Google has made its WeatherNext AI model publicly available, offering the meteorological community a tool that can predict a tropical cyclone’s path and strength up to two weeks in advance. The model, developed by Google DeepMind and Google Research, was trained on nearly 20 terabytes of global atmospheric data and historic storm records from the International Best Track Archive for Climate Stewardship.

WeatherNext’s ability to produce a single 15‑day forecast in less than a minute on a Tensor Processing Unit (TPU) marks a significant speed improvement over traditional forecasting pipelines. "We can now generate a single 15‑day forecast in less than a minute on a TPU, empowering forecasters to quickly evaluate the probability distribution of potentially devastating tail‑risks," the research team wrote in a paper published in Nature.

The model’s development was a collaborative effort. The National Hurricane Center, the Cooperative Institute for Research in the Atmosphere, the UK Met Office and several other weather agencies contributed data and expertise. By unifying the tasks of track and intensity prediction—traditionally handled by separate global and local models—WeatherNext aims to streamline the forecasting process without sacrificing accuracy.

Google has placed both the source code and the trained model weights on GitHub, inviting scientists and operational forecasters to adapt and extend the system. A layperson‑focused blog post on Google’s site explains the model’s capabilities in plain language, complementing the technical details presented in the peer‑reviewed study.

WeatherNext arrives as the second generation of the platform, following an initial release last year. The same research groups are also applying AI techniques to forecast flash floods, suggesting a broader push to embed machine learning across extreme‑weather prediction.

Industry observers note that open‑sourcing the model could accelerate innovation in climate resilience, especially for regions where forecasting resources are limited. By lowering the barrier to high‑performance, AI‑driven forecasts, Google hopes to help communities prepare for storms before they make landfall.

This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.