Planning: AI-Ready - Planning the Expansion of NOAA's Hurricane and Ocean Testbed to an AI-Ready Testbed

Extreme weather events are a critical challenge for today’s society, endangering life and property. Hurricanes are a prime example, with a broad range of hazards and impacts. Recent hurricanes have highlighted the complexity of their combined hazards, such as storm surge, inland flooding, and damaging winds, including tornadoes. The National Oceanic and Atmospheric Administration (NOAA) established the Hurricane and Ocean Testbed (HOT) in 2021 to test new models and products to improve hurricane forecasts, but the HOT testbed was not designed to handle artificial intelligence (AI) models. Promising new AI models have the potential to improve current forecasts of hurricanes and their related hazards, but the HOT testbed was not designed with the computing or personnel needed to test these promising AI methods and use them operationally. Furthermore, there is a disconnect between the AI community that develops such promising models and the operational community that could benefit from them. The research team’s objective is to bridge this divide by planning the expansion of NOAA’s HOT testbed to include AI models, working closely with the AI community. Bringing the AI and operational hurricane forecasting communities together to improve hurricane forecasts benefits society, addressing NSF’s mission to advance the national health, prosperity and welfare. To achieve these goals this project engages an interdisciplinary research team from academia, the National Center for Atmospheric Research, the National Hurricane Center, and NOAA’s Global Systems Lab. The team includes experts in hurricanes, AI, risk communication, and research-to-operational transitions, and is thus well equipped to bridge this divide; that is, to address technical challenges, as well as design and development challenges. Proposed activities fall into two general categories. Firstly, the project will connect guest AI researchers and their models to the HOT testbed by inviting AI guest researchers, serving as intermediary between AI guest researchers and the testbed, developing a tiered test protocol, identifying AI infrastructure needs for the expansion, and extracting AI model insights. Secondly, the project will prepare the expansion of the HOT testbed by conducting meetings with testbed users, holding a workshop to design a framework for expanding the HOT testbed, planning the expansion, and assessing generalizable aspects for other testbeds. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.