The majority of the existing seismic risk studies use a deterministic approach to define vulnerability functions, despite the well-recognized large variability in the probability of loss ratio conditional on ground-shaking intensities. This study explored a statistical framework to simulate this variability, considering the existing correlation between assets separated by a given distance. The impact that these vulnerability modeling approaches may have in probabilistic seismic risk assessment is evaluated considering three fictitious building portfolios with distinct characteristics. To this end, loss exceedance curves and average annualized losses are compared, and recommendations are drawn regarding the optimal vulnerability modeling approach.

You do not currently have access to this article.