Google WeatherNext 3 AI launch

Google DeepMind and Google Research have launched WeatherNext 3, an artificial intelligence model designed to enhance global weather forecasting precision. The system utilizes real-time satellite data and ground station observations to provide hourly updates at high resolutions, including a 5-kilometer scale for temperature and humidity.

Research indicates the model offers significant improvements over traditional numerical forecasting methods. It can reduce precipitation error metrics by up to 50% and provides enhanced accuracy for predicting cyclones, hurricanes, and typhoons, potentially offering an extra day of warning. By delivering three-day forecasts with the accuracy previously seen in two-day predictions, the model aims to mitigate delays associated with traditional numerical weather prediction models.

The technology is being integrated into various Google services, including Search, Maps, Gemini, Google Earth Engine, the Maps Platform Weather API, and Google Cloud services. It also provides specialized data for the renewable energy sector, such as wind speeds at 100-meter turbine heights and solar irradiance predictions to assist grid operators.