Google has officially unveiled WeatherNext 3, its new AI-powered weather forecasting system, poised to revolutionize weather prediction. Developed by Google DeepMind and Google Research, this model is trained using real-time geostationary satellite data and local station observations, enabling it to generate hourly forecasts with a resolution of up to 5 kilometers. The company describes it as its most advanced and accurate global weather model to date, continuing its commitment to more precise weather forecasting through AI.
The main difference from its predecessor, WeatherNext 2, is that it no longer relies exclusively on numerical weather prediction (NWP) models, which are physical simulations requiring supercomputers and carrying a six-hour delay. This lag caused biases in rapidly changing variables, such as rainfall or surface temperature. Now, by ingesting a mosaic of live satellite data, the system obtains a continuous and up-to-date view of the atmosphere, resulting in earlier and more detailed warnings when storms or fronts form.
Higher resolution and better accuracy in rain
One of the most significant advances is the improvement in precipitation prediction. Global models have always struggled to anticipate rain and snow because these phenomena depend on small-scale cloud processes. To address this, WeatherNext 3 is trained using two high-quality data sources: NASA's IMERG system and a proprietary reanalysis based on satellite radar. According to internal assessments, the improvement in the CRPS (Continuous Ranked Probability Score) reaches up to 60% compared to IMERG, 30% compared to MRMS, and 10% compared to ground-based rain gauges.
Boost to renewable energies
Another new feature is the incorporation of specific predictions for clean energy production. The model can forecast wind speed at 100 meters, approximately the height of wind turbines, as well as cloud cover and solar radiation levels. This data is essential for electricity grid operators and wind and solar farm developers to accurately estimate how much energy they will generate and adjust supply to demand, integrating with specialized wind forecasting apps.
This advancement is especially important for regions like Latin America, Africa, and Asia-Pacific, which have historically been underserved by high-resolution models due to the high computational cost of traditional regional systems. With WeatherNext 3, billions of people and local businesses in these areas can access detailed and reliable forecasts without needing their own supercomputer infrastructure.
Availability and applications
Since its launch, WeatherNext 3 has been integrated into Google products such as Search, Gemini, Google Maps with weather forecast , Google Maps Platform, and Google Cloud. In practice, users will see significantly more accurate long-term forecasts, with up to a 50% improvement in precipitation predictions for the coming days. The greatest improvements are seen in regions where forecasts were previously less reliable, making this tool a valuable resource for both daily planning and emergency management.
In short, WeatherNext 3 represents a significant step towards more accurate, accessible, and useful meteorology for the energy transition. Its ability to learn from real-time data, combined with unprecedented resolution, positions it as a key tool for addressing the climate and energy challenges of the future.
