Applying Machine learning to Operational Meteorology

Machine learning is revolutionizing meteorology by offering new approaches to weather and climate forecasting. The computational efficiency of Machine learning techniques addresses the limitations of current numerical weather prediction and climate models, particularly in resolution and complexity. Machine learning is also significantly changing meteorology by offering more efficient and accurate forecasting methods, improving model performance, and enabling the quantification of the impact of meteorological factors on various phenomena from extreme events to long term climate change.
This meeting invites key players, in the public and private sectors such as the Met Office, ECMWF, and Google to illustrate how these organisations have been actively exploring the use of machine learning for weather forecasting or as part of their subseasonal to seasonal predictions (S2S). The MetOffice as well as the ECMWF utilizes machine learning (ML) for weather and climate forecasts by leveraging historical data to train forecasting models and improve prediction accuracy. Machine learning-based forecasting systems have been developed to enhance temperature and precipitation predictions, particularly in regions with traditionally low predictability and research has shown that such methods yield better prediction results for weather forecasting compared to conventional physics-based models. Our workshop aims to explore all these aspects and the latest advancements and applications of ML techniques in weather prediction. Additionally, the workshop will address the challenges and limitations associated with Machine learning-based weather forecasting, emphasizing the need for trustworthy and interpretable algorithms, as well as the incorporation of physical knowledge of the atmosphere into Machine learning techniques. Last the workshop will highlight the relevance of Machine learning in addressing the uncertainty of weather forecasts, offering an alternative approach to predict the uncertainty of weather conditions based on large-scale atmospheric data.