John D. Landis is a Penn IUR Faculty Fellow and Crossways Professor and Chair of the Department of City and Regional Planning at the University of Pennsylvania School of Design.
Superstorm Sandy, coming just seven years after Hurricane Katrina, served to again remind America how vulnerable its large coastal populations are to extreme weather events, and of the need to better prepare for future such events. Better preparation is especially important in the foreshadow of climate change, which many meteorologists expect to generate more severe hurricanes and tropical storms. Congress’s slowness in appropriating post-Sandy rebuilding funds rekindled a periodic policy conversation about the government’s after-the-fact responsibilities to pay for rebuilding versus its before-the-fact opportunity to encourage or even require more resilient building forms. Promoting more of the latter to reduce the former is at the heart of Rebuild by Design (RBD) an initiative of the U.S Department of Housing and Urban Development (HUD), the Hurricane Sandy Rebuilding Task Force, and the Rockefeller Foundation.
Announced as a combination policy-design competition in June 2013, Rebuild by Design drew 148 competitors representing the top engineering, architecture, design, landscape architecture, and planning firms as well as research institutes and universities worldwide. Ten RBD finalists were selected, including a team led by Dean Marilyn Jordan Taylor, and Landscape Architecture faculty Ellen Neises, and Lucinda Sanders representing the University of Pennsylvania School of Design together with OLIN Partners. Working at a particular location, each RBD team is to develop new design and policy approaches to promoting more resilient housing, commercial, and infrastructure development models. The PennDesign/OLIN team is currently focusing its efforts on the Hunts Point Food Market area in New York City.
In theory, efforts like RBD should focus on locations where: (1) there is significant hazard potential; (2) the population is highly vulnerable to hazard-based danger, damage, or disruption; and (3) the population will have greater difficulty recovering on its own, or where resilience is lacking. This high hazard/high vulnerability/low resilience model has long been used in the disaster planning community to prepare precautionary disaster mitigation programs.
Operationalizing these concepts is harder than it might seem. Hurricanes, unlike earthquakes, floods, and even tornados don’t occur at known intervals or follow easily predictable pathways, especially where they are infrequent (as in the Northeast). Tides also matter a great deal: high tides can dramatically magnify storm surge potential, exacerbating coastal flooding and building damage. Despite its less-than-hurricane force wind levels, Superstorm Sandy was particularly devastating because it moved slowly and occurred when tide levels were high.
Vulnerability and resilience can be similarly difficult to gauge. Building and infrastructure vulnerability is largely a matter of location, elevation, and construction quality. Human vulnerability, by contrast, depends mostly on people’s ability to evacuate in a timely fashion. Resilience is mostly a function of economic resources. Recovery from disasters is far easier for higher-income households and profitable businesses with access to savings and insurance than for lower-income households and small businesses that are far less able to tap into reserve funds or insurance payments. The real estate market magnifies these discrepancies because poorer households and more marginal business are more likely to occupy lower-quality—and therefore more vulnerable and less resilient—structures.
To put its design efforts into the proper context, the PennDesign/OLIN RBD team began by using commonly available census and economic data to identify high vulnerability/low resilience locations along the Atlantic Coast from Maryland to Cape Cod. Based on data from the 2010 Census and the Department of Commerce’s County Business Patterns series, we identified census tracts with higher levels of socio-demographic vulnerability as those with year-round resident populations greater than 3,000; more than 500 residents younger than five years old or older than seventy-five; a household poverty rate in excess of 10 percent; and the percentage of households lacking available public transit and access to a car at 20 percent or more. We identified census tracts with higher levels of economic vulnerability as those with more than 100 establishments and 1,000 employees. Building on the relationship between housing vulnerability and housing costs, we identified census tracts with higher levels of housing vulnerability as those with more than 1,000 dwelling units; those where the median value of owner-occupied homes was less than 80 percent of county-median home value; and those where the median rent level was less than 80 percent of the county-median rent level.
The PennDesign/OLIN team took a similar approach to identifying census tracts with higher levels of demographic, economic, and housing resilience as those with residents or businesses with the personal and economic resources to bounce back on their own, and tracts with lower resilience levels as those that lacked such resources. Census tracts with higher levels of socio-demographic resilience were identified as those with unemployment rates below 8 percent; median household incomes above $50,000; and which had lower-than-typical percentages of one-person households, non-English-speaking households, and renter households. Census tracts with higher levels of economic vulnerability were identified as those with average payroll levels (per establishment) above $500,000 per year, and average annual payroll per employee levels above $40,000. Census tracts with higher levels of housing resilience were identified as those with more new housing, fewer numbers of blighted dwelling units, rates of homeownership above 50 percent, and median home values higher than 95 percent of the county median home value.
Map A below, in red, summarizes socio-demographic vulnerability by census tract for the counties around New York City harbor, one of the areas hardest hit by Superstorm Sandy. Map B, in green, summarizes the level of socio-demographic resilience. Map C, in green and red, combines the vulnerability and resilience ratings.
These vulnerability and resilience ratings are independent of particular hazards or storms. Viewed together, they suggest that combination of poverty, dependent physical and social isolation, and the deficiencies of an older building stock are at least as important as location and physical proximity when developing new models of disaster resilience.