Multihazard Analysis: Integrated Engineering and Social Science Approach

Abstract

Reducing the potential impacts from a future disaster can be accomplished through decreasing the hazard exposure and reducing the community’s vulnerability. Moreover, communities have both physical and social vulnerabilities that deserve attention; however, most engineering studies focus on assessing and mitigating the physical infrastructure without fully considering the social infrastructure. This paper offers a more holistic examination of vulnerability. Specifically, a two-stage analytical approach is presented that treats both an earthquake and a community’s socioeconomic and demographic makeup as hazards. The first stage addresses the physical vulnerability of a community through retrofitting the residential building stock using an inventory of woodframe building archetypes. The second stage incorporates the social characteristics of a community through modeling six social vulnerability variables. A social disaster factor (SDF) is introduced to offer a quantifiable approach for understanding the intersections between physical and social vulnerabilities. Case studies are presented for three communities: a middle-class ZIP code, the poorest ZIP code, and the wealthiest ZIP code, all in Los Angeles County, California. The SDF is computed and compared for the case studies during both stages of the analysis. The analyses demonstrate that when only physical vulnerabilities are modeled, one might incorrectly conclude that the impacts of the event are virtually eliminated. However, when social vulnerabilities are modeled as a hazard alongside the physical vulnerabilities, the projected impacts of the disaster are severe, especially for the most vulnerable populations, in terms of injuries, fatalities, posttraumatic stress disorder diagnoses, and number of dislocated households. In the combined model, these impacts run along racial and economic fault lines, with the most marginalized communities experiencing the most extreme projected losses. These results may have implications for both theory and practice.