Model Output Discussion

[Hammer]

Perhaps a model discussion thread for each model would be an idea, it all gets rather pointless that in the space of a few hours some people seem to go from and forgive the pun the highs of one model to the lows that another might show in terms of an outcome. It can make keeping track of the current trend rather difficult with the frequency of posts that occur, esp after what might be a negative model run.

[snowlover2009]

I couldn't agree with you more.

I understand why people are concerned, usually the ECM is a very good model. Although this time around its consistency has been rubbish and thats what you really need to look for when predicting cold spells like this. The charts that show more consistency, which in this case would be the GFS and UKMO (bar the wobble the UKMO had yesterday) are more likely to be the accurate ones. The ECM has chopped and changed quite a lot in the last few days and I think the ECM will be back on board come tomorrow. We still have the big 2 on board, so I'm not worried either.

i agree the ecm is actually shocking even at 72 hrs imo.

[Thundery wintry showers]

I recall the ECMWF was often too progressive last winter- for example in late November many ECMWF runs were bringing the Atlantic in as early as 2/3 December, but in reality it did not arrive until 4/5 December. I don't remember if it happened within T+72, but there were certainly errors as close as T+96.

What all of this means is that we are still as uncertain as ever, even closer to the period, and I still give 50% probability of a significant snowy spell across eastern Britain. It will be pretty cold, regardless.

[Harsh Climate]

http://www.cpc.noaa.gov/products/precip/CWlink/pna/nao.mrf.gif

Even the NAO is expected to be well negative over the next 14 days or so, In relation to the MJO in phase 4 and current telleconections I find it very hard to believe our expected greenland high is just guna slip south to be replaced by PV or depressions...

I think the ecmwf might just be performing badly as I have seen from time to time in the past. If the GFS does out peform the ecmwf in the next week or so, there is guna be a lot of people with egg on their face..

Anyone know when the next ecmwf esembles are out?

[weather eater]

One thing for certain either the ECM or the GFS is going to look very poor after this. I would have thought. given the length of time that this pattern has been modelled for, that we would have seen some variation in model outputs from one or the other of the GFS runs, and also the ECM would have come up with a plausible evolution and stuck with it, that hasn’t happened, it chopping and changing every run, so for me the ECM looks suspect.

[Met4Cast]

I'm only 18, but even I know you cant pin everything onto one single run, specially when the two other main models are going for colder weather still

[J10]

I don't really think that judgements on ECM/GFS can rally be done until after the weather has happened. Claiming victory for either now is a bit premature.

[Solar Cycles]

One dodgy run and the christmas pudding fanatics are out in full force, laughable really

Yep, you can almost hear them wishing this to go the way of the pear. Not going to happen this time. GFS is on the ball with this one!

Edited December 10, 2009 by Solar Cycles

[snowlover2009]

this is ecm 12z at 72 hrs

ecm 00z at 72 hrs i know ecm doesnt update as regularly as gfs, but still would not be too much difference in 12 hrs imho

gfs 12z at 72

then the gfs 6z 72 hrs

[Devon-Nelly]

The ECM run has hardly changed all day keeping high pressure in the mid atlantic.

But the latest GFS run i have to say looks abit messy and all over the place towards 180hrs, i think with further runs it may become clearer but i can see that low to the south becoming a headache for alot of the models.

Also the AO and NAO arn't falling away as sharply as they were predicted to, which i think may also be adding to the confusion with the ECM maybe?

[NorthantsSnow]

I'm only 18, but even I know you cant pin everything onto one single run, specially when the two other main models are going for colder weather still

The GFS could fall into line with the ECM at 18z tonight and the UKMO at 00z tomorrow and that would be that.

So I'm afraid there's no room for complacency here. 2/3 is not good enough- we need all 3 to be on board and fast.

[Methuselah]

Yep, you can almost hear them wishing this to go the way of the pear. Not going to happen this time. GFS is on the ball with this one!

I certainly hope so!!!!

[snowlover2009]

please does anyone know when the ecm ensembles are due?

thanks

[Nights King]

I'm only 18, but even I know you cant pin everything onto one single run, specially when the two other main models are going for colder weather still

It is only one run but it is a bad one and they have a habbit of going down like dominoes. i wouldnt be surprised if the ukmo charts follows suit next...

We should know by the morning and then met office will update the long range again to give us a frosty more normal outlook.

Edited December 10, 2009 by mullender83

[Met4Cast]

The GFS could fall into line with the ECM at 18z tonight and the UKMO at 00z tomorrow and that would be that.

So I'm afraid there's no room for complacency here. 2/3 is not good enough- we need all 3 to be on board and fast.

True - But still, 2/3 is better than just one model going at it alone, like the ECM seems to be doing at the moment. We have the AO being forecasted to fall off the charts, and NAO being forecasted for negative too - I find it very hard to believe that we'll see nothing, specially as the GFS and UKMO have been very consistent so far.

Tonight and tomorrows runs will be crucial.

[PersianPaladin]

Another interesting point I'd like to make about the ECMWF and the GFS is at T96.

GFS: -

http://www.wetterzentrale.de/pics/Rhavn961.gif

ECMWF: -

http://www.wetterzentrale.de/pics/Recmnh961.gif

What are the notable differences? Deeper wave extending down from the cold upper-trough on the GFS; resultingly its more amplified down towards Newfoundland and the vorticity tends to follow it

dispersing the energy away from the subpolar regions. The ECM on the other hand; less amplified and the tropospheric vortex producing more instability and vorticity into the poles with shortwave-genesis into Greenland. Generally, this upper-level divergence will erode the block from there.

Whats my take and which model do I trust to model that part of the world? Well, it depends which one has been highest on the verification list in Canada and the USA. Perhaps those models are best suited to handling the upper-trough and the sub-arctic west of Greenland.

[cold snap]

The most anticipated 18Z GFS i can ever remember.

Will it side with ECM and leave UKMO on its own?

I Will be heartbroken If GFS pulls the plug on this.

[VillagePlank]

Yep, you can almost hear them wishing this to go the way of the pear. Not going to happen this time. GFS is on the ball with this one!

I've seen it go pear-shaped at a closer distance than this ... particularly, that the features we are hoping to be modelled correctly are all based in areas where there simply isn't a great collection of data gathering points. It can be problematic 'filling' in the details, and, of course, given the dimensions of our country, and the dimensions of where each and every house of ours is, a very small error, can translate into a bigger error that is larger than the size of our country, let alone our county, or house.

That said (sorry for the spoiler all) consistently modelled over days of runs is probably the key, here. I noticed that the last two runs (6z,12z) were both on the low end of the min/max range and the standard error was certainly always showing positive figures for the entire period.

Still room for caution ... however, a cold spell, I would say is locked in: the meteorological details that go with that are best played out at T+72, where, IMHO, the MetO comes into it's own.

Edited December 10, 2009 by VillagePlank

[Solar Cycles]

certainly hope so, I certainly dont want this to go the way of the pear, im the biggest snow/cold fan on here

You have to remember, the models are always going to struggle with evolutions like this. They have never been programmed to deal with Northern blocking during the winter months, these sort of set ups are so difficult to forecast.

[Cheese Rice]

I think some are forgetting this is a week away, so until Tuesday (at the very earliest) we will have a better idea of what may (or may not) happen. Some members are jumping the gun at a few dodgy runs, as members keep saying there will be highs and lows from each run, all the models are chopping and changing so there is little if any certaincy as to what will happen at then end of next week. All we know is it will get colder.

[Nights King]

The GFS could fall into line with the ECM at 18z tonight and the UKMO at 00z tomorrow and that would be that.

So I'm afraid there's no room for complacency here. 2/3 is not good enough- we need all 3 to be on board and fast.

All three will be on board but which ship.....

[Mondy]

Sorry for OT. Where do the "off-topic" weather thoughts some of us inadvertently post in the model thread go to? Are they simply deleted without reason or shifted somewhere? My own recent one has disappeared - it mentioned the GFS (model) a few times.

Just curious.

[chrisbell-nottheweatherman]

Sorry for OT. Where do the "off-topic" weather thoughts some of us inadvertently post in the model thread go to? Are they simply deleted without reason or shifted somewhere? My own recent one has disappeared - it mentioned the GFS (model) a few times.

Just curious.

Probably cold snap discussion or whatever it's called.

[Iceberg]

The ECM starts at 0 like every model. We don't have access to it. All I am saying is we all fund this model in Europe to forcast for Europe. It cannot be just overlooked. It is the main model. The GFS is not the main model for this part of the world. I just think it is incredulous for anyone to describe the ECM as simply "wrong" at 72hrs. In terms of absolute detail, yes it is wrong, every model is wrong at just 6hrs out. Something is always not right. But in general my point is the ECM is more likely to be right for us then any other model. That is why it is there.

I don't think anybody is ignoring the ECM, it's being looked and and discounted for the following reasons.

-The short-term models are used upto 96 hrs and these do not follow ECM at 72 and 96

-For ECM to be right at 72, 96, 120 or 144 it requires all of the other models to be wrong.

-We don't know what the ECM EPS is showing.

-No model accept the ECM has gone down this route in the last 24 hrs.

-ECM has been chopping and changing twice a day at 144 or below range.

On the plus side for ECM is has continued it's trend from this morning.

To be the honest this is why I think the chances of ECM being right are probably less than 20% tbh.

[Harsh Climate]

Where did you get the time machine from, Ian?

quote from Ian Brown:

''A very sorry episode for the GFS in particular''

lol, didnt know that the GFS had got it all badly wrong!

Edited December 10, 2009 by Jed Bickerdike