Estimating Joint Return Periods for Flooding Due to Storm Surge

Exposure to coastal flooding has traditionally been studied by estimating the likelihood of flooding at a single location using return level curves. While a cornerstone of flood risk analysis, this approach neglects the spatial impact of flood events capable of affecting assets over large areas simultaneously. Understanding the joint likelihood of flooding at multiple locations is important when studying critical infrastructure systems comprising interdependent networks of energy, fuel, and water systems. Recent events such as Superstorm Sandy (2012), which led to power outages to over 8.5 million customers and severe fuel shortages, highlight the devastating consequences of cascading failures both within and across infrastructures systems due to flooding. In this study, a bivariate copula is used to estimate the joint return period of storm surge flooding at two locations flooded during Sandy (the East 14th Con Edison electrical substation and the Port Reading Hess refinery). The analysis is based on simulated storm surge from a database of 566 synthetic cyclone events.