Font Size: a A A

Landscape-level analysis of suburban white-tailed deer overabundance

Posted on:2006-06-13Degree:Ph.DType:Dissertation
University:State University of New York College of Environmental Science and ForestryCandidate:Gorham, Dawn AFull Text:PDF
GTID:1459390008462985Subject:Agriculture
Abstract/Summary:
Efficient means of assessing wildlife habitat in suburban environments is needed in order to better understand how suburban community design affects conflicts between humans and wildlife. I use an interdisciplinary approach to address the problem of suburban white-tailed deer (Odocoileus virginianus) overabundance; combining the fields of wildlife biology, systems ecology and landscape architecture. I developed a methodology for using digital aerial orthoimagery imagery to create high resolution classifications in two suburban communities: Irondequoit and Amherst, NY. Classification of 1m digital orthoimagery for the two communities provided overall classification accuracies of 83.8% and 88.9%, respectively. I then created a suburban deer energy flow diagram for use in determining which composition and configuration variables may be important in creating a suburban deer habitat assessment model. I used regression analyses to create separate models for each study site as well as an overall model. Model creation resulted in the Irondequoit model, with an adjusted R2 = 0.68, the Amherst model, with an adjusted R2 = 0.44, and the combined model with an adjusted R2 = 0.52. Tree mesh was the variable most highly correlated to deer density in both communities. To examine issues of scale associated with creating a suburban deer model I examined different resolution imagery (1m, 10m, and 30m) for differences in both composition of classes and performance of the Irondequoit deer model. I found that high resolution imagery better represented the scale and complexity of the suburban environment. Finally, I used the 3 suburban deer models to compare patterns of suburban development in a third suburban community; Camillus, NY. I classified 60cm digital orthoimagery for Camillus, at an overall accuracy of 93.7%, and applied the 3 deer models to areas with different types of development. I found that predicted deer densities were significantly increased in conventional suburban (sprawl type) development as compared to traditional (neighborhood type) development. These findings show that suburban landscape design affects deer density, and highlight the need for integrative programs for community growth and development. This research also introduces the use of a systems ecology approach to examine suburban environments with respect to wildlife.
Keywords/Search Tags:Suburban, Deer, Wildlife, Adjusted R2, Development, Model
Related items