Skillful Kilometer-Scale Regional Weather Forecasting via Global and Regional Coupling

2026-03-30

AI Summary

High-resolution regional weather forecasting struggles to effectively couple large-scale atmospheric circulation with fine-scale processes such as complex terrain and coastal effects. This work proposes a global–regional coupled framework that bidirectionally links a pretrained Transformer-based global model with a 5-km-resolution regional network through an innovative ScaleMixer module, enabling cross-scale feature interaction. ScaleMixer incorporates adaptive key-location sampling and a dedicated attention mechanism to dynamically focus on meteorologically sensitive regions, substantially enhancing coupling efficiency. Experiments demonstrate that the proposed method outperforms operational numerical weather prediction systems and existing AI baselines on both reanalysis data and surface observations, accurately capturing fine-scale phenomena such as terrain-induced wind patterns and foehn warming while maintaining large-scale consistency and high-fidelity small-scale detail.

Abstract

Data-driven weather models have advanced global medium-range forecasting, yet high-resolution regional prediction remains challenging due to unresolved multiscale interactions between large-scale dynamics and small-scale processes such as terrain-induced circulations and coastal effects. This paper presents a global-regional coupling framework for kilometer-scale regional weather forecasting that synergistically couples a pretrained Transformer-based global model with a high-resolution regional network via a novel bidirectional coupling module, ScaleMixer. ScaleMixer dynamically identifies meteorologically critical regions through adaptive key-position sampling and enables cross-scale feature interaction through dedicated attention mechanisms. The framework produces forecasts at $0.05^\circ$ ($\sim 5 \mathrm{km}$ ) and 1-hour resolution over China, significantly outperforming operational NWP and AI baselines on both gridded reanalysis data and real-time weather station observations. It exhibits exceptional skill in capturing fine-grained phenomena such as orographic wind patterns and Foehn warming, demonstrating effective global-scale coherence with high-resolution fidelity. The code is available at https://anonymous.4open.science/r/ScaleMixer-6B66.

Problem

• regional weather forecasting
• multiscale interactions
• high-resolution prediction
• terrain-induced circulations
• coastal effects

Innovation

• global-regional coupling
• ScaleMixer
• cross-scale interaction
• kilometer-scale forecasting
• adaptive key-position sampling