This online course explores the technical foundations of machine learning (ML) for weather and climate prediction within the Destination Earth (DestinE) initiative. As the second course in ECMWF’s three-course ML learning pathway, it focuses on how modern AI forecasting systems are designed, trained, and evaluated using real-world Earth system data and workflows.
Participants will gain insight into the architectures, datasets, infrastructure, and evaluation methods that underpin data-driven weather prediction systems. The course combines conceptual explanations with practical examples relevant to operational forecasting and the development of next-generation Earth digital twins.
The training provides a structured pathway from data handling and preprocessing to advanced AI forecasting workflows, uncertainty quantification, and benchmarking approaches used in contemporary Earth Systems modelling.
Main topics
The course will cover the following themes:
- Foundations of ML for weather prediction
- Deep learning architectures for atmospheric dynamics
- Datasets, data handling, and preprocessing workflows
- Compute infrastructure and distributed systems for ML forecasting
- AI forecasting systems (AIFS and AIFS-ENS) and operational workflows
- Uncertainty quantification and probabilistic prediction
- Evaluation frameworks and benchmarking methods
- The Anemoi framework for data-driven weather forecasting
Target audience
This course is part of a series designed for:
- Researchers and practitioners in meteorology and climate science
- Technical specialists from meteorological services and climate centres
- Academic and industry professionals working with Earth Systems data and AI workflows
This second course, Architectures, Data, and Prediction, specifically targets technical learners with experience in Earth system sciences or related disciplines who want to deepen their understanding of machine learning workflows for weather prediction and forecasting systems.
Requirements
Participants are expected to have:
- A background in meteorology, climate science, Earth Systems Sciences, or related fields
- Basic programming experience (preferably Python)
- Familiarity with statistics and introductory machine learning concepts
- Experience interpreting geophysical datasets and Earth system data workflows
Prior completion of the first course in this series, Foundations and New Frontiers is recommended, but not required.
Register now!
Register for this course through our eLearning portal.