Google DeepMind Unveils WeatherNext 3 for Global Forecasts
Google DeepMind has introduced WeatherNext 3, a new meteorological artificial intelligence model designed to generate high resolution global weather forecasts, according to a report by MarkTechPost. The system is trained directly on ground weather station observations, enabling it to deliver forecasts at a five kilometer resolution globally.
According to the details outlined by MarkTechPost, WeatherNext 3 updates its predictive outputs every hour, providing builders and meteorological researchers with significantly tighter refresh cycles compared to traditional numerical weather prediction models. By leveraging machine learning over extensive observation datasets, the architecture aims to capture localized atmospheric variations more accurately than previous iterations.
The release highlights the ongoing transition within operational meteorology toward deep learning approaches that bypass traditional compute heavy physics simulations while maintaining granular spatial accuracy. WeatherNext 3 is positioned as a tool for applications requiring rapid data ingestion and high fidelity forecasting across global topologies.
Based on reporting by www.marktechpost.com.
