Author : Nicholas G. Walton
Publisher :
ISBN 13 :
Total Pages : 274 pages
Book Rating : 4.:/5 (953 download)
Book Synopsis Spatial Patterns of Avian Species Diversity in Forest Landscapes of the Western Great Lakes Region, USA by : Nicholas G. Walton
Download or read book Spatial Patterns of Avian Species Diversity in Forest Landscapes of the Western Great Lakes Region, USA written by Nicholas G. Walton and published by . This book was released on 2016 with total page 274 pages. Available in PDF, EPUB and Kindle. Book excerpt: Predictive species distribution models based on long-term, regional bird monitoring data can be used to assess the local distributions of forest breeding birds in the western Laurentian Great Lakes region of the United States. I modeled species distributions, species richness, and total abundance for 65 forest/forest edge nesting bird species that are abundant enough to permit statistical analysis. Predictor variables were derived from publicly available GIS land cover databases. Zero-inflated Poisson or negative binomial linear models were used to account for the excess number of zeros in the species distribution models, whereas Poisson general linear models were adequate for species richness and abundance models. Model selection was based on likelihood ratio tests for a backwards elimination process. My predictions provide predictive maps with an unprecedented fine grain, although the underlying approach parallels methods from other projects. While the predicted distributions were unique to each species, three general patterns can be recognized: 1) Twenty species showed higher abundances in the northern portion of the study area, especially Superior National Forest and adjacent parts of northern Minnesota. 2) Sixteen–eighteen species were more abundant in the southern part of the region; these species tended to avoid northeastern Minnesota but often had relatively high abundances along Lake Superior’s south shore. Some of these species were characteristic of forest edge habitats. 3) Most of the remaining 26 species showed some degree of geographic localization, but their predicted abundances exhibited no distinct north-south pattern. National forests were important for many of the species modeled. At least 10 species had distinctly high predicted abundances in one or more national forest. Northern hardwoods and mixed boreal forest were the most frequently selected positive covariates (41 and 40 out of 65 models, respectively). Pasture/hay was the most frequent negative predictor (19 models). Superior National Forest generally had the highest species richness. Overall abundances of individuals were more diffuse and generally greater in southern areas, including parts of central Minnesota and the north-central Upper Peninsula of Michigan. My analysis indicates that forest bird populations in the study area are generally patchy with locally abundant concentrations interspersed among uninhabited or sparsely inhabited areas. These results suggest that conservation efforts need to account for geographic context in addition to habitat. For example precipitation and temperature were significant predictors for 32 and 25 of 65 models respectively. Future work should include estimates of detection probability and tests of spatial predictions presented here.
