European hot and dry summers are projected to become more frequent and expand northwards

Heatwaves and dry spells pose significant climate hazards impacting human health, economies, agriculture, and ecosystems. The compounding effect of simultaneous heatwaves and droughts results in amplified negative consequences. Europe has witnessed a series of such CHDs in recent years (e.g., 2003, 2015, 2018), leading to increased mortality, economic losses (crop failure, transport disruptions), decreased worker productivity, and severe stress on ecosystems. Quantifying the probability of CHDs is complex due to the interplay of temperature and precipitation, both influenced by large-scale atmospheric drivers like anticyclonic conditions. A self-intensifying feedback loop exists: heatwaves dry the soil, reducing evapotranspiration and leading to higher air temperatures and further drying. This feedback is enhanced during CHDs, with a higher Bowen ratio (sensible to latent heat) and lower soil moisture. Studies show a strong correlation between summer temperatures and precipitation, indicating more frequent CHDs than expected from univariate probabilities. Future climate projections indicate rising temperatures and drying trends, potentially intensifying these feedback mechanisms. Understanding the probability of historical CHDs under various global warming levels (GWLs) is crucial for stakeholders (policymakers, engineers, farmers) to manage resources, adapt agricultural practices, and mitigate impacts. Increased CHD probability necessitates developing alternative solutions for cooling power plants, water conservation measures, alternative transportation methods, and creating green spaces in urban areas.

Previous research has highlighted the significant impacts of compound extreme weather events like CHDs. Studies have documented the high mortality rates associated with past European heatwaves such as the 2003 heatwave, and the economic losses resulting from reduced crop yields and disrupted transportation. Existing literature explores the complex interactions between temperature and precipitation, emphasizing the self-reinforcing feedback loop between heat and drought. While some studies have examined the probability of individual extreme events (heatwaves or droughts), fewer have focused on the joint probability of CHDs, particularly under future climate scenarios. The use of large ensembles of climate model simulations to assess the probability of extreme events is gaining traction, but this approach has not been extensively applied to compound events like CHDs. Previous studies using various methods and datasets have indicated an increased risk of CHDs in Europe under future warming, but there is a need for a comprehensive analysis using a high-resolution large ensemble to provide more robust and regionally specific probability estimations.

This study utilizes a Single Model Initial Condition Large Ensemble (SMILE) approach, specifically the Canadian Regional Climate Model version 5 Large Ensemble (CRCM5-LE), to overcome limitations associated with short observational records. The CRCM5-LE consists of 50 ensemble members, each with the same forcing but different initial conditions, allowing for robust statistical estimation of extremes while distinguishing between internal climate variability and forced climate change responses. The analysis focuses on identifying the most extreme CHDs in Europe between 2001 and 2022. The study uses the ERA5 reanalysis dataset to identify past extreme summers and evaluates their probabilities of occurrence in three periods from the CRCM5-LE representing different GWLs (+1.2 °C (PRES), +2 °C (GWL2), +3 °C (GWL3)). The CHD definition is based on seasonal summer averages (June-July-August, JJA) of temperature and precipitation. To account for the interdependency of temperature and precipitation, the study uses copulas, specifically focusing on the Survival Kendall probability (Psk), which estimates the probability of an event at least as rare as the observed event. Spatial clustering is applied to identify nine European sub-regions with similar CHD patterns. The study compares the probabilities of the identified extreme CHDs across the three GWL periods, analyzing both bivariate (temperature and precipitation) and univariate (temperature and precipitation separately) probabilities. Kernel density estimation is employed to visualize the changes in the bivariate distributions of temperature and precipitation anomalies. To ensure the robustness of results, the analysis is repeated with another regional climate model large ensemble (CCLM), driven by CESM, allowing for model comparison. To investigate the future climatology of CHDs, the study matches the CHD distributions under GWL3 with those of the present climate to identify analogous conditions and assess the northward shift in CHD zones.

The analysis identified the most extreme CHDs on both the European and regional scales from 2001-2022. The summer of 2003 was identified as the most extreme on a continental scale. The study found that the probability of the identified CHDs significantly increases under GWL2 and GWL3 scenarios. Probabilities rise considerably (up to 5-6 times) when moving from GWL2 to GWL3. For some CHDs (2002, 2003, 2018), the probability under GWL3 reaches values above 5%, indicating an occurrence nearly every other summer. Other CHDs (2006, 2012, 2015), characterized by extremely low precipitation, show only a moderate increase in probability, remaining relatively rare events. The analysis reveals a clear northward shift in CHD climatology under GWL3. Regions currently experiencing relatively mild CHDs will see an intensification, while the most extreme CHD conditions currently observed in Southeastern Europe are expected to extend into substantial parts of Central Europe, the Baltic Sea coast, and even Scandinavia. This northward expansion of CHD-like climates is illustrated by comparing GWL3 distributions with present-day climate conditions. The study uses Kullback-Leibler divergence to identify analogous regions, demonstrating the northward shift.

The study's findings strongly support the assertion that CHDs are projected to become significantly more frequent and expand northwards in Europe under future climate change scenarios. The significant increase in probability, especially under GWL3, underlines the importance of mitigating global warming to limit the severity of these events. The twofold pattern observed – some CHDs becoming much more frequent while others remain relatively rare – highlights the regional variations in the impact of future warming. This variation is primarily attributed to differences in the relative contributions of temperature and precipitation extremes. The northward shift of CHD climatology indicates a major redistribution of climate zones, leading to unprecedented conditions in previously less-affected northern regions. This underscores the need for adaptation strategies tailored to the specific challenges posed by intensified and geographically shifted CHDs.

This research demonstrates the increasing frequency and northward shift of compound hot and dry summers in Europe under future climate change. Using a high-resolution large ensemble, the study provides robust probability estimations for historical CHDs under different global warming levels, revealing a marked increase in frequency, especially under a +3°C warming scenario. A significant northward expansion of CHD-like conditions is projected. Future work could include a finer temporal resolution (monthly instead of seasonal), investigation of other contributing factors to CHD formation, and an expanded analysis of other types of compound extreme events.

The study's use of a summer-season definition might exclude CHDs that begin earlier or extend beyond the typical summer months. The analysis could also include seasons with high internal variability in temperature and precipitation. The study employs a CHD definition relative to local climatology, which is appropriate for local ecological impacts but might not capture all aspects related to broader societal impacts. Model biases and uncertainties inherent in climate models should be considered when interpreting the results, and further research with other high-resolution large ensembles is needed. Focusing on only temperature and precipitation neglects other extreme variables that might influence CHD probability.