ECM declared world's best model following NOAA cuts

[The Eagle]

The US media and meteorlogical experts have declared the ECMWF as the world's foremost weather model given the accuracy of forecasting Irma's track against the GFS which was hopeless.

US media is blaming cuts to NOAA. US media also says had they followed ECM without GFS they would have known days in advance almost to the point now of the hurricane track claiming the ECM model ensemble handles atmospheric conditions far better.

Do you agree? Real frustration stateside. I have to say it's hard to disagree.

Irma brought unusual heavy focus on the models to the general public like never before.

Edited September 9, 2017 by The Eagle

[jvenge]

Well, it's of no surprise here that the GFS ranks behind the ECM and also UKMETO. It's not about money, though. They have spent huge amounts chasing the ECM and raw $$$ doesn't solve everything.

Even if their funds were cut this year, the model itself has been behind the ECM since I joined here.

[summer blizzard]

The gulf between models is nothing new, in 2014 the NOAA judged that the UKMO model had performed best that hurricane season and in 2015 and 2016 the Euro rained supreme. It seems then that in the tropics the GFS has been behind the curve for some years. 

The scenarios where the GFS appears to be more accurate anecdotally are in the northern hemisphere winter and more especially, zonal patterns (it seems to have a good grip on the dynamics of the various barlonic lows). 

[Seasonality]

This needs to be pinned somewhere for the upcoming winter silly season.

[Thundery wintry showers]

I recall that when we've had northerlies and little troughs running down the western flank of the northerly flow with embedded pools of warmer air, the GFS has tended to outperform the UKMO and ECMWF, as the latter two models tended to underestimate the extent of the warmer air and so forecast snow instead of rain or sleet.  The other main advantage of the GFS model is of course much of the data being freely available.

But the ECMWF has been the best-performing model overall since at least 2006.  There was a time when the UKMO model didn't do significantly better than the GFS, particularly at day 6, but I recall that it gained ground several years ago and these days tends to be a little behind the ECM overall at days 4-6 but a fair way ahead of the GFS.

Edited September 16, 2017 by Thundery wintry showers

[bluearmy]

In other news, water confirmed wet and ice is indeed cold ..............

[Cheshire Freeze]

Last winter, in terms of the UK, none of the models covered themselves in glory. In fact from memory, the GFS was ahead of the ECM (and it's longer range modelling) many times....anyone remember the phantom Euro troughing that never materialised?

So whilst the ECM  may be the best globally, it's less clear cut for the UK.

Edited September 16, 2017 by CreweCold

[bluearmy]

10 hours ago, CreweCold said:

Last winter, in terms of the UK, none of the models covered themselves in glory. In fact from memory, the GFS was ahead of the ECM (and it's longer range modelling) many times....anyone remember the phantom Euro troughing that never materialised?

So whilst the ECM  may be the best globally, it's less clear cut for the UK.

Depends which model you are looking at Crewe - the operational ec was generally good. It was the extended ens and 30 Dayer which drifted the wrong way - luckily exeter's own extended ens model didn't see it so we were never convinced enough to declare winter !