The first online training course in the Destination Earth (DestinE) Machine Learning for Earth Systems Modelling series has successfully concluded. Course 1 attracted more than 4,000 participants, highlighting the growing international interest in artificial intelligence (AI) and machine learning (ML) for weather and climate science. Developed under the Destination Earth (DestinE) initiative of the European Commission’s DG Connect, the training series aims to support researchers, practitioners, and technical professionals in understanding how ML is transforming Earth-system modelling and weather prediction workflows. Registrations for Course 2 are now open.

Course 1 – Foundations and New Frontiers, introduced participants to the conceptual foundations of ML in Earth system science, including AI forecasting systems, digital twins, ethics and regulation, and the evolving role of ML within the DestinE ecosystem. The course combined lectures, quizzes, webinars, and panel discussions featuring experts from ECMWF, academia, operational forecasting centres, and industry. Live discussions explored topics including explainable AI, hybrid Earth-system modelling, uncertainty quantification, operational forecasting, and the future role of machine learning in weather and climate prediction. More than 780 learners obtained a certificate of completion following the live course run. 

Speakers highlighted that future systems are likely to combine AI-driven methods with physics-based modelling approaches, rather than replacing them entirely. As discussed during the panel sessions, building trust, explainability, and physical consistency into AI forecasting systems will remain a key challenge for the coming years.  

While the moderated live run of Course 1 has now concluded, the course remains available in self-paced mode via the ECMWF learning platform. Participants can continue following Course 1 while preparing for the more technical and hands-on focus of Course 2. 

Building on the foundations established in Course 1, registrations are now open for Course 2: Machine Learning for Earth Systems Modelling – Architectures, Data, and Prediction, starting on 1 June 2026. 

While Course 1 focused on concepts and context, Course 2 will take a more technical and hands-on approach. Participants will explore how modern AI-based prediction systems are designed, trained, evaluated, and deployed for operational weather forecasting and Earth-system applications. 

Topics covered in Course 2 include: 

• Neural architectures for atmospheric dynamics 

• Data handling and preprocessing workflows 

• Compute infrastructure for ML forecasting 

• AI forecasting systems and uncertainty quantification 

• Evaluation frameworks and benchmarking 

• The Anemoi framework for AI-driven weather prediction, developed by ECMWF and several meteorological services across Europe  

The free online course is designed for technical learners with a background in Earth system sciences or related fields and includes practical examples, notebooks, and workflows relevant to Destination Earth. The course is offered in a self-paced format.  

The course series is developed by ECMWF in collaboration with the Karlsruhe Institute of Technology (KIT) and Wageningen University, with support from a network of European scientists. 

Register now for Course 2.  

Find more information on our training page .