Learning Kilometer-Scale Weather Prediction with Global-Regional Alignment

2026-10-08

AI Summary

This study addresses the challenges of grid alignment and global-local interaction when repurposing global pretrained models for kilometer-scale regional weather forecasting by proposing the ScaleCast framework. The method introduces a global-to-regional transformation module to achieve cross-grid representation alignment and employs dynamic block fusion to deeply integrate global guidance with local dynamics. A core innovation lies in its ability to adapt to diverse global driving sources, such as Pangu-Weather, without requiring retraining. Experiments conducted on ERA5 and CERRA datasets demonstrate that the proposed framework significantly improves forecast accuracy for both surface and upper-air variables, refines storm track and central pressure estimation, and exhibits strong generalization capability across varying meteorological conditions.

Abstract

Kilometer-scale regional weather forecasting is essential for local weather warnings and weather-sensitive decisions. Existing data-driven approaches often rely on numerical forecasts for large-scale guidance or require additional training of global forecasting components. Pretrained global weather models offer an efficient source of large-scale forecasts, motivating their reuse to guide high-resolution regional prediction. However, this coupling requires aligning global and regional representations across different grids and integrating global guidance with local interactions to advance regional states. We propose ScaleCast, a regional forecasting framework that addresses these challenges through Global-Regional Alignment. Its Global-Regional Conversion module aligns joint global and regional representations with regional locations, while the Global-Regional Alignment and Dynamics block combines aligned guidance with regional neighborhood interactions. Experiments using ERA5 global analyses on a 0.25-degree grid and CERRA regional reanalysis at 5.5 km spacing demonstrate improved regional forecasts across surface and upper-air variables, with a single trained model supporting multiple global forecast drivers (i.e., Pangu-Weather, GraphCast, and HRES) without specific retraining. Fine-tuning on HRRR at 3 km spacing further demonstrates the framework's adaptability to a different regional domain and spatial resolution. Windstorm case studies show improved cyclone positioning and core-pressure estimates, while comparisons with HadISD station observations show closer agreement with local temperature and humidity changes.

Problem

• kilometer-scale weather prediction
• regional forecasting
• global-regional alignment
• data-driven weather models

Innovation

• Kilometer-scale weather prediction
• Global-Regional Alignment
• ScaleCast
• Pretrained global models
• Cross-resolution adaptability