View article

To assess the risk of natural disasters to complex urban infrastructure networks, researchers commonly use variations on Monte Carlo simulation because the the link between natural disasters and the network performance and community impact is difficult to express analytically. Particularly if evaluating the network performance is computationally intensive, a challenge then is to choose an appropriate suite of scenarios to capture the risk posed by natural disasters that allows for a computationally tractable evaluation. In seismic risk assessment, for example, efforts to choose a representative suite of scenarios, as expressed by weighted ground motion intensity maps, have traditionally focused on consistency with the seismic hazard only, whereas in some applications consistency with the exceedance curves of a performance measure may be more relevant. Here we show a method for selecting and evaluating a reduced suite of ground motion intensity maps consistent with both ground motion intensity hazard and performance measure exceedance curves for a regional network of bridges, highways and minor roads. We first use optimization to select a reduced suite of ground motion intensity maps. We then evaluate the consistency with the ground motion hazard. In the second stage, we choose a computationally efficient performance measure that is representative of a metric of greater interest and evaluate how well the reduced suite of ground motion intensity maps matches the performance measure exceedance curves obtained using an extensively-sampled catalog of ground motion intensity maps. To illustrate our approach, we present an …