Dear useRs,
I've been trying to run a Mann-Kendall test in my data in order to detect trends.
I studied the examples given at the Kendall package and I can understand pretty well how it works on time-series data.
However, my data consists of values in different sites per year, as I display below;
Year 1 | Year 2 | Year 3 | ...
Site 1 x x x ...
Site 2 x x x ...
Site 3 x x x ...
... ... ... ... ...
(where 'x' represents different values)
There's the MannKendall() function on package 'Kendall' and the tau() function on package 'pheno', and I guess they should do the trend detection I need.
The problem is I don't know how to manipulate my data in order to get the results.
Should I run the M-K test on each Site, on each Year or on the entire dataset?
Also, there are some probabilities I should take into account when running a M-K test, but I can't seem to find out how to obtain them.
Thanks in advance,
Rafael.
Veja quais são os assuntos do momento no Yahoo! +Buscados
[[alternative HTML version deleted]]