Weather Prediction with Diffusion Guided by Realistic Forecast Processes
This addresses reliability concerns for the weather forecasting community by creating a more trustworthy deep learning system that integrates with existing numerical methods.
The paper tackles the challenge of making deep learning weather forecasting models more flexible and reliable by introducing a diffusion model approach that can perform both direct and iterative forecasting while incorporating numerical weather prediction outputs. The method demonstrates feasibility and generalizability without requiring retraining.
Weather forecasting remains a crucial yet challenging domain, where recently developed models based on deep learning (DL) have approached the performance of traditional numerical weather prediction (NWP) models. However, these DL models, often complex and resource-intensive, face limitations in flexibility post-training and in incorporating NWP predictions, leading to reliability concerns due to potential unphysical predictions. In response, we introduce a novel method that applies diffusion models (DM) for weather forecasting. In particular, our method can achieve both direct and iterative forecasting with the same modeling framework. Our model is not only capable of generating forecasts independently but also uniquely allows for the integration of NWP predictions, even with varying lead times, during its sampling process. The flexibility and controllability of our model empowers a more trustworthy DL system for the general weather community. Additionally, incorporating persistence and climatology data further enhances our model's long-term forecasting stability. Our empirical findings demonstrate the feasibility and generalizability of this approach, suggesting a promising direction for future, more sophisticated diffusion models without the need for retraining.