trendLogisPPY: Performing logistic regression to test for trend in number of...

In martindurocher/floodRFA: A R-package for regional frequency of floods

Description

Return the statistics and the p-value of a null trend. By defaut a t-test verify that the slope is null otherwise a F-test is used to verify that the model is better than random. A variable dispersion parmeter is used to account for potential autocorrelation.

Usage

trendLogisPPY(obj, span = NULL, method = "T", trend = "poly",
  degree = 2)

Arguments

obj	Output from fitPot.
span	List
method	Type of test to use. Eiter 'T' or 'F'
trend	Type of trend. Must be with method 'F'. Either poly for polynomial or 'spline' for b-spline polynomial with equally space knots
degree	Degree of the polynomial or dimension of spline basis .

References

Frei C, Schär C. (2001), Detection Probability of Trends in Rare Events: Theory and Application to Heavy Precipitation in the Alpine Region. J Climate. Apr 1;14(7):1568–84.

See Also

mannKendall, glm.

Examples

fit <- fitPot(flow~date, canadaFlood$daily, u = 1000,
               declust = 'flood', r = 14)

mannKendallPeaks(fit, lag.k = 0)

trendLogisPPY(fit, span = canadaFlood$daily$date)  ## linear trend
trendLogisPPY(fit, method = 'F', trend = 'spline')

martindurocher/floodRFA documentation built on June 5, 2019, 8:44 p.m.