mannKendall: Mann-Kendall test for trend
In martindurocher/floodRFA: A R-package for regional frequency of floods

Description Usage Arguments Details See Also Examples

Return the results of the Mann-Kendall test for trends. In case of autocorrelation, the p-value of the test is obtained by block bootstrap. This function is a wrapper for the function MannKendall.

mannKendall(obj, ...)

## S3 method for class 'atSite'
mannKendall(obj, lag.k = NULL, lag.max = 3, nboot = 1000,
  use.rank = TRUE, ...)

mannKendallPeaks(obj, ...)

`obj`	Output of function `atSite` or `fitPot`.
`...`	Other parameter pass to `tsboot`.
`lag.k, lag.max`	Lag size for the bootstrap blocks. If `lag.k` is `NULL` it is automatically selected and force to be inferior to `lag.max`.
`nboot`	Number of bootstrap sample.
`use.rank`	Does ranks should be used to transform data into uniform [0,1] variable or the model parameter should be used. Necessary for the selection of `lag.k`.

The automatic selection of the block bootstrap is chosen according to the best order of an autoregressive model using the AIC criteria and Yule-Walker estimates. See ar for more details. Previously, the ranks of the original data are transformed into normal variable using either the estimated distribution or the ranks.

The function mannKendallPeaks is performing the Mann-Kendall on the magnitude of the peaks for a POT analysis.

atSite, MannKendall, ar.

library(lmomco)

# TS with trend
xset <- rlmomco(100, vec2par(c(100,20,0), 'gev')) + (1:100)/5

ax <- atSite(xset)
mannKendall(ax)

mannKendall(ax, lag.k = 0)
mannKendall(ax, lag.k = 3)