AI4ClimateRRR: AI for Climate Risk and Resilience at Rice
Climate change is intensifying extreme weather events, posing unprecedented risks to urban infrastructure and communities. Managing these risks requires predictive understanding of both natural hazards and the response of human systems. While traditional hazard models shed light on many quesitons, they are often subject to sever limitations in resolution, computational efficiency, or geographic transferability.
The AI4ClimateRRR research cluster integrates artificial intelligence and machine learning with physics-based models to advance understanding of climate risks and enhance infrastructure resilience. We are working to develop open-source, computationally efficient, and high-resolution models to inform the management of complex, interconnected systems facing extreme weather in a changing climate.
Cluster Members
- Lead PI: James Doss-Gollin (Civil & Environmental Engineering, Rice University)
- Avantika Gori (Civil & Environmental Engineering, Rice University)
- Arlei Lopes da Silva (Computer Science, Rice University)
- Jamie Padgett (Civil & Environmental Engineering, Rice University)
- Noemi Vergopolan (Earth, Environmental and Planetary Sciences, Rice University)
Collaborators
- Silvia Dee (Earth, Environmental and Planetary Sciences, Rice University)
- Xinwu Qian (Civil & Environmental Engineering, Rice University)
- Lu Zhang (Computational Applied Mathematics and Operations Research, Rice University)
- Yanmo Weng (Postdoctoral Fellow and Cluster Coordinator)
Research Areas
The cluster's initial pilot project focuses on real-time flood forecasting and transportation accessibility in the Houston, TX region.
Synthetic Hazard Generation
Creating large datasets of synthetic weather patterns and employing AI for high-resolution downscaling to urban scales.
Infrastructure Systems Response
Characterizing hazard impacts on critical systems using AI- and optimization-enhanced approaches.
Multiscale Earth Observation and Data Assimilation
Leveraging heterogeneous datasets from diverse sources to improve model performance and enable real-time initialization.
Trustworthiness and Validation
Advancing the transparency and reliability of physics-informed AI methods through rigorous evaluation.
-
Selected News
-
Rice News Coverage of OpenSafe Fusion (August 26, 2024)
Rice University recently highlighted research from our AI for Climate Risk and Resilience at Rice (AI4CRRR) cluster. The feature story covers work led by cluster co-PI Jamie Padgett, focusing on an AI-driven system for real-time sensing of flooded roads.
The featured study introduces OpenSafe Fusion, an automated data fusion framework that leverages existing public data sources to predict road flooding conditions. By combining insights from traffic alerts, cameras, and traffic speed data, the system demonstrated promising results in tests using historical data from Hurricane Harvey. We hope this approach can significantly advance flood situational awareness without requiring substantial investment in new infrastructure, potentially improving public safety and mobility during increasingly frequent urban flooding events.
For more details, read the full story or access the research study.
Get Involved
Join the AI4ClimateRRR mailing list (link works best when opened on private window) to receive updates and announcements. The cluster holds monthly research discussions, with the next meeting occuring on November 20, 2024 at 12:00 PM CT. Contact James or Yanmo for additional information.
Updates will also be posted to the following web page, managed by the cluster: https://ai4climaterrr.rice.edu/
