Dr Peter Battaglia, Google DeepMind provided a presentation on the use of AI in medium range global weather forecasting.
Peter Battaglia is a Research Scientist/Director at DeepMind. His work focuses on approaches for modelling and reasoning about complex systems, by combining richly structured knowledge with flexible learning algorithms.
In this webinar, Peter presented his recent work on “GraphCast”, a new machine-learning (ML) based weather simulator. GraphCast is based on graph neural networks, and was trained from 40+ years of re-analysis data to forecast various surface and atmospheric variables over 10 days, at 0.25-degree latitude/longitude resolution. It can generate a forecast in 60 seconds, and outperforms ECMWF’s deterministic operational forecasting system, HRES, on around 90% of target fields. These results represent a key step forward in complementing and improving weather modelling with ML, opening new opportunities for fast, accurate forecasting.
Further information and details, this and the wider webinar series is online at https://digitalenvironment.org/webinars/cde-webinar-series-upcoming/#DeepMind .