Sensitivity of Heatwave Characteristics to the Climatic Baseline

Abstract

Temperature extremes are usually defined relative to conditions during a past baseline period. Due to global warming, different baselines produce different results for heatwave frequency, intensity, spatial extent and duration. This study investigates how the choice of baseline affects summer heatwave metrics across Europe from 1950 to 2023. We evaluate regional and temporal differences in heatwave characteristics using both grid-point analysis and a case study of the 2003 European heatwaves using observational daily maximum summer temperature for three different baselines (1961–1990, 1981–2010 and 1991–2020). Results show that more recent baselines yield a substantial reduction across Europe in observed heatwave frequency, of up to 12 days per summer in some southern European regions. The average intensity of heatwaves relative to the 90th percentile decreases by up to 2.0°C when using more recent baselines. The largest reductions are observed in southern and central Europe. The spatial extent of the 2003 European heatwave declines by more than 0.5 million km² if using more recent baselines, and the duration shortens by up to 2 days as conditions that would have been extreme in the past are nearer the new normal. However, 2003 is still classified as a mega-heatwave under both baselines. Comparisons of fixed, detrended and rolling reference approaches show that methods that vary baselines, assuming linear warming—despite its unrealistic nature—yield similar results to rolling reference windows from 1960 onward.