Estimating extreme environmental conditions remains a key challenge in the design of offshore structures. This paper describes an exercise for benchmarking methods for extreme environmental conditions, which follows on from an initial benchmarking exercise introduced at OMAE 2019. In this second exercise, we address the problem of estimating extreme metocean conditions in a variable and changing climate. The study makes use of several very long datasets from a global climate model, including a 165-year historical run, a 700-year pre-industrial control run, which represents a quasi-steady state climate, and several runs under various future emissions scenarios. The availability of the long datasets allows for an in-depth analysis of the uncertainties in the estimated extreme conditions and an attribution of the relative importance of uncertainties resulting from modelling choices, natural climate variability, and potential future changes to the climate. This paper outlines the methodology for the second collaborative benchmarking exercise as well as presenting baseline results for the selected datasets.