Bird-habitat relationships and the utility of multiscale methods (New Mexico, Vireo bellii)

My research investigates the relationships between habitat and birds and how those relationships change over different spatial scales. In this dissertation I present results from both field research and a study using theoretical modeling.; My field study looked at the relationship between habitat and nest success in a small songbird called the Bell's Vireo. I investigated nest placement and patch characteristics at two locations in New Mexico. I found that, when possible, Bell's Vireos place nests in locations with denser vegetation and canopy cover than what is available. However, nest success does not correlate with these conditions except at the landscape scale. This suggests that large-scale habitat selection, getting to the ‘right’ habitat patch in the first place, may be more important than the decisions Bell's Vireos make once they arrive.; The data from my field study were also used to examine biogeographic variation in Bell's Vireo habitat. Timothy Parker and I compared nest placement data at my New Mexico sites with a third location in Kansas. We found a general trend toward use of denser vegetation but discovered substantial variation in nest characteristics between localities. We evaluate these differences in light of their conservation significance and argue that conservation plans must be based on data collected at the location where the population of interest occurs.; My theoretical work evaluates the way different methods for rescaling data influence the degree of correspondence we can detect between birds and their habitat. Using a model developed at fine scales for the Mexican Spotted Owl, Bruce Milne and I systematically rescaled habitat (forest density, slope, and cover type) and distribution maps according to a variety of different rescaling rules, including spatial averaging, thresholding, presence/absence, and majority. We found that the spatial autocorrelation of habitat data interacts with rescaling rules to alter the correspondence between species presence and habitat across scale. Different rules lead to substantially divergent and sometimes opposite correlations among owl and habitat features on the landscape. We argue that such differences in interpretation due to variation in methodology have important implications for ecology and conservation.