Estimating Evacuation Vulnerability of Urban Transportation Systems using GIS

Abstract

The focus of this thesis is to estimate evacuation vulnerability in a non-specific evacuation context. Three measures were calculated to evaluate the evacuation vulnerability of locations within Frontenac County, Ontario: population/capacity ratios, link removal analysis and the composite vulnerability measure. By utilizing both population/capacity ratios, as typified by Cova and Church (1997), and link removal analysis, a composite estimate of vulnerability can be determined. The composite vulnerability measure integrates the population/capacity ratios and the link removal analysis to examine the effect of link unavailability on evacuation difficulty. Subsequently, a linear regression analysis is utilized to estimate these complex measures incorporating simple variables. These include population, road and dead-end densities. The advantage of a regression analysis is that it can be used in any situation and for any given area. This research demonstrates that evacuation vulnerability of a network is of key importance for identifying which parts of a populated area may need plans put into place before the necessity of an evacuation occurs. Certain road configurations are shown to be more vulnerable than others, particularly neighbourhoods with a limited number of exits. Overall, the composite vulnerability measure identifies areas of increased and decreased vulnerability throughout Frontenac County. This regression analysis demonstrates a spatial pattern of calculated vulnerability values similar to those predicted based on the regression equation. Importantly, the regression analysis has demonstrated that it might be used to identify vulnerable areas in a simpler manner.