Ultee et al. (2026) CMIP6 climate model spread outweighs glacier model spread in 21st-century drought buffering projections

Identification

• Journal: ˜The œcryosphere
• Year: 2026
• Date: 2026-02-25
• Authors: Lizz Ultee, Finn Wimberly, Sloan Coats, Jonathan Mackay, Erik Holmgren
• DOI: 10.5194/tc-20-1339-2026

Research Groups

• Cryospheric Sciences Lab, NASA Goddard Space Flight Center, Greenbelt, MD, USA
• GESTAR-II Cooperative Agreement, Morgan State University, Baltimore, MD, USA
• Woods Hole Oceanographic Institution, Woods Hole, MA, USA
• Department of Earth Sciences, University of Hawaii at Manoa, Honolulu, HI, USA
• Environmental Science Centre, British Geological Survey, Keyworth, UK
• Department of Space, Earth and Environment, Chalmers University of Technology, Gothenburg, Sweden

Short Summary

This study quantifies 21st-century glacial drought buffering across 75 major river basins using an ensemble of three global glacier models forced by 11 CMIP6 climate models, finding that climate model uncertainty significantly outweighs glacier model uncertainty in projections.

Objective

• To examine changes in glacial drought buffering for 75 major river basins in the early, mid-, and late 21st century, evaluating the sensitivity of these projections to the choice of global glacier model and climate model forcing.

Study Configuration

• Spatial Scale: 75 large river basins globally (basin area ≥ 3000 km², glacier cover ≥ 30 km²), spanning North and South America, Europe, Asia, and New Zealand. Glaciers are simulated on a per-glacier basis.
• Temporal Scale:
  • Historical period: 2000–2014 (glacier model forcing), 1900–1979 (SPEI reference period).
  • Projection period: 2015–2100 (SSP2-4.5 scenario).
  • Analysis periods: Early century (2000–2020), Mid-century (2040–2060), Late century (2080–2100).

Methodology and Data

• Models used:
  • Glacier Models: Global Glacier Evolution Model (GloGEM), Python Glacier Evolution model (PyGEM), Open Global Glacier Model (OGGM).
  • Climate Models: 11 CMIP6 global climate models (GCMs) for forcing, compared against a larger 123-member CMIP6 ensemble (from 28 distinct GCMs) for context.
  • Drought Index: Standardized Precipitation-Evapotranspiration Index (SPEI), specifically SPEIN (no dynamic glacier runoff) and SPEIG (with glacier runoff).
• Data sources:
  • Glacier outlines: Randolph Glacier Inventory v6 (RGI Consortium, 2017).
  • Basin outlines: Global Runoff Data Centre (GRDC, 2020).
  • Observed glacier mass change: Hugonnet et al. (2021).
  • Reanalysis products for bias correction: ERA5 (Hersbach et al., 2020) and W5E5 (Lange et al., 2021).
  • CMIP6 climate model simulations (SSP2-4.5 scenario).

Main Results

• Glacial drought buffering is projected to continue through the 21st century and is robust across different GCM forcings and glacier models.
• The strength of buffering correlates with basin glacier coverage, being strongest in moderately glacierized basins and decreasing for the most heavily glacierized ones.
• Glacial buffering generally persists or increases through the late 21st century; about 45 % of glacier model-basin pairs show no change, and 38 % show increased buffering.
• Qualitative patterns of glacial drought buffering over time are consistent across the three glacier models.
• Inter-GCM spread in glacial drought buffering is substantial and always exceeds inter-glacier-model spread, with ratios ranging from 1.7 to 133.
• The 11-member GCM ensemble used for glacier forcing under-samples the full CMIP6 archive's uncertainty, showing a more limited range and biased low mean changes in precipitation and temperature compared to a 123-member CMIP6 ensemble.

Contributions

• Provides the first comprehensive intercomparison of glacial drought buffering projections using multiple global glacier models forced by a common CMIP6 climate model ensemble.
• Demonstrates that large absolute differences in glacier runoff among models may be relatively unimportant for downstream drought metrics (SPEI).
• Highlights that uncertainty from climate model forcing (GCM spread) is significantly more important than uncertainty from glacier model physics for large-scale hydrological projections in glacierized river basins.
• Offers guidance for future glacio-hydrological studies, emphasizing the imperative of analyzing a representative ensemble of climate models.

Funding

• Natural Environment Research Council (grant no. NE/W004895/1)

Citation

@article{Ultee2026CMIP6,
  author = {Ultee, Lizz and Wimberly, Finn and Coats, Sloan and Mackay, Jonathan and Holmgren, Erik},
  title = {CMIP6 climate model spread outweighs glacier model spread in 21st-century drought buffering projections},
  journal = {˜The œcryosphere},
  year = {2026},
  doi = {10.5194/tc-20-1339-2026},
  url = {https://doi.org/10.5194/tc-20-1339-2026}
}