Model Output Discussion

[TEITS]

Reading some of the posts on this thread and I do agree with Stu that the models have downgraded the potential of a prolonged cold spell. However I strongly disagree that just because they have downgraded doesn't mean the downgrades will continue. I have to laugh at such statements because have some forgotten what happened prior to our recent cold spell. We had a 24hr period where the models seriously downgraded the cold spell and I remember a particularly poor 12Z run. However as the cold spell came closer the models upgraded especially with regards to the intital E,ly.

Another point I wish to raise is about this missing data. I very much doubt that those who say this isn't the case followed the models avidly during last xmas. I did and I remember how the model output on Xmas day/Boxing day was very odd indeed until normality returned on the 27th. Now im not saying this was due to missing data but its a bit of a coincidence.

In summary I am disappointed with the trend from the 12Zs. However I have learn't not to make too many assumptions and would prefer to keep an open mind with regards to tomorrows model output. I suggest that others do the same especially those who should know better!

[Harsh Climate]

Reading some of the posts on this thread and I do agree with Stu that the models have downgraded the potential of a prolonged cold spell. However I strongly disagree that just because they have downgraded doesn't mean the downgrades will continue. I have to laugh at such statements because have some forgotten what happened prior to our recent cold spell. We had a 24hr period where the models seriously downgraded the cold spell and I remember a particularly poor 12Z run. However as the cold spell came closer the models upgraded especially with regards to the intital E,ly.

Another point I wish to raise is about this missing data. I very much doubt that those who say this isn't the case followed the models avidly during last xmas. I did and I remember how the model output on Xmas day/Boxing day was very odd indeed until normality returned on the 27th. Now im not saying this was due to missing data but its a bit of a coincidence.

In summary I am disappointed with the trend from the 12Zs. However I have learn't not to make too many assumptions and would prefer to keep an open mind with regards to tomorrows model output. I suggest that others do the same especially those who should know better!

I recall last year or the year before where the gfs brought out ludicris runs on xmas day/boxing day.. I think missing data maybe an issue

[nick sussex]

But as I mentioned earlier, the link you posted showed the UKMO verifying than the GFS over the Southern Hemisphere but not the Northern Hemisphere (did you get the two confused?). UKMO over the last few days was a poor 4th over the Northern Hemisphere with GFS and ECM both doing a lot better.

The UKMO has consistently overdone the northern extent of Atlantic lows over the past two weeks. I don't think it's worth cold/snow lovers getting worried over it unless the ECMWF falls into line with it. Regarding the long-term potential of its output, I think the main issue is that the PFJ straddles the British Isles keeping the colder air up north, so there would be more chance of the cold air never reaching the south under that scenario despite the large Greenland High, if lows kept tracking across the British Isles. But I think the UKMO probably has the lows too far north just like it often has done recently.

I'm not disagreeing with recent problems with the ukmo output but i think you've read the link wrong! have you been on the brandy? the link clearly shows for the northern hemipshere ukmo 0.812, gfs 0.797 both behind the ecm. What seems strange is that the ukmo gets criticized for it's 144hrs output but over the last month verifies above every other model bar the ecm. Yet everyone seems happy to go along with the gfs.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz6.html

Further to tonights output happy to see the ecm disagreeing with both the gfs and ukmo, this isn't a surprise but personally until the low track is settled then all the output past this point can be binned. Anything past 96hrs is just liable at present to alot of changes.

Edited December 26, 2009 by nick sussex

[Matt12]

Two things to remember folks:-

1) The charts are fairly decent at the mo. Granted, they don't paint an armageddon type scenario, but it's not at all bad.

2) Charts change rapidly. There will probably be some pretty major changes in the coming days - whether they be good or bad.

All in all, I'd would have settled for the current situation.

[snowlover2009]

Models look a wet drab for the south at least to the start of ecm, then gets better, whereas midlands and north england and parts of wales look really great for some really interesting weather.

Edited December 26, 2009 by snowlover2009

[igloo]

of course there is missing data for all the models do you think for 1 minute everbody is maning all there computers from the info from satellites or buoys of course not but should be by monday

[Thundery wintry showers]

I'm not disagreeing with recent problems with the ukmo output but i think you've read the link wrong! have you been on the brandy? the link clearly shows for the northern hemipshere ukmo 0.812, gfs 0.797 both behind the ecm. What seems strange is that the ukmo gets criticized for it's 144hrs output but over the last month verifies above every other model bar the ecm. Yet everyone seems happy to go along with the gfs.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz6.html

Further to tonights output happy to see the ecm disagreeing with both the gfs and ukmo, this isn't a surprise but personally until the low track is settled then all the output past this point can be binned. Anything past 96hrs is just liable at present to alot of changes.

Ah I see what you mean now- you're looking at December as a whole whereas I was looking at the verification stats for the past week.

I agree with the sentiments of your last paragraph- things are just way too uncertain at the moment to be able to promise anything, be it a cold snowy Armageddon from the north-east, some milder air sticking around in the south or whatever!

[Harsh Climate]

Ah I see what you mean now- you're looking at December as a whole whereas I was looking at the verification stats for the past week.

I agree with the sentiments of your last paragraph- things are just way too uncertain at the moment to be able to promise anything, be it a cold snowy Armageddon from the north-east, some milder air sticking around in the south or whatever!

With the exception of Joe laminate floori, nobody knows whats going to happen..

edit: his names in the swear filter lmao

Edited December 26, 2009 by Jed Bickerdike

[Duane S.]

Another point I wish to raise is about this missing data. I very much doubt that those who say this isn't the case followed the models avidly during last xmas. I did and I remember how the model output on Xmas day/Boxing day was very odd indeed until normality returned on the 27th. Now im not saying this was due to missing data but its a bit of a coincidence.

I am amazed that we don't have a definitve answer on this by now. I would imagine it is some sort of middle ground, i.e. that some data is missing but it doesn't have a massive effect. Although maybe in retrospect that sentence (mine) is a bit daft as in theory just a small piece of missing data at the start can make a massive difference later on. Who knows...

What I do think though is that the ECM is very strange after +168. Looking forward to the FAX later on. I'm fairly hopeful it will not back the METO, instead be a lot closer to the ECM evolution.

[ribster]

Blimey, with the exception of a a few posters, not a lot of model discussion going on, just a load of whining (now I've gone and added to the non-model discussion - doh!)

[Ian Fergusson]

Deep breaths....

There's an aptly-named Facebook group titled: Meteorology is an imperfect science: Get over it.

Equally, my own branch of original scientific specialism - marine biology - is also imperfect. And almost by definition ( = the ocean is HUGE and largely unexplored.)

Example: I can happily stand at a symposium and describe to you the (known, proven bits of ) behavioural ecology and biogeography I've discovered about great white sharks in the Mediterranean Sea. And I can equally assure you that what I'm telling you is a mere minority of the genuine facts.... so: ask me to speculate on how many white sharks reside there - either right now; or for that matter, in three days time - and I won't even bother offering you an answer.

Why?

Because it's mere speculation. Blarney. Nonsense. And that's hardly the sort of 'science' that survives proper peer review. ... versus our self-publication on a web forum.

And anyone can do the latter. Therein lies the beauty of sharing information, ideas and thoughts in our modern web world...

We've witnessed on this elongate thread some very learned speculation about how the weather will resolve itself towards New Year. However, certain models - and especially the most widely-scrutinised one in most public forums ( = GFS) - have a cold / snowy bias that needs special discrimination and due emphasis / caution. Especially when discussing time frames of T+240, +144 or whatever. Even T+48 can become a forecasting shambles at a regional level - especially when faced with knife-edge snowfall scenarios and the like.

Thus, be it heavy rain (as the UKMO currently believes, incidentally - at least for most of southern England until post-New Year) or snow (Wales / Midlands, if UKMO analysis to be believed).... it makes little odds for those of us interested in the science, versus the 'ramping'.

Thanks to those of you who post well-judged, well-considered, non-bickering analysis on here; these are always read with great interest and with HUGE respect for your very evident combined scientific / forecasting expertise.... Merry Christmas to you all. Cheers, Ian

Edited December 26, 2009 by Ian Fergusson

[TEITS]

I am amazed that we don't have a definitve answer on this by now. I would imagine it is some sort of middle ground, i.e. that some data is missing but it doesn't have a massive effect. Although maybe in retrospect that sentence (mine) is a bit daft as in theory just a small piece of missing data at the start can make a massive difference later on. Who knows...

What I do think though is that the ECM is very strange after +168. Looking forward to the FAX later on. I'm fairly hopeful it will not back the METO, instead be a lot closer to the ECM evolution.

What I find strange with the output is the indications of a persistant, strong negative NAO/AO and yet this isn't really illustrated in the medium range of the GFS/ECM. I think we need GP to hopefully put our minds at ease.

[stratty]

..........the thing is, as soon as the Express have the weather as a headline, you know it's not gonna happen!

[keithlucky]

727 UPDATE POSSIBLE 3 DAYS OF SNOW M4 NORTHWARDS

[Duane S.]

TBH, if I was to be non-biased and look for recent trends in all the models over the past day or two, I would say there has been a slight tendancy to raise heights in the Med and for the cold air boundry line in the South to be delayed in its Southward march down through the UK. Not good but no disaster, it is not a trend I would like to see continued throughout tomorrow though.

Regards the models that are showing a quick breakdown of the Greenland high though, I don't buy it. It won't of course last forever but the upstream signal is not there for it to just get sucked away within a few days.

[MR EXTREMES]

i find it very difficult to understand how the met office have come out with the kind of medium term forecast considering models are not in agreement,

although we must remember they may have been using earlier or even yesterdays outputs if so then things are good.

but we will have to wait until tomorrows meto update to see where they go from this i think they to will backtrack there forecast,

but it is true our recent cold event before christmas had a major wobble but everything was back on track in 24 hours.

but its certain if tomorrows projections are the same as today then for the southern half fairly mild and very wet oh well still atleast 8 weeks for more cold lol,

im not sure i can handle 8 weeks of a rollercoaster ride:drinks:

[The PIT]

Joe wotsit his name is normally 100% wrong so if he says cold expect warm. Although he will strike lucky sooner or later.

Ecm is good run if you like snow that is as is the GFS. UKMO isn't bad either.

Anyway a few less cold days to come first of all so time to put away the sore heads and wine influcened posting and have a group hug.

[TEITS]

i find it very difficult to understand how the met office have come out with the kind of medium term forecast considering models are not in agreement,

I have some answers to your question.

1. Far more access to data.

2. The art of forecasting isn't just quoting what the models say. The art of forecasting is knowing whether the models are correct and if not where are they going wrong. These computers are excellent forecasting tools but you will always need this human touch. This is why until this model uncertainity is sorted the best available data for us is the Met O fax charts.

[nick sussex]

I am amazed that we don't have a definitve answer on this by now. I would imagine it is some sort of middle ground, i.e. that some data is missing but it doesn't have a massive effect. Although maybe in retrospect that sentence (mine) is a bit daft as in theory just a small piece of missing data at the start can make a massive difference later on. Who knows...

What I do think though is that the ECM is very strange after +168. Looking forward to the FAX later on. I'm fairly hopeful it will not back the METO, instead be a lot closer to the ECM evolution.

Hi Lancia

In terms of data the last few gfs runs have been missing some automated aviation data from ARINC, this would probably effect all the models though. NOAA talk about critical and non critical data, the question is though is critical here meaning a model run has to be dumped or in terms of non critical does this mean the model run can go ahead and be accepted but with a likely higher error rate or forecast variability.

The variable here is that the Arctic area of the North Atlantic has a lack of observational data anyway and this has been commented on previously by NOAA, the upcoming pattern relies on the models calling this area correctly, they could overplay or underplay the extent of blocking in those regions. Given that the models can't even agree on events upto 96hrs i think many members are expecting too much from them, we need to wipe the slate clean here and start again focussing on the small changes within the 96hrs range that will snowball into large scale differences later.

[snowlover2009]

if there is missing data, i would think it would be at least in FI.

[Harsh Climate]

if there is missing data, i would think it would be at least in FI.

I think after the next 72 hours is complete FI, why people are getting concerned at 180h+ is laughable.

[MR EXTREMES]

to be honest i think its unlikey theres missing data i think eye was spot on with his reply.

if we like it or not nothing is going to change what will be.

and fax is my favourate form of model watching athough i dont look at them as much as i should.

so can anyone tell me what time the ecm comes out?

[Mark Bayley]

to be honest i think its unlikey theres missing data i think eye was spot on with his reply.

if we like it or not nothing is going to change what will be.

and fax is my favourate form of model watching athough i dont look at them as much as i should.

so can anyone tell me what time the ecm comes out?

Its already out Unless you mean the ensembles?

http://www.netweather.tv/index.cgi?action=ecm;sess=

[Lewis]

so missing data at xmas always results in us seeing dream scenarios. can i ask why 'missing data' doesnt mean mild crap. if it wa smissing data, the models would be showing all sorts. yes the low track isn't nailed. but i don't see any models predicting the eye of the low in central france and one low in the far north of scotland. if data was missing then there would be no agreement surrely and not just showing the same scenarios as they are now regarding cold weather.

No need to swear in nearly every post you have made tonight, i thought last night you said you understand the charts??

Lewis

[Snowyowl9]

Bitterly disappointed with ukmo this evening.

I think the meto/bbc are going to do some serious backtracking tonight/tomorrow unless ecm pulls a rabbit out of the hat.

ukmo is just rain,and more rain,totally flies in the face of bbc calling cold next week.

What`s wrong with charts like this I must be missing something.

http://91.121.93.17/pics/Rukm1201.gif

http://91.121.93.17/pics/Recm1441.gif

http://91.121.93.17/pics/Recm2161.gif

They`re brilliant,mild air is struggling you have to remember that,not doom and gloom posts.

Edited December 26, 2009 by Snowyowl9