On the application of decadal climate predictions to generate a catalogue of heat extremes

Heat extremes represent a serious threat for society due to their strong impacts on mortality and economic losses. Amongst the investigation tools that rely on numerical weather predictions to gain insights about weather extremes, the UNSEEN (UNprecedented Simulated Extremes using Ensembles) method has been employed extensively with promising results: it consists in building large and realistic ensembles that help understanding the features of extreme events. In this work a literature review has allowed to comprehend the advantages and drawbacks of this method and a new workflow based on the UNSEEN has been proposed. Through a cluster analysis, this new approach aims to build a catalogue of extreme events that is made up by the most recurrent events, the most severe ones and exotic scenarios. This new methodology is tested in a case study about summer heat extremes over a region in Northern Italy, by using a set of decadal hindcasts provided by the intermediate-complexity SPEEDY-NEMO model within the BONSAI (BOosting eNsemble Size for Advanced Insights into climate predictability) project. The added value of using decadal predictions from a lower complexity model consists both in the opportunity to enlarge ensemble size to improve the forecast skill, without lacking the resolution needed to analyze the large-scale processes leading to extremes. Once performed the clustering and evaluated the frequent, severe and exotic events, the catalogue obtained shows the onset of anticyclonic conditions near the regions with high 2m air temperature anomalies and confirms that an Atlantic Low pattern favors heat extremes. A comparison with the ERA5 reanalysis demonstrates how the model succeeds in reproducing past hot events, but with some discrepancies outside the euroatlantic sector. A more in-depth single event study leads to the detection of some SST anomalies signals in the Mediterranean and the North Atlantic that could enhance intensity and duration of heat extremes.