Archived

This topic is now archived and is closed to further replies.

johnholmes

GFS Checks For T+168 Hours

150 posts in this topic

Hi

This is the place I will put the checks on GFS T+168 hours. My first one seems to have gone adrift in the new forum, never mind. The next one is due tomorrow afternoon. It was started on 31st October 2004 and is the check on what GFS was suggesting for tomorrow, Sunday at 1200Z 7th November. I hope to do this most weeks and so build up a picture of how GFS treats differing weather patterns over the next few months. The next one, starting tomorrow could be interesting as GFS goes for a cold Northerly, time will tell.

Share this post


Link to post
Share on other sites

Hi

At the moment there is a problem, well I have one in that I had hoped that anyone who wished could call up any of the GFS progs and see the charts for themselves. I'm unable to do this today, so anyone who has ideas on how it can be done please e mail me with your thoughts.

So for the time being you will have to take my word for what developed over the past week with the GFS runs from T+168 hours down to T+00.

This is the T+00 Sunday 31st October 12Z chart for the start of the T+168 houir, out to same time Sunday 7th November 2004.

Uk under large area of HP, centre>1025 WSW Eire(150+miles):

First run: This is the T+168.

and shows a flabby high from the central Atlantic over the SW approaches and out to southern Sweden, central pressure there >1030. A trough is evident off to West of Ireland.

2nd run is the T+144

This shows the HP >1030 centred over southern UK, with east/west extension into Atlantic and south Sweden again; little sign of trough to west of UK other than very shallow feature into Hebrides(centre of low <970 off NE Greenland.

3rd run is for T+120 hours

Similar position and centre but with some signs of ridging to Greenland high which was not there on T+144, lows below 965 off SW Greenland and one near Norwegian coast(on Lat with N Iceland), centre <980.

4th run is for T+96 hours

UK with HP, >1030, just to sw of UK. Ridge aligned e-w with minor troughing off nw Scotland. Low<975 off sw Greenland with another <980 near Jan Meyen. Slack wnw’ly over England and Wales, with somewhat stroner w or wsw’ly over much of Scotland

HP similar value and similar position. More noticeable trough into nw Scotland from low(<970) off nw Nowray.; HP>1025 over northern Greenland with another low, <975 off sw Greenland. Slack w-nw flow England/Wales with mod-fresh w-sw over much of Scotland.

6th run is for T+48 hours

very similar except the trough is now forecast to be down the east side of Scotland, into the North Sea. Everything else looks very similar to the previous run.

7th run is for T+24 hours

very similar to T+48.

T+0 is the actual for Sunday 12Z 7th November 2004.

This shows a very similar pattern, bearing in mind how far ahead the original forecast was. High pressure to the wsw of Ireland, the trough it had into the west of Ireland is in fact off the nw of Scotland, otherwise its a pretty good forecast in my opinion.

This is the 2nd good one, although both were fairly 'simple' weather types.

Its first big test will be the next 7 days(T+168) from the 12Z today. Its predicting, as everone knows on this forum, a cold, if brief, northerly.

Hopefully I shall have sorted out the problem in being able to call up each chart.

many thanks.

hopefully this will have the two pics to view I suspect from another window!

Share this post


Link to post
Share on other sites

thanks John,I think GFS has done quite well this week with getting todays weather correct.

as you say the first real changllange will be this northerly next Saturday.

Share this post


Link to post
Share on other sites

help!!

Paul or anyone, not any good, so can a Mod pse delete the last post from me please. Back to drawing board!

Edited by JACKONE : Last post deleted as Requested.

Share this post


Link to post
Share on other sites

(scarp that,I see what you mean now!

Share this post


Link to post
Share on other sites

Hi again, here is this week's check on the GFS, hope you all enjoy.

Share this post


Link to post
Share on other sites

hi again

not quite sure why its appeared like that, can the mods sort it? If not I'll try and sort it out, but hope it does not detract too much from reading it.

regards

Share this post


Link to post
Share on other sites

Hi again

Just trying to be too clever with the first one. Moderators/Admin would you please delete the first one?

many thanks.

Hope you all enjoy and it stimulates comments.

Summary of how the GFS performed from the T+168 hours based on the 12/18Z chart for Sunday 7 November

Centres of pressure are in millibars, LP=low pressure and HP=high pressure

It started with, at T+00, with HP, just off sw Ireland >1030, slight ridging into near continent; low <970 off sw Greenland with marked trough to its south; second low <960 off nw coast of Norway

The T+168; the first forecast run and the one being checked showed;

HP, still > 1030, but further sw of Ireland than at T+00; low <1000 off sw

Greenland with low off west Norway <975 troughing back to north of Iceland; slackish wnw flow much of UK but fresh

in far north. . The original plunge of cold air into Europe produced a classic low in the Mediterranean.

During the various runs from T+168, 144, 120 etc down to T+24 the HP off sw UK remained in much the same place and almost the same central pressure. The various lows, off Iceland, north Russia and in the Med did change position and depths, but not by more than relatively small distances or central pressures. There was also an increasing tendency to increase pressure over Greenland through the runs. By T+24 this was over 1040 over central Greenland.

The actual today at 12Z gave; HP mid way off sw coast of Ireland and Brest(France), the Icelandic low, split in two, both <980, and both north of Iceland, the Russian low, <975 and well off the north Russian coast, with the Med low <995 and just south of the toe of Italy.(Ref my post on the Mistral last night). The UK being in generally light to moderate w’ly flow, stronger in the nw.

My Summary

Yet another very good forecast from GFS, all the centres being predicted, except the high over Greenland. This makes it the third consecutive ‘correct’ T+168, in my view.

Anyone like to comment please?.

NB: I am away next weekend so am unable to do a full round up on the probably interesting developments, but I will do a resume next Friday, which may be ‘crunch’ day for the model in terms of do we get a northerly or not. Normal service will resume on Monday 22nd November 2004!

Enjoy the weather whatever.

Share this post


Link to post
Share on other sites

Thanks John, an interesting read as always!

Share this post


Link to post
Share on other sites

another attempt to load it!

sorry cannot get the T+168 and the T+00 for comparison to attach or load??

Once again all will have to simply take my words.

Summary of the T+168 starting on Friday 12th November 2004hours to Friday 19th November 2004

We started with the coldish set up of; the HP to nw of Spain, LP Norway with a new LP formed se of Greenland. The UK, at least the southern half in a milder wsw flow, much of Scotland still in a coldish COL.

The first two runs, T+168 and T+144 and to some extent the T+120 each varied with a cold to mild scenario. Since running checks on the GFS it’s the first time I had seen this. Not sure what to make of it.

By T+120 it had ‘settled’ to the following:-

HP>1025 on north coast of Portugal, with HP over Greenland >1030. LP off south Greenland <980 with a secondary to its south of <995. Low on borders of Baltic states and west Russia <980, giving a northerly flow over the eastern half of the UK with a COL over the rest of the UK.

The T+96 had, a rather flabby HP from Spain, >1025, all the way up to what is now given as the main centre >1040 over Greenland. LP off southern Greenland <985 with trough down to south, well down into Atlantic, the other lo(s) now triple centered; <975 northern Russia, <995 off nw coast of Norway and <980 eastern Baltic states. The UK, as before, east in rather strong nnw’ly with the rest in a COL area.

On the T+72; has the same HP centres and positions, the LP off southern Greenland is more complex, main centre <985 off se Greenland with secondary about 400-500 miles south., other LP’s about the same. Over the UK the northerly is rather further to the west with only the extreme wsw in a small COL area.

The T+48 shows; is pretty similar to the T+72 in almost all major centres and the flow over the UK.

Down to T+24 and; again much the same for the centres and the UK, quite consistent.

Then to the T+00

Like I said in my preamble it’s the 06Z not 12Z as I am just about to set off into snowy Derbyshire for the weekend! It could be summed up in very few words, very accurate.

HP over Greenland >1045 with slackish ridge down to another high off Portugal >1030. LP’s south of Greenland <985 and its secondary to the south <985. Three LP’s to the east of the UK; <990 se Germany (our LP from yesterday), <985 off northern Norway and <980 in Russia. So the UK in a moderate northerly, all of it.

Only the extent of the northerly over the UK was not well forecast once the GFS settled down. Just as a matter of interest over several runs around T+24 and out to T+48 BEFORE the event of yesterday it had an uncannily accurate picture of the events that actually happened.

So once again, aside from its initial hiccups around T+168 and T+144 it’s a very good forecast for 7 days in advance. The next one for checking is the T+168 12Z start time is 12Z Monday 22 November 2004.

best wishes

John

Share this post


Link to post
Share on other sites

Thanks John,

Very interesting to read, and notable that GFS had a bit of a "wobble" this timel. :mellow:

Share this post


Link to post
Share on other sites

This is the 6th check run on the GFS T+168 hours. It starts on Monday 29 November and is valid for Monday 29 November at 06Z

The first chart is T+120 which is when the computer settled down to a consistent ‘story’ the final one is the T+00(see subsequent post)

The T+168 has HP from Newfoundland, >1025 into towards sw approaches all >1025. LP’s are <985 west of Greenland and <990 just east of Greenland, with other main LP <995 west of Paris and also <990 north Russia with another LP <1005 north of Mumansk. The UK in cyclonic east/ne flow, strong in south with ridge trying to push into west Scotland. All UK at 850mb <0C with small area se Scotland/ne England <MS05C.

By T+144 the centres were HP >1030 centred about 450 miles wsw of Eire with a ridge towards the sea between Greenland and Newfoundland; LP’s west of Greenland and wsw of Iceland with trough into northern Ireland, <980 with LP over north Russia <1005 and another LP (the one near Paris) is now just off the east coast of Italy <1005. This leaves the UK in a sw’ly over all areas.

Similarities to T+168=not much especially the UK!!

T+120 shows the HP now about 48n 40w, >1035 ridging into Greenland. LP sw of Iceland <990 with twin LP off north Norway/Russia, both <990, LP<1005 over western Germany. Slack ridge over sw UK Col for rest. Fairly similar to T+144, so, maybe, its settling down (T+120=5 days)

T+96 and it does seem to have settled down. The only real difference from the last 2 runs is the low over Germany is now over north Italy, still, 1005. The isobar pattern over the UK is almost identical to T+120

T+72 shows, at the surface, a very similar pattern with just the ridge over the UK rather more pronounced. This gives more of ridge conditions over most areas with a sw’ly over nw Scotland.

T+48 and T+24 show almost identical charts in all major aspects.

Finally to T+00 which gives an almost identical final ‘picture’ to what has been consistently shown from about T+120 hours.

So to sum up: - Yet again the GFS almost a week out has given a pretty accurate surface and 500mb pattern. To compare the two charts T+120 and T+00, I'm sorry but you will have to refer to my second post on this. My ability to post other than text remains zilch!!

I am becoming a convert to the GFS from about T+120 hours down. Beyond that, the jury is still out. It does sometimes give a first indication of a change occurring but only a pattern. Each run (for the same time validity needs looking at to see if the trend is confirmed). It is not very productive if each successive run at 00Z, 06Z, 12Z and 18Z are compared. Part of the reason for this is the data input at each of those times has a weighting which varies for different parameters for each run (see an earlier post I made to comments – ‘model discussion thread part two’ post 155)

Thanks for reading this post, and for any comments.

John

Share this post


Link to post
Share on other sites

Hi

Hopefully here with the two charts, first one is T+144 and the second is T+00 today at 12Z

Share this post


Link to post
Share on other sites

i give up both are for the same time. just read the words I'm sorry about this. I've spent many hours with a great deal of help from Paul but its still wrong.

John

Share this post


Link to post
Share on other sites

the next check will be valid for 12Z on Monday 6th December 2004.

Share this post


Link to post
Share on other sites

Thanks john, an interesting read and one I imagine must've took some time to do.

:D

Share this post


Link to post
Share on other sites

Hi Essex

Thanks, making the notes each day is no problem its just trying to get the d****d thing presented as I want that is so infuriating, even after all the help Paul has given. In future it will be without any attempt to put any charts in the text - never mind.

Share this post


Link to post
Share on other sites

Hi again

Here with T+168 Mon 6th December 2004; the 7th check on the GFS

Once again no pics, Ive given up with that. I do have copies if anyone wants an e-mail of them.

This first run had a ‘flabby ridge of HP from the Black Sea, >1030 out west and sw off the coast of Portugal. The main LP’s are <960 west of Iceland and <985 off nw Norway with a LP in the Med near Corsica, <1015.A fairly minor trough also shows off Newfoundland. The UK under the HP with a sw’ly over much of Scotland.

During the next runs, at T+144 and T+120 GFS suggested the HP splitting into two, then just the one giving a fairly mild and breezy wsw’y flow over the UK.

On that basis I did a brief outlook for where I live:-

dry with variable cloud and a light s-sw breeze. No frost with temps ranging from about 7C overnight to 11-12C by day.

The actual was:-

dry, about half cover of cloud, day max of 11.8C and a night min of 4.8C

The remaining runs down to T+00 showed not much change to the flow over the UK.

Now, in my view, assuming my forecast ability is not wildly in error, apart from the overnight Min, then to be able to get a forecast pretty near to what happened at T+120 hours, which is 5 days out is not bad. Yes, it would be great if it had been so at T+168, 7 days, and even better if it always allowed that accuracy at T+384, which is 16 days in advance, would be, well beyond description. Still I am getting more and more confidence in GFS at T+144 and nearer. Beyond that as I have said several times its not ever so good. However, it does sometimes pick up general trends, even when they are very different from what one might expect.

On the broader scale it did have some changes to both LP and HP in different parts of the chart, but they did not materially alter what one would expect the weather to be like over the UK.

Thanks for reading this and I welcome any feedback.

The next T+168 starts with the 12Z today, so for 12Z on Monday 13th December 2004. What does that show………………… sadly nothing for snow lovers to get excited about. Just a minor Atlantic trough easing through with HP over Europe and north of the Azores re-establishing itself!!

regards

John Holmes

Share this post


Link to post
Share on other sites

Personally, I believe the GFS to be the most accurate of the models these days- due in no small part to the improvements made in the model this year, for last winter I trusted the UKMO and ECMWF more.

These interesting summaries do nothing to change my current view, in fact they reinforce it- I haven't seen any evidence that the other models are this reliable for 4-7 days out.

Share this post


Link to post
Share on other sites

Thanks John, excellent summaries. :)

I agree Ian, the GFS is the best performer really for the 4-7 day period IMO.

Share this post


Link to post
Share on other sites

Hi

part of this is from the Christmas 2004 forecast but applies to this thread.

On the other thread I am running, T+168 checks. Its almost certain that the 8th consecutive run over the past 8-10 weeks is going to show a pretty close pic to the actual. The next one could be even more interesting as it will go to 20 December.

I'll post this also in the other thread. Mods if this is wrong pse delete just one!

thank you

John

Share this post


Link to post
Share on other sites

Wow.

It seems quite consistant for T+168. It will be quite interesting when (or if :) ) the cold weather enters that time frame.

Thanks for the update John.

Share this post


Link to post
Share on other sites

Thanks John.

The GFS is doing quite well really at the T+168 timeframe.

Share this post


Link to post
Share on other sites

Hi folks

Well amidst all the mounting excitement over will it/ won't it be white at Christmas, let us not forget what enables us to even begin to try to forecast that far ahead. I give you, ladies and gentlemen, the GFS T+168 forecast check.

This one runs from the start time of 12Z Monday 6th December 2004 to T+168 on Monday 12Z 13th December 2004

I suppose it is possible, if you want to, to 'nit-pick' at this to find some fault. Personally I find it very good at this time frame for reasons I've regularly posted about.

The wind flow, even its main source, and the flow strength, are all accurate by my measuements.

Hopefully by clicking on the attachment below you will be able to see for yourself what I mean.

So I now have eight checks on GFS T+168 from late October to now, most with differing synoptic start situations, and all accurate.

The next check will have to be over 144 hours as I go away next Wednesday morning, it will run from 12Z today until 12Z Tuesday next week.

regards

John

Share this post


Link to post
Share on other sites

okay Mods pse tell me how i get the attachment to open please??

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.