in direction - eg: a change to a much flatter pattern beyond day 9 (this is merely an example and NOT related to any of the current output - more on that shortly). This is particularly useful when this possible (or probable) broader pattern has been predicted in some of the longer term forecasts and/or by those using and interpreting the sophisticated background and teleconnection signals which drive all the model output. We can look out to see whether (or when) the models seem to be starting to "sniff out" and explore these changes and factoring in the changing background signals. Then we can look at later runs and see whether they stick with that evolution (ie: part of a developing trend), drop it temporarily but then bring it back again or drop it altogether. We can also look for the much shorter term model output (from all the models) to see if there are any hints supporting the longer term solutions offered by the GFS. As these possible scenarios move forward, we can see if they make it into the middle term (D7 to D10) and steadily within the scope of other models. Then this trend may (or may not) be picked up by some (or all) of the other models allowing us to assess the potential for and likelihood of a particular evolution.

In a number of my posts during the last month, I have completed some pretty detailed cross-model analyses where I have often examined three periods to cover the shorter, medium and longer term output. In the same (or a separate) post I've also examined some of the key factors that have been driving (or interfering with) possible broader pattern changes down the line. The real "winter on a knife-edge" crunch point has been getting steadily closer. This is the battle between a very disrupted tropospheric PV and the status quo - a return to more of a climatological norm with the seasonal strengthening of the PV and the jet stream (I've covered many of the factors behind this in some recent posts and again in my last post on page 47) but this has actually been very consistently signalled for quite a while by the "teleconnection brigade" (completely contrary to what a few posters have been stating). The likes of @Glacier Point, @chionomaniac@Tamara and @Catacol have frequently updated us on the changes, the timing issues and the downside risks with some very high quality posts and thorough explanations. They went out of their way to describe the the key "players" behind these outcomes which (very importantly were totally ignored by some) almost always kept reminding us all of the downside risks. I very strongly feel that, just as for last winter, they have been "extraordinarily accurate" in describing the two possible routes - it's simply not a question of what one might want to believe (eg: a change to a prolonged colder regime) it's seeing (with the benefit of hindsight) how accurate these predictions are and not taking any particular parts of a post out of context. In fact it's not a question of right or wrong or success or failure -  it's using the extremely useful guidance provided as a yard stick. I have based my "search for this potential colder regime" this winter on focusing on these factors and signals and I'm so grateful for these posts. There are a few who ask why I keep "ramming this point home". Well, I feel so strongly about this, that whenever I see unjustified criticism without any meaningful evidence, I feel that I must address the balance. I could easily provide some conclusive evidence here but I doubt that the moderators would allow me to disrupt this thread - if pushed might produce that type of post for the "moans" thread! If that tiny minority stopped this quite sickening practise, then I would very happily move on.

Rant over and back to the nitty-gritty of this post...the evidence! I will focus solely on GFS 12z runs. As @johnholmes frequently reminds us, it's far better to compare like with like. This is because different data comes out at different times of the day and certain updates often have a quite regular time slot.  For instance, the 12z might have some additional input that was not updated in time for the 6z run. I shall work back from yesterday's 12z run using the Meteoceil archive run facility. I begin with set dates at the same time (1300) and adjust these to 24 hours earlier from the previous day's 12z run. Obviously, the T+360 can only go back 24 hours to the T+384 to get the same date comparison. So for Dec 26th backwards, I show the T+384 charts for the "then" furthest date showing. This might sound confusing but it should become clear when you go through the charts below:

 GFS Dec 28th 12z:     T+24 for 1300 Dec 29th                T+120 for 1300 Jan 2nd                T+240 for 1300 Jan 7th              T+360 for 1300 Jan 12th    

                                         gfsnh-0-24.png?12   gfsnh-0-120.png?12   gfsnh-0-240.png?12   gfsnh-0-360.png?12 

 GFS Dec 27th 12z:     T+48 for 1300 Dec 29th                T+144 for 1300 Jan 2nd                T+264 for 1300 Jan 7th               T+384 for 1300 Jan 12th   

                                          gfsnh-0-48.png?12  gfsnh-0-144.png?12   gfsnh-0-264.png?12   gfsnh-0-384.png?12

 GFS Dec 26th 12z:     T+72 for 1300 Dec 29th                T+168 for 1300 Jan 2nd                T+288 for 1300 Jan 7th            T+384 for 1300 Jan 11th  

                                 gfsnh-0-72.png?12  gfsnh-0-168.png?12  gfsnh-0-288.png?12  gfsnh-0-384.png?12

                                       

 GFS Dec 25th 12z:     T+96 for 1300 Dec 29th                T+192 for 1300 Jan 2nd                T+312 for 1300 Jan 7th            T+384 for 1300 Jan 10th   

                                           gfsnh-0-96.png?12  gfsnh-0-192.png?12  gfsnh-0-312.png?12   gfsnh-0-384.png?12

 GFS Dec 24th 12z:     T+120 for 1300 Dec 29th              T+216 for 1300 Jan 2nd                T+336 for 1300 Jan 7th            T+384 for 1300 Jan 9th       

                                           gfsnh-0-120.png?12  gfsnh-0-216.png?12  gfsnh-0-336.png?12   gfsnh-0-384.png?12

 GFS Dec 23rd 12z:     T+144 for 1300 Dec 29th              T+240 for 1300 Jan 2nd                T+360 for 1300 Jan 7th            T+384 for 1300 Jan 8th    

                                           gfsnh-0-144.png?12  gfsnh-0-240.png?12  gfsnh-0-360.png?12  gfsnh-0-384.png?12

 GFS Dec 22nd 12z:     T+168 for 1300 Dec 29th             T+264 for 1300 Jan 2nd                T+384 for 1300 Jan 7th 

                                  gfsnh-0-168.png?12  gfsnh-0-264.png?12   gfsnh-0-384.png?12    

 GFS Dec 21st 12z:      T+192 for 1300 Dec 29th             T+288 for 1300 Jan 2nd                T+384 for 1300 Jan 6th    

                                  gfsnh-0-192.png?12   gfsnh-0-288.png?12   gfsnh-0-384.png?12

Looking at the charts above and viewing down the columns for 1300 December 29th (today) you can see the extraordinary consistency on the GFS operational runs from as far back as December 21st when that run was a T+192 chart. If you cast your minds back, some of the models were showing some disrupted PVs for the middle period (and the GFS for the longer term) up until around December 10th or so. Then all the models started to move away from this scenario and the later runs started to show far less disruption for much of the extended output for the following few days. Then the GFS was the first of the models which started to pick up a new signal for disruption again (or dropped another "interfering" signal). This steadily progressed from the extended period and into the mid-term period and several other models started to show this evolution again. Then the disruption with WAA into the Arctic from the Alaskan side actually verified! Then (around Christmas time) the models started to show far less disruption once more, with a steadily strengthening PV but again, GFS operational run lead the way just a couple of days later, by showing renewed disruption. I was busily posting regular updates from the GEFS ensemble suites which only dropped this signal on one run but then immediately resurrected it on the next run. Apart from that single run, there has always been a cluster showing this disrupted PV outcome, sometimes a majority and sometimes a minority of members. In recent days, the other models have been joining the GFS party again with a renewed impetus towards further disruption. 

Now, none of this means that the GFS is correct - in might be very wrong and leading the way would become irrelevant. My point is that the GFS has demonstrated some remarkable consistency. In fact, even the operational run has only shown smaller adjustments which is highly unusual for the extended period. Of course, as a "coldie" I'm hoping that we continue to a see thoroughly disrupted PV rather than a strengthening and more organised PV. The suggestion that a build of heights to our north-east is definitely NOT a new one and some (at times, many) of the GEFS ensemble members have toyed with this idea - just as @blizzard81 has pointed out this morning. There is still an awful lot to be resolved and to play for - how many times do we model watchers say that? LOL!!! I am definitely not writing off the chance of a more prolonged cold spell and while we continue to see some very mixed signals, I will continue my searches for colder patterns.     

 

Edited by Guest
Correct typos and check charts

Share this post


Link to post
Share on other sites
12 minutes ago, Big Rain said:

Surely everyone can express an opinion on a forum whether for or against teleconnections? As a long-time lurker I find the above post unnecessarily inflammatory. Teleconnections are often quoted on here and appear to be very often of no use to our small island in certain setups...there's also a huge lack of commentary on what controls/drives these teleconnections from certain posters...hats off to those who try and predict > 7-10 days out and who understand the drivers not just the measure of what a modelled activity may or may not equate to in terms of the weather in the UK. Do we really think convection on the other side of the world is that important? Our geographical position ultimately results in tiny features having a big impact and until a wedge/shortwave can be accurately forecasted more than a few days out there's no chance to accurately forecast mid-long term for the UK (which is what the vast majority of us are interested in - not the long wave pattern). 

But surely the long wave pattern gives you a "theme" and the short term models look to give the detail. I say let's have both, learn and move forward. I would agree that unnecessarily inflammatory comments are not helpful by anyone.

Edited by That ECM

Share this post


Link to post
Share on other sites

Well there are so many changes in the near term at the moment that it is quite staggering. The GFS now has a vicious low pressure crossing the country tomorrow night- further south on this run and if it occurred then it would be a pretty devastating storm for central areas. 

Share this post


Link to post
Share on other sites

A good start to the afternoon output from the ICON. 

http://www.meteociel.fr/modeles/icone_cartes.php?&ech=180&mode=0&map=0&archive=0

It shows an amplified Atlantic pattern and the jet stream digging into the continent leading to a weak easterly flow for the UK. Hopefully, this is a good indication for what the ECM will show later.

The METO update also talks about colder weather for next weekend. 

Share this post


Link to post
Share on other sites

Yes Steve, UKMO looks more amplified in 500mb flow at t+144 than GFS on 12z. Quite marked difference on wetterzentrale.

98AF143B-9C92-4F04-BA7F-1AA15B5DE260.thumb.png.60c4d6fbe7404d6cbdf07f6abbd0fac9.pngGFS

739FDE6D-DC98-4765-B0D1-9C711195FF96.thumb.png.3c993e85c235e9d5fa1480a6281e43aa.pngUKMO

 

Share this post


Link to post
Share on other sites

Ukmet is simpler and more amplified. However the angle on gfs at t144 isn’t to be sniffed and a few frames later will probably show some nice blocking the lp should pump up the waa readily on the gfs t144. 

Share this post


Link to post
Share on other sites
4 minutes ago, chionomaniac said:

Yes, everyone can express an opinion and I am sorry that you found mine inflammatory.

You ask "Do we really think convection on the other side of the world is that important? " and the answer is of course yes. I opened a thread many rears ago for the placement of teleconnective papers. For those inclined to find it, the studies have a wealth of information of how teleconnective activity is linked to the long wave hemispheric pattern. To highlight the point you make about convective activity on the other side of the world - just look at the MJO and how certain phases bring a better chance of Northern blocking. To be more interested in the shortwave pattern to me is akin to looking at how a pebble may make a splash in a pond whilst ignoring the boulder that has dropped in behind it

 

https://www.netweather.tv/forum/topic/73911-technical-teleconnective-papers/

Share this post


Link to post
Share on other sites

gem looks good but also further on is building a high pressure cell further on.

i suspect the ecm will be similar tonight.

gem-0-180.thumb.png.4f40fb49f2f741c5ecaf945eded2faca.png

gem-0-240.thumb.png.5d31b8cd371b9069b974c776ad84a2f6.png

worth noting 2013....... shallow heights can be affective lets see if under cuts happen.

 

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...