Comparative evaluation of data-driven weather forecast models performance for medium- to extended-range weather forecasting and tropical cyclone genesis in 2024
Évaluation comparative des performances des modèles de prévisions météorologiques basés sur les données pour les prévisions à moyen et long terme et la genèse des cyclones tropicaux en 2024
Ho, C. H. ; Leung, Marco Y. T. ; He, Y. H. ; Chan, P. W.
Année de publication
2026
This study examines the performance of state-of-the-art artificial intelligence (AI) weather forecasting models in 2024. For tropical cyclone genesis forecasts in the western North Pacific, AI models generally perform better than the physics-based European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS). Furthermore, higher-resolution AI models, such as a fine-tuned version of Pangu-Weather, improved anomaly correlation scores and Fengwu-GHR produced more realistic synoptic-scale structures for medium-range forecasts compared with lower-resolution versions. Moreover, certain AI models could match or outperform ECMWF IFS on daily rainfall forecasts over Hong Kong, especially for heavier rainfall events. These findings highlight the potential of AI-based methods for operational applications on high-impact weather.</div>
Accès à la notice sur le site du portail documentaire de Météo-France