Ah, we just found out why: spcorr doesn't allow missing values, which we do have in one of the time series. Now fixed, sorry for the bother.
Emilie
On Dec 30, 2010, at 15:03, Emilie wrote:
> Hello,
>
> I have been using some correlation functions with NCL for quite a while now, and a recent comparison of 'spcorr' (Spearman's rank correlation) with the equivalent function in R show that the correlation factors obtained are different between both programs.
>
> An example:
>
> Site R Spearman R Spearman
> --------------------------------
> Site 1 0.45 0.47
> Site 2 0.39 0.50
> Site 3 0.50 0.51
>
> Any idea as to why this happens?
>
> Happy New Year 2011 to everyone,
> Emilie
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu Dec 30 15:39:16 2010
This archive was generated by hypermail 2.1.8 : Tue Jan 04 2011 - 09:16:52 MST