October Pattern Index! Predicting Winter AO from October, with 90+% accuracy?

[Crocodile23]

I'm surprised that weather forums aren't flooded with discussions about the new OCTOBER PATTERN INDEX.

 

Perhaps some of you don't know what this even is. And myself just learnt about it before 2 days.

So here is some read(unfortunately it is in Italian so you will have to use a translator whatever that means):

 

http://www.meteogiuliacci.it/articoli/october-pattern-index-opi-un-nuovo-indice-altamente-predittivo-per-la-stagione-invernale.html

http://forum.meteonetwork.it/meteorologia/150667-per-i-curiosi-disponibile-paper-opi.html

http://www.centrometeotoscana.it/forum/index.php?topic=7356.0

http://www.centrometeotoscana.it/forum/index.php?topic=7364.msg

http://forum.meteonetwork.it/meteorologia/150683-opi-october-pattern-index.html

 

 

The original link with the complete researce paper(which is 28 pages and it is also in Italian) was available for downloading but now it isn't. The reason if i understood correctly from the crappy google translation, is that the 3 authors will publish their research to a scientific journal and in order to do so, a free available copy is forbidden to exist.

 

The deal is simple. But its results(if correct) ARE AMAZING! EXTRAORDINARY!

It is based on a research job of Cohen with his snow advance index(SOI) where he managed to create this index to correlate it with AO(Arctic Oscillation) with very good results.

But the October Pattern Index(OPI) goes many steps further and manages to predict the winter AO(allegedly, by all the data the authors give in their paper and anyone can repeat and see by themselves) since it is an index we obtain fully at the end of October and then we can predict the next winter's AO with an amazing accuracy! They talk about 90% accuracy for the years(if i recall correctly) 1979 to 2012.

 

 And i repeat, that we can verify these results by applying the method the authors give in their paper(that is now not available, but some people had managed to get it in time) to search for possible mistakes. Note that the authors are not a joke, they are scientists and hobbyists at weather/meteorology, the one is astrophysicist for example.

 

So any thoughts?

I find the result highly annoying!  Because in a nonlinear chaotic system like the weather we should not be able to predict the next 4 months of October, or at least such a critical index of the next 4 months(like Arctic Oscillation), by just doing some "statistical investigations" on October. We shouldn't be able to do that!

Furthermore there is the logical flaw i find with this amazingly high correlation(between the October OPI and the Winter AO), that the solar activity on the winter has only a minor, if at all, role on how the winter will go.

If find it amazingly hard to believe that just an index, a statistical index which is rather simple to find or conceive the idea, by knowing its value at the end of October, is able to predict the whole winter circulation of the atmosphere, to be able to predict how the massive amount of different things that drive the climate like solar activity, QBO, AO, NAO, ENSO, SSTs, AMO, etc, etc, will combine with each other. This is utter nonsense if you ask me. It can't be done!

 

What the authors say with all these(if i understood the google translations good enough), is that the OPI at the end of October gives a distinct special stamp of the state of the atmosphere and its OPI will lead to a specific AO value/winter state. And this is ridiculous to say since it assumes an EXTREME OVER-OVERSIMPLIFICATION of the way the climate works.

But the problem is the data supports them!

 

So please help....

Edited October 9, 2013 by Crocodile23

[BornFromTheVoid]

Can't say too much about it without being able to see how they construct the index, but very interesting if it turns out to be correct.

 

Do any of those links explain how the index is created?

[Crocodile23]

Can't say too much about it without being able to see how they construct the index, but very interesting if it turns out to be correct.

 

Do any of those links explain how the index is created?

 

Not in full details, not even in small details. Only the original link with the complete paper of 28 pages that gave free yesterday for some time contains the whole method. I've managed to grab it in time but unfortunately i guess it's not legal to provide it as the authors don't provide it. I guess when it gets publiced it will be in English so it will get known and examined by experts and non experts much better....