Google's WeatherNext 3 AI Model Boosts Forecast Accuracy
Google released WeatherNext 3, an AI weather model enhancing prediction accuracy, especially for rain and snow.
"Forget your umbrella? Not anymore. Google's new AI weather model is here to make sure you're always prepared. No excuses!"
Google has launched WeatherNext 3, an updated AI weather model, which the company states is more accurate, particularly for predicting rain and snowfall. This model will integrate into Google Search, Google Maps, and Gemini, and will also be accessible to users and researchers via Google's cloud platforms. Google describes WeatherNext 3 as a significant advancement in meteorology, leveraging deep learning techniques.
WeatherNext 3 is designed to provide forecasts with "unprecedented resolution." It has demonstrated improved accuracy, with evaluations on rain showing a 60% improvement over its predecessor, WeatherNext 2. The model can now produce hourly forecasts, a shift from the previous standard of predictions every six hours. Furthermore, it can predict down to a resolution of 5 km on key variables.
This new model represents a "sea change" in meteorology, moving beyond traditional supercomputer-based forecasting methods. Its ability to ingest real-time hourly weather satellite data contributes to more accurate forecasts. The model is also trained to target specific weather data stations, allowing for more granular predictions and better evaluation against ground-truth data.
This advancement in AI weather forecasting could lead to more reliable operational planning for businesses sensitive to weather conditions. Improved accuracy in predicting rain and snowfall can impact logistics, event planning, and resource allocation, reducing weather-related disruptions.
Find the right AI tool for your business
Chat with Insta and get matched to the right tool in seconds.
Try Insta Tool Finder →