EUVIMEDEUVIMED entdecken
Open Access

Predicting regional gray swans via translocation: AI weather models and Dubai’s unprecedented 2024 rainfall

Y Sun iD, Pedram Hassanzadeh iD, Tiffany Shaw iD, Hamid A Pahlavan iD, Adam Marchakitus iD

Science Advances · 09.10.2026 · 12 (41)

DOI: 10.1126/sciadv.ady9406 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science Advances, erschienen am 09.10.2026. Verfasst von Y Sun und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, USA. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 94 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NSF, NSF ACCESS program, AI for Climate Initiative, Data Science Institute, University of Chicago.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Artificial intelligence (AI) models have transformed weather forecasting, but their skill for unprecedented weather extremes is unclear. Here, we analyze GraphCast, AIFS, and FuXi forecasts of the unprecedented 2024 Dubai storm, which had twice the training set’s highest rainfall in that region. GraphCast and AIFS accurately forecast this event up to 8 days ahead. FuXi forecasts the event but underestimates the rainfall. Fine-tuning and receptive field analyses suggest that these models’ success stems from “translocation”: learning from comparable/stronger dynamically similar events in other regions during training. Evidence of “extrapolation” (learning from weaker events) is not found. Even events within the global distribution’s tail are poorly forecasted, which is not only due to data imbalance (generalization error) but also spectral bias (optimization error). These findings demonstrate the potential of AI models to forecast “regional” gray swans and the opportunity to improve them through understanding the mechanisms behind their successes/limitations.

Autor:innen und Institutionen

  • Department of the Geophysical Sciences, University of Chicago, Chicago, IL 60637, USA.
  • State Key Laboratory of Severe Weather Meteorological Science and Technology, Nanjing University, Nanjing 210023, China.
  • Committee on Computational and Applied Mathematics, University of Chicago, Chicago, IL 60637, USA.
  • NorthWest Research Associates, Boulder, CO 80301, USA.

ORCID-Profile: Y Sun, Pedram Hassanzadeh, Tiffany Shaw, Hamid A Pahlavan, Adam Marchakitus

Förderung

  • NSF
  • NSF ACCESS program
  • AI for Climate Initiative, Data Science Institute, University of Chicago
  • NCAR'S CISL

Zitieren

APA

Sun, Y., Hassanzadeh, P., Shaw, T., Pahlavan, H. A., & Marchakitus, A. (2026). Predicting regional gray swans via translocation: AI weather models and Dubai’s unprecedented 2024 rainfall. Science Advances, 12(41). https://doi.org/10.1126/sciadv.ady9406

Vancouver

Sun Y, Hassanzadeh P, Shaw T, Pahlavan HA, Marchakitus A. Predicting regional gray swans via translocation: AI weather models and Dubai’s unprecedented 2024 rainfall. Science Advances. 2026;12(41). doi:10.1126/sciadv.ady9406

BibTeX
@article{Sun2026,
  author = {Sun, Y and Hassanzadeh, Pedram and Shaw, Tiffany and Pahlavan, Hamid A and Marchakitus, Adam},
  title = {Predicting regional gray swans via translocation: AI weather models and Dubai’s unprecedented 2024 rainfall},
  journal = {Science Advances},
  year = {2026},
  volume = {12},
  number = {41},
  doi = {10.1126/sciadv.ady9406}
}
Kostenlos starten →