Camilletti et al. (2025) AI Reconstruction of European Weather From the Euro‐Atlantic Regimes

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

Identification

• Journal: International Journal of Climatology
• Year: 2025
• Date: 2025-12-09
• Authors: Alessandro Camilletti, Gabriele Franch, Elena Tomasi, M. Cristoforetti
• DOI: 10.1002/joc.70216

Research Groups

Not explicitly mentioned in the abstract.

Short Summary

This study develops a non-linear AI model to reconstruct monthly mean anomalies of European temperature and precipitation using Euro-Atlantic Weather Regime (WR) indices, demonstrating its potential for sub-seasonal to seasonal forecasting by capturing complex non-linear relationships and showing improved or comparable skill to a state-of-the-art seasonal forecast system.

Objective

• To develop and evaluate a non-linear AI model capable of reconstructing monthly mean two-meter temperature and total precipitation anomalies in Europe based on Euro-Atlantic Weather Regime (WR) indices, thereby capturing complex non-linear relationships not addressed by current linear methods.

Study Configuration

• Spatial Scale: Europe
• Temporal Scale: Monthly mean anomalies; sub-seasonal to seasonal forecasting (winter and summer).

Methodology and Data

• Models used: Non-linear Artificial Intelligence (AI) model; ECMWF operational seasonal forecast system (SEAS5) for comparison and providing predicted WR indices.
• Data sources: Euro-Atlantic Weather Regime (WR) indices; Monthly mean two-meter temperature and total precipitation anomalies (implied for training/evaluation, likely from observations or reanalysis).

Main Results

• The developed AI model successfully captures complex non-linear relationships between Euro-Atlantic WR indices and corresponding surface temperature and precipitation anomalies in Europe.
• When the mean absolute relative error in WR indices is below 80%, the AI model yields improved seasonal reconstruction compared to the ECMWF SEAS5 operational seasonal forecast system.
• When evaluated using WR indices predicted by SEAS5, the AI model demonstrates slightly better or comparable skill relative to the SEAS5 forecast itself.
• WR-based anomaly reconstruction, powered by AI tools, offers a promising pathway for sub-seasonal and seasonal forecasting.

Contributions

• Introduces a novel non-linear AI model for estimating ground-level meteorological variables (temperature and precipitation) from Euro-Atlantic WR indices, addressing a gap where previous methods were largely linear.
• Demonstrates the ability of AI to capture complex non-linearities in the relationship between atmospheric circulation states (WR) and surface weather anomalies.
• Provides evidence that WR-based anomaly reconstruction can achieve improved or comparable skill to state-of-the-art seasonal forecast systems (like SEAS5) under certain conditions.
• Highlights the practical applicability and potential of AI tools for sub-seasonal and seasonal forecasting based on atmospheric circulation regimes.

Funding

Not explicitly mentioned in the abstract.

Citation

@article{Camilletti2025scpAIscp,
  author = {Camilletti, Alessandro and Franch, Gabriele and Tomasi, Elena and Cristoforetti, M.},
  title = {<scp>AI</scp> Reconstruction of European Weather From the Euro‐Atlantic Regimes},
  journal = {International Journal of Climatology},
  year = {2025},
  doi = {10.1002/joc.70216},
  url = {https://doi.org/10.1002/joc.70216}
}