FRUIT ABUNDANCE AND LOCAL DISTRIBUTION OF WINTERING HERMIT THRUSHES (CATHARUS GUTTATUS) AND YELLOW-RUMPED WARBLERS (DENDROICA CORONATA) IN SOUTH CAROLINA

Auk, The, Jan 2004 by Kwit, Charles, Levey, Douglas J, Greenberg, Cathryn H, Pearson, Scott F, Et al

ABSTRACT.-We conducted winter censuses of two short-distance migrants, Hermit Thrushes (Catharus guttatus) and Yellow-rumped Warblers (Dendroica coronata), over seven years in five different habitats to determine whether their local abundances could be predicted by fruit pulp biomass. Sampled habitats were stands of upland and bottomland hardwood, loblolly pine (Pinus taeda), longleaf pine (P. palustris), and young (

RESUMEN. - Durante un periodo de siete anos realizamos censos invernales, en cinco ambientes diferentes, de dos migrantes de corta distancia, Catharus guttatus y Dendroica coronata, para determinar si las abundancias locales pueden ser predichas a partir de la biomasa de pulpa de fruto. Los ambientes muestreados fueron plantaciones de especies de madera dura en sitios altos y bajos (inundables), de Pinus taeda, de P. palustris y bosques jovenes (

THE DISTRIBUTION AND abundance of birds is a central theme in avian biology (MacArthur 1972, Cody 1985, Root 1988, Wiens 1989, Jones 2001). Habitat use, in particular, has drawn attention because of its relevance to conservation and management (Verner et al. 1986, Martin and Finch 1995, Sherry and Holmes 1996, Kilgo et al. 2002). Although most studies have focused on the habitat requirements of breeding birds, recent studies have begun to emphasize nonrandom habitat associations of species, especially long-distance migrants, when they are not breeding (Sherry and Holmes 1996, Marra et al. 1998, Marra 2000, Strong and Sherry 2000, Marra and Holmes 2001). To truly understand such patterns and their ecological consequences, one must first uncover the ecological mechanisms that generate them.

Availability of food is a likely mechanism underlying nonrandom distributions of wintering birds (Fretwell 1972, Jansson et al. 1981, Hutto 1985, Leisler 1990, Newton 1998, Johnson and Sherry 2001). However, winter bird distributions may not always mirror food distribution. Potential reasons include overabundance or rapid fluctuation of food, inability to accurately assess food abundance, high risk of predation, social interactions, and preference for habitats that mirror those used during the breeding season (Hutto 1985, Greenberg 1986, Recer et al. 1987, Rappole et al. 1989, Lindstrom 1990, Greenberg et al. 1993, Herrera 1998). Those alternative explanations for nonbreeding-season distributions are often difficult to separate. A first step is to focus on a single factor that seems important to determine if it alone can explain habitat use. Here we focus on food abundance.

A primary challenge in testing for a link between food abundance and wintering bird abundance is the difficulty of quantifying food resources in a way that accurately reflects food availability from a bird's perspective. In North America, many migrant passerines consume fruits, seeds, and insects during winter (Martin et al. 1951). Insects are difficult to sample; different methodologies yield different results and are appropriate for different types of in sectivores (Cooper and Whitmore 1990, Wolda 1990, Johnson 2000). In contrast, fruits are conspicuously displayed and relatively easy to census and are often consumed by many species (Moermond and Denslow 1985).

Using seven years of data across five habitats, we examined whether local abundances of Hermit Thrushes (Catharus guttatus) and Yellowrumped Warblers (Dendroica coronata) are related to fruit abundance. We selected those species because they are common and highly frugivorous and differ markedly in social behavior (and, therefore, may respond differently to variation in fruit abundance). In particular, many Hermit Thrushes defend territories on their wintering grounds (Brown et al. 2000), whereas nonbreeding Yellow-rumped Warblers usually occur in flocks (Hunt and Flaspohler 1998). Hermit Thrushes and Yellow-rumped Warblers also differ in the types of fruits they consume: Hermit Thrushes take a large variety of fruits (Martin et al. 1951), especially those that are lipid-rich (e.g. Cornus, Lindera, Viburnum; Whitmer and Van Soest 1998); whereas Yellow-rumped Warblers specialize on Myrica spp. fruits (Place and Stiles 1992, Hunt and Flaspohler 1998). A second goal was to test, when possible, whether variation in local abundance of those migrants could also be explained by other general factors, such as habitat type and year.

METHODS

STUDY AREA

The study was conducted from 1996 through 2002 at the U.S. Department of Energy's Savannah River Site, a National Environmental Research Park located in Aiken and Barnwell counties, South Carolina, USA (33�18'N, 81�37'W). The site lies within the Sandhill and upper Coastal Plain physiographic provinces. Forested areas cover >80% of the site (Workman and McLeod 1990). The majority of that area consists of a patchy mosaic of managed longleaf pine (Pinus palustris) and loblolly pine (P. taeda) stands and, to a lesser extent, upland and bottomland hardwood stands (Odum 1991, White and Gaines 2000).

As part of a long-term study begun in 1994 to address spatiotemporal patterns of fleshy-fruit and hard-mast production, a total of 56 plots, 0.1 ha (50 � 20 m) each, were established in 56 stands of 5 structurally and floristically distinct habitat types. Plots are separated by at least 600 m and are assumed to be independent. Ten plots each were placed in bottomland hardwood, upland hardwood, and stands clearcut in 1993. Thirteen plots each were placed in longleaf pine and loblolly pine stands that were at least 40 years old. Bottomland hardwoods on the site have a nearly continuous canopy layer of trees (e.g. Nyssa biflora, Magnolia virginiana, Qiicrcus spp.) and canopy-reaching vines (Rhus radicans, Smilnx spp.), a well-developed understory stratum (e.g. Ilex spp., Persea borbonia, Vaccinium spp.), and a pronounced ground layer of Sphagnum, woody stems, and herbs (e.g. sedges, Mitchella repens, Arisaema triphyllum). Upland hardwoods are characterized by a welldeveloped canopy (e.g. Quercus spp., Carya spp., Pinus taeda) and understory (e.g. Cornus florida, Ilex opaca, Vaccinium arboreum), and a relatively sparse ground layer. Although clearcuts initially consisted of short, second-growth vegetation (e.g. Phytolacca americaiia), by the end of our stvidy they had a "canopy" of densely planted longleaf pine (maximum height = 6 m) and often included a well-developed understory (e.g. M. cerifera, R. copallina, Prunus spp.) and a ground layer (e.g. grasses, sedges, Vaccinium stamincum, R. toxicodendron, Opuntia compressa); hereafter, we will refer to those former clearcuts as "regeneration habitat." Longleaf and loblolly pine habitats have a sparse canopy layer of pines, a few understory shrubs and trees (e.g. M. cerifera, Quercus spp.), and a sparse ground layer that often includes low-lying vegetation (e.g. V. stamineum, R. toxicodendron) and vines (e.g. Vitis rotundifolia, Smilax spp.). Successional development undoubtedly typified regeneration habitats in our study; however, prescribed fires in pine habitats, treefall gaps, and other natural processes contributed some degree of vegetatiortal change in all study plots.