Google's WeatherNext 3 uses satellite data for forecasts, bypassing physics

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Google Research and DeepMind have introduced WeatherNext 3, a weather forecasting model that learns directly from live satellite data instead of using traditional physics simulations. This model aims to provide more detailed hourly forecasts, potentially improving predictions in regions that have historically lacked accurate weather information.

Key facts

• WeatherNext 3 is a new weather model developed by Google Research and DeepMind.
• The model learns weather patterns directly from real-time satellite data, bypassing traditional physics simulations.
• WeatherNext 3 can generate hourly forecasts with a resolution of up to five kilometers.
• This new model is expected to significantly improve weather predictions in regions like Africa, Latin America, and the Asia-Pacific.
• The reliance on observed data rather than physical principles raises questions about accuracy and verifiability.