Abstract. With climate change, extremes such as heatwaves, heavy precipitation events, droughts and extreme fire weather have become more frequent in different regions of the world. It is therefore crucial to further their physical understanding, but due to their rarity in both observational and climate modeling samples, this remains challenging. For numerical simulations, one way to overcome this under-sampling problem is Ensemble Boosting, which uses perturbed initial conditions of extreme events in an existing reference climate model simulation to efficiently generate physically consistent trajectories of very rare extremes in climate models. However, it has not yet been possible to estimate the return periods of these simulations, since the conditional resampling alters the probabilistic link between the boosted simulations and the underlying original climate simulation they come from. Here, we introduce a statistical framework to estimate return periods for these simulations by using probabilities conditional on the shared antecedent conditions between the reference and perturbed simulations. We validate this framework with a simple red-noise process and find the typical time scale at which one could expect to sample stronger extremes. This is then applied to simulations of the fully-coupled climate model CESM2: first for a pre-industrial control simulation, and then in present-day conditions, where, as an example, we estimate the return period of the record-shattering 2021 Pacific Northwest heatwave to be 2500 2000–4000 years. Our evaluation of the method shows that return periods estimated from Ensemble Boosting are consistent with those of a 4000-year control simulation, while using approximately 6 times less computational resources. We thus outline the usage of Ensemble Boosting as an efficient tool for gaining statistical information on rare extremes. This could be valuable as a complement to existing storyline approaches, but also as an additional method of estimating return periods for real-life extreme events within a climate model context.
Bloin-Wibe et al. (Tue,) studied this question.