Our approach

We treat regional risk as an end-to-end problem. The same methodological chain applies whether the driving hazard is seismic, wind, or fire — which is what allows methods developed for one to transfer to the others.

1. Hazard characterization

   Physics-based simulation of earthquake ground motions at regional scale, coupled fire–atmosphere modeling of wildfire spread, and characterization of hurricane wind and surge fields — each with explicit treatment of the uncertainty that propagates downstream.

2. Building and infrastructure inventories

   Regional risk assessment is only as good as the inventory behind it. We develop automated methods that fuse satellite and street-level imagery with public records to generate attribute-rich inventories of the built environment.

3. Physics-based simulation

   High-fidelity modeling of structural and geotechnical response — soil-structure interaction, nonlinear site response, bridge and pipeline performance — resolved at the scale of individual assets rather than aggregated classes.

4. Data-driven models

   Machine learning models that carry physics-based results to regional scales where direct simulation is intractable, spanning surrogate modeling, transfer learning for structural health monitoring, and deep learning for rapid damage assessment from imagery.

5. Risk assessment and mitigation

   Translating estimated damage into risk and loss metrics that support retrofit prioritization, network resilience planning, and public policy. Much of this work is carried out with state and municipal agencies and with utility and infrastructure operators.

Funding

Talks and presentations

Selected publications

• Zsarnóczay A, Deierlein GG, McKenna F, et al., including Taciroglu E (2025). An open-source simulation platform to support and foster research collaboration in natural hazards engineering. Frontiers in Built Environment, 11, 1590479.
• Yen CH, Pitarka A, Tang H, Nakata R, McCallen D, Taciroglu E (2026). Regional earthquake ground motion simulations for Southern California with EQSIM. Earthquake Spectra, 42(3), e70079.
• Wang CF, Law K, McKenna F, Yu S, Taciroglu E, Zsarnóczay A, Elhaddad W, Cetiner B (2021). Machine learning-based regional scale intelligent modeling of building information for natural hazard risk management. Automation in Construction, 122, 103474.Code
• Shaik RU, Alipour M, Rowell E, Balaji B, Watts A, Taciroglu E (2025). FUELVISION: A multimodal data fusion and multi-model ensemble algorithm for wildfire fuels mapping. International Journal of Applied Earth Observation and Geoinformation, 138, 104436.Code
• Shamsaei K, Juliano TW, Roberts M, Ebrahimian H, Kosović B, Lareau NP, Taciroglu E (2023). Coupled fire-atmosphere simulation of the Camp Fire using WRF-Fire. International Journal of Wildland Fire, 32(2), 195–221.
• Tackie-Otoo NO, Askari M, Hadinata P, Davidson RA, Taciroglu E, Hardy G (2026). Hurricane wind loss modeling using insurance claims data. Natural Hazards, 122(7), 300.
• Zhang W, Restrepo D, Crempien JGF, Erkmen B, Taborda R, Kurtulus A, Taciroglu E (2021). A computational workflow for rupture-to-structural-response simulation and its application to Istanbul. Earthquake Engineering & Structural Dynamics, 50(1), 177–196.
• Dogan G, Cetinkaya MY, Soleimani-Babakamali H, Yilmaz Z, Taciroglu E (2026). Automated generation of model and seismic fragility inventories of school buildings in Türkiye. Journal of Building Engineering, 128, 116419.