The Wharton School
University of Pennsylvania
Over the past few decades scientific advances in hurricane prediction coupled with policy changes related to construction codes and evacuation procedures have reduced mortality and morbidity associated with tropical cyclones in the U.S.. Yet, the substantial costs of hurricanes when they make landfall suggests that this increase in awareness of hurricane risk has not been matched by commensurate increases in our ability to adapt to the risk, either through improved decisions about how and where to build or short-term decisions about how to prepare. We 1) report the findings of a program of research that tries to explain this adaptation paradox. We draw on evidence gathered from a unique program of survey research that measured coastal residents’ risk perceptions and preparation plans as they were being made while four hurricanes—Earl, Irene, Isaac, and Sandy---were approaching the United States coast during the 2010- 2012 hurricane seasons. The surveys measured subjective probabilities of different kinds of impacts, objective storm knowledge, the media channels through which this knowledge was being gained, and most importantly, the kind of preparation actions that had been taken and/or were planned.
And further as hurricane losses are often best mitigated from a local perspective it is important to delineate all the potential factors driving hurricane losses at the relatively local level. Thus we develop 2) A deeper understanding of the potential drivers of hurricane losses through a case study analysis contrasting two recent Category Three US landfalling hurricanes (Ivan in 2004 and Dennis in 2005) that, although similar in terms of maximum wind speed at their proximate coastal landfall locations, caused vastly different loss amounts. We show that the commonly used approach of making simplifying assumptions of loss confined to coastal counties and normalizing loss by specific exposure factors that typically represent only a single business line, can significantly misrepresent the true underlying localized loss, exposure, and vulnerability data. We also show that (in terms of the physical characteristics of the hurricane) size/area and duration of winds are at least as important as wind speed as potential drivers of damage; and (in terms of exposure and vulnerability attributes) building count, building density and building age are all potential drivers of damage. Appropriate consideration of these potential drivers of hurricane loss in statistical modeling and normalization techniques is essential for improved historical loss assessments and future projections of hurricane losses under climate change.
Thursday, 30 May 2013, 3:30 PM
Refreshments 3:15 PM
3450 Mitchell Lane
Bldg 2 Main Auditorium, Room 1022