Rice University researchers Avantika Gori and Guha Balakrishnan are developing an open-source AI tool that simulates thousands of potential tropical cyclones to improve estimates of hurricane hazards and insurance risk.
The project, called “Leveraging Large Earth Foundation Models for Probabilistic Hurricane Hazard Assessment,” will use NASA’s Prithvi-WxC weather and climate foundation model, which was developed with IBM.
The researchers aim to address a fundamental challenge in hurricane risk modeling: major storms occur relatively infrequently, limiting the historical data available to estimate the probability and potential severity of future events.
The Rice team plans to create a dataset of approximately 10,000 synthetic cyclones representing storms that could occur rather than simply recreating past events.
The researchers will test how many storms can be generated within different computing periods and use the resulting dataset to study a wider range of potential storm tracks, wind conditions and rainfall patterns.
An average Atlantic hurricane season produces 14 named storms, including seven hurricanes and three major hurricanes, according to the National Oceanic and Atmospheric Administration.
The project will use Prithvi-WxC to simulate storms from their early development through dissipation while accounting for surrounding weather conditions.
Researchers will then refine the simulations with physics-informed models to generate more detailed estimates of wind and rainfall, including complex hurricane characteristics such as spiral rainbands.
The framework is also being designed to evaluate hurricane hazards across different climate scenarios, potentially allowing researchers and insurers to better understand how risk could change over time.
The simulations could help estimate the probability of hurricane landfalls and the severity of wind and rainfall that individual locations may experience.
Results could also support portfolio-level risk assessments involving the combined financial exposure across groups of insured properties.
The research is being conducted in collaboration with CERCat, an industry consortium involving Rice and Lehigh universities.
Gori, an assistant professor of civil and environmental engineering, is the project’s principal investigator and will contribute expertise in the physical processes governing hurricanes.
Balakrishnan, an assistant professor of electrical and computer engineering, will bring expertise in computer vision.
The team plans to make the resulting tool open source so other researchers can examine, use and extend the framework.
KEY QUOTES:
“AI gives us an opportunity to explore thousands of realistic storms that history alone cannot provide. Our goal is to turn those simulations into practical information that helps researchers and insurers better understand hurricane risk and how it may change.”
Avantika Gori, Assistant Professor of Civil and Environmental Engineering at Rice University
“By enhancing the accuracy of natural hazard risk estimates, we hope to reduce uncertainty for insurers and reinsurers, enabling more transparent pricing and greater willingness to insure complex risks.”
Avantika Gori, Assistant Professor of Civil and Environmental Engineering at Rice University
“Computer vision techniques can identify and learn complex patterns in data, and this project gives us an opportunity to combine those methods with our physical understanding of hurricanes. Bringing those approaches together can help us build models that are useful while remaining grounded in how storms behave.”
Guha Balakrishnan, Assistant Professor of Electrical and Computer Engineering at Rice University
“Better risk estimates can reduce uncertainty about the hazards facing homes and communities. More accurate and transparent information can help the insurance and reinsurance industry make better-informed decisions about complex risks, particularly in hurricane-prone areas.”
Avantika Gori, Assistant Professor of Civil and Environmental Engineering at Rice University
