Resilience against natural hazards can only be achieved at community and regional scales, and risk assessment must match this requirement. The first part of this seminar (Taciroglu) presents regional risk-assessment workflows that fuse automated metadata harvesting, data-driven mod-els, and high-performance computing—from fault-to-structure earthquake simulations of Istanbul’s building stock to AI-driven wildfire fuel and risk mapping for California.
The second part (Dutta) presents WUIgnis, a fire-spread model for the wildland–urban interface that resolves structure-to-structure ignition and ember-driven spread at the parcel scale. It describes the model’s coupling with NSF NCAR’s Coupled Fire Behavior Model (CFBM), the newly completed standalone WUIgnis implementation, and hindcast simulations of the January 2025 Eaton Fire evaluated against observed damage patterns in Altadena.