General Model Output Discussion

[cyclonic happiness]

FAX 96-

S

So the latest fax charts are showing the low where most of us and all the other models think it should be, so why is the gfs being so progressive?

I don't care what anyone says there must be SOMETHING wrong with their data or human input error?

[Guest]

Even I'm getting annoyed by how far north the low is encroaching - even here the snow showers would be marginal!! Move it 20 miles south and it's exceptional here, and even further south looks good for eastern Scotland, as well as giving frontal snow to one lucky area in the north Midlands/northwest England, but the pub run just looks out of kilter with all the other models. It's a bit silly really, even someone who this outlook potentially benefits hugely can see how ridiculous this run is.

I can see how silly it is by looking at the GFS based forecast for my area for Wednesday night into Thursday. Sustained wind speeds of 36mph from the NE (this would give gusts up to possibly even 60mph) and heavy snow. The GFS has overcooked the low which has sent it too far north, plain and simple. I have seen this so often with Atlantic wind/rain storms which the GFS forecasts as severe storms for here at t+72hrs and when it comes to it the low is not as strong as predicted.

[Guest North Sea Snow Convection]

The differences between the UKMO and GFS are especially different from t60. It is down to the orientation of the low. The UKMO (and other models other than the GFS) elongate the low laterally and do not pivot it vertically in the way that GFS does. The effect of the vertical pivot on the GFS is send the warm air advection (warm front) significantly further north than the other models.

Thereafter the GFS continues to dumbell this low around on a coastal walk of the south of england and then as nick s so amusingly says, it runs out of ideas and appears to wait for the next low to catch it up. Total )apparant) over elaboration.

But the main point where the rot sets in is the big difference in low orientation from t60. The centre of the lows looks similar but GFS appears to want to make a secondary pulse on its southern flank that flips the low 90 degrees into that vertical oval shape and causes the extra southerlies that ruin the set-up by sending the warm front much further north.

It all goes completely wrong from there...

Edited December 27, 2009 by North Sea Snow Convection

[Rollo]

Whilst it is out of kilter with all the others, as a few have said ignore it at your peril as we cannot be sure it is wrong at this stage-another sleepless night for some I reckon-night all.

[snowmadchrisuk]

Well... GFS is going mad whilst ECM and FAX charts look stunning. What is going on with the GFS!?

well if you read the link Nick just put up http://www.hpc.ncep.noaa.gov/discussions/pmdhmd.html

They do admit there are initial data errors .

[Optimus Prime]

They do admit there are initial data errors .

Lol I was wondering when the data loss would be brought up!

[snowlover2009]

Here's a scarey though....maybe the gfs is right??!!

After all everyone thought it was a fortnight agao when it was showing the currant cold spell and the other models were having a wobble.

no all the the models started to agree anyway, the gfs is taking its time and i mean taking its time, we are 2 days or even less than that for this upcoming low pressure and the gfs is still not agreeing for some reason. when this cold spell was predicted, yes gfs was the one who discovered it first and the models didn't agree, but it did not take this long for all the models come into line 2 days before the event, like what is happening now. trust it to all go wrong, but i doubt it. so mcuh is on the line, ukmo have made so much of this in the forecasts and on the website, and all the models including the ecm ens are in good agreement, minus one model which is gfs. i am very annoyed. even gem is better atm, at this rate the gfs will be discarded as one of the nogaps type models. i would rather no one tell me gfs could be wrong because gfs is out on its own and very very alone.

[TEITS]

Very nice fax charts which go with the ECM. However if you look very closely at the Fax for +120 it isn't an exact copy of the ECM. This is the sort of tweaking I was referring to about requiring the human touch.

[snowmadchrisuk]

Lol I was wondering when the data loss would be brought up!

Well the NOAA saying it confirms that there are current data issues.

[Shunter]

Yet another GFS run fit for the bin. The programmers of the GFS need to take long hard look at why it has a tendency to be over progressive. It seems utterly incapable of dealing with subtle energy distributions.

The ECM ensembles paint the picture and it is certainly not a mild one.

I simply cannot see all the other models coming in line with the GFS. It simply wont happen. I have always viewed the GFS with a sceptical eye and the rest of this week, will, hopefully, reinforce this view. I am starting to rate the GEM more highly than the GFS.

Somebody ought to check the forum archives, I'm absolutely sure we had identical issues last Xmas and the Xmas before that with the GFS model. IF it is omitted data then I would expect to see GFS fall back into line for tomorrows 12z run. Also I seem to recall model differences during the previous New Year schedule.

Edited December 27, 2009 by Shunter

[SMU]

if you got excited over the T96, i guess you fell over and knocked yourself out when you saw the T120 - otherwise you'd have posted that one aswell !!

LOL Yes I just got off the floor-

THe GFS 18z isnt all bad-- Its still getting the Low wrong in terms of orientation to start with but the ensemble mean is catching up now-

T 90 V t96 is around 100 miles further south with a sharper Easterly-

96 12z

90 18z

Cheers

S

[yamkin]

The main point here is that the MetO are NOT utilising their out forecast via the GFS model. Their 'Unified Model' which inputs ALL model systems for analysis, is not going along with the GFS model output.

[kold weather]

FWIW I don't expect the ensembles too be too good, generally when they will broadly reflect the op run in terms of something in the data that is making it churn out good or bad results.

I'd suspect the 0z will be key, if it continues the 18z and the rot we see on that then I'll be a little concerned.

One thing I'm not at all worried about is the -ve NAO...the GFS IS overly progressive in that respect, I have utterly no doubt in that respect at all...

[cyclonic happiness]

Well the NOAA saying it confirms that there are current data issues.

Well let's not starting to fall out with one another about this, how long do these data issues (if they are), take to resolve?

Or perhaps let's just wait until morning as, if it's till showing woe, then we can get a bit more worried?

[snowlover2009]

Yet another GFS run fit for the bin. The programmers of the GFS need to take long hard look at why it has a tendency to be over progressive. It seems utterly incapable of dealing with subtle energy distributions.

The ECM ensembles paint the picture and it is certainly not a mild one.

I simply cannot see all the other models coming in line with the GFS. It simply won't happen. I have always viewed the GFS with a sceptical eye and the rest of this week, will, hopefully, reinforce this view. I am starting to rate the GEM more highly than the GFS.

we all seem to take gfs really seriosuly like ecm and ukmo, but when you think about it is quite pants, even gem and no gaps probably perform better out to 96 hrs. i think if gfs is wrong which i am very much sure it will be then i don't see why we should always get too uptight about what it is showing when all other models are better. it is very updown all the time, sometimes it can be very good, i remember the end of octobers mild spell and about 9 days out the gfs just kept showing the same scenario very accurately which bugged me because i didn't want to see mild weather lol

[Skyraker]

Whilst it is out of kilter with all the others, as a few have said ignore it at your peril as we cannot be sure it is wrong at this stage-another sleepless night for some I reckon-night all.

Which they also point out does not affect the short range forecast.

Lurker from TWO saying hello.

[kold weather]

Amazingly the 18z ALSO once again backs up the idea that the track of the low is vital, the runs that do a similar thing to the GFS with the low evolve in a very similar way to the GFS 18z op.

Most are once again along with the op run WAY too progressive with dropping the -ve NAO, that will not be the case. Simple as that IMO!

[SMU]

FWIW I don't expect the ensembles too be too good, generally when they will broadly reflect the op run in terms of something in the data that is making it churn out good or bad results.

I'd suspect the 0z will be key, if it continues the 18z and the rot we see on that then I'll be a little concerned.

One thing I'm not at all worried about is the -ve NAO...the GFS IS overly progressive in that respect, I have utterly no doubt in that respect at all...

Hi Darren-I think you might be wrong there- I suspect there will be a few more -10's/-12's popping up at around 144- esp on the Manchester ens & some on the london-

S

[casparjack]

we all seem to take gfs really seriosuly like ecm and ukmo, but when you think about it is quite pants, even gem and no gaps probably perform better out to 96 hrs. i think if gfs is wrong which i am very much sure it will be then i don't see why we should always get too uptight about what it is showing when all other models are better. it is very updown all the time, sometimes it can be very good, i remember the end of octobers mild spell and about 9 days out the gfs just kept showing the same scenario very accurately which bugged me because i didn't want to see mild weather lol

I think the point is that all the models handle mild weather much better, as they are more familiar with forecasting it.

Synoptics such as we have at the moment have probably cropped up less than ten times in the past 25 years, there is no wonder they are struggling with the details...

[kold weather]

Possibly Steve, there will be some cold runs of course as the models suggest a better tilt to the upper high on the 18z runs to drag in the colder air, it just comes down to the track of the orginal low and how long it holds back near the UK, we need the low to get shunted to our east, if that happens the milder runs will vanish faster then can be imagined...

Very hard way to get into the cold spell according to the 18z ensemles.

[snowlover2009]

I think the point is that all the models handle mild weather much better, as they are more familiar with forecasting it.

Synoptics such as we have at the moment have probably cropped up less than ten times in the past 25 years, there is no wonder they are struggling with the details...

yes i have thought about synoptics that havent been shown for a very long time and that the models would struggle, buts its been confirmed that gfs has data problems.

at this rate we will be at 3 hrs and gfs will still be out on its own.

[TEITS]

I was wondering if anyone could supply me with the links to the ECM precip forecasts. Also I know longer have access to the UKMO 06/18Z as the website no longer supplies these. So could someone please post me a link to the UKMO.

Thanks ever so much.

[Yeti]

FAX 96-

its totally inline with the ECM- & identical to the PPN distribution I posted earlier-

S

HAHAHA this made me laugh. Don't think many people noticed the cheeky comment there! :lol:

[SMU]

I was wondering if anyone could supply me with the links to the ECM precip forecasts. Also I know longer have access to the UKMO 06/18Z as the website no longer supplies these. So could someone please post me a link to the UKMO.

Thanks ever so much.

I Havent got the UKMO 18z- it used to be on the tropical Experimental site which was handy- but its gone from there- So I need that one as well-

ECM uppers & PPN Charts-

http://en.vedur.is/weather/forecasts/atlantic/#type=prec

S

[Snowaddict]

I was wondering if anyone could supply me with the links to the ECM precip forecasts. Also I know longer have access to the UKMO 06/18Z as the website no longer supplies these. So could someone please post me a link to the UKMO.

Thanks ever so much.

Hi TETIS,

Some debacle this is this evening with GFS!

Here is tonight's ECMWF Precip evolution, just change the url for the appropriate timeframe:

http://en.vedur.is/p...227_1200_72.png

SA :lol:

Edited December 27, 2009 by Snowaddict