Google has introduced WeatherNext 3, a new artificial intelligence model designed to make weather forecasts more detailed and more frequently updated, producing hourly forecasts at higher resolution with applications ranging from weather alerts to renewable energy planning.
Developed by Google DeepMind and Google Research, WeatherNext 3 uses data from geostationary satellites and ground weather stations to produce new forecasts every hour. Google said the model can forecast surface temperature and moisture at a resolution of up to 5 kilometers, while surface winds can be forecast at 10 kilometers.
For people checking the weather, the more frequent updates could provide a clearer picture as conditions change. For businesses and organizations, the higher-resolution forecasts could also provide useful information for planning around storms, heavy rain and other weather events.
WeatherNext 3 is already being used in Google Search, the Gemini app, Google Maps, Google Maps Platform Weather API and Google Earth Engine.
More frequent weather updates
IMAGE CREDIT: WeatherNext 3
Weather forecasting has traditionally relied on numerical weather prediction, which uses complex calculations to model how the atmosphere is expected to behave.
These systems remain an important part of forecasting, but they can require significant computing power and time to produce results.
WeatherNext 3 uses AI to process weather information and generate forecasts more frequently. Google said the model can produce a new forecast every hour, allowing forecasts to be updated as weather conditions develop.
The company also reported improvements in precipitation forecasting. On selected benchmarks, WeatherNext 3 reduced forecast errors by as much as 50% compared with traditional numerical weather prediction systems.
Google said the improvement is particularly noticeable in forecasts made a day or more in advance and in areas where weather forecasts have historically been less accurate.
Potential impact on the Philippines
IMAGE CREDIT: WeatherNext 3
More detailed and frequently updated forecasts could be useful in the Philippines, where typhoons and heavy rainfall regularly affect communities, businesses and transportation.
Weather information can also be important for industries that have to make decisions based on changing conditions.
A logistics company, for example, may need to adjust delivery schedules during severe weather, while farmers and businesses can use forecasts when planning operations.
The technology also has applications in the energy sector.
WeatherNext 3 provides information including wind speeds at around 100 meters above the ground, along with cloud cover and solar irradiance. These are factors that can affect the amount of electricity generated by wind and solar facilities.
Google said the information can help energy producers and grid operators better understand expected renewable energy output.
Bringing AI forecasting to everyday services
One of the more notable aspects of WeatherNext 3 is that the technology is not being offered only as a standalone forecasting system.
Google is incorporating the model into products that consumers and businesses already use, including Search, Maps and Gemini, while making the technology available through its Weather API and Earth Engine platform.
For consumers, the benefit may be as simple as getting a better indication of whether rain is likely tomorrow.
For businesses and organizations, however, more detailed forecasts can influence decisions that depend on weather conditions.
The rollout also reflects Google’s broader effort to apply AI to areas that have traditionally depended on large-scale scientific and computational models.
As weather patterns become harder to predict and extreme weather continues to affect communities and businesses, the ability to process more data and update forecasts more often is becoming increasingly important.
WeatherNext 3 is Google’s latest move in that direction, bringing AI-powered forecasting into the services people already use to make everyday decisions.
