Outline
ECMWF will host a live panel on 3 July 2026, under the Destination Earth (#DestinE) initiative of DG-CNECT at the European Commission. The session will explore the role of machine learning (ML) in advancing regional high-resolution weather forecasting, as part of the online training course Machine Learning for Earth Systems Modelling: Architectures, Data, and Prediction.
This second course in ECMWF’s three-course ML learning pathway focuses on the technical foundations of machine learning for weather prediction. It explores how modern AI-based prediction systems are designed, trained, and evaluated using real-world Earth system data and workflows.
The panel will bring together experts from operational centres, research institutions and international organisations to discuss how ML can support regional and high-resolution weather forecasting. The discussion will address how ML can support regional forecasting through limited-area modelling, statistical and ML-based downscaling, transfer learning, and the integration of heterogeneous Earth-system data. The panel will also explore the role of ML in delivering more local, impact-relevant, and decision-ready forecasts within the Destination Earth Digital Twin framework.
The panel will be recorded and made available as part of Module 8 of the online training course.
Speakers
-
Håvard Homleid Haugen (MET Norway)
-
Joel Oskarsson (ETH Zurich)
-
Roland Potthast (DWD)
-
Lucy Mtilatila (Department of Climate Change and Meteorological Services Malawi)
-
Corentin Carton de Wiart (ECMWF)
Who should attend?
This event is aimed at researchers, practitioners, technical specialists, operational forecasters, climate and weather service professionals, and industry experts interested in the use of ML for regional weather forecasting and downscaling applications.
The event is open to the public and to learners of the course Machine Learning for Earth Systems Modelling: Architectures, Data, and Prediction. If you are following the course, you will still need to register here to attend the live panel.
Registration
Participation is free of charge. To register for the event, click the "Registration" link on the left side of this page.
More information
To find out more about our ML training under Destination Earth, take a look at our ML training page.
If you are interested, join the free online course Machine Learning for Earth Systems Modelling: Architectures, Data, and Prediction.
The panel will be recorded and made available shortly after the event on our YouTube channel, and accessible through our eLearning platform as part of the course.