Probabilistic Earthquake Scenarios: Extending Risk Analysis Methodologies to Spatially Distributed Systems

This paper proposes a methodology by which probabilistic risk analysis methods can be extended to the assessment of urban lifeline systems. Probabilistic hazard information is commonly used for site-specific analysis. However, for such systems as highway networks, electric power grids, and regional health care delivery systems, the spatial correlation between earthquake ground motion across many sites is important in determining system functionality. The methodology developed in this paper first identifies a limited set of deterministic earthquake scenarios and evaluates infrastructure system-wide performance in each. It then assigns hazard-consistent probabilities to the scenarios in order to approximate the regional seismicity. The resulting probabilistic scenarios indicate the likelihood of exceeding various levels of system performance degradation. A demonstration for the Los Angeles study area highway network suggests that there is roughly a 50% probability of exceedance of Northridge-level disruption in 50 years. This methodology provides a means for selecting representative earthquake scenarios for response or mitigation planning.