Electronic Theses and Dissertations

Identifier

3757

Author

Cem Akkus

Date

2016

Date of Award

7-29-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Earth Sciences

Committee Chair

Esra Ozdenerol

Committee Member

Ramin Homayouni

Committee Member

Hsiang-te Kung

Committee Member

George Relyea

Abstract

Review of the recent literature suggests that integration of geographical information systems (GIS) into childhood lead exposure studies significantly enhances identifying lead hazards in the environment and determining at risk children. The purpose of this study is to find at-risk areas of childhood lead poisoning as well as determining risk factors in Shelby County, Tennessee. The two common deduplication methods: the first blood lead level (BLL) test result and the highest BLL test result were compared. Kappa statistic was used to investigate the effect of residential mobility on hot spots. Global and local spatial autocorrelations, Moran’s I and Getis and Ord’s Gi, were used to test the existence of global spatial autocorrelation as well as to find local pockets of high BLLs, and their trends. BLLs were grouped into four time periods during the 20 years between 1994 and 2013. Ordinary Least Squares (OLS) and Geographically Weighted Regression (GWR) were used to model risk factors. Kappa agreement results indicated that residential mobility had an approximate effect of 10% agreement change for multiple-screened children. Spatial autocorrelation statistics indicate that there is a strong global spatial autocorrelation within the BLL dataset. Local statistics showed that local clusters of high BLLs are concentrated in the western part of the county in the first period: 1994-1998 and moved to a more disperse pattern towards the east and south. The global and local statistical models showed that there is a significant relationship between the percent of children with elevated blood lead levels (EBLLs) and the percent of screening, median construction year, old housing, median income, monthly rent, African American population, education attainment, public assistance, poverty, and median income. Population density and vacancy were not found to be associated with the percent of children with EBLLs.

Comments

Data is provided by the student.

Library Comment

dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Share

COinS