extRemes (version 1.65)

gpdrlgradient: Gradient for the GPD Return Level

Description

Calculate the gradient for the GPD return level for computing delta method confidence intervals.

Usage

gpdrlgradient(z, m)

Arguments

z

a list object output from the ismev function gpd.fit

m

the N*npy return period (i.e., the actual return period for the N-year return level). May be a vector.

Value

• a 3 by np matrix whose columns give the (three) gradient values for each return period.

Details

Does not allow covariates in the parameter estimates. See Coles (2001) section 4.3.3, p. 82 for more information.

References

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

See Also

gevrlgradient, From the ismev package: gpd.fit, gpd.diag

Examples

data(rain)
fit <- gpd.fit(rain, 10)
gpdrlgradient( fit, m=c(10*fit$npy, 20*fit$npy, 50*fit$npy, 100*fit$npy))