extRemes (version 1.61)

return.level: Computes m-year return levels for GEV and GPD with confidence bounds.

Description

Computes confidence limits for m-year return levels for GEV and GPD with the confidence levels using the ``delta" method.

Usage

return.level(z, conf = 0.05, rperiods= c(10,100,210,510,810,980), make.plot = TRUE)

Arguments

z

An object of class "gev.fit" or "gpd.fit" from the gev.fit or gpd.fit functions from the ismev library of Stuart Coles (ported to R by Alec Stephenson). Note that z must be assigned one of the two classes or nothi

conf

The confidence level(s) for (delta) confidence bounds. Confidence intervals are not trustworthy for return periods beyond the range of the data. Check skewness of profile likelihood for a given return level--symmetric profile likelihoods imply tha

rperiods

Return periods at which to compute confidence limits.

make.plot

logical, if 'TRUE' plots the return levels with confidence bounds.

Value

• Creates a plot if 'make.plot' is TRUE and returns a list invisibly with components.
• return.level
  The m-year return levels.
• return.period
  The return periods, m.
• confidence.delta
  The confidence bounds computed by the delta method.
• conf.level
  The confidence level.

Details

The function also allows for the plotted values to be returned (invisibly).

Confidence levels are computed using the delta method as described in Coles (2001).

For the GEV return levels, return periods must be greater than 1 year.

References

Coles, Stuart. "An Introduction to Statistical Modeling of Extreme Values", Springer 2001. ISBN: 1852334592

Gilleland, Eric and Katz, Richard W. Tutorial for the 'Extremes Toolkit: Weather and Climate Applications of Extreme Value Statistics.' http://www.assessment.ucar.edu/toolkit, 2005.

See Also

gevrlgradient, gev.effective.rl, From the ismev package: gev.fit, gpd.fit, gev.diag, gpd.diag, gev.rl, gpd.rl

Examples

# Must have the 'ismev' package loaded.
require( ismev)
data( ftcanmax)
fit <- gev.fit( ftcanmax[,"Prec"])
class( fit) <- "gev.fit" # 'gev.fit' does not actually assign this class,
			 # but it must be done.
return.level( fit)