Florence Rabier (ECMWF Director-General)
Peter Dueben (Royal Society University Research Fellow at ECMWF)
Machine learning allows to learn complex, non-linear behaviour from data which is useful for many application areas across the workflow of numerical weather prediction and climate services. This talk provided an update on the activities at ECMWF to explore the potential of machine learning, and in particular deep learning. Peter Dueben introduced the machine learning roadmap that identifies challenges, provides potential solutions, and defines steps to channel the many distributed science and technology projects that study machine learning for weather and climate prediction into a coordinated effort. The roadmap will help to make the most of machine learning for weather and climate predictions in the years to come.
Machine learning is an important part of ECMWF’s new Strategy 2021–30. The idea is to combine the best of what data-driven approaches can provide with the strengths and physical understanding encapsulated in our existing forecasting systems.