This five-day training offers a deep dive into Earth system forecasting using state-of-the-art deep learning approaches, the Anemoi framework, and other modern machine learning tools.

This in an advanced course aimed at meteorologists, Earth system modellers and other researchers who already have a good foundational knowledge of machine learning. See the Requirements section below.

Alongside lectures, there will be hands-on coding sessions and discussions.

Main topics

The main topics of the training will be as follows:

• An overview of deep learning flavours underlying Earth systems modelling, such as graph neural networks and transformers
• A deep dive into the methodology and operation of the AIFS-Single and AIFS-ENS global weather models
• Use of the Anemoi software framework for data-driven forecasting
• Advanced applications of machine learning in weather and Earth systems modelling

All lectures will be held in English.

Requirements

In order to follow this course, participants should have:

• A good understanding of foundational ML concepts and workflows (training, evaluation, generalisation, etc)
• A general understanding of (or experience with) neural networks and deep learning models
• Working knowledge/experience in deep learning tools, such as PyTorch, Keras and similar
• Knowledge/experience of weather forecasting and/or Earth systems modelling

More ML training

If you are looking for more machine learning training, or this course is not the right level for you, consider the following ECMWF online courses and resources:

• Machine Learning for Earth Systems Modelling: Foundations and New Frontiers - a high-level tour of the latest developments in ML for Earth systems modelling and ECMWF's Destination Earth programme (through DG CNECT and the European Commission)
• An Introduction to ML in Weather and Climate - a MOOC-style course introducing many key concepts in machine learning (note this course ran in 2023)
• Concepts of ML in Weather and Climate - a follow up to the previous course with more technical learning material on ML, deep learning, and more

See also our training catalogue with many more resources and courses on machine learning and other topics.