Abstract
Urban centers are complex systems, formed and influenced by both external environmental variables like climate and geography, and by the residents who participate in the systems. As the global climate warms, these already complex systems are likely to change in ways that will negatively impact their residents, and certain resources will be required to mitigate and abate these changes. However, these resources are not always fairly distributed amongst participating communities within these systems, particularly populations considered socially vulnerable. One such resource is urban green infrastructure (UGI), the vegetation growing throughout a city that has the potential help mitigate flooding, reduce excessive heat, and improve the quality of life of urban residents. Although studies have examined the amount of UGI in relation to social vulnerability indicators (SVIs), few studies have examined the connectivity of UGI, a variable that can bolster ecosystem services provided by UGI, in relation to SVIs. This dissertation provides a novel method to quantify UGI connectivity inequity in urban centers by examining four US case studies and utilizing a series of Landscape Metrics, SVIs, Principal Component Analysis, and Mann Whitney U Tests to empirically test for inequity. For the first case study, Washington, DC, results indicate that there are disparities in tree areal coverage and connectivity between the upper 50th percentile and the lower 50th percentile regarding Minority %, with plots with a higher percentage of Minority residents having a significantly (p < 0.05) lower mean rank in Tree PLAND, Tree LPI, Tree PLADJ, and Tree Cohesion and a higher mean rank in Tree ENN_MN. The second case study, Phoenix, AZ, indicated disparities in tree and shrub connectivity, with plots in the upper 50th percentile of Poverty %, No High-school Diploma %, and Age 17- % all having a lower mean rank in Tree PLADJ, Tree Cohesion, Shrub PLADJ, and Shrub Cohesion than did the plots in the lower 50th percentile for those variables. In the third case study, Detroit Metropolitan Area, MI, there were no significant disparities facing vulnerable populations over the entire area. However, when comparing the city of Detroit to nearby Oakland County suburbs, results suggested that the city of Detroit, which is known to have a relatively high proportion of vulnerable residents compared to the surrounding suburbs, has a significantly lower mean rank for Tree PLAND, Tree LPI, Tree PLADJ, and Tree Cohesion and a significantly higher mean rank for Tree LSI than the suburbs of Oakland County. This suggests disparities based on city boundaries that were not exposed via this study's initial methodology. In the final case study, New York City, NY, results indicated a significantly higher mean rank for Tree ENN_MN in plots in the upper 50th percentile for the SVI variables Unemployed %, Uninsured %, and Minority % than in plots in the lower 50th percentile. The results of this dissertation empirically demonstrate disparities in UGI coverage in 3 of the 4 cities, and hints at disparities in the 4th, continuing the research into the identification of environmental injustices and providing a new method to quantify this facet of urban ecosystem dynamics through the lens of UGI connectivity.
