What Matters? The Peak, the Volume and the Duration of Floods and Their Coincidence Across the Globe

Abstract:

The bulk of the literature on floods relates to the computation of the flood frequency curve, which is based on the annual maximum or threshold exceedance characteristics of flows at a site. Practically, the duration of a flood may be just as important for response and recovery, and the volume of the flood for determining the adequacy of storage for flood control. Where many assets are at risk, either for property owners, or for a supply chain, or for an insurance portfolio, the coincidence of floods in certain areas may be of greatest interest. A trivial calculation shows us that as the number of assets considered increases, the probability of exceedance of multiple assets being flooded increases. Given climate variations such as ENSO that have impacts across the globe, global portfolio risk may be substantially higher than indicated by an analysis that does not conisder such spatial variability. Of course many assets are located in the same river basin and their correlated risk for flooding is well recognized but rarely modeled.

Today, as we consider resilience and other performance metrics for infrastrucutre systems, the correlation of flood risk in terms of peak, volume and duration, in space and in time needs to be considered. In this talk, I will present some evidence of the importance of this risk, and propose a hierarchical framework for the analysis of flood risk in a changing climate from this perspective.