● Covered by 3 sources · 3 reportsMedium impact2 neutral1 positive
Google DeepMind open-sources WeatherNext AI for 15-day cyclone forecasts
Key points
- WeatherNext AI open-sourced by Google DeepMind and Google Research.
- Provides 15-day tropical cyclone track and intensity forecasts.
- Uses Functional Generative Networks (FGNs) for rapid predictions.
- Offers one day more lead time than current models.
- Code and model weights are publicly available on GitHub.
AI Model for Cyclone Forecasting Released
Google DeepMind and Google Research have open-sourced their WeatherNext AI model, designed to improve tropical cyclone forecasting. The model can generate 15-day predictions for a storm's track and intensity, providing an average of one additional day of lead time compared to existing models. This means its predictions three days out are as accurate as previous models' predictions two days out.
Technology and Performance
WeatherNext utilizes Functional Generative Networks (FGNs) to produce ensembles of predictions, capturing the inherent uncertainty of weather. It can generate a single 15-day forecast in less than a minute on a TPU. The model also demonstrates the ability to capture rare events, such as rapid intensification, as observed during Hurricane Melissa in 2025.
A notable aspect of WeatherNext is its ability to achieve accurate predictions with significantly coarser data resolution. It operates effectively with data at 28x28km resolution, which is 100 times coarser than traditional models. A smaller version, WeatherNext 2-mini, also performs well at 111x111km resolution, a finding that has surprised scientists.
Open-Sourcing for Broader Impact
Both the code and model weights for WeatherNext are now available on GitHub. This open-source release aims to enable other scientists and weather agencies, including contributors like the National Hurricane Center, the Cooperative Institute for Research in the Atmosphere, and the UK Met Office, to use and further develop the model. The goal is to accelerate weather research and enhance operational forecasting capabilities globally.
Significance for Early Warnings
The increased lead time provided by WeatherNext can significantly impact preparedness for communities in the path of hurricanes and typhoons. An extra day, or even a few hours, can be crucial for organizing evacuations, staging supplies, and allocating resources for disaster response. The model's ability to predict with higher confidence earlier could mitigate the catastrophic impacts of severe weather events.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
How outlets covered it
Google DeepMind and Google Research developed WeatherNext, an AI model that predicts cyclones with greater accuracy, providing forecasters an additional day of lead time compared to existing models. This advancement allows for earlier warnings and better preparation for communities in the path of hurricanes, potentially mitigating catastrophic impacts.
Google has open-sourced its WeatherNext AI model, which provides 15-day forecasts for tropical cyclone track and intensity. This release allows other scientists to use and build upon the model, potentially improving early warning systems for hurricanes and typhoons.
Google has open-sourced its WeatherNext AI model, which uses Functional Generative Networks (FGNs) to produce 15-day cyclone forecasts in under a minute, capturing rare events like rapid intensification. The model achieves high accuracy with significantly coarser data resolution than traditional physics models, and its code and weights are now publicly available to accelerate weather research and operational forecasting.