Despite its success, GraphCast has limitations, including not outperforming conventional models in all scenarios and not being able to create forecasts as detailed as traditional ones. It also has transparency issues as meteorologists can't yet see why it makes the forecast it does. However, the Google DeepMind researchers see their AI-based approach as a complement to current weather prediction techniques. Looking ahead, ECMWF plans to develop its own AI model and explore integrating it with its numerical weather prediction system, while the UK Met Office is also developing a graph neural network for weather forecasting.
Key takeaways:
- Google DeepMind's AI meteorology model, GraphCast, has outperformed conventional weather forecasting methods in predicting global weather conditions up to 10 days in advance, according to a study published in the journal Science.
- GraphCast demonstrated superior performance over the world's leading conventional system, operated by the European Centre for Medium-range Weather Forecasts (ECMWF), in 90 percent of 1,380 metrics, including temperature, pressure, wind speed and direction, and humidity at various atmospheric levels.
- Despite its success, GraphCast has limitations such as not outperforming conventional models in all scenarios and not being able to create forecasts as detailed or granular as traditional ones. It also has transparency issues since meteorologists can't yet look inside the "black box" of the AI model and see exactly why it makes the forecast it does.
- Looking ahead, ECMWF plans to develop its own AI model and explore integrating it with its numerical weather prediction system. The UK Met Office, in partnership with the Alan Turing Institute, is also developing a graph neural network for weather forecasting to be incorporated into its supercomputer infrastructure in the future.