A Study of Cost of Hurricane Evacuations in Florida

The purpose of the study was to analyze variations in idealized costs of evacuation due to hurricane landfall. Evacuation populations were chosen as a forecast metric since evacuation zones combine information about population density and the areas of greatest vulnerability. Differences between forecasts and the actual storm track and intensity framed in terms of idealized evacuation statistics capture differences in storm track and intensity, as well as the susceptibility of a region to incurring damages at landfall.

For the initial analysis, the study focused on forecasts of hurricane landfalls in Florida during the 2004 and 2005 hurricane seasons as released by the National Hurricane Center. Forecasts of minimum pressure, wind intensity, and landfall coordinates starting from the initial tropical watch declaration up to landfall were compared with conditions at landfall. An index value was computed for the total cost for an evacuation, taking into account both the cost of evacuating the population of a region and the costs incurred by not evacuating the population actually impacted at landfall (overwarning). These measures were compared to the “cost” of a perfect forecast.

In addition, the ratio between the cost as predicted by a forecast and the cost at landfall was used to compare the seven storm tracks. From the seven storm tracks studied, there is no clear correlation between cost and time of forecast or between the ratio of forecast and landfall costs and time of forecast. Compared to errors in forecasted intensity, errors in forecasted hurricane location were more frequently responsible for excess costs from incorrect predicted evacuation areas. In the future, the analysis will be expanded to include data from computer model runs backing the National Hurricane Center forecasts.