Data-Driven Mesoscale Weather Forecasting Combining Swin-Unet and Diffusion Models

Abstract

Data-driven weather prediction models show promising performance and are continuously advancing. In particular, diffusion models represent fine-scale details without spatial smoothing, which is crucial for mesoscale predictions, such as heavy rainfall forecasting. However, the applications of diffusion models to mesoscale predictions remain limited. To address this gap, this study proposes an architecture that combines a diffusion model with Swin-Unet as a deterministic model, achieving mesoscale predictions while maintaining flexibility. The proposed architecture trains the two models independently, allowing the diffusion model to remain unchanged when the deterministic model is updated. Comparisons using the Fractions Skill Score and power spectral analysis demonstrated that incorporating the diffusion model improved accuracy compared to predictions without it. These findings highlight the potential of the proposed architecture to enhance mesoscale predictions, including heavy rainfall, while maintaining flexibility.