The Effects of Landscape Composition on Avian Community Diversity

Personnel: Carl J. Anderson, David E.Capen

Cooperators: Vermont Military and Wildlife Management Institute


Introduction
This study explores changes in forest breeding bird diversity associated with differing landscape pattern in a forested landscape in northern Vermont. Specifically, it addresses the question: what is the relative contribution of spatial clustering or aggregation of similar patch types on changes in avian diversity? A better understanding of the influence of landscape pattern on the distribution and abundance of forest birds can help us make more informed decisions as we alter the landscape through development, agriculture, forestry, etc.

Objectives
Objectives of this study are: (1)To quantify the effects of differences in landscape composition on breeding forest birds- specifically the effects of spatial clustering of similar habitat patch types, (2)To identify the principle landscape attributes influencing variability in forest avian diversity, (3)To identify the principle landscape attributes influencing variability in specialized avian guilds and single species' abundance, and (4)To compare the influence of local habitat vs. landscape level factors on avian diversity.

Methods

Mapping
Aerial photographs were scanned, ortho-rectified using Erdas Imagine (Ortho-max) software, then interpreted to create an accurate and current land-use/ land-cover map for the firing range (see right and top-right of poster).

Bird Surveys
We developed a systematic/ stratified sampling design with bird survey transects located within >30 ha patches of similar habitat. Transects were stratified across deciduous, mixed and coniferous habitat types. Point-counts used a modified 50m fixed-radius circular plot method, after Blondel et. al. 1977. 10 minute counts were executed from approximately 5:00 am through 10:00am varying with local area sunrise. We established five point-count sites per transect with four to ten surveys per site. Approximately 925 point-counts were executed over 1999 and 2000 field seasons (33 transects, 163 point locations). Transect order and observers were rotated to ensure consistent sampling effort. The number of surveys varied per site depending on location, year and habitat type.

Vegetative Data
A suite of vegetative measurements were taken for all year 2000 transect locations to allow for local habitat characterization.

Data Preparation
We used FRAGSTATS software (Mcgarigal and Marks, 1996) to analyze the 250 ha circular landscapes around each bird transect. Avian diversity per transect was calculated using species richness, Simpson's and Shannon-Wiener indices (EstimateS 6.0 software, R Colwell, 2000). We ran rarefaction analyses on the bird diversity data for each transect to standardize species richness and index values to the number of samples.

Data Analysis
We compiled all bird survey data, then computed species richness and diversity indices for each transect. Rarefaction analyses were run to standardize bird diversity values for sampling effort. We used FRAGSTATS to quantify landscape metrics for circular landscapes around each transect location. Additional statistics yet to be run: We will run multivariate-multiple regression analyses to assess the relative contributions of landscape metrics against bird diversity values, focusing on those metrics related to spatial clumping of habitat types. We will use multi-colinearity diagnostics to eliminate covarying parameters. Principle components analysis will identify significantly contributing metrics (landscape and local) to variations in avian diversity. Spatial auto-correlation analyses will ensure statistical independence of each 250ha landscape.

Management Implications
Currently, our understanding of the effects of differing landscape pattern on the communities of species that inhabit those landscapes is low. To be sure, it is a complex topic, with many potential factors influencing changes in diversity at a variety of scales. We hope to add to the knowledge base of these influences by focusing on the landscape attribute of spatial clumping of patches, and by identifying the landscape metrics most influential to variations in forest breeding bird diversity on our study site. By better understanding the influence of these attributes under naturally heterogeneous landscape conditions, we can begin to better understand the influence that our artificial modifications of landscape pattern have on species' distribution and abundance, and hopefully make more informed land management decisions.

       Updated: 13 November 2002