EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Optimal scenario design for climate emulation

Christopher B Womack iD, Shahine Bouabid, Andrei Sokolov, Popat Salunke, Glenn Flierl iD, Sebastian D Eastham, Noelle E Selin iD

Proceedings of the National Academy of Sciences · 17.11.2026 · 123 (46)

DOI: 10.1073/pnas.2620731123 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 09.11.2026. Verfasst von Christopher B Womack und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge 02139, MA. Die Arbeit nennt 78 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Schmidt Sciences (Schmidt Futures).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As deep learning for physical systems continues to grow in popularity, efforts to improve generalizability have primarily focused on designing architectures that embed physical constraints. However, for machine-learning surrogate climate models (emulators), we show that the low structural diversity in existing scenarios commonly used to generate training data places a ceiling on predictive skill. Here, we examine whether training datasets themselves can be optimized to improve generalization. We introduce a method to create datasets that produce emulators capable of generalizing to new, structurally different scenarios absent from the training data. We use a differentiable simple climate model (SCM) to calculate the sensitivity of emulator loss to perturbations in the training data, iteratively updating the training data to maximize emulator skill. For an SCM, an emulator trained on one scenario optimized in this fashion outperforms an emulator trained on six standard Scenario Model Intercomparison Project (ScenarioMIP) pathways. We achieve this higher predictive skill despite training on a smaller dataset, finding that our emulator successfully isolates distinct physical behaviors of different climate forcing agents (e.g., greenhouse gases vs. aerosols) without single-forcing runs. We then demonstrate that scenarios optimized using an SCM, when used to drive an intermediate-complexity climate model, produce a training dataset that yields a more skillful emulator than training on ScenarioMIP outputs. Our results suggest that, in the compute-constrained environment of running full-scale climate models, generating a small number of dynamically rich scenarios provides greater marginal value for emulation and characterizing system responses than expanding the suite of traditional emissions pathways.

Schlagwörter

Stichwörter der Autor:innen

climate dynamicsclimate emulationmachine learningsurrogate modelingtraining data

Autor:innen und Institutionen

  • Department of Aeronautics and Astronautics, Massachusetts Institute of Technology, Cambridge 02139, MA.
  • Center for Sustainability Science and Strategy, Massachusetts Institute of Technology, Cambridge 02139, MA.
  • Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge 02139, MA.
  • Department of Aeronautics, Brahmal Vasudevan Institute for Sustainable Aviation, Imperial College London, London SW7 2AZ, United Kingdom.
  • Institute for Data, Systems, and Society, Massachusetts Institute of Technology, Cambridge 02139, MA.

ORCID-Profile: Christopher B Womack, Glenn Flierl, Noelle E Selin

Förderung

  • Schmidt Sciences (Schmidt Futures)

Zitieren

APA

Womack, C. B., Bouabid, S., Sokolov, A., Salunke, P., Flierl, G., Eastham, S. D., & Selin, N. E. (2026). Optimal scenario design for climate emulation. Proceedings of the National Academy of Sciences, 123(46). https://doi.org/10.1073/pnas.2620731123

Vancouver

Womack CB, Bouabid S, Sokolov A, Salunke P, Flierl G, Eastham SD, et al. Optimal scenario design for climate emulation. Proceedings of the National Academy of Sciences. 2026;123(46). doi:10.1073/pnas.2620731123

BibTeX
@article{Womack2026,
  author = {Womack, Christopher B and Bouabid, Shahine and Sokolov, Andrei and Salunke, Popat and Flierl, Glenn and Eastham, Sebastian D and Selin, Noelle E},
  title = {Optimal scenario design for climate emulation},
  journal = {Proceedings of the National Academy of Sciences},
  year = {2026},
  volume = {123},
  number = {46},
  doi = {10.1073/pnas.2620731123}
}
Kostenlos starten →