In the latest innovation to NOAA’s operational weather forecast model technology, the agency has announced that it is moving its high-performance supercomputing operations to the commercial cloud.
This transition of NOAA’s Weather and Climate Operational Supercomputing System (WCOSS) supports current and future improvements to NOAA’s suite of weather forecast models. The move will result in a modeling system that is more flexible and easy to update, laying the foundation for NOAA to become the world leader in numerical weather prediction.
Ultimately, the weather forecasting community, emergency managers and the public can expect to see earlier and more precise forecasts and warnings for extreme weather events.
“Cloud-based high-performance computing will accelerate the transition of research into operations by eliminating traditional bottlenecks of on-premise systems,” said NOAA Administrator Neil Jacobs, Ph.D. “Not only will this new approach provide access to the latest chips, it will advance weather forecasting by aligning operational and research compute and provide capabilities for elastic on-demand model development.”
Computing power that drives scientific innovation
Transferring NOAA supercomputers from on-premises hardware to cloud infrastructure will create a computing environment that’s more easily adaptable for scientific innovation. This new cloud technology will eliminate infrastructure bottlenecks and make weather models more nimble and flexible by:
- Enabling rapid updates and adaptations at the pace of scientific advancement;
- Making the system more resilient and prone to fewer and shorter periods of downtime; and
- Instantly increasing computing power when it’s needed, like during tropical storm season.
NOAA will transition operations, including the Global Forecast System and Global Ensemble Forecast System, to the new cloud environment through early 2027. The transition to the cloud is scheduled to be completed by December 2027.
“We are charting a course to global leadership in cloud-based numerical weather prediction,” said Ken Graham, director, NOAA’s National Weather Service. “This is a significant achievement for NOAA; one that will ensure our critical modeling operations can seamlessly adapt to emerging science, and strengthen our mission to protect life and property.”
A strategic partnership with Google
Google Cloud will be the primary provider of high-performance computing infrastructure for the WCOSS program, supporting NWS’s gold-standard weather forecasting and decision support capabilities. Through this partnership, NOAA will benefit from Google’s expertise to optimize workloads with modern technologies and innovations in artificial intelligence.
NOAA’s partnership with Google extends to other AI applications, including the use of Google DeepMind as the framework for the new AIGFS model suite, NOAA’s first set of weather forecast models powered by machine learning. Next-generation machine learning weather prediction tools created by NOAA and Google will significantly reduce the time required to generate a global forecast from hours to mere seconds.
NOAA to lead in adoption of cloud technology
NOAA will be one of the first operational numerical weather prediction centers in the world to transition its operations to a secure, commercial cloud infrastructure. The initial deployment in Phase 1 will establish our supercomputing capabilities in the cloud, ensuring a successful and timely transition while integrating the flexibility provided by the cloud environment. Phase 2, the implementation phase, will involve migrating individual weather forecast models to the cloud for operational use.
NOAA’s migration to cloud infrastructure for its supercomputers follows the announcement earlier this year of contracts to develop cloud-based technology for NWS forecasters at our 122 local Weather Forecast Offices and other NWS Centers around the country.
NOAA’s weather forecasting systems and models are developed and managed by NWS Modeling and Development in College Park, Maryland, a division of the NWS’ National Centers for Environmental Prediction.
The original announcement can be found here.
