extRemes (version 1.61)

gevrlgradient: GEV return level function gradient.

Description

Calculates the gradient of the GEV return level function for computing the GEV return level confidence intervals by way of the delta method (e.g., Coles, 2001 section 3.3.3, pp. 56-57)

Usage

gevrlgradient(z, p)

Arguments

z

list object output from the gev.fit function from ismev.

p

Vector of 1/p return periods (e.g., if block maxima are annual maxima, and the 20-year return level is desired, then use p=1/20).

Value

• A '3 X np' matrix whose columns give the (three) gradient values for each return period.

References

Coles S, 2001 (2nd prt, 2003). An Introduction to Statistical Modeling of Extreme Values. Springer-Verlag, London. 208pp. ISBN: 1852334592

See Also

return.level, From the ismev package: gev.fit, gev.diag

Examples

data(portpirie)
fit <- gev.fit(portpirie[,2])
gevrlgradient( fit, p=c(1/10, 1/20, 1/50, 1/100, 1/1000))
class( fit) <- "gev.fit"
return.level( fit, conf=0.05, rperiods=c(10, 20, 50, 100, 1000))