Efficient bootstrap estimates for tail statistics

Bootstrap resamples can be used to investigate the tail of empirical distributions as well as return value estimates from the extremal behaviour of the sample. Specifically, the confidence intervals on return value estimates or bounds on in-sample tail statistics can be obtained using bootstrap techniques. However, non-parametric bootstrapping from the entire sample is expensive. It is shown here that it suffices to bootstrap from a small subset consisting of the highest entries in the sequence to make estimates that are essentially identical to bootstraps from the entire sample. Similarly, bootstrap estimates of confidence intervals of threshold return estimates are found to be well approximated by using a subset consisting of the highest entries. This has practical consequences in fields such as meteorology, oceanography and hydrology where return values are calculated from very large gridded model integrations spanning decades at high temporal resolution or from large ensembles of independent and identically distributed model fields. In such cases the computational savings are substantial.

Efficient Bootstrap Estimates for Tail Statistics

Bootstrap resamples can be used to investigate the tail of empirical distributions as well as return value estimates based on the extremal behaviour of the distribution. Specifically, the confidence intervals on return value estimates or bounds on in-sample tail statistics can be estimated using bootstrap techniques. However, bootstrapping from the entire data set is expensive. It is shown here that it suffices to bootstrap from a small subset consisting of the highest entries in the sequence to make estimates that are essentially identical to bootstraps from the entire sequence. Similarly, bootstrap estimates of confidence intervals of threshold return estimates are found to be well approximated by using a subset consisting of the highest entries. This has practical consequences in fields such as meteorology, oceanography and hydrology where return estimates are routinely made from very large gridded model integrations spanning decades at high temporal resolution. In such cases the computational savings are substantial.