Estimating Confidence Intervals of Long Return Period Design Values by Bootstrapping

Abstract

This paper describes a method for extrapolation of extreme value data for estimating long return period characteristic values. It is based on using yearly extreme value data subjected to a transformation which is derived from analysis of the underlying all-year data. The problem of establishing confidence intervals for the predicted return period values is discussed. It also demonstrates how the method of bootstrapping can be used for this purpose.