Model Output Discussion - Early November Onwards

[bobbydog]

It's definitely the same model, uses the same physics on every run and so on. The only difference is that there are differing amounts of some of the data types going into each run - for instance the 06z run may have less upper air data from the atlantic because there are less flights over the atlantic at that time of day.

interesting point there paul, that would suggest some runs are more accurate than others. could that be the case?

[Paul]

interesting point there paul, that would suggest some runs are more accurate than others. could that be the case?

I'm not so sure, there are differing amounts of data going in, but if one run perhaps doesn't have as much atlantic flight data, it may have more pacific flight data so things will tend to even themselves out. You have to bear in mind that no model run has perfect starting conditions - ie there will be holes in the data and this is one of the reasons why the ensembles are run as the adjustments made to the starting conditions in each ensemble run are put in place to simulate what differences those imperfections may make.

[Strider]

I'm not so sure, there are differing amounts of data going in, but if one run perhaps doesn't have as much atlantic flight data, it may have more pacific flight data so things will tend to even themselves out. You have to bear in mind that no model run has perfect starting conditions - ie there will be holes in the data and this is one of the reasons why the ensembles are run as the adjustments made to the starting conditions in each ensemble run are put in place to simulate what differences those imperfections may make.

Interesting, I would've thought the same amount of data went into each run.

On the link you edited into John's post you talked about different amounts of flight data per run. Do you have evidence to back this up?

Edited November 4, 2011 by Strider

[Cheshire Freeze]

How anyone can look at the last 6/7 GFS runs and say there is no liklihood of a pattern change bemuses me slightly. The trend is there for all to see.

There appears to be a persistent trend to high latitude blocking showing in the 10 day timeframe.

Edited November 4, 2011 by CreweCold

[bobbydog]

I'm not so sure, there are differing amounts of data going in, but if one run perhaps doesn't have as much atlantic flight data, it may have more pacific flight data so things will tend to even themselves out. You have to bear in mind that no model run has perfect starting conditions - ie there will be holes in the data and this is one of the reasons why the ensembles are run as the adjustments made to the starting conditions in each ensemble run are put in place to simulate what differences those imperfections may make.

would be interesting to find out, maybe, as has been said, it would be better to treat each run as a different set of data as per the time of day, i.e 06z's and 12z's as separate data sets, if the input sources differ from run to run (if that makes sense!)

Edited November 4, 2011 by bobbydog

[wolvesfan]

HI all, just finished viewing GFS barely a single low pressure system to transfer west to east across England in the next 16 days if it verifys.what an extremly benign autumn we are experiencing.i'd like to think we might now be seeing the begining of the end also being progued for the euro high,hope the weekends following runs bring its demise a touch closer,high pressure never far from our shores though.

Edited November 4, 2011 by wolvesfan

[Fozfoster]

I'm not so sure, there are differing amounts of data going in, but if one run perhaps doesn't have as much atlantic flight data, it may have more pacific flight data so things will tend to even themselves out. You have to bear in mind that no model run has perfect starting conditions - ie there will be holes in the data and this is one of the reasons why the ensembles are run as the adjustments made to the starting conditions in each ensemble run are put in place to simulate what differences those imperfections may make.

So the 18z is sometimes called the pub run so maybe less data goes into that run yeah.

[rich1]

How anyone can look at the last 6/7 GFS runs and say there is no liklihood of a pattern change bemuses me slightly. The trend is there for all to see.

There appears to be a persistent trend to high latitude blocking showing in the 10 day timeframe.

on the 12z this is progged in about 2 weeks time

http://176.31.229.228/modeles/gfs/run/gfs-0-300.png?0

just for fun at that stage but some food for thought. Its not that cold yet but on that run it does point to a repositioning of the block in a much better place in the longer run. Until it appears in the High Res though it remains out of the reliable, but a few more runs of this sort of evolution could spark up some interest later on. Worth keeping half an eye on

[phil nw.]

How anyone can look at the last 6/7 GFS runs and say there is no liklihood of a pattern change bemuses me slightly. The trend is there for all to see.

There appears to be a persistent trend to high latitude blocking showing in the 10 day timeframe.

Yes i note that GFS is modelling more Ht. rises further North as we go into week 2.

With little energy off the Atlantic the trough to our West is looking more negatively tilted here at day 10 encouraging more of an East rather than South flow.

This is something i am monitoring following my post re. the MJO state yesterday.

Of course it doesn`t mean a sudden shift to cold but if such an evolution occurs then we would be better placed for this sooner than if for example we had an Azores High pushing in from the West.

Edited November 4, 2011 by phil n.warks.

[snowfall09]

I'm not so sure, there are differing amounts of data going in, but if one run perhaps doesn't have as much atlantic flight data, it may have more pacific flight data so things will tend to even themselves out. You have to bear in mind that no model run has perfect starting conditions - ie there will be holes in the data and this is one of the reasons why the ensembles are run as the adjustments made to the starting conditions in each ensemble run are put in place to simulate what differences those imperfections may make.

Just out of interest was the GFS more unreliable when the Icelandic volcano erupted as there where hardly any European flights meaning loss of a fair amount of data???

[Radiating Dendrite]

How anyone can look at the last 6/7 GFS runs and say there is no liklihood of a pattern change bemuses me slightly. The trend is there for all to see.

There appears to be a persistent trend to high latitude blocking showing in the 10 day timeframe.

I stated this before as a developing trend and got shot down (well blown to smitherines)! Good luck

[Aaron]

I stated this before as a developing trend and got shot down (well blown to smitherines)! Good luck

There's quite a lot of talk on TWO about a pattern change emerging in the distant future (say 2 - 4 weeks time), even from well respected, more knowledgeable members. I guess people just view the models differently.

[Isolated Frost]

I stated this before as a developing trend and got shot down (well blown to smitherines)! Good luck

But then again, it's a good 2 days into FI, and it's basically an over-optimistic ridge from Greenland eastwards with no polar jet distruption. If that signifies a pattern change, I am Glacier Point.

Wet, windy, southerlies and easterlies and the odd westerly as the subtropical ridge continues to set itself up in the 'death to cold weather' area, eastern europe with a n-s positioning. Mild, mild, mild with the odd north-easterly possible in 16 November+ range, with some wintry showers easily dispossessed thanks to troughs lining up against that block.

I do foresee a cooler outlook after 20 November though- with winds from the north and east quadrant thanks to the block subsiding and heights to our northwest rising. But again I reiterate the short-term, mild with the odd average snap.

[johnholmes]

t

Just ended up deleting about 10 posts made this pm, can we keep this thread to the models please....

I added it to your post for you

thanks for that

Edited November 4, 2011 by johnholmes

[Cheshire Freeze]

hanks for that

Who's hank?

[Paul]

Interesting, I would've thought the same amount of data went into each run.

On the link you edited into John's post you talked about different amounts of flight data per run. Do you have evidence to back this up?

It's fairly common knowledge that differing amounts of data types go into each run and flight data was just an obvious example of that as there are most certainly less atlantic flights at different times of the day.

You can see a real time summary of the data going into the GFS here:

12z run:

http://www.nco.ncep....x.summary.shtml

06z run:

http://www.nco.ncep.noaa.gov/pmb/nwprod/realtime/gfs/t06z/index.summary.shtml

As you'll see there are different percentages next to each data type, signifying the different levels of that data dumped into that run.

[Cheshire Freeze]

But then again, it's a good 2 days into FI, and it's basically an over-optimistic ridge from Greenland eastwards with no polar jet distruption. If that signifies a pattern change, I am Glacier Point.

Wet, windy, southerlies and easterlies and the odd westerly as the subtropical ridge continues to set itself up in the 'death to cold weather' area, eastern europe with a n-s positioning. Mild, mild, mild with the odd north-easterly possible in 16 November+ range, with some wintry showers easily dispossessed thanks to troughs lining up against that block.

I do foresee a cooler outlook after 20 November though- with winds from the north and east quadrant thanks to the block subsiding and heights to our northwest rising. But again I reiterate the short-term, mild with the odd average snap.

I think we've been rather spoilt over the past couple of years, if the pattern shown in GFS FI over the past couple of days was an output from the GFS at t120 for example during the 90's and 00's years, the servers would have gone into meltdown!

[johnholmes]

I knew you would reply to this John!!

One can always agree to disagree but I think it is worth discussing and is model related. Keeping this discussion to the GFS, I was under the impression that the data variation from run to run was almost insignificant compared to what is inputted. It is with this respect that I feel that one should look at every run with the latest being the most accurate. One should look at overall FI solution of every run to get an accurate picture as possible to what is the likely outcome. Why would the operators of the GFS bother running the programme 4 times a day if they too believed that it is just better to look at only one of them?

To me it makes better sense to combine the output of every run and ensemble direction. One can see where divergence from one solution to another begins when doing this and work out for oneself up to which point the run can be trusted!

edit: Oh and to add further following Paul's insert if you concentrate on one run then you may inadvertently concentrate on the poorest run . I know that the data is compared somewhere to evaluate each run.

I think we will just have to disagree about the relative merits of each others' view chiono-no great problem there.

Re the input from Paul-its not what I was looking for, I had found that, there is a post somewhere from one of the team on just what goes into each GFS run, or rather did about 12 months ago.

I'll keep searching on Net Wx and within my own pc files.

[MP-R]

Not the most desirable of model runs at the moment. The flimsy Atlantic comes up against a flimsy Euro high and similar to February this year we end up in the cloud mush in the middle, whether it be with winds from the SE or S.

Really need the Atlantic to fire up and blow itself out like 2009. The best chances of cold are more likely to be from the north as the continent is still relatively warm so easterlies will most likely just be cool and cloudy with nothing much of interest - the same going for southeasterlies.

I hope future runs bring more in the way of Atlantic weather.

[Paul]

I think we will just have to disagree about the relative merits of each others' view chiono-no great problem there.

Re the input from Paul-its not what I was looking for, I had found that, there is a post somewhere from one of the team on just what goes into each GFS run, or rather did about 12 months ago.

I'll keep searching on Net Wx and within my own pc files.

All data types go into every run John, just differing amounts based on availability - have a look at the data summary links I posted above for clarification on this.

[rich1]

I think we've been rather spoilt over the past couple of years, if the pattern shown in GFS FI over the past couple of days was an output from the GFS at t120 for example during the 90's and 00's years, the servers would have gone into meltdown!

Is true

for me the interesting thing is the Atlantic activity not being all that...well, active really. its just at the moment the blocking is in the wrong place. There isnt the 90's type lows tracking across us and in to Russia type of set up. The LP's are not getting all that far against the block there. If the signal towards height rises to the north intensifies then that will pay off eventually. Wont be anytime soon but worth keeping watch nonetheless

may be a combination as you say of a mixture of very cold weather in recent winters raising expectations perhaps combined by a feeling that a 4th coldish winter on the bounce isnt seen as statistically probable which may be in part pointing to a lack of any real interest or enthusiasm right now. That and a long spell of mostly above average and relatively nondescript weather dominating this autumn

One thing we have learned in recent years is to not rule out anything and the form horses have changed track, but would say at this time that its all speculation right now and in the medium timeframe its generally mild and relatively uneventful unless you are looking for the odd unusually warm early November day when the sun pops out

[phil nw.]

Not the most desirable of model runs at the moment. The flimsy Atlantic comes up against a flimsy Euro high and similar to February this year we end up in the cloud mush in the middle, whether it be with winds from the SE or S.

Really need the Atlantic to fire up and blow itself out like 2009. The best chances of cold are more likely to be from the north as the continent is still relatively warm so easterlies will most likely just be cool and cloudy with nothing much of interest - the same going for southeasterlies.

I hope future runs bring more in the way of Atlantic weather.

Idealy we need that Atlantic trough to eject an undercutting low to our South--under the block further down the line.

However there`s little energy in the jet at the moment.

I feel that any change will be gradual because of this.

[Snowmad79]

I knew you would reply to this John!!

One can always agree to disagree but I think it is worth discussing and is model related. Keeping this discussion to the GFS, I was under the impression that the data variation from run to run was almost insignificant compared to what is inputted. It is with this respect that I feel that one should look at every run with the latest being the most accurate. One should look at overall FI solution of every run to get an accurate picture as possible to what is the likely outcome. Why would the operators of the GFS bother running the programme 4 times a day if they too believed that it is just better to look at only one of them?

To me it makes better sense to combine the output of every run and ensemble direction. One can see where divergence from one solution to another begins when doing this and work out for oneself up to which point the run can be trusted!

edit: Oh and to add further following Paul's insert if you concentrate on one run then you may inadvertently concentrate on the poorest run . I know that the data is compared somewhere to evaluate each run.

If that is true then why do NOAA and other bodies use only the 00z runs for forcasts such as AO/ NAO teleconnections ? I mean , they dont use the 00z one day then the 12z the next nor do they use a combintation of all 4 runs, which kind of makes Johns post seem like more sense regarding following one run per day and compairing that run to tomorrows version rather than trying to paint a picture from all 4 runs daily which is messy becuase they vary so much. NOAA do it after all when looking at the longer range.

What about the MJO forcasts, do they not use one opertational run rather than a combination of all 4 runs ?

[Frosty.]

Another very mild Gfs 12z operational run for the first 276 hours but then trending a little cooler but no sign of anything cold. The cold plunge into northeast europe has downgraded somewhat but it has the effect of displacing the euro high northwest into scandinavia, at the same time, low pressure systems makes more progress towards the uk from the atlantic and bring an unsettled spell but a fine although rather cloudy first half to next week is looking inevitable with diminishing sunshine for nw britain and then much milder air pumping north across all areas with +10 T850 hPa approacing the south so another dose of +18c seems likely later next week although nearer 15c in the north and with some rain and increasing S'ly windflow but the 12z does then show the very mild uppers being replaced by something more seasonal but still in the range of 10 to 13c from north to south. For scotland & n. ireland..enjoy the sunshine this weekend, you deserve it following a dismal summer and autumn so far..the dross is reserved for the bulk of england and wales this weekend for a change followed by a cloudy but dry start to next week.

[Snowmadsam]

I currently don't mind the models at the moment, but nevertheless I will be watching from now until the last snowfall

But until any wintry prospects come, I can be found on here or watching youtube videos of snow falling