Skilful predictions of the winter North Atlantic Oscillation one year ahead

[xmetman]

Here is news on a breakthrough in the latest research regarding the forecasting of the NAO from the Met Office website:-

Here's the preamble to it on the Nature Geoscience website:-

If this is such an important breakthrough why don't the Met Office make it freely accessible after all we paid for it?

Why fork out $32 to read it when the Daily Mail seem to have done it for me - here is their take on the findings:-

You can imagine what most Mail readers comments were, and I tend to agree with them, 62% accuracy isn't particularly conclusive or a Earth shatteringly high percentage to base a forecast for the coming winter on. 

The statement I have outlined above from the Mail is incorrect as well as you might expect, the Met Office researchers have said that they can forecast the NAO for the coming winter to an accuracy of 62%. The NAO is one thing, translating that into what it means for the actual weather in the UK is something else entirely.

I find that seasonal forecasting are always so bland anyway, It may benefit the transport sector but not in any specific way. It's not going to say months in advance for instance, watch out for the second week of January especially on the M1. It would give local authorities a heads up and allow them to buy in some extra rock salt, but surely that shouldn't be necessary if you're managing it correctly.

[Len Wood]

Yes Bruce, 62% is not much to get very excited about.
I seem to remember Rodwell and Folland (2002) had a similar accuracy when they found a correlation between

May SST in the N. Atlantic and the following winter NAOI.

It was only just statistically significant though.

Reference:

Rodwell, MJ and Folland, CK (2002) Atlantic air-sea interaction

And seasonal predictability.

Q.J.R. Meteorol. Soc., 128, p1413-1443.

So 'current signals suggest the start to winter (November?) is likely to cooler and drier than 2015.'

Not a bad projection given that November 2015 was wetter and a good bit milder than average.

We shall see.

Len

Wembury

[John Hall]

62% doesn't sound that great, but the degree to which the forecast and actual NAO match is quite impressive whem you look at the results in graphical form. The main graph is reproduced in the online Telegraph article that covers it (though not in the one on the front page of today's paper version):

http://www.telegraph.co.uk/science/2016/10/17/met-office-can-now-predict-winter-weather-one-year-in-advance/

The good prediction of the low value for 2009-10 is particularly noteworthy.

[Richard Dixon]

I'm not much of a fan of these type of "showy" statement that underscore the latest bit of research - especially when the sort of research we're looking into will typically not find the magic bullet and it's more a case of small advances in the science that sound like major breakthroughs when run through a marketing department. Although I'm rather used to this in the industry I work in where the latest science is used to sell catastrophe risk models and things are always trumpeted...

Richard