Kelly at work in the field
|
|
Kelly at work in the field |
|
|
KELLY
SOMERLOT
Department of Biology
Mentor: Terry
Derting
Title of Presentation: Landscape Patterns of Terristrial Vertebrate
Speicies Richness in Kentucky: Factors Influencing Future Planning
Presented at: the Annual Research Symposium at MSU in 2003 where
Kelly won the Outstanding Graduate
Research Award in Biology The research was supported by:
ABSTRACT As human populations continue to expand, increased development of the landscape reflects changes occurring in our natural environment. The challenge to preserve the diversity of species and the habitats that they require becomes more complex as human populations place heavier demands on the environment for resources and space. In order to quantify biodiversity in regard to human land-use demands found in Kentucky, I evaluated the relationship between wildlife diversity and human development. Twenty-one extrinsic variables characterizing land use in Kentucky were analyzed with regard to biodiversity indices of native terrestrial vertebrates, including species richness, taxonomic diversity, and frequency of special status species (i.e., endangered and threatened species). Variables used to characterize land use patterns were quantified for each county in the state using a Geographic Information System. Species richness analyses were also conducted on a county level using the predicted species distributions from the Kentucky Gap Analysis Project. Utilizing extrinsic land use variables to model landscape level patterns and predict associated biodiversity was found to be an efficient approach to understanding large-scale ecological implications of land use patterns with minimum data and experimental limitations. Eighteen of the 21 explanatory variables were found to be significant predictors of biodiversity (P= 0.10). Only four predictors (area of free water, habitat heterogeneity, land variation, and size and shape of forest patches) explained = 30% of the variation in biodiversity. Other variables, such as amount of area protected in federal land and wetland patch size and shape, were consistently used in many of the models but never explained more than 9% of the total biodiversity variation. Amount of area protected in state land and wetland patch edge were the poorest predictors of biodiversity, explaining less than 1% of the variation in 50% of the models in which they were used. Physical variables that follow a gradient pattern across the landscape, such as topography, were found to be strong predictors of biodiversity and also have an inherent relationship to the current status of the species. Predictors directly associated with human activity and development (i.e., population density and road density) were found to be secondary to those characterizing existing habitat (i.e., habitat heterogeneity, and amount of federally protected land) in modeling biodiversity. The resulting models provide a better understanding of the regional impacts that habitat fragmentation, human population growth, and land developments have on biodiversity in Kentucky. Land use patterns influencing biodiversity can be incorporated into Kentucky's future planning by modeling land changes and evaluating the effects on wildlife diversity. My study shows that freely available data (i.e., from State Gap Analysis Projects and the U.S. Census Bureau) can be used to effectively model large-scale ecological relationships between wildlife diversity and human land use.
 Date Modified July, 2003
|