Deep learning and a changing economy in weather and climate prediction

Summary: Artificial intelligence-based methods have shown potential in accelerating weather forecasting, but their accuracy is still lower than numerical weather prediction (NWP) methods. In this study, we introduce an AI-based method for accurate medium-range global weather forecasting. Trained on global data, our program outperforms the world's best NWP system (ECMWF) in deterministic forecast results and works well with extreme weather and ensemble forecasts.

Summary: The use of machine learning in meteorology has increased greatly in recent years. However, the lack of formal instruction and requirement for meteorology students has led to a perception that machine learning methods are opaque and intimidating. To address this, this paper provides a survey of common machine learning methods and uses plain language and meteorological examples to explain them, aiming to enable readers to apply machine learning to their own datasets.