Extreme heat and drought typical of an end-of-century climate could occur over Europe soon and repeatedly

The study investigates how soon extreme heat and drought conditions characteristic of an end-of-century climate might occur in Europe, including the likelihood of such conditions repeating in successive years and occurring in compound form. It focuses on the role of internal climate variability—particularly the Atlantic Multidecadal Variability (AMV)—in modulating risks under continued warming. Prior events (e.g., the extraordinary 2010 summer heat in Europe) demonstrated that internal variability can bring forward extremes expected only in much warmer climates, yet a systematic quantification for Europe of the timing, frequencies, and compounding/successive nature of such end-of-century-like conditions has been lacking. The purpose is to quantify near-term risks for single and compound heat and drought extremes, their potential multi-year successions, and the decadal accumulation of stress, providing impact-relevant insights for preparedness and adaptation.

Past work shows increasing frequency and intensity of heat extremes with warming and the potential for record-shattering events to become far more likely. Documented impacts include substantial heat-related mortality (e.g., Europe 2003), ecological damage, reduced labor productivity, large wildfires, and agricultural losses. Compounding stressors (humidity, hot nights, drought) amplify impacts. Europe’s climate appears sensitive to multi-year North Atlantic variability (AMV), which can modulate European temperatures and hot-dry events via ocean–atmosphere teleconnections (e.g., barotropic wave trains leading to warm, dry anticyclonic conditions). Idealized and observational studies link positive AMV phases to warmer European conditions and altered precipitation. Large ensembles have been emphasized as crucial for low-probability, compound, and successive extreme event analysis. There is high confidence in projected increases of European heat extremes across models, while drought projections show larger inter-model spread and lower signal-to-noise, underscoring uncertainty for precipitation-related extremes.

Data and models: The study uses the 100-member Max Planck Institute Grand Ensemble (MPI-GE; MPI-ESM1.1, climate sensitivity ~2.8 °C) with fully coupled transient simulations under historical (1950–2005) and RCP4.5 (2006–2099) forcing. Observations from E-OBS v24.0e (1950–2021) are used for evaluation, regridded to model resolution, over European land [35–63°N, 10°W–55°E]. Excess metrics: Four single-season, cumulative excess metrics are defined to capture both intensity and persistence beyond region- and time-specific thresholds: - Excess Heat: JJA daily maximum temperature exceedances above the pooled daily 90th percentile (1950–1999 reference) summed over days and grid cells. - Humid Heat: JJA wet-bulb temperature-based exceedances above the daily 90th percentile (daily mean T and RH used; relative, not absolute values, due to daily mean RH availability). - Night Heat: JJA daily minimum temperature exceedances above the daily 90th percentile. - Rain Deficit: May–October precipitation deficits calculated as monthly 10th percentile threshold minus actual monthly precipitation, summed over months and grid cells. Only positive exceedances are accumulated. Sensitivity to stricter thresholds (95th for temperature, 5th for precipitation) yields qualitatively similar results because both end-of-century levels and exceedances are defined relatively. Compound metrics: Three compound measures assess co-occurrence: (i) Compound Heat Stress (Excess Heat with concurrent Humid Heat and/or Night Heat); (ii) Compound Heat and Drought (Excess Heat and Rain Deficit in the same year); (iii) Drought–Rain Volatility (years with extreme Rain Deficit preceded and/or followed by extreme Rain Excess, defined analogously with monthly precipitation above the 90th percentile). Definition of extremes and decades: Extreme years are those exceeding end-of-century typical levels, defined as the 50th percentile of the ensemble distribution averaged over 2090–2099 for each metric. Decadal metrics are computed as 10-year sums of annual excess metrics. Distance to typical end-of-century decade is calculated as the decadal value minus the 2090–2099 ensemble median (or mean for grid-cell maps), and shown as a percentage ratio in figures. Internal variability and AMV: An AMV index is defined as the 10-year running mean SST over [20–60°N, 70–20°W], detrended by removing the ensemble mean at each grid cell and normalized by its standard deviation. Positive (AMV+) and negative (AMV−) phases are set by thresholds of ±0.5 standard deviations. Analyses compare likelihoods of reaching/exceeding end-of-century levels under AMV+ vs AMV− for decades starting 2030–2049. Evaluation: MPI-GE performance is assessed against E-OBS for 1950–2021 in magnitude, variability, and forced trends of excess metrics; observations generally fall within the ensemble spread and perfect-model range. Notable observed extremes include 2010 for heat (~10× 1950–1999 average across heat metrics) and 2015 for Rain Deficit (~ensemble 90th percentile), with ensemble maxima exceeding observed rain deficits. Multi-model framing: A simplified cross-ensemble comparison (ACCESS-ESM1.5, CanESM5, MIROC6) at ~2 °C GMST above pre-industrial situates MPI-GE results: temperature extremes broadly consistent across models; larger spread for precipitation deficits.

- Rapid emergence of end-of-century extremes: - Heat metrics that were virtually impossible two decades ago reach ~1-in-10 likelihoods by 2030–2039. - By 2050–2074, the likelihood of a single year with end-of-century compound extremes (Compound Heat Stress, Compound Heat & Drought) rises to ~30–35% (about 1 in 3). - Successive extremes become plausible: - Two successive summers with end-of-century heat increase from near-zero historically to >15% in 2050–2074, with up to ~3% likelihood for 5 consecutive years of end-of-century heat. - For Rain Deficit, the chance of a single end-of-century level year is ~20% in recent decades, rising to >30% in the next 25 years; two consecutive years approach ~15%. Five-year continental-scale megadroughts become plausible by 2050–2074 (~>3.5%). - Magnitude of extremes by century’s end: - Typical end-of-century summers: Excess Heat ~10× the 1950–1999 average; extreme upper-tail summers (90th percentile and beyond) reach ~20–35× for Excess Heat and Night Heat; Humid Heat upper tail ~10–15×. Rain Deficit central tendency ~2×, with most extreme years ~3–8× the historical average. - Decadal accumulation and timing: - Decadal variability widens markedly under warming; 5–10% of decades exceed typical end-of-century levels for heat starting ~2040, and already in 2020–2029 for Rain Deficits. By 2060, chances of decades exceeding end-of-century levels rise to >10%. - Whole decades of end-of-century heat stress could begin by ~2040; for drought, by ~2020. - Spatial patterns and compounding: - Early likelihood increases for end-of-century heat are highest over Central and Northern Europe in the next 25 years. - Drought–Rain Volatility (extreme deficit adjacent to extreme excess) roughly doubles in the next 25 years and could occur about one-third of the time by 2050–2074. - Role of North Atlantic variability (AMV): - Positive AMV phases more than double the likelihood of reaching typical end-of-century decadal heat stress levels beginning in the 2030s, with the strongest and most widespread effects for Humid Heat and Night Heat. - Under AMV+, there is >1-in-10 chance that decades starting 2030–2049 exceed end-of-century levels for compound day–night heat, humid–night heat, and heat–drought. Heat- and drought-loaded decades are roughly twice as likely under AMV+ as under AMV−. - Observational context: - The 2010 European summer reached heat levels ~10× the 1950–1999 average across heat metrics, consistent with ensemble upper-tail behavior; 2015 was the most extreme rain-deficit year observed, near the ensemble 90th percentile.

The study shows that internal climate variability, superimposed on anthropogenic warming, can bring end-of-century-like heat and drought extremes to Europe much earlier than expected. By quantifying conditional and successive probabilities, it demonstrates that both single and compound extremes can repeat in consecutive years with non-negligible likelihoods by mid-century, elevating systemic risk for health, ecosystems, agriculture, and infrastructure. The decadal lens reveals that accumulated heat and drought stress can reach or exceed typical end-of-century levels as early as the 2030s–2040s, especially under positive AMV phases. Because the AMV exhibits comparatively higher decadal predictability than weather-scale extremes, its linkage to multi-year heat and drought stress offers a pathway to improved risk outlooks. These findings directly address the research question by specifying when and how often end-of-century levels may arise (including successions and compounding), and by isolating AMV’s modulation of these risks, thereby informing preparedness and adaptation planning across Europe.

Using a 100-member Earth system model ensemble and impact-relevant cumulative excess metrics, the study quantifies the near-term emergence of end-of-century-like heat and drought over Europe, including the likelihood of successive and compound extremes and decadal accumulations. Key contributions include: (i) identifying that end-of-century heat and drought levels can become ~1-in-10 events by the 2030s; (ii) showing that successive extremes (two or more consecutive years) become non-negligible by mid-century, with plausible 5-year continental-scale megadroughts; (iii) demonstrating that entire decades can reach end-of-century stress levels starting around 2040 for heat and 2020 for drought; and (iv) attributing a strong modulation to North Atlantic decadal variability, which roughly doubles the likelihood of such decades under warm phases. Future research should extend these analyses across multiple models and scenarios—especially for precipitation where inter-model spread is large—refine compound-event metrics and regional thresholds, improve representation of humidity and wet-bulb processes at sub-daily scales, and leverage decadal predictability (e.g., AMV) for anticipatory adaptation strategies.

- Single-model primary analysis (MPI-GE): while temperature extremes align well with multi-model assessments, precipitation and drought responses exhibit larger inter-model spread and lower signal-to-noise, making drought-related findings more model dependent. - Observational dataset caveat: E-OBS v24.0e includes some non-homogenized station series, warranting caution for trend evaluation. - Wet-bulb calculation uses daily mean relative humidity rather than instantaneous values, potentially overestimating absolute wet-bulb temperatures; analyses rely on relative changes, not absolutes. - Internal variability is sampled via initial-condition ensemble, but structural model uncertainty and scenario uncertainty (only RCP4.5) are not fully explored. - Decadal predictability remains imperfect; attributions of successive/compound extremes to AMV are statistical within the model framework and may vary across models.