Vulnerability and Hazard Modeling for Coastal Flooding due to Storm Surge

Abstract

Coastal flooding due to storm surge poses a persistent threat to coastal communities around the world, accounting for a majority of cyclone-related losses and fatalities. This significant existing exposure is compounded by sustained coastal development and the expected consequences of sea-level rise and climate change on future flooding. With more people living in areas increasingly more susceptible to flooding, better understanding storm surge risk is a critical concern.

Two important considerations when assessing storm surge risk are (1) accurately quantifying the vulnerability of buildings to damage and (2) reliably predicting the flood conditions likely to affect building performance. Previous flood studies have relied heavily on depth-damage curves relating the overall loss in a structure's value to flood inundation. This approach, however, does not capture important damage features occurring on different scales and neglects dynamic flood characteristics like wave action which are the primary cause of building damage. Using a detailed study of Hurricane Sandy (2012), this dissertation addresses both of these limitations by developing a multi-scale analysis of flood vulnerability and a framework for simulating inland flooding.

At the most detailed level, the multi-scale analysis presents manual and remote sensing methods for assessing flood damage to critical structural components like foundations and building envelopes. The obtained data is used to develop fragility curves quantifying component vulnerability as a function of house characteristics like building elevation. The multi-scale analysis next considers the correlated performance of nearby structures caused by interaction mechanisms like waterborne debris. Findings from two statistical methods proposed for studying correlated damage are used to discuss the implications of performance interdependencies on mitigation strategies like insurance pricing. Lastly, the multi-scale analysis presents methods for extracting community characteristics like dune erosion and relating them to overall community performance. These methods are used to investigate observed community performance during Sandy along 100 km of coastline in New Jersey and New York. Finally, a framework for efficiently simulating inland flooding using a pair of hydrodynamic models is presented. Using a hindcast simulation of Sandy, the framework is used to incorporate dynamic effects like wave action into vulnerability estimation and flood hazard mapping.