Nie et al. (2026) Diagnosing Forecast Error Propagation and Large‐Scale Dynamics of Weather Extremes With an AI Weather Model

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Geophysical Research Letters
• Year: 2026
• Date: 2026-09-14
• Authors: Yanbo Nie, Agniv Sengupta, Jorge Baño‐Medina, Luca Delle Monache
• DOI: 10.1029/2026gl123505

Research Groups

• Department of Atmospheric Science, University of California, Los Angeles (UCLA)
• National Center for Atmospheric Research (NCAR)

Short Summary

This study proposes a true-state constraint method to diagnose error propagation in artificial intelligence (AI) weather models and explores its application to predict regional extremes. The approach effectively captures critical atmospheric processes responsible for extreme event development.

Objective

• Investigate the feasibility of using AI weather models as research tools for understanding mechanisms driving weather extremes

Study Configuration

• Spatial Scale: Global, with a focus on regional extremes
• Temporal Scale: Hourly to daily time scales, with a focus on short-term forecasting and extreme event development

Methodology and Data

• Models used: GraphCast, a leading global AI model
• Data sources: Reanalysis data (e.g., ERA5), satellite observations, and ground-based measurements

Main Results

• The true-state constraint method effectively captures critical atmospheric processes responsible for the development of two extreme events.
• The model's responses to adjustments are consistent with behaviors in dynamical nudging experiments.

Contributions

• This study provides a novel approach to diagnosing error propagation in AI weather models and highlights their potential as research tools for understanding mechanisms driving weather extremes.
• The proposed climatology constraint method can quantify remote impacts of evolving atmospheric circulation anomalies, providing new insights into the dynamics of extreme events.

Funding

• National Science Foundation (NSF) grant number: NSF-AGS-1924979
• National Oceanic and Atmospheric Administration (NOAA) Cooperative Agreement number: NA19OAR017035

Citation

@article{Nie2026Diagnosing,
  author = {Nie, Yanbo and Sengupta, Agniv and Baño‐Medina, Jorge and Moore, Benjamin and Monache, Luca Delle},
  title = {Diagnosing Forecast Error Propagation and Large‐Scale Dynamics of Weather Extremes With an AI Weather Model},
  journal = {Geophysical Research Letters},
  year = {2026},
  doi = {10.1029/2026gl123505},
  url = {https://doi.org/10.1029/2026gl123505}
}