Testing Habitat-Relationship Models for Forest Birds of the Southeastern United States

Abstract

The bird-habitat relationships proposed by Hamel (1992) represent the most comprehensive description of habitat suitability for all birds that breed and/or winter in forests of the southern United States. These relationships were developed from compilations of published census and natural history data, field experience, and expert opinion. As such, these relationships can be considered untested models of bird-habitat associations. We tested the ability of these models to predict the distribution of 25 common breeding birds among the forest habitat types of the Southern Blue Ridge physiographic province. We used point-count survey data from national forests in Virginia, Tennessee, and Georgia to test the models. Thirteen of 25 models (52%) performed well (i.e., positive association between ranks from model predictions and ranks from the observed data) on the Tennessee data, while only 33% and 23% of the models performed well on the Georgia and Virginia data, respectively. Models for some species with distributions restricted to mid-aged to mature deciduous habitats (e.g., black-and-white warbler [Mniotilta varia] and black-throated blue warbler [Dendroica caerulescens]) and/or restricted to high elevations (e.g., veery [Catharus fuscescens]) performed well across study sites. However, models for other mature deciduous forest species did not perform as well because these species (e.g., ovenbird [Seiurus aurocapillus], scarlet tanager [Piranga olivacea], and wood thrush [Hylocichla mustelina]) tended to occur more frequently than predicted in early age classes or mixed forest types. For these species, we suggest that early-aged deciduous habitat types be considered marginal habitat (rather than unused, as defined in Hamel's publication) for these relatively widespread forest birds in the Southern Blue Ridge region. Models for habitat generalists (e.g., Carolina chickadee [Poecile carolinensis], Carolina wren [Thryothorus ludovicianus], and red-eyed vireo [Vireo olivaceus]) did not perform well. We suggest that Hamel's models be used cautiously for most birds in mature forest habitats. Further tests of these models are needed to clarify the discrepancies between the predicted and observed patterns of habitat use we observed.