Fatal tradeoff? Toward a better understanding of the costs of not evacuating from a hurricane in landfall counties

摘要: Given the significant population increases in United States hurricane-prone coastal areas since 1970, we empirically model U.S. hurricane fatalities from 1970 to 2007. The findings suggest that an apparent decline fatality risk may be overstated, especially considering impact of evacuation levels achieved, and various factors believed affect rate have divergent outcomes. research also illustrates necessity understanding a seminal storm such as Hurricane Katrina. In analyses, account for versus inland county location raw data, well role using zero-inflated specifications handle excessive number zero observations. Initial results, with without Katrina, show expected count by factor 1.7–3.4 one category level increase strength, begin provide bias data varying achieved. Finally, find higher more frequently struck counties, those Gulf Mexico; lower counties percentages over 65, less than 18, poverty. We found forecasting technology improvements had relatively minor effect on fatalities.