Re: Cross-Correlation with NCL (function esccr)

You will be sent a fortran version of the code used
in a separate email.

On 11/28/11 7:29 AM, daniel.buettner wrote:
> Hello everyone,
>
> I would like to know about the exact direction of the lag when doing a
> cross-correlation between variables x and y. More precisely, I want to
> know what the following (taken from the discription of the esccr
> function) really means:
>
> x_Lead_y = esccr(x,y,mxlag)
>
> Assume, just for simplicity, both x and y have values for 5 timesteps
> (t0,t1,t2,t3,t4). For lag=0 this would mean a correlation between
>
> x(t0),x(t1),x(t2),x(t3),x(t4)
> y(t0),y(t1),y(t2),y(t3),y(t4)
>
>
> If we have mxlag=1, to which direction is y shifted when applying
> x_Lead_y = esccr(x,y,mxlag)?
>
> Is it in the way that the following values are correlated?
>
> x(t0),x(t1),x(t2),x(t3)
> y(t1),y(t2),y(t3),y(t4)
>
> Or the other way round:
>
> x(t1),x(t2),x(t3),x(t4)
> y(t0),y(t1),y(t2),y(t3)
>
>
> I would suggest it's done in the former sense because this is how I
> would interpret the term "leading". But I would be happy if someone can
> make this perfectly clear to me.
>
>
> Thanks for any help!
>
>
> Cheers,
> Daniel
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Nov 28 08:05:02 2011