Urban centers exhibit differential vulnerability to terrorism
Two US academics, Walter W. Piegorsch and Susan L. Cutter, have rated the vulnerability of 132 US cities to terrorist attacks, with some surprising results. Their method could be applied to the UK, providing a useful tool for homeland security planning
Urban vulnerability to terrorism is a highly complex phenomenon, with many factors underlying the occurrence and incidence of each event. While it is a challenge to distill terrorist vulnerability down to simple explanatory components, efforts to do so can provide guidance to numerous interested parties—emergency managers, urban first-response teams, insurance underwriters, banking and property administrators, to name a few—when assessing their locality's level of vulnerability to a terrorist event.
Towards this end, we recently studied the connections between terrorist incidence and a series of place-based vulnerability indices developed by the Hazards & Vulnerability Research Institute (HVRI) at the University of South Carolina; we combined the indices into a single summary measure to quantify the vulnerability to terrorism of 132 of the largest urban centers in the U.S. Our goal was to 'benchmark' an urban locality's level of vulnerability, using previous data on terrorist events and the other multidimensional factors that describe vulnerable conditions in that community.
Our methods hinged on manipulation of three separate placed-based vulnerability metrics; in a sense, we distilled the complex, multivariate data on place-based vulnerability collected by the HVRI into three summary dimensions: socio-economic vulnerability, built-environment (e.g., infrastructure) vulnerability, and geophysical hazard vulnerability. We then combined these statistically into a single Place Vulnerability Index, or PVI. Each of the tripartite constituents of the PVI was designed to represent a different component of the overall level of place-based vulnerability; their combination then served as a single summary measure that could model an urban location's exposure vulnerability to adverse events from a wide range of sources—natural hazards to willful human acts.
We related the PVIs to data on urban terrorist incidence and to casualties from such incidents in our 132 U.S. cities. The terrorism data were taken from two separate sources: The Terrorism Knowledge Base (http://www.tkb.org) produced by the Memorial Institute for the Prevention of Terrorism, and the Global Terrorism Database compiled by the U.S. National Center for the Study of Terrorism and Responses to Terrorism (START; http://www.start.umd.edu/). The terrorism incidence and casualty data, along with the underlying social, infrastructure, and geophysical hazard information comprising the PVI measure, spanned the 35-year period 1970–2004.
To perform the vulnerability benchmarking, we borrowed from a growingly-popular dose-benchmarking technology in cancer risk assessment: there, risk assessors use laboratory dose-response data to identify when exposures to carcinogens exceed a pre-specified level of risk for developing the cancer. By viewing the PVI as a form of 'dose' and the occurrence of a terrorist incident (or casualties from a terrorist incident, etc.) as 'cancer', the statistical methodology developed to set 'benchmark dose' exposure limits for adverse carcinogenic risk could be translated to setting 'benchmark index' limits for adverse terrorism vulnerability on our PVI scale.
We applied this benchmarking technology to both terrorist incidence data and terrorist casualty data for our 132 U.S. urban communities. This produced to two PVI benchmark points (technically, lower 95% confidence limits on the benchmarked PVI for each outcome). If a community was below both benchmarks, we gave it a 'low' vulnerability label; if it was below one and above the other, we gave it a 'medium' vulnerability label; and, if it was above both benchmarks we gave it a 'high' vulnerability label. Color-coded on a map of the U.S. as green, yellow, and red, respectively, this produced a visual, geographic indication of the pattern of urban vulnerability to terrorism among these large U.S. cities.
'Red' locations generally located in the eastern half of the U.S., and concentrated in a swath from Houston, TX, up through the U.S. Southeast and mid-Atlantic states, ending in the New York metropolitan area. Interestingly, all of the top five cities most vulnerable to terrorist attack based on our analysis—New Orleans; Baton Rouge, LA; Charleston, SC; New York-Newark; and Norfolk, VA—were port cities. Few western locations received 'red' scores, the only exception being Boise, ID. Indeed, the lower (relative) vulnerability urban centers appeared almost exclusively in the West and along the northern tier of states; no such lower-vulnerability cities appeared in the deep South.
Our results illustrated a robust methodology for the potential allocation of national and regional funding to support homeland security preparedness and response in U.S. cities. We found that not all urban areas are equally at risk, nor do they have the same underlying vulnerabilities. In a sense, "place matters," so any one-size-fits-all strategy of resource allocation and training will ignore the reality of geographic differences in the social, infrastructural, and geophysical environments, and will correspondingly affect an urban area's ability to prepare for and respond to terrorist events.
Similar sorts of analyses could be performed on any nation's major urban centers, including the U.K. If our "place matters" message were to hold true in Great Britain to the same extent that it does in the U.S., instructive guidance could be available to British emergency management officials considering the vulnerability of large centers such as London, Birmingham, or Glasgow, and of smaller communities whose underlying vulnerabilities might require more careful, national attention to terrorist threats.
Reference:
Piegorsch, W.W., Cutter, S.L., and Hardisty, F. (2007). Benchmark analysis for quantifying urban vulnerability to terrorist incidents. Risk Analysis 27, 1411-1425; http://www.blackwell-synergy.com/doi/abs/10.1111/j.1539-6924.2007.00977.x.
Walter W. Piegorsch
BIO5 Institute
University of Arizona
Tucson, AZ, 85721
Susan L. Cutter
Hazards & Vulnerability Research Institute
University of South Carolina
Columbia, SC 29208
