GenCast: AI-Powered Weather Forecasting
Product Introduction
GenCast is a cutting-edge AI weather forecasting model designed to deliver high-precision weather forecasts for up to 15 days. Developed by Google DeepMind, it utilizes advanced diffusion models to analyze historical weather data and predict future weather scenarios with remarkable accuracy. The model's release aims to support a broader community in enhancing weather forecasting capabilities.
Key Features
- Provides high-precision weather forecasts for up to 15 days.
- More accurate than existing top systems like the ECMWF ENS.
- Combines 50 or more forecasts to represent potential weather trajectories.
- Adapts to the spherical geometry of the Earth for precise probability distributions.
- Trained using 40 years of historical weather data from ECMWF's ERA5 dataset.
- Outperforms ENS in 97.2% of test targets, especially beyond 36 hours.
Rapidly generates predictions, with a single Google Cloud TPU v5 producing a 15-day forecast in only 8 minutes.
Product Data
- High-resolution model (0.25°)
- Developed by Google DeepMind
- Utilizes historical data for training
- Prediction speed: 8 minutes for a 15-day forecast on TPU v5
- Target audience: meteorologists, data scientists, renewable energy companies, and disaster response organizations. ## Product Link
GenCast Official Website
