UVM Theses and Dissertations
Format:
Print
Author:
Anderson, Carl J.
Dept./Program:
Wildlife and Fisheries Biology Program
Year:
2004
Degree:
M.S.
Abstract:
Changes in landscape patch composition and configuration are among the most frequent and dramatic consequences of human alteration of habitat. Due to the scale of these changes and the complexity of the landscape mosaic, interpretation of their effects on various species and communities is difficult, yet the importance of understanding these effects cannot be overstated. This study addressed the following questions regarding landscape-scale habitat use by forest birds: Do avian species richness and abundance vary with landscape patch configuration and composition? If so, which metrics best predict these variations, and what management implications do these metrics suggest? My study site was the 4409-ha Ethan Allen Firing Range (EAFR) in Underhill, Vermont. This site's extensive topographic relief, coupled with a richly heterogeneous mosaic of patch types within a deciduous forest matrix, provided an excellent opportunity to research avian diversity and landscape pattern.
I surveyed 33 transects representing different patch configurations at EAFR. Each survey consisted of 50-m fixed-radius bird point-counts, usually 5 per transect, with 4 to 10 replicates per point over the duration of the study. I executed 925 1O-minute point-counts at 163 point-count locations during the 1999 and 2000 field seasons. I applied sample-based rarefaction analyses to point-count data to standardize sample size. I then calculated species richness and abundance values for overall diversity as well as for habitat-use guilds and various single-species analyses. Using GIS, I selected a suite of landscape-level variables describing patch composition and configuration, and calculated these values for 250-ha circular landscapes around each transect. I eliminated parameters with high levels of covariance, then ran multiple regression analyses to evaluate whether avian diversity and abundance varied with measures of landscape configuration. I also assessed effects of other landscape variables.
I controlled for contiguous matrix area and distance from survey point to edge. High-contrast edge appeared as a significant independent landscape variable in more models than any other index, explaining 48% of the variability of avian species richness on EAFR, 27% of total variation in member abundance of the edge-open guild, and 42 and 52% of the abundance variability of black-throated blue warblers (Dendroica caerulescens) and black-throated green warblers (Dendroica virens) respectively. Other important metrics within models included measures of landscape core area and cover- type composition. Regression models explained more than 60% of the variability in EAFR avian species richness and some guild abundances, and over 50% of the abundance variability for some single-species models. Vegetative characteristics showed weak-to-moderate relationships to species abundance and overall species richness, but did tend to support either known characteristics of individual species, or habitat associations also reflected in the landscape-level analyses.
I surveyed 33 transects representing different patch configurations at EAFR. Each survey consisted of 50-m fixed-radius bird point-counts, usually 5 per transect, with 4 to 10 replicates per point over the duration of the study. I executed 925 1O-minute point-counts at 163 point-count locations during the 1999 and 2000 field seasons. I applied sample-based rarefaction analyses to point-count data to standardize sample size. I then calculated species richness and abundance values for overall diversity as well as for habitat-use guilds and various single-species analyses. Using GIS, I selected a suite of landscape-level variables describing patch composition and configuration, and calculated these values for 250-ha circular landscapes around each transect. I eliminated parameters with high levels of covariance, then ran multiple regression analyses to evaluate whether avian diversity and abundance varied with measures of landscape configuration. I also assessed effects of other landscape variables.
I controlled for contiguous matrix area and distance from survey point to edge. High-contrast edge appeared as a significant independent landscape variable in more models than any other index, explaining 48% of the variability of avian species richness on EAFR, 27% of total variation in member abundance of the edge-open guild, and 42 and 52% of the abundance variability of black-throated blue warblers (Dendroica caerulescens) and black-throated green warblers (Dendroica virens) respectively. Other important metrics within models included measures of landscape core area and cover- type composition. Regression models explained more than 60% of the variability in EAFR avian species richness and some guild abundances, and over 50% of the abundance variability for some single-species models. Vegetative characteristics showed weak-to-moderate relationships to species abundance and overall species richness, but did tend to support either known characteristics of individual species, or habitat associations also reflected in the landscape-level analyses.