TitleThe spatial distribution of lead in urban residential soil and correlations with urban land cover of Balitmore, Maryland
NameSchwarz, Kirsten (author), Pickett, Steward (chair), Lathrop, Richard (co-chair), Weathers, Kathleen (internal member), McCay, Bonnie (internal member), Pouyat, Richard (outside member), Cadenasso, Mary (outside member), Rutgers University, Graduate School - New Brunswick,
SubjectEcology and Evolution,
DescriptionLead contamination of the urban environment is not a new phenomenon. A great deal of research has focused on the health effects of lead-based paint. Less attention, however, has been given to the potential problem of soil contaminated with lead from the past use of lead-containing products such as lead-based paint and leaded gasoline. Identifying areas of high contamination is necessary in order to prioritize soil remediation and public health efforts. This requires a comprehensive understanding of a highly heterogeneous and dynamic system.
This research addresses whether land use or land cover is a better predictor of lead concentrations in soil. Specifically, this research addresses whether landscape features, including trees, lawns, buildings, and roads, can be used to predict lead concentrations in soil. Through a method of rapid assessment of soil lead concentrations, I gathered spatially explicit data from urban residential yards to generate several models that predict the spatial distribution of lead in soil. Using the results of these models, potential inequities associated with the modeled spatial distribution of lead in soil and socio-demographic features were explored.
The results of this study suggest that the distribution of lead in urban residential soils is more closely correlated with features of urban land cover compared to metrics of land use. Specifically, the spatial distribution of lead in urban residential soils is strongly influenced by three factors: housing age, distance to the major road networks, and distance to built structures. Through the comparison of various spatial models, this research demonstrates that a greater amount of variation in the data is explained by machine learning techniques compared to traditional modeling techniques. In addition, important correlations between the modeled distribution of lead in soil and socio-demographic features such as race and poverty have been identified. Specifically, a greater amount of soil contamination is predicted to be present in high poverty areas.
This research contributes to the growing field of urban ecology by advancing our knowledge of how spatial heterogeneity affects the distribution of a critical pollutant in urban systems. This work also tests the suitability of using land cover as a predictive ecological variable.
NoteIncludes bibliographical references
Noteby Kirsten Schwarz
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.