[R] Kendall v MannKendall Functions

Peter Ehlers ehlers at ucalgary.ca
Tue Mar 22 09:31:52 CET 2011


On 2011-03-21 14:16, armstrwa wrote:
> Hi,
>
> I am running a correlation analysis on a temporal dataset.  I was wondering
> if you would receive the same tau and p values running the function:
>
> MannKendall(x), where x is the dependant variable that changes with time
>
>   as you would running:
>
>   Kendall(d,x), where x is the exact same dataset as the x entered into
> MannKendall and d is the date on which the observation was made (assuming
> that the order is the same as the data I entered into the Kendall function).
>
> Can anyone elucidate this for me?
>

Assuming that you're referring to the (unstated) package Kendall,
a quick look at the code for MannKendall tells you that
MannKendall(x) just does Kendall(1:length(x), x).
Does that answer your question?

Peter Ehlers



More information about the R-help mailing list