Climate Change on Extreme Winds Already Affects Wind Energy Availability in Europe

Climate change is one of the most urgent challenges that humankind confronts nowadays. In order to mitigate its effects, the European Union aims to be climate-neutral, i.e. set the Greenhouse Gas (GHG) emissions to zero, by 2050. In this context, renewable energies (REs) play a key role: on the one hand their development and extensive usage can help to reduce the GHG emissions, on the other hand substantial local changes in atmospheric conditions could modify, for better or for worse, their efficiency. Extreme atmospheric events, in particular, can badly affect the efficiency of the RE infrastructures, preventing them from working or even damaging them. In this work, we focus on wind energy off shore, on the European panorama, with the purpose of estimate the behavior of extreme high winds, over the period 1950-2020, and their impact on wind energy availability. Indeed, the potential wind power production, according to the working regimes of a wind turbine, depends only on the wind speed and, over a certain wind speed threshold, called cut-off speed (25 m s^-1), the turbine stops working. By using 6-hourly ERA5 reanalysis data-set and convection permitting simulations, covering the European domain and a period from 1950 to 2020 and from 2000 to 2009 respectively, we analysed the 100 m wind speed over the cut-off threshold and its relation with the geopotential height at 500 hPa, in order to investigate the large-scale weather regimes related to these extreme events. We focused especially on five regions, where high winds flow more frequently: United Kingdom, Denmark, Greece, and the areas off the south of France and north of Spain. By using the Mann-Kendall test, we analysed the trends in the occurrence of extreme events, and we detected significant increasing trends in large areas of the regions selected, particularly during the winter period (DJF). Finally, considering only the events over the 99^th percentile, we found that they are often concurrently with storms, and, by means of the K-means clustering algorithm, we identified the different weather regimes at which they occur.