Bevacqua et al. (2026) Moderate global warming does not rule out extreme global climate outcomes

Identification

• Journal: Nature
• Year: 2026
• Date: 2026-03-25
• Authors: Emanuele Bevacqua, Erich M Fischer, Jana Sillmann, Jakob Zscheischler
• DOI: 10.1038/s41586-026-10237-9

Research Groups

• Department of Compound Environmental Risks, Helmholtz Centre for Environmental Research - UFZ, Leipzig, Germany
• Institute for Atmospheric and Climate Science, Department of Environmental Systems Science, ETH Zurich, Zurich, Switzerland
• Research Unit for Sustainability and Climate Risks, University of Hamburg, Hamburg, Germany
• CICERO Center for Climate Research, Oslo, Norway
• Department of Hydro Sciences, TUD Dresden University of Technology, Dresden, Germany

Short Summary

This study reveals that extreme global climate outcomes for several sectors (e.g., droughts, precipitation, fire weather) may occur even under a moderate 2 °C global warming, often exceeding model-averaged projections for 3 °C or 4 °C warming, primarily due to large uncertainties in climate model projections.

Objective

• To assess whether global climate-sensitive sectors may face potential high-impact climate outcomes even under moderate warming levels (e.g., 2 °C), challenging the common perception that worst-case futures are only associated with extremely high global warming levels.

Study Configuration

• Spatial Scale: Global, with a focus on specific critical regions: highly populated areas, global key breadbasket regions, world's forests, low-income countries, and global fisheries. Data were interpolated to a 2.5° spatial grid for analysis.
• Temporal Scale: Preindustrial period (1851–1900) as a baseline. Future warming levels of 2 °C, 2.5 °C, 3 °C, and 4 °C above preindustrial, identified using the earliest 30-year windows where global-mean temperature exceeded the target warming level.

Methodology and Data

• Models used: Coupled Model Intercomparison Project Phase 6 (CMIP6) climate models (one ensemble member per model, typically r1i1p1f1). Single Model Initial-condition Large Ensembles (SMILEs) (e.g., CanESM5, MIROC6) were used to quantify the contribution of internal climate variability.
• Data sources:
  • CMIP6 climate data (daily precipitation (pr), monthly total column soil moisture (mrso), daily maximum temperature (tasmax), monthly sea surface temperature (tos), monthly global-mean temperature (tas)).
  • Fire Weather Index (FWI) annual maxima data.
  • Gridded Population of the World (GPWv4) dataset for highly populated areas (2020 data).
  • Global breadbasket definitions from existing literature (union of maize, wheat, soybean, and rice breadbaskets).
  • European Space Agency Land Cover dataset for forest grid cells.
  • Low-income countries list from existing literature.
  • The Union of World Country Boundaries and EEZs (version 3) dataset for fisheries.

Main Results

• For several sectors, extreme global climate outcomes at 2 °C global warming can be much more severe than the multimodel mean projections at 3 °C or 4 °C warming.
• Precipitation Extremes (Highly Populated Areas): Worst-case 2 °C warming scenarios project increases in maximum consecutive 5-day precipitation (Rx5day) across populated areas that exceed the multimodel mean at 3 °C warming. Projections vary widely across models, from a 4% to 15% increase.
• Droughts (Global Breadbaskets): Worst-case 2 °C warming scenarios show an increase in drought frequency across breadbaskets that is well beyond the multimodel mean at 4 °C warming (e.g., a shift from 20% to over 70% frequency in some models).
• Fire Weather Extremes (Global Forests): Worst-case 2 °C warming scenarios project increases in the Fire Weather Index (FWI) annual maxima across forests that are larger than the multimodel mean projection at 3 °C warming (e.g., a range from +1.5 to +6.5 relative to preindustrial conditions).
• Heatwaves (Low-Income Countries): The worst-case 2 °C warming scenario for annual maximum daily maximum temperature aligns with the multimodel mean at 3 °C warming.
• Marine Heatwaves (Global Fisheries): Projections for 10-year return level of sea surface temperature show a narrower global-scale uncertainty at 2 °C warming (ranging from 1.50 °C to 1.78 °C increase relative to preindustrial), but still exhibit very different spatial patterns between best- and worst-case outcomes.
• The uncertainty in globally averaged projections (quantified by the global climatic impact-driver 'f') at 2 °C warming is substantial, often comparable to or even 12 times larger than the difference between the multimodel mean at 4 °C and 2 °C warming, depending on the sector.
• Spatially incoherent worst-case scenarios (assuming the worst local outcomes from different models at each location) significantly overestimate global-scale uncertainty (by approximately 80% to 290%).
• Simultaneous worst cases across all five considered sectors are unlikely, suggesting sector-specific drivers and regions.

Contributions

• Introduces a novel, bottom-up approach to identify sector-specific, spatially consistent potential high- and low-impact global climate outcomes by spatially averaging projected sector-relevant climatic impact-drivers across key global regions.
• Demonstrates that extreme global climate outcomes are possible even under moderate 2 °C warming, challenging the common perception that worst-case scenarios are only associated with very high warming levels.
• Highlights that large uncertainties in climate model projections, primarily stemming from model differences rather than internal climate variability, are evident even at moderate warming and lead to significant sector-specific high-impact outcomes.
• Underscores the urgency of rapid mitigation to limit warming well below 2 °C, emphasizing that even a 2 °C warmer world does not necessarily safeguard against severe sectoral global climate impacts.
• Provides a methodology adaptable to a wide range of sectors to improve climate risk assessment and inform climate policy, particularly for stress-testing adaptation measures.
• Suggests a more targeted use of climate models in impact modeling, moving beyond uniform subsets to identify sector-specific extreme outcomes that merit explicit simulation.

Funding

• Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) via the Emmy Noether Programme (grant ID 524780515).
• European Union’s Horizon 2020 research and innovation programme (grant agreement number 101003469).
• Helmholtz Initiative and Networking Fund (Young Investigator Group COMPOUNDX, Grant Agreement VH-NG-1537).

Citation

@article{Bevacqua2026Moderate,
  author = {Bevacqua, Emanuele and Fischer, Erich M and Sillmann, Jana and Zscheischler, Jakob},
  title = {Moderate global warming does not rule out extreme global climate outcomes},
  journal = {Nature},
  year = {2026},
  doi = {10.1038/s41586-026-10237-9},
  url = {https://doi.org/10.1038/s41586-026-10237-9}
}