Masters Theses

Date of Award

8-2004

Degree Type

Thesis

Degree Name

Master of Science

Major

Wildlife and Fisheries Science

Major Professor

David Buehler

Committee Members

David Buckley, Frank van Manen, Kathleen Franzreb, Arnold Saxton

Abstract

The recent decline of eastern hemlock (Tsuga canadensis) is a dramatic example of the effects exotic insect pests can have on forest composition and structure. Hemlock decline is attributed to defoliation caused by hemlock woolly adelgid (HWA, Adelges tsugae), an exotic Hemipteran introduced to the northwest U.S. from Asia during the 1920s and the eastern U.S. during the early 1950s. Within the past several years, HWA infestations have reached the southern portions of the Appalachian mountains and threaten to eliminate both eastern and Carolina hemlock (Tsuga carolinensis) from the region.

Recognized as nationally important bird areas by the American Bird Conservancy, lands administered by the U.S. Forest Service and National Park Service in the southern Appalachians form one of the largest networks of contiguous forested habitat in the eastern United States and provide great quantities of high quality habitat for numerous breeding bird species. Changes in southern Appalachian forest ecosystems resulting from loss of hemlock may have impacts on the distribution and demography of several avian species in the region. This study had two primary objectives designed to investigate the potential effects of eastern hemlock decline on southern Appalachian breeding birds: 1) develop models of avian associations with eastern hemlock in southern Appalachian national forests comparable to existing models developed for Great Smoky Mountains National Park (GSMNP) (Shriner 2001) and 2) compare nest success and the mechanisms influencing productivity of three hemlock-associated species [black-throated blue warbler (Dendroica caerulescens), dark-eyed junco (Junco hyemalis), and blue-headed vireo (Vireo solitarius)] in hemlock and hardwood-dominated stands.

Using the U.S. Forest Service R8BIRD database containing point-count data for Jefferson and George Washington National Forests (JGWNF), Virginia, and Cherokee National Forest (CNF), Tennessee, I developed logistic regression models predicting occurrence probabilities as a function of hemlock presence for 29 avian species. Two sets of models were developed based on single-year and multiple-year occupancy criteria for classifying a point-count location as a presence point for a given species. On JGWNF, the presence of Acadian flycatcher (Empidonax virescens), blue-headed vireo, black-throated blue warbler, black-throated green warbler (Dendroica virens), Canada warbler (Wilsonia canadensis), dark-eyed junco, northern parula (Parula americana), and veery (Catharus fuscescens) had significant positive associations with hemlock presence at a Bonferroni-corrected α-level of 0.0017. Eastern towhee (Pipilo erythrophthalmus) and indigo bunting (Passerina cyanea) displayed negative associations. On CNF, four species—Acadian flycatcher, black-throated blue warbler, black-throated green warbler, and northern parula—were positively associated with hemlock presence and one species—indigo bunting—was negatively associated at an α-level of 0.0017.

Three species showed a consistent positive association with eastern hemlock in JGWNF, CNF, and GSMNP: Acadian flycatcher, black-throated blue warbler, and black-throated green warbler. Although these species may experience negative impacts from hemlock decline, analysis of distributional data ignores mechanisms driving the observed species associations with hemlock, and, consequently, potential plasticity in habitat selection and other life history traits that may mitigate the impacts of hemlock loss are unknown.

A possible proximate mechanism for the hemlock associations identified by the models for JGWNF, CNF, and GSMNP may be greater reproductive fitness afforded by hemlock sites for hemlock-associated species. My second objective was to test the prediction that nest success of three hemlock-associated species in the southern Appalachians—black-throated blue warbler (BTBW), dark-eyed junco (DEJU), and blue-headed vireo (BHVI)—was greater in breeding habitats containing hemlock than habitats without hemlock. Nest success of the three focal species was compared between hemlock and hardwood breeding sites at two spatial scales: a 15-ha stand scale and a 0.04-ha nest site scale. Data for the stand-scale comparisons were collected in two 15-ha stands in Cataloochee Valley, GSMNP, North Carolina—one in an eastern hemlock stand and one in a rich cove hardwood stand. Data for the nest-site scale comparisons were collected at the Cataloochee site and at Coweeta Long-term Ecological Research Station in Nantahala National Forest, North Carolina. In addition to nest success, I investigated potential mechanisms influencing productivity by comparing Lepidoptera larvae abundance, parental provisioning rates of nestlings, and predator abundance in hemlock and hardwood-dominated stands. The potential impacts of hemlock decline on nest sites of my primary focal species, BTBW, were examined by developing models of nest-site selection using selected vegetation composition and structure variables sensitive to hemlock loss and relevant to BTBW nest placement.

Mayfield BTBW nest success at Cataloochee in 2002-2003 (pooled) was greater in the hemlock plot than in the hardwood plot (hemlock plot = 59.4 %, hardwood plot = 30.5 %, χ2 = 5.50, P = 0.0190). DEJU nest success did not differ between forest types (hardwood plot = 20.5 %, hemlock plot = 17.8 % , χ2 = 0.033, P = 0.8554). BHVI nest success did not differ between forest types (hardwood plot = 20.8 %, hemlock plot = 29.7 %, χ2 = 0.237, P = 0.6267). Nest success at 0.04-ha nest sites containing hemlock did not differ from nest success at nest sites without hemlock for BTBW (Cataloochee P = 0.4859; Coweeta P = 0.5879), DEJU (P = 0.9319), and BHVI (P = 0.2864).

Although Lepidoptera larvae biomass tended to be lower on the hemlock plot compared to the hardwood plot, the difference did not appear to affect the amount of biomass delivered to BTBW nestlings. Total biomass and caterpillar biomass delivered to BTBW nestlings did not differ between plots (hardwood total biomass = 13.16, hemlock total biomass = 15.29, total biomass P = 0.3302; hardwood caterpillar biomass = 5.90, hemlock caterpillar biomass = 9.00, caterpillar biomass P = 0.1050). Surveys of diurnal predators on the hemlock and hardwood plots did not detect any differences in relative abundance for any mammalian or avian predator except American crow (Fisher’s exact P = 0.0135), which had a greater number of detections on the hemlock plot. Using Akaike’s information criterion-based best-subsets variable selection, our BTBW nest-site selection logistic regression models revealed nest sites were positively associated with high shrub stem density and negatively associated with increasing overhead midstory cover.

The greater BTBW nest success observed in hemlock stands compared to hardwood stands is consistent with the hypothesis that hemlock-dominated sites in the southern Appalachians are “optimal” breeding habitats for this species. Although nest success in this study was not related to any measured microhabitat variables, observed differences in vegetation structure of the 15-ha plot scale may explain the nest success difference between forest types. The hemlock plot had greater means for shrub stem counts, horizontal visual obscurity (measured with a cover board), and overhead shrub cover. Therefore, based on the results of BTBW nest site selection models, the hemlock plot contained more potential nest sites and greater nest concealment, two characteristics that may have decreased predator foraging efficiency. This study suggests that impending hemlock decline from hemlock woolly adelgid may have negative reproductive consequences for black-throated blue warblers. However, these deleterious effects may be mitigated by an increase in shrub density from changing understory light conditions as canopy hemlock trees die.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS