(Georgia Institute of Technology, 2016-07-20)
Rao, Arthi
This research creates a consistent, scalable approach for incorporating health considerations into regional land planning. A prototypical framework is presented for the Atlanta region. Determinants of healthy places from Social/Landscape Epidemiology, Urban Planning and Landscape Ecology are incorporated into defining the landscape and its associated pattern metrics. Key research objectives are to — 1) provide a new method to measuring urban form and health relationships through the use of landscape metrics 2) analyze urban form to understand configuration, mix, spatial distribution and proportions of land uses and socioeconomic factors and their association with health outcomes. Methodologically, this research examines associations between landscape patterns at nested scales (county and tract) with mortality rates across chronic conditions such as cancer, diabetes and heart disease. Two primary research questions are explored— 1) Are landscape patterns significant determinants of mortality rates? 2) At what scale do landscape patterns matter for reduced mortality rates?
Landscape Pattern metrics are generated using GIS software. Random Forest, Hierarchical Clustering and other classification techniques are used to identify preliminary landscape signatures and associations. Hierarchical impacts of county and tract-level determinants on local health outcomes are examined through multilevel logistic modeling. The aim is to present a succinct set of landscape metrics to inform land use planning for healthy communities. The framework developed can be used for multiple applications including Transportation Planning and sustainable Comprehensive Planning at multiple scales.