Loss Estimates for Large Soft-Story Woodframe Buildings in San Francisco

As part of a study for the City of San Francisco entitled Community Action Plan for Seismic Safety (CAPSS), we estimated the consequences of 4 large Bay Area earthquakes for large soft-story woodframe dwellings in San Francisco to inform City risk-management policy. A survey by the City's Department of Building Inspection found 2,800 residential woodframe buildings with 3 or more stories and 5 of more housing units with soft-story conditions. As proxies for these 2,800 unique buildings, four index buildings and three retrofits for each building were designed and analyzed. We developed nonlinear pseudostatic structural models for each building and retrofit. From lab tests and other observations, we created fragility functions for straight-sheathed woodframe walls, lath and plaster interior wall finish, brick veneer, and stucco veneer, and related the observed damage to HAZUS-MH damage states. We estimated the distribution of drift at which buildings would likely receive a red tag under ATC-20 post-earthquake safety inspection procedures. From observations of similar buildings in the San Francisco Marina District after the 1989 Loma Prieta earthquake, we estimated the transient drift associated with collapse. Using a procedure based on the HAZUS Advanced Engineering Building Module, we estimated the economic loss, ATC-20 safety tag color, and collapse probability for each of the 2,800 buildings in each of the 4 scenario earthquakes. Loss estimates were reviewed by an expert panel. We conclude that, in a M7.2 rupture of the San Andreas Fault, between 30 and 50% of the 2,800 buildings would be rendered unsafe to enter or occupy (i.e., they would be red-tagged), and an additional 10 to 30% would collapse. Seismic retrofit costing $11,000 to $17,000 per dwelling unit could reduce red tags to between 6 and 12% of the subject buildings, with 0.3 to 0.7% collapsing—substantial improvements on both counts—though repairs would still be costly.