Advances in understanding extreme meteorological events: a review of the UNSEEN methodology and its applications

Extreme meteorological events represent a serious threat to society, due to their strong impacts on mortality and economic losses. However, a deep comprehension of the features of such events is often limited by the scarce availability of reliable observations from the past. The UNSEEN (UNprecedented Simulated Extremes using ENsembles) method relies on the use of large ensemble simulations to investigate both the statistical and the dynamical features of events more severe than those seen in the historical record (Kelder et al. 2022). The study presents a literature review of the UNSEEN method, with particular attention to implications for downstream applications and conclusions reached on model realism. It is shown that the majority of studies addresses rainfall-related extremes with seasonal reforecasts, but the method is used for multiple hazards and with simulations ranging from medium-range weather predictions up to decadal and multi-decadal simulations. We emphasize that UNSEEN studies can rely on an unprecedented of high-resolution models and large ensembles. One of the main advantages emerged from the use of the UNSEEN method involves the possibility to reduce the uncertainty associated to the estimate of the statistics of extreme events, with respect to the only use of observations. Moreover, if the simulated events are considered physically consistent, this approach enables to better comprehend the dynamics of the processes that lead to the occurrence of extremes, including the possible role of remote teleconnection patterns. We discuss potential applications of the UNSEEN method that are currently underrepresented in the scientific literature and the opportunity to apply the UNSEEN method with a hierarchy of climate models to better understand the role of climate teleconnections in driving the emergence of the extreme event. An example application toward this goal has been conducted through the use of a coupled general circulation model of intermediate complexity (SPEEDY-NEMO) for the study of heat extremes in a region in Europe. Finally, we propose a workflow to generate a low-dimension catalogue of simulated plausible events that can be relevant to support adaptation studies and downscaling exercises.