Simulating storm surge waves for structural vulnerability estimation and flood hazard mapping

Nat Hazards DOI 10.1007/s11069-017-3001-5 ORIGINAL PAPER Simulating storm surge waves for structural vulnerability estimation and flood hazard mapping Adam Hatzikyriakou1 • Ning Lin1 Received: 30 November 2016 / Accepted: 18 July 2017  The Author(s) 2017. This article is an open access publication Abstract Wave action during storm surge is a common cause of building damage and therefore a critical consideration when estimating structural vulnerability and mapping flood risk. Traditional depth-damage curves, however, relate building vulnerability solely to inundation depth and therefore neglect an important damage mechanism. Similarly, flood mapping studies typically emphasize expected inundation rather than wave conditions. In this study, we consider the impact of wave effects on vulnerability estimation and flood mapping using a pair of hydrodynamic models (ADCIRC ? SWAN and BOUSS1D) to simulate inland storm surge flooding. The models are used to simulate flooding in a heavily impacted coastal community (Ortley Beach, New Jersey) during Hurricane Sandy (2012) and to estimate inland hazard parameters characterizing inundation, wave and velocity effects. To quantify structural vulnerability, fragility curves are developed by statistically relating the simulated hazard parameters to surveyed building damage. The results indicate that dynamic hazard characteristics such as significant wave height are the dominant predictors of severe structural damage. The flood simulation is also used to map the variation of surge and wave effects in the community. Comparing this analysis to flood zones delineated by the Federal Emergency Management Agency in the community’s Flood Insurance Rate Map reveals severe wave action and building damage in a significant portion of the community deemed least exposed to flood impact. It is suspected that this misrepresentation of risk resulted from overconfidence in the performance of the community’s frontal dune under severe surge and wave actions. Keywords Storm surge  Structural vulnerability  Flood mapping  Waves  FEMA  FIRM & Adam Hatzikyriakou Ning Lin 1 Department of Civil and Environmental Engineering, Princeton University, Engineering Quad, Princeton, NJ 08540, USA 123 Nat Hazards 1 Introduction More than 6.6 million homes in the USA, valued at nearly $1.5 trillion, are at risk of storm surge damage (Botts et al. 2015). This significant exposure to losses is compounded by expected increases in surge flooding due to sea level rise and climate change. In places like New York City, these combined effects are expected to change a 400-year flood event like Hurricane Sandy to a 20–130-year event by the end of the century (Lin et al. 2016). Given this widespread and increasing risk, accurately assessing storm surge impact is essential. Two important steps for doing so are estimating structural vulnerability and mapping flood hazards. 1.1 Structural vulnerability estimation Estimating structural vulnerability consists of relating a building’s performance during storm surge to its exposure to flooding. This has traditionally been treated using depthdamage curves which give the expected loss in a building’s economic value as a function of flood inundation depth (Davis et al. 1988). The most common approach for constructing such curves is by empirically relating post-event surveys of damaged structures to their estimated inundation depth using observed highwater marks (Davis and Skaggs 1992). Frequently used depth-damage curves include those developed by Federal Insurance Administration (FIA), which are the basis of the popular flood loss model HAZUS (Davis and Skaggs 1992; Scawthorn et al. 2006). Depth-damage curves have been pivotal for estimating structural vulnerability and are essential for quantitative storm surge risk assessment. Recent applications include predicting future flood losses due to climate change in major coastal cities (Hallegatte et al. 2013) and performing cost–benefit analyses of flood mitigation strategies (Aerts et al. 2014). Quantifying structural vulnerability using depth-damage curves, however, can be problematic in several respects. First, highwater marks used as proxies for the maximum water level at a structure are invariably sparse and can introduce large uncertainties into vulnerability estimation. Second, interpreting this maximum water level is frequently ambiguous. Whereas for slow-rise flooding (e.g., riverine and pluvial floods) highwater marks are typically characteristic of stillwater inundation, highwater marks during storm surge include the variable contribution of waves and flood velocity. Since wave and hydrodynamic loads on coastal structures are often orders of magnitude greater than hydrostatic loads (FEMA 2011), two structures with the same inundation depth can have considerably different damage responses depending on the relative magnitude of these dynamic effects. As a result, ignoring wave action and flood velocity may significantly misrepresent a structure’s performance during storm surge, particularly if the primary concern is estimating structural damage losses as opposed to content losses. To overcome these challenges, studies of recent surge events have used hindcast simulations of inland flooding as the basis for vulnerability estimation. One of the first such studies was an analysis of building damage on the Bolivar Peninsula during Hurricane Ike (2008), which confirmed that wave height and the associated slamming load were important predictors of building collapse (Tomiczek et al. 2014). Additional studies of Hurricane Sandy using a high-resolution inundation model determined that the depthaveraged velocity was also a significant damage predictor (Tomiczek et al. 2017). These results parallel similar studies of tsunamis, which have used numerical models to construct fragility curves based on flow velocity and hydrodynamic loads (Suppasri et al. 2011). Despite this progress, however, a formal framework for simulating flooding locally at a 123 Nat Hazards structure and extracting relevant hazard predictors when estimating structural vulnerability remains to be proposed. 1.2 Flood hazard mapping In addition to estimating structural vulnerability, accurately modeling inland storm surge flooding is also a fundamental step in flood hazard mapping. As the primary tool for communicating flood exposure, flood maps quantify the spatial variation of flood effects. Combining such maps with the estimated vulnerability of buildings to damage can be used to determine the risk storm surge poses to the built environment. In the USA, flood mapping dictating insurance premiums which ideally reflect this risk is governed by the Federal Emergency Management Agency (FEMA) through its National Flood Insurance Program (NFIP) (Crowell et al. 2007). Flood Insurance Rate Maps (FIRMs) delineate flood risk by defining flood zones based on the estimated 100-year base flood elevation (BFE). Since (...truncated)