Timo Kelder , Tim Marjoribanks , Louise J. Slater , Christel Prudhomme, Rob Wilby, Julia Wagemann, Nick Dunstone

Ensemble members from weather and climate predictions can be used to generate large samples of simulated weather events, allowing the estimation of extreme (hitherto unseen) events. Here, we provide a protocol and open workflow for applying the ‘UNSEEN’ method for hydro-climatic extremes globally, based on Copernicus Climate Change Services (C3S) seasonal predictions but also considering other compatible modelling systems. We discuss common challenges and potential solutions using three examples of extreme events that caused severe damage in 2020 (extreme rainfall, heat, and wildfire danger). These case studies demonstrate the potential of the method to inform decision-making with maximum credible events used for stress-testing adaptation measures and to anticipate unprecedented extremes in a changing climate. As such, this paper may be used to guide the generation of large ensembles that are a credible resource for evaluating otherwise unforeseen hydro-climatic risks.