NOAA has announced it is transitioning its high-performance supercomputing operations from on-premises hardware to the commercial cloud, in the latest update to its operational weather forecast model technology.
The move affects NOAA’s Weather and Climate Operational Supercomputing System (WCOSS) and is intended to support current and future improvements to the agency’s weather forecast models, resulting in a system NOAA says will be more flexible and easier to update. The agency expects the change to lead to earlier and more precise forecasts and warnings for extreme weather events.
NOAA administrator Dr Neil Jacobs said, “Cloud-based high-performance computing will accelerate the transition of research into operations by eliminating traditional bottlenecks of on-premise systems. 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.”
NOAA said the shift to cloud infrastructure will make its systems more adaptable to scientific innovation by enabling faster updates, improving resilience with fewer and shorter periods of downtime, and allowing computing power to scale on demand, such as during tropical storm season.
The agency will transition operations – including the Global Forecast System and Global Ensemble Forecast System – to the new cloud environment through early 2027, with the full migration scheduled for completion by December 2027.
Ken Graham, director of NOAA’s National Weather Service, said, “We are charting a course to global leadership in cloud-based numerical weather prediction. 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.”
Google Cloud will serve as the primary provider of high-performance computing infrastructure for the WCOSS program. NOAA said that through the partnership, it will draw on Google’s expertise to optimize workloads using modern technologies and AI innovations.
The collaboration also extends to NOAA’s use of Google DeepMind as the framework for its new AIGFS model suite, the agency’s first set of machine learning-powered weather forecast models, which NOAA said will cut the time required to generate a global forecast from hours to seconds.
NOAA said it will be one of the first operational numerical weather prediction centers in the world to move its operations to commercial cloud infrastructure. Phase one will establish supercomputing capabilities in the cloud, and phase two will involve migrating individual forecast models for operational use.
The migration follows NOAA’s earlier announcement this year of contracts to develop cloud-based technology for NWS forecasters at its 122 local Weather Forecast Offices and other NWS centers nationwide.
Related news, NOAA and SpectraRep to evaluate broadcast datacasting for weather radio resiliency