Google DeepMind and Google Research introduced WeatherNext 3 on September 3, a global AI weather model that Google said is being integrated into Search, the Gemini app, Google Maps, Google Maps Platform, Google Cloud, and Google Earth Engine.

Google said WeatherNext 3 uses live geostationary satellite data to generate hourly forecasts.

It models surface variables such as temperature and moisture at 5-km resolution, improving on WeatherNext 2, which produced forecasts on a 25-km grid every six hours.

The new model also adds renewable energy-related variables such as 100-meter wind speeds, cloud cover, and solar radiation, and Google said it improves precipitation forecasting.

The company cited independent live evaluations by Brightband in calling WeatherNext 3 its most accurate global weather model to date.

Google added that the model’s medium-range precipitation forecasts showed a Continuous Ranked Probability Score improvement of up to 60% against NASA’s Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG) dataset.

Google also published developer documentation and a guide to access forecasts, showing that WeatherNext is available to developers beyond Google’s consumer services.

🔗 Source: Google

Recent Google developments