Selection of representative natural hazard scenarios for engineering systems

Abstract

Representative hazard scenarios are essential for many tasks in risk management, such as preparedness and emergency response planning. However, criteria and methods for systematically selecting such scenarios for natural hazards are lacking. From a risk perspective, such scenarios should be selected considering the losses they incur. Hence, we propose to define a scenario that is representative for a certain degree of loss, for example, the 100-year loss, as the most likely one among all possible scenarios leading to this loss. Taking basis in a generic model of natural hazards and their impact on engineering systems, we formally introduce the representative scenarios. We then develop algorithms that enable an efficient evaluation of these scenarios. The method and algorithms are demonstrated on a hypothetical example considering a spatially distributed infrastructure system subjected to earthquakes.