ABSTRACT
The underlying philosophy for the statistical analysis of extreme values is to “let the tail speak for itself.” In practice, this is often done by analyzing only data that are considered to be extreme. Additionally, the goal of an extreme value (EV) analysis is often extrapolation; one may have data with a record length of 50 years, but need to estimate the magnitude of an event which has an annual exceedance probability (AEP) of.01; that is the so-called ‘100-year event’. The foundation of EV theory is results which characterize the limiting distribution of sample maxima. GCMs (formally ‘general circulation models’, but colloquially ‘global climate models’) are tools for understanding Earth’s climate via simulation. As it is impossible to perform experiments on the Earth itself, GCMs are the best tools available for learning how different forcings, such as increased greenhouse gas concentrations, affect climate.
