Speakers
The Domains
Data driven modelling is now well established for short to medium range weather forecasting, with a number of data-driven models shown to outperform existing physics-based models. Ocean modelling poses different challenges to short term atmospheric weather prediction, nevertheless there is increasing research into this area, with a growing number of data-driven ocean models. In this talk we discuss some of the challenges and recent advances in this area. We then present recent work from ECMWF developing a data-driven ocean model, discussing different design choices and model performance
Rachel Furner is a data-driven ocean modeller at ECMWF, where she is developing a machine learning-based global ocean model. This model will be used as part of a fully machine-learning based earth system model funded by the Destination Earth initiative.
Rachel has a strong foundation in mathematics and over a decade of experience in ocean modelling, having previously worked at the UK Met Office, contributing to the development and operational use of the NEMO physics-based model for storm surge and coastal ocean forecasting. She recently completed a PhD at the University of Cambridge and the British Antarctic Survey using machine learning to emulate idealised channel ocean configurations.