Date of Graduation


Document Type


Degree Type



Davis College of Agriculture, Natural Resources and Design


Division of Forestry and Natural Resources

Committee Chair

Petra B. Wood

Committee Co-Chair

Donald J. Brown

Committee Member

Donald J. Brown

Committee Member

James Rentch

Committee Member

Michael Strager

Committee Member

Adam Duerr


In this study, I assessed the response of Cerulean Warblers (Setophaga cerulea) and 5 additional songbird species to timber harvests prescribed through operational silviculture. The research took place in relatively contiguous mature deciduous forests in 4 states in the central Appalachian region—Kentucky, Pennsylvania, Virginia, and West Virginia, USA.

For the first part of the study, I collected Cerulean Warbler abundance and territory data through point counts and territory mapping, respectively. I used the point count data to model Cerulean Warbler abundance pre- and post-harvest at 5 study areas (Kentucky [n=1], Virginia [n=2], West Virginia [n=2]) and post-harvest at an additional 2 study areas (West Virginia [n=2]). I analyzed territory data from 2 of the 4 study areas in West Virginia. The primary objective was to determine Cerulean Warbler response to timber harvests, implemented based on the Management Guidelines for Enhancing Cerulean Warbler Breeding Habitat in Appalachian Hardwood Forests (“Guidelines”), published in 2013, but covering a broader range of topographic conditions than those described in the Guidelines. The harvests at my study areas encompassed all available slope positions (i.e., lower to ridge) and aspects (i.e., Beers aspects 0–2). Two of my study areas were within the Ridge and Valley physiographic region, not included in the original study. I used 3 point types—harvest interior, harvest edge, and reference—to assess Cerulean Warbler change in abundance pre- and post-harvest and by years-post-harvest. I used dynamic, open population N-mixture models with point count data from the 5 study areas sampled pre- and post-harvest to estimate modeled abundance and population growth and change in abundance pre- to post-harvest, while accounting for detection bias. Using only the post-harvest data from these 5 study areas and the count data from the 2 study areas sampled only post-harvest, I used static N-mixture models to estimate modeled abundance and change in abundance among years-post-harvest. The primary results of this analysis indicated that point type, basal area of their preferred tree species, and basal area of large diameter trees were the most important drivers of Cerulean Warbler abundance. Models including slope position and Beers aspect had limited support for the data, which is important in comparing my findings to the Guidelines in which harvests were implemented within Cerulean Warbler preferred habitat characteristics of north- to northeast-facing aspects and upper slopes and ridgelines. Territory density increased 100% between pre-harvest and 2 years post-harvest. The greatest increases in abundance and territory density occurred where pre-harvest numbers were low.

For the second part of my study, I used logistic regression and resource selection functions to assess male Cerulean Warbler territory habitat selection with matched used-available habitat data. I used vegetation points at male Cerulean Warbler singing locations as “used” locations, which I delineated from detections during 7–8 territory mapping visits, and systematically sampled vegetation points within the territory mapping plot grids as “available.” The basal area data were organized into 8 categories reflecting Cerulean Warbler preferred or avoided tree species and structure. I also compared territory size and territory clustering among pre- and years-post-harvest. Territory size decreased post-harvest, suggesting an improvement in quality of breeding habitat. Territory size should increase when resources or quality are less than adequate and the bird must move around more to acquire resources. I did not observe territory clustering in any year. Pre-harvest, male Cerulean Warblers selected for increasing Beers aspect, whereas tree species composition was not important for territory habitat selection. Conversely, post-harvest, Beers aspect was the least important variable to cerulean habitat selection (largest ΔAICc value). Male Cerulean Warblers selected for increasing percentage of basal area that was trees intermediate to or overtopped by the canopy trees (i.e., midstory vegetation). Post-harvest, males selected breeding habitat across a wider range of available vegetation and topographic conditions than they did pre-harvest.

Finally, for the third part of my study, I analyzed the response of 5 additional focal species to the harvests at the same 7 study areas I used for Cerulean Warbler analysis, plus 2 additional study areas in Pennsylvania. The focal species are Eastern Towhee (Pipilo erythrophthalmus), Indigo Bunting (Passerina cyanea), Hooded Warbler (S. citrina), Scarlet Tanager (Piranga olivacea), and Wood Thrush (Hylocichla mustelina). I selected focal species that use the range of seral stages available in the central Appalachian region or that can be created and maintained through operational silviculture (i.e., early successional to mature forest). I again used static N-mixture models to estimate abundance and population growth of the 5 focal species, pre- and post-harvest, and among years post-harvest. I analyzed territory density for 4 of the 5 focal species. Eastern Towhee, Indigo Bunting, and Hooded Warbler increased post-harvest and continued to increase 2 and 3 years-post-harvest. Eastern Towhee and Indigo Bunting abundance increased most at harvest interior points followed by harvest edge points and both decreased in abundance at reference points. Hooded Warbler abundance increased at all 3 point types, indicating an overall positive response to the harvest mosaics rather than just to the reduction of basal area or opening of the canopy. Scarlet Tanager and Wood Thrush abundance initially decreased post-harvest, but subsequently increased at harvest edge and reference points or reference points, respectively. Although Wood Thrush remained on the study areas, their territory density declined by 50% during the first year post-harvest.

Within the third part of this study, I used the avian community data from 5 years of point counts to assess differences in community structure among point types within pre-harvest and each post-harvest year, and among pre- and the post-harvest years by point type. I used analysis of similarity and an ordination technique to evaluate the data. The avian community structure did not differ among point types pre-harvest nor did it differ among years at reference points, but the community structure did differ increasingly by year-post-harvest and among years at harvest interior and harvest edge points.

In summary, my findings corroborate the results of the Management Guidelines for Enhancing Cerulean Warbler Breeding Habitat in Appalachian Hardwood Forests, but show that harvests designed with Cerulean Warblers in mind can create breeding habitat on otherwise less desirable topographic conditions. Further, topographic conditions were not important influences on abundance of the five focal species from the third part of my study. Where mature forests are not available, silvicultural prescriptions can be used to enhance forest stands for Cerulean Warblers and a broad suite of other birds, particularly those that breed in shrubby early successional habitat, young forest, or mature forest gaps. Benefits to closed canopy, mature forest species may be limited, but my results show that they were not extirpated from harvest mosaics that retain reference stands or at least lighter harvests with higher basal area.