Consistent geographical patterns of changes in high-impact European heatwaves

Climate-change projections suggest that European summer heatwaves will become more frequent and severe during this century1–4, consistent with the observed trend of the past decades5,6. The most severe impacts arise from multi-day heatwaves, associated with warm night-time temperatures and high relative humidity. Here we analyse a set of highresolution regional climate simulations and show that there is a geographically consistent pattern among climate models: we project the most pronounced changes to occur in southernmost Europe for heatwave frequency and duration, further north for heatwave amplitude and in low-altitude southern European regions for health-related indicators. For the Iberian peninsula and the Mediterranean region, the frequency of heatwave days is projected to increase from an average of about two days per summer for the period 1961–1990 to around 13 days for 2021–2050 and 40 days for 2071–2100. In terms of health impacts, our projections are most severe for low-altitude river basins in southern Europe and for the Mediterranean coasts, affecting many densely populated urban centres. We find that in these locations, the frequency of dangerous heat conditions also increases significantly faster and more strongly, and that the associated geographical pattern is robust across different models and health indicators. Heatwaves often lead to higher morbidity and mortality predominantly in the elderly, infants and persons with pre-existing cardiovascular and respiratory disease7. Substantial excessmortality has been observed during several recent heatwaves8,9, including the devastating summer of 2003, with around 40,000 heat-related deaths across Europe10,11. The climatic factors contributing to enhanced morbidity and mortality mainly relate to a combination of extremely high dayand night-time temperatures12, to high relative humidity13,14 and often to an extended duration (several days) of the heatwave9,15. Herewe analyse future changes in impact-relevant summer heatwave indices based on six high-resolution regional climate models (RCMs) of the ENSEMBLES multi-model scenario experiment16. The RCMs are driven by three different general circulation model runs (see the Methods section). The findings are compared to eight RCM simulations of the previous multi-model experiment PRUDENCE (Supplementary Information). The methodology involves analysis of the statistics and spatial patterns of daily summer temperature and heatwave indices in 30-year time slices. The selected indices account for the roles of extended duration, night-time temperatures and relative humidity of heatwaves.