http://modeles.meteociel.fr/modeles/cfs/runs/2014011706/run2/cfsnh-0-798.png?06

http://modeles.meteociel.fr/modeles/cfs/runs/2014011706/run2/cfsnh-0-1410.png?06

Yep, spring is early

Indeed, nothing more to add.

February could deliver something though......

What is the point in producing pressure charts for 2 months hence?

We cannot forecast with any accuracy what the weather in March will be even in general terms, let alone being able to pinpoint the position of individual systems on a particular day.                                                                                    

are they always that innaccurate yhough? has anyone studied how a chart produced two months early compares to the actual chart on the day? perhaps some are fairly close. i dont know.

The verification figures for CFS charts remain above 0 right out to 9 months. After about 2 months the forecats stop decreasing in accuracy to any measurable extent. For our latitude verification corralation tends to be about 1 or 2 out to 9 months, (ECM 3 day is in excess of 85). But in the tropical latitudes verification statistics are much higher, in the equatorial regions they can be as high as 30, even out to 9 months! Sometimes our 10 day models drop below 30! 

If any one is interested this is where the number for verifications come from

Sum( forecast data points - avg forcast data points) * ( observed data points - avg observed data points)

All divided by

Standard deviation of forecast points * standard deviation of observed points.

There is also a correction factor that accounts for latitude.

The bit on the top is actually the covariance of forecast vs obersvation which tells you how much one variable changes wrt the other. The bit on the bottom basically normalises it so everything is on a scale between -1 and 1

-1 is a perfect negative coralation

1 is a perfect positive coralation. 

a corralation of 0.6 and above is recommended by the ECMWF as useful for long wave patterns

and 0.8 and above for short wave patterns.

Based on this for 500mb heights (which is the most accurate parameter)

'short wave' FI is 168 hours for the ECMWF and 159 hours for the GFS

'long wave' FI is  216 hours for the ECMWF and 204 hours for the GFS

For surface pressure, probably take 12 hours off these figures. 

EDIT: sorry posted this in the media thread by mistake :(

EDIT: just realised I made a careless error in the top part of the equation. This ofc should be covarience. 

Note: The FI figures are based on the most recent model verification data averaged over the last 30 days. Sometimes the 10day ECMWF can have verification figures as high as 0.8, other times as low as 0.25.

Most recently the ECMWF and GFS have been performing better than usual for long timescales. 

Ermmmmm called long range models as Gav has mentioned in his Video, people look at them for trends etc,  I bet you clicked on the link though

Of course you don't have to watch Gav's vids or click on LR links , there is that choice

http://modeles.meteociel.fr/modeles/gfs/runs/2014011806/gfsnh-1-138.png?6  The -20 isotherm gets into Florida and other deep south states. Not good for the oranges!

Originally Posted by: JACKO4EVER 

I will miss the beautiful deep cold and snow when I leave behind that day on chart so it will be a big contrast when I arrive to flooded deary landscape. I will have my fix for snow by then so hope February to be drier and early spring if there no prospect of real winter weather on offer. Those zonal pattern had been on place since September so surely it must end at some point?

Originally Posted by: Gooner 

Probably by the end of Summer Jiries! It will be good to see you regularly posting again. Looking at Gavin P's latest video we could well be in for an early Spring if some long range models are to be believed.

Originally Posted by: JACKO4EVER 

http://modeles.meteociel.fr/modeles/cfs/runs/2014011706/run2/cfsnh-0-798.png?06  http://modeles.meteociel.fr/modeles/cfs/runs/2014011706/run2/cfsnh-0-1410.png?06  Yep, spring is early

Originally Posted by: Jiries 

Ever the optimist Marcus! LOL

Originally Posted by: The Beast from the East 

I have to be Jacko, been horrendous so far

If things that didnt work were never produced, things would never progress to actually working!!

I give you early versions of the wheel and aeroplane as my examples sir 

 LLTNP 105m asl 

Agreement between UKMO and GFS

GFS goes onto  explode a polar bomb

http://modeles.meteociel.fr/modeles/gfs/runs/2014011812/gfsnh-0-204.png?12

stronger block than the 06z

http://modeles.meteociel.fr/modeles/gfs/runs/2014011812/gfsnh-0-252.png?12

Bank!

I must admit for once GFS fi is of some interest.

http://www.wetterzentrale.de/pics/Rtavn2161.png

http://www.wetterzentrale.de/pics/Rtavn2401.png

http://www.wetterzentrale.de/pics/Rtavn2641.png

http://www.wetterzentrale.de/pics/Rtavn2881.png

http://www.wetterzentrale.de/pics/Rtavn3121.png

http://www.wetterzentrale.de/pics/Rtavn3361.png

Does seem a growing trend for something cold and possibly wintry towards the end of the month.

Point taken but 2 *months* away? No doubt the models if programmed to do so could churn out pressure charts for the next 10 years but would be wholly meaningless.

Anything beyond T+240 is highly speculative and unlikely to verify with any degree of accuracy so producing specific charts for T+1440 is just plain daft. You might indeed get it right once in a while but at that range it would be pretty much by luck.

I don't think that would bring snow to lowland areas, but what a cracker of a chart from a meterological perspective!! A real stunner, a bringer-of-carnage

The verification figures for CFS charts remain above 0 right out to 9 months. After about 2 months the forecats stop decreasing in accuracy to any measurable extent. For our latitude verification corralation tends to be about 1 or 2 out to 9 months, (ECM 3 day is in excess of 85). But in the tropical latitudes verification statistics are much higher, in the equatorial regions they can be as high as 30, even out to 9 months! Sometimes our 10 day models drop below 30! 

thanks for the information quantum, i thought someone here would know.