Professor Oregon State University
Modeling Overland Flow of Hurricane Surge and Wave and the Impact on the Buildings
Co-Author: Mehrshad Amini (Univ. Rhode Island), Aaron Anton (Oregon State University) Stefano Biondi (NIST), Keenan Hubbard (Oregon State University), Rick Luettich (University of North Carolina), Dylan Sanderson (NIST), Don Slinn (NIST), Erick Velaso (Oregon State University)
Abstract: Traditional storm surge models generally neglect the influence of buildings, representing the urban landscape as a static, bare-earth surface with parameterized roughness. However, field and laboratory evidence shows that structures redirect currents, accelerate flow through gaps, create sheltered wake zones, and have significant impact on the wave field. Therefore, to estimate the overland surge and wave accurately, it is necessary to incorporate the effects of the built environment in a physically realistic way, including the progressive damage and collapse of structures.
This talk draws from two events: the impacts of Hurricane Ian on Fort Myers Beach in 2022 and the impacts of Hurricane Helen Florida’s Big Bend in 2024. This talk provides a background of the reconnaissance efforts, observed damages, Virtual Damage Assessments (VDA), and first floor elevations estimates all of which are used to inform the numerical modeling. The modeling efforts are comprised of regional scale modeling (ADCIRC/SWAN; DELFT3D) and wave-resolving models (eg. XBeach) for the overland conditions. The dynamic coupling of hazard to damage leads to building failure which in turns alters flood hazard. Validation with VDA estimates from the two testbeds will be discussed. The use of high-resolution models (OpenFOAM, DualSPHysics) to estimate wave loads on structures will be discussed in the context of a large-scale (1:3) physical model test of progressive damage and collapse of light-frame wood structures at the NHERI wave lab. The talk will conclude with a summary of recommendations from two NIST workshops held in 2025 for future research needs on surge/wave impacts in the built environment.
Associate Professor UC Berkeley
Fire Reconstruction and Risk Assessment in the Wildland-Urban Interface
Abstract: Over the last twenty years almost 130,000 structures have been destroyed by wildfires in the United States, devastating communities, disrupting insurance markets, and claiming over 300 lives. These Wildland-Urban Interface (WUI) fires move beyond natural vegetation into urban communities, turning buildings into fuels that spread conflagrations through neighborhoods. In this study, we apply the recently-developed Wildland–Urban Extension (WU-E) framework to the Eulerian Level Set Model for Fire Spread (ELMFIRE) that treats buildings as urban fuels ignited by embers, surrounding wildland fire, and structure-to-structure fire spread. Reconstruction of past WUI fires in California will be presented to demonstrate the model’s capability to better understand the behavior of past events. The model will then be extended to simulate thousands of stochastic events and perform risk analysis of communities. Mitigation strategies, including fuel treatments and home hardening, will be evaluated for their impact on structure loss probability across WUI regions. Despite its limitations, this study offers insights into the relative effectiveness of mitigation strategies and the design of WUI fire risk assessments, providing lessons to support evidence-based planning in fire-prone communities.
Assistant Professor New York University
Identify–Measure–Manage: AI-Driven Modeling of Infrastructure Risk and Adaptation under Climate Extremes
Abstract: Climate extremes increasingly propagate through interconnected infrastructure systems, causing cascading disruptions across transportation, energy, housing, and economic systems. While hazard simulations have advanced significantly, translating these simulations into an actionable understanding of infrastructure risk and adaptation remains a major challenge. This talk introduces the Identify–Measure–Manage (IMM) framework, an integrated modeling approach for infrastructure risk under climate extremes. The Identify pillar develops vision-language modeling approaches that map infrastructure exposure from satellite imagery. Projects such as RoofNet infer building materials and structural characteristics from imagery at a global scale, enabling exposure modeling even in data-scarce regions. The Measure pillar develops fast computational models to quantify hazard impacts and cascading failures across infrastructure systems, including high-resolution urban flood simulations and infrastructure interdependency models. The Manage pillar develops decision models that evaluate adaptation strategies through return-on-investment analysis, reinforcement learning under uncertainty, and interactive 3D visualization tools that support risk communication and education. These models integrate infrastructure exposure, system-level impacts, and adaptation decisions, enabling more effective analysis and management of climate risks in collaboration with stakeholders.