GraphCast is already being used by weather agencies, including the European Centre for Medium-Range Weather Forecasts (ECMWF), and its code has been open-sourced to benefit scientists and forecasters worldwide. The model is part of a broader effort to use AI in weather forecasting and climate research, with the aim of addressing major environmental challenges and benefiting billions of people in their everyday lives.
Key takeaways:
- GraphCast, a state-of-the-art AI model, can make medium-range weather forecasts with unprecedented accuracy and speed, predicting weather conditions up to 10 days in advance more accurately and much faster than the industry gold-standard weather simulation system.
- GraphCast can also offer earlier warnings of extreme weather events, predicting the tracks of cyclones, identifying atmospheric rivers associated with flood risk, and predicting the onset of extreme temperatures.
- GraphCast is based on machine learning and Graph Neural Networks (GNNs), and makes forecasts at the high resolution of 0.25 degrees longitude/latitude, predicting five Earth-surface variables and six atmospheric variables at each of 37 levels of altitude.
- GraphCast is now the most accurate 10-day global weather forecasting system in the world, and its code has been open sourced to make AI-powered weather forecasting more accessible, with ECMWF already experimenting with GraphCast’s 10-day forecasts.