Landscape Metrics to Assess Habitat Suitability for Conversation Bird Species in the Southeastern United States

摘要

Habitat fragmentation is a major factor in the decline of biological diversity and is an example of how changes in spatial parameters of a habitat can impact species survival. The degree to which a given species is affected by habitat fragmentation is dependent on the complex interaction of the habitat requirements of the species and the shape, size, and makeup of the fragmented habitat. Conservation of the biological diversity of a landscape would be facilitated if there was a way to determine the impact of habitat changes on species of interest. The objective of this study was to use existing U.S. Geological Survey (USGS) Land Use Land Cover (LULC) and the Breeding bird Survey (BBS) data from the 1970-1976 time frame to determine if kilometer-resolution horizontal spatial pattern metrics are suitable indicators of habitat suitability for conservation birds. The study included 15 conservation bird species with 53 BBS routes per species. It focused on using existing data in predicting bird abundance and evaluating the sensitivity of predictive models to varied sizes of landscape analysis units. Landscape structure was quantified using 12 spatial pattern metrics calculated from USGS LULC data. The metrics were summarized into three unique variables using principal components analysis techniques. Multiple regression analyses of bird abundance, as a function of the three variables, were used to explore the sensitivity of each bird species to landscape structure at various distances from the BBS route. Variables computed from the nearest distance were the most useful. Five of the species studied had models with R(sup 2) values greater than 35 percent. Of these, the wood thrush, Kentucky warbler, and prothonotary warbler, were sensitive to the habitat composition and forest configuration variables, while the hooded warbler and white-eyed vireo were sensitive to the forest configuration and landscape diversity/interspersion variables.