International Conference on Applications of Statistics and Probability in Civil Engineering (ICASP) (12th : 2015)

Local measures of disruption for quantifying seismic risk and reliability of complex networks Baker, Jack W.; Miller, Mahalia K.; Markhvida, Maryia


This paper presents a study of seismic risk to a complex transportation system while quantifying disruption at the local level. The San Francisco Bay Area transportation system is considered as a case study. The network consists of 32,858 road segments, 3152 bridges subject to damage, and 43 transit modes. A refined model of this network’s performance under damage, incorporating features such as transportation mode choice and dynamic demand, is used to predict disruption. Disruption is caused by earthquake shaking, where a full suite of earthquake scenarios in the region (with associated occurrence rates) are considered in order to obtain a fully probabilistic description of risk. Several strategies to manage the computational cost of this analysis are discussed. A number of network performance metrics are presented to provide insight into the disruption risks faced by residents of the region. Mode-destination accessibility, a metric based on network users’ utility functions, is used here as a limit state to evaluate the potential disruption to individual users of the transportation system as it is a performance metric of interest to urban planners. Additionally, local measures of disruption, such as changes in the number of trips in and out of individual locales, are used to identify regions where users may be at particularly high risk of disruption. Using this complex network model, computationally efficient analysis strategies, and refined measures of disruption, we obtain new insights about users’ risk, and obtain results in formats that are usable by urban planners responsible for long-term management of the transportation system’s risk.

Item Media

Item Citations and Data


Attribution-NonCommercial-NoDerivs 2.5 Canada