Playback surveys and breeding habitat characteristics of golden-winged warblers (Vermivora chrysoptera) on high-elevation pasturelands on the Monongahela National Forest, West Virginia

Across the breeding range of the Golden-winged Warbler (Vermivora chrysoptera), significant population declines have been documented even before the implementation of the North American Breeding Bird Survey (BBS) and continue as such today. The plight of this species has been blamed on the loss and degradation of early successional habitat, interactions with the closely related Blue-winged Warbler (Vermivora pinus), and Brown-headed Cowbird (Molothrus ater) brood parasitism. Most breeding Golden-winged Warblers are found in the upper Midwest and Great Lakes regions of the United States, but there are considerable populations persisting at higher elevations (\u3e600 m) in the central Appalachian region. States in this part of the range are witnessing some of the steepest population declines. In West Virginia, a state that is predominantly forested, Golden-winged Warblers breed on actively grazed high-elevation pasturelands composed of a matrix of grasses, forbs, Rubus, shrubs, few saplings and trees, and a forest edge. The habitats in this study are unique to the published literature on Golden-winged Warblers because they remain in a suitable state due to annual low-intensity (1.2--2.4 ha usable forage/animal unit) livestock grazing, require little additional management, and Blue-winged Warblers and hybrids are rare or absent. My primary objectives were to (1) test the effectiveness of using two types of Golden-winged Warbler song playback surveys at different times of day and season to monitor known populations and (2) quantify breeding habitat characteristics of this species on high-elevation pasturelands on the Monongahela National Forest, West Virginia.;In 2008 and 2009 on the Monongahela National Forest, I conducted point counts with Golden-winged Warbler song playback on grazing allotments where the number of territorial males was known. I used the proportion of males detected during song playback surveys (detection probability) as a measure of the effectiveness of each survey type. Behavioral responses were noted to further assess the value of playback. Results have implications for monitoring efforts and recent attempts to estimate absolute population size for this species.;I measured habitat characteristics at Golden-winged Warbler nests, territories, and random points associated with each. Habitat characteristics were grouped and analyzed using classification trees at the microhabitat, macrohabitat, and global (all variables) scales.;Regardless of song type or season, more Golden-winged Warbler males were detected during morning surveys than evening surveys. Detection probability was higher during playback of type 2 songs, associated with territorial interactions, than with type 1 songs, associated with mate attraction. Season did not have a significant effect on detection probability, although we only detected 40% of the known male population using 3-min passive point counts during the season that coincided with the BBS. In most instances, playback elicited behavioral responses that would allow positive identification between Golden-winged Warblers, Blue-winged Warblers, and their hybrids; however, with regards to detection probability, playback surveys performed similarly to passive point counts. Morning point counts with 1--1.3 mins of Golden-winged Warbler song playback followed by passive listening conducted from 22 May to 2 June can maximize detection probability for this species on pasturelands in West Virginia. Because detection rates were low, attempts to estimate absolute abundance from point count data will underestimate population size for this species.;Classification trees correctly identified nests, territories, and random plots at a rate better than would be expected at random. Vegetation density, microhabitat-scale woody ground cover, and macrohabitat-scale grass cover all emerged in multiple models as important explanatory variables. Most nests had woody vegetation supporting the nest or were built less than one meter from a shrub stem. The remaining nests were characterized by greater Rubus and herbaceous cover. Global classification trees combined variables from microhabitat and macrohabitat scales; however, no final global models incorporated variables from both scales. Thus, microhabitat and macrohabitat may be the important scales for categorizing nests and territories, respectively. Characteristics such as woody, grass, and vine cover, height of the shrub layer, and aspect may predispose a nest to success or failure. Low intensity livestock grazing can be used to manage habitat for Golden-winged Warblers breeding on high-elevation sites in West Virginia and perhaps elsewhere