FOREST DISTURBANCE: BREEDING ECOLOGY RESPONSE OF
SONGBIRDS
by
JILL M. WICK
A THESIS
Submitted in partial fulfillment of the requirements
for the degree of Master of Science
in the Department of Natural Resources and Environmental Science
in the School of Graduate Studies
Alabama A&M University
Normal, AL 35762
May 2008
Submitted by JILL M WICK in partial fulfillment of the requirements
for the degree of MASTER OF SCIENCE specializing in WILDLIFE SCIENCE.
Accepted on behalf of the Faculty of the Graduate School by the
Thesis Committee:
____________________________ Major advisor
____________________________
____________________________
____________________________
____________________________ Dean of the Graduate School
____________________________ Date
ii
Copyright by
JILL M. WICK
2008
iii
This thesis is dedicated to my biggest supporters: my parents, Norm and Diane
Wick; and my siblings, Joel, Jason, and Molly Wick.
iv
FOREST DISTURBANCE: BREEDING ECOLOGY RESPONSE OF SONGBIRDS
Jill Wick, B.S. University of Wisconsin at Stevens Point, 2002
Thesis Advisor: Yong Wang
Many migratory songbird species have experienced declines over the past four decades,
possibly due to loss of habitat. Forest management practices have the potential to create
suitable avian habitat. My research examined the effect of tree thinning and prescribed
burning on the breeding songbird community structure and avian individual fitness
(breeding success measured by home range size and nesting success) with a replicated
field experiment instead of focusing only on species abundance and richness, as much
past research has. The research was part of a comprehensive ecosystem research project
through collaboration with other researchers at Alabama A&M University and the USDA
Forest Service. I collected field data for one year pre treatment and one year post
treatment on the bird community composition and structure, microhabitat characteristics,
and microclimate conditions. One year after treatment, I collected data on arthropod
availability (avian food source) and home range and habitat selection of two songbird
species, the hooded warbler (Wilsonia citrina Boeddart) and the worm-eating warbler
(Hemithoeros vermivora Gmelin). Thinning had a greater impact on the bird community
than burning, although burning affected the bird community on a smaller scale. There
was an increase of shrub nesting and foliage foraging species on thinned plots, as well as
an increase in interior/edge species that use early successional habitats. However, the
thinning treatment did not displace most of the interior forest birds. The low intensity
prescribed burns had less of an effect on the bird community, however diversity increased
following treatment. Thinning and burning in combination resulted in decreased
v
diversity, a decrease in tree and cavity nesting and foliage foraging species. Interior/edge
species decreased and open/edge species increased. It appears that thinning and burning
in combination has more negative effects on the bird community than either treatment
alone. Hooded warbler home ranges ranged from 3.41 ha to 13.05 ha; worm-eating
warbler home ranges ranged from 4.38 ha to 8.81 ha. Core areas ranged from 2.19 ha to
5.84 ha for hooded warblers and 1.88 ha to 3.01 ha for worm-eating warblers. Hooded
warblers selected habitat with high herbaceous ground cover and vertical cover. Wormeating warblers selected habitat with an increased presence of slopes and high vertical
cover. It appears that understory forest structure is a key habitat feature for both species.
Keywords: Avian community, bird community, Cumberland Plateau, habitat selection,
hooded warbler, home range, prescribed burning, silviculture, thinning, worm-eating
warbler.
vi
TABLE OF CONTENTS
CERTIFICATE OF APPROVAL………………………………………………….. ii
ABSTRACT AND KEYWORDS…………………………………………………. v
LIST OF TABLES…………………………………………………………………. vii
LIST OF FIGURES.……………………………………………………………….. x
ACKNOWLEDGEMENTS……………………………………………………….. xvi
CHAPTER 1: INTRODUCTION, OBJECTIVES, HYPOTHESES, AND LITERATURE
REVIEW
Introduction……………………………………………………………………. 1
Hypotheses……………………………………………………………………... 2
Literature Review……………………………………………………………….4
Forest management…………………………………………………..…….. 4
Home range size………………………………………………………......... 8
Importance of the Study………………………………………………….…….. 11
Bibliography…………………………………………………………………… 12
CHAPTER 2: MICROCLIMATE, MICROHABITAT, AND BIRD COMMUNITY
CONDITIONS BEFORE STAND TREATMENT
Introduction…………………………………………………………………….. 16
Study Area and Methods……………………………………………………….. 17
Study Area…………………………………………………………………. 17
Sampling…………………………………………………………………… 19
Microclimate……………………………………………………….. 19
Microhabitat………………………………………………………... 19
Bird sampling………………………………………………………. 20
Data Analysis………………………………………………………………. 20
vii
Results………………………………………………………………………….. 24
Microclimate……………………………………………………….. 24
Microhabitat………………………………………………………... 26
Bird Community…………………………………………………… 26
Canonical correspondence analysis………………………………... 33
Discussion……………………………………………………………………… 36
Bibliography…………………………………………………………………… 39
CHAPTER 3: MICROCLIMATE, MICROHABITAT, ARTHROPOD AVAILABILITY,
AND BIRD COMMUNITY IN STANDS TREATED WITH THINNING AND
BURNING.
Introduction…………………………………………………………………….. 42
Study Area and Methods……………………………………………………….. 43
Study Area…………………………………………………………………. 43
Sampling…………………………………………………………………… 47
Microclimate………………………………………………………. 47
Microhabitat………………………………………………………... 47
Arthropod Abundance……………………………………………… 48
Bird sampling………………………………………………………. 49
Data Analysis………………………………………………………………. 49
Results………………………………………………………………………….. 54
Microclimate……………………………………………………….. 54
Microhabitat………………………………………………………... 62
Arthropod availability……………………………………………… 67
Bird Community…………………………………………………… 70
Canonical correspondence analysis………………………………... 79
Discussion……………………………………………………………………… 83
Bibliography…………………………………………………………………… 87
CHAPTER 4: HOME RANGE SIZE, HABITAT USE, AND REPRODUCTIVE
SUCCESS OF HOODED WARBLERS AND WORM-EATING WARBLERS IN
THINNED FOREST STANDS
Introduction……………………………………………………………………. 92
Study Area and Methods………………………………………………….... 93
Study Area………………………………………………………………..... 93
viii
Target Species…………………………………………………………….... 96
Sampling…………………………………………………………………… 96
Radiotelemetry……………………………………………………... 96
Reproductive success…………………………………………......... 97
Microhabitat………………………………………………………... 97
Data Analysis…………………………………………………………......... 98
Home range delineation………………………………………......... 98
Reproductive success…………………………………………......... 99
Microhabitat………………………………………………………... 100
Results………………………………………………………………………….. 100
Home range………………………………………………………… 100
Microhabitat ………………………………………………………. 102
Reproductive success……………………………………………… 106
Discussion……………………………………………………………………… 108
Bibliography…………………………………………………………………… 113
CHAPTER 5: DISCUSSION AND CONCLUSION
Conclusion……………………………………………………………………... 117
Management recommendations………………………………………………... 119
Bibliography…………………………………………………………………… 120
ix
LIST OF TABLES
Table 2.1
Guild memberships of detected bird species………………………… 22
Table 2.2.
Microclimate characteristics………………………………………… 25
Table 2.3.
Principle component analysis loadings for microclimate variables….. 27
Table 2.4.
Microhabitat characteristics…………………………………………. 28
Table 2.5.
Principle component analysis loadings for microhabitat variables…...29
Table 2.6.
Significance values and means for pre-treatment bird community
characteristics……............................................................................... 30
Table 2.7.
Species of continental importance detected on plots………………… 31
Table 2.8.
Morisita’s similarity index for bird community…………………....... 32
Table 3.1.
Guild memberships of detected bird species…………………………51
x
Table 3.2. Results of two-way analysis of variance for microclimate……………... 55
Table 3.3.
Results of two-way analysis of variance for changes in
microclimate………………………………………………………… 57
Table 3.4.
Results of one-way analysis of variance for microclimate…………... 59
Table 3.5.
Results of one-way analysis of variance for changes
in microclimate…………………………………….………………… 60
Table 3.6.
Principle component loadings for microclimate variables………...... 61
Table 3.7.
Results of one-way analysis of variance for microhabitat variables… 65
Table 3.8.
Results of one-way analysis of variance for changes in
microhabitats………………………………………………………… 66
Table 3.9.
Principle component loadings for microhabitat variables…………… 68
Table 3.10. Results of one-way analysis of variance for arthropod index………... 69
Table 3.11. Results of one-way analysis of variance for bird community……….. 74
xi
Table 3.12. Results of one-way analysis of variance for
changes in bird community………………………………………….. 76
Table 3.13. Species of continental importance detected on plots………………… 77
Table 3.14. Morisitia’s similarity index for bird community one
year after treatment………………………………………………….. 78
Table 3.15. Morisitia’s similarity index for bird community before treatment
and one year after treatment…………………………………………. 78
Table 4.1.
Distribution of radio tracked birds among treatments……………….. 101
Table 4.2.
Significance values and means of core area
size and home range size…………………………………………….. 103
Table 4.3.
Principle component loadings for habitat variables…………………. 105
Table. 4.4. AIC scores for logistic regression models…………………………… 107
xii
LIST OF FIGURES
Fig. 1.1.
Effect of forest disturbance on avian fitness through alteration of available
resources…………………………………………………………….. 2
Fig 2.1.
Location of study sites in Bankhead National Forest, Alabama…...... 18
Fig. 2.2.
Canonical correspondence analysis of bird species abundance and
microhabitat variables…………………….......................................... 34
Fig. 2.3.
Canonical correspondence analysis of nesting guild abundance and
microhabitat variables…………………………………....................... 35
Fig. 2.4.
Canonical correspondence analysis of foraging guild abundance and
microhabitat variables……………………………………………….. 37
Fig. 3.1.
Location of study sites in Bankhead National Forest, Alabama…….. 45
Fig. 3.2.
Experimental design…………………………………………………. 45
xiii
Fig. 3.4.
Interaction between burning and thinning in litter depth……………. 63
Fig. 3.5.
Interaction between burning and thinning in the change
in forest level 3………………………………………………………. 64
Fig. 3.6.
Interaction between burning and thinning in the
change in diversity index……………………………………………. 64
Fig. 3.7.
Interaction between burning and thinning in the change in cavity nesting
abundance……………………………………………………………. 71
Fig. 3.8.
Interaction between burning and thinning in the change in foliage foraging
abundance……………………………………………………………. 71
Fig. 3.9.
Interaction between burning and thinning in the change in resident species
abundance……………………………………………………………. 72
Fig. 3.11.
Canonical correspondence analysis of bird species abundance and
microhabitat variables……………………………………………...... 73
Fig. 3.12.
Canonical correspondence analysis of nesting guild abundance and
microhabitat variables………………………………………………... 81
xiv
Fig. 3.13.
Canonical correspondence analysis of foraging guild abundance and
microhabitat variables. ……………………………………………… 82
Fig. 4.1.
Location of study sites in Bankhead National Forest, Alabama……... 94
Fig. 4.2.
Experimental design…………………………………………………. 94
Fig. 4.3.
Example home range distribution for a hooded warbler and worm-eating
warbler……………………………………………………………….. 104
xv
ACKNOWLEDGEMENTS
I thank my advisor, Yong Wang, and my committee members, Callie Schweitzer,
William Stone, and Ken Ward for their guidance and advice throughout my course of
study. Zach Felix and Bill Sutton provided invaluable support and advice throughout my
research and for that I thank them. I also thank fellow graduate students, field
technicians, and others who have provided their help in the field and in the office:
Carleen Bailey, Matthew Bolus, Rachel Bru Bolus, John Carpenter, Drew Fowler,
Heather Howell, Lisa Gardner-Barillas, Daryl Lawson, Dawn Lemke, Kimi Sangalang,
Ryan Sisk, Molly Wick, Kelvin Young, and Joel Zak.
This publication “SONGBIRD BREEDING ECOLOGY: RESPONSE TO
FOREST MANAGEMENT” was developed under GRO Research Assistance Agreement
No. MA916706 awarded by the U.S. Environmental Protection Agency. It has not been
formally reviewed by the EPA. The views expressed in this document are solely those of
Jill Wick and the EPA does not endorse any products or commercial services mentioned
in this publication.
xvi
CHAPTER 1
INTRODUCTION, OBJECTIVES, HYPOTHESES, AND LITERATURE REVIEW
Introduction
Many migratory bird species have experienced population declines in the past
four decades, primarily due to loss of suitable habitat from anthropogenic environmental
disturbances (Askins et. al. 1990, Donovan and Flather 2002, Rappole and McDonald
1994). One of the factors contributing to the loss of suitable habitat for many bird
species in the United States is the prevention of disturbances, such as fire, wind throw,
beaver activity, floods, and forest management, which create and sustain early
successional forests (Askins 2001, Trani et al. 2001).
Canopy reduction and prescribed burning have been used to simulate natural
disturbance. Early research shows that such disturbance can affect the abundance and
availability of the resources on which birds rely (Weins 1989) (Fig. 1.1). Disturbance
can trigger changes in microclimate, habitat structure, food and nest-site availability,
predation, and nest-parasitism (Wiens 1989). Such alterations in turn affect the
likelihood of breeding success and fitness (an individual's contribution to the breeding
population in the next generation) of birds.
I studied the effect of forest disturbances, specifically thinning and prescribed
burning, on the avian community. I examined the effects of these disturbances on avian
1
species richness and abundance. In addition, I examined the mechanisms (microclimate,
habitat structure and composition, food availability, and brood parasitism) responsible for
changes in avian population demographics. My objectives were to (1) examine
differences in microclimate and microhabitat among disturbance levels, (2) determine
relationships between microhabitat and avian community structure, (3) determine the
effect of forest disturbance on food availability, (4) determine relationships between
forest disturbance and avian territory size, and (5) determine the relationship between
forest disturbance and avian breeding success.
Forest
environmental
disturbance
Resource
Abundance
Foraging
Territory
Nest site
Resource
Availability
Productivity
Survivalship
Resource
Use
Resource
Allocation
Individual
Performance
Figure 1.1. Effect of forest disturbance on avian fitness
through alteration of available resources. Modified based
on Wiens (1989).
Population
Patterns
“Fitness”
Community
Patterns
Research Predictions
Microhabitat and microclimate parameters will be affected by changes in
canopy density. Hypothesis 1: (1) Vegetation structure will be more complex in stands
thinned than those burned. (2) Vegetation composition will be more complex in stands
burned than those thinned. (3) Temperature will be indirectly related to the amount of
canopy reduction, whereas moisture will have a direct relationship with canopy
reduction.
2
Avian species richness and relative abundance will vary throughout
treatment types and will be correlated with habitat heterogeneity. Hypothesis 2: (1)
Avian species richness and abundance will be greatest in thinned stands due to increased
understory habitat heterogeneity and complexity (Wiens 1989). (2) Abundance of
canopy nesting and foraging birds will be lower on thinned stands due to decreased
overstory habitat complexity. (3) Abundance of sub-canopy nesting and foraging birds
will be higher on treated stands due to increased understory habitat complexity. (4)
Abundance of ground nesting and foraging birds will be lower on burned stands due to
decreased litter layer.
Food availability will be affected by changes in forest structure. Hypothesis
3: (1) Foliage feeding insect abundance will be directly related to increased leaf growth.
Avian home range size will vary among the treatments and will be correlated
with food availability, and size of individuals. Hypothesis 4: (1) Home range size for
ground nesting birds will be directly related to canopy reduction, due to decreased food
resources with high disturbance (thin*burn). (2) Sub-canopy nesting birds will have
larger home ranges on treated increased due to decreased habitat complexity. (3) Larger
individuals will have larger territories than smaller birds (Fretwell and Lucas 1970).
Nest success will vary among the treatments and will be correlated with food
availability, presence of predators, and presence of brood parasites. Hypothesis 5:
(1) Increased brood parasitism will be seen in thinned stands due to a more open canopy
(Chambers et al. 1999, Evans and Gates 1997). (2) Ground nesting birds will have lower
nesting success in burned stands due to the reduction of nesting substrate and food
3
availability. (3) Sub-canopy nesting birds will have lower nesting success on treated
stands due to decreased habitat complexity and food resources.
Literature Review
Forest management
Forest managers have the ability to employ a variety of silvicultural techniques,
depending on their management goal. Intermediate stand treatments are used to enhance
stand composition, structure, growth, health and quality by regulating and controlling tree
growth through adjusting tree density or species composition (Smith et al. 1997). These
treatments do not direct any effort at regeneration. Thinning treatments impact stand
growth, development and structure; and enhances vigorous tree growth of selected trees
through the removal of the competitors (Smith et al. 1997). If sunlight is the main
limiting factor for the desired trees, thinning from above is used to reduce competition for
sunlight (Smith et al. 1997). If water is the main limiting factor for the desired trees,
thinning from below is used to reduce the competition for water in the soil (Smith et al.
1997). Free thinning removes trees to control stand spacing (density) and favor desired
trees without strict regard to crown position (Smith et al.1997).
Prescribed fires that are slow and low-burning are used as a silvicultural tool in
many western systems and in savannah habitats however; in eastern forests; the use of
prescribed fire is less common. In the eastern United States, fire has historically been
used to keep forest open for hunting (by Native Americans) and to clear land for pastures
and grazing (by early farmers). Forest managers use fire to reduce or eliminate fuel loads
and undesirable vegetation, to prepare seedbeds for regeneration of wind-disseminated
4
species, to control competing vegetation, and control pests (Smith et al. 1997). Fire also
stimulates herbaceous species and sprouts of woody plants, so it is used to create or
improve wildlife habitat (Smith et al. 1997).
The response of birds to forest management varies by the degree of disturbance
and type of disturbance. Density and productivity of birds that use mature hardwood
forests tend to decline with tree removal, whereas early successional species tend to
experience an increase in density and productivity (Gram et al. 2003). More intensive
cutting results in more dramatic changes in abundance, species richness, and density.
Clearcuts cause species composition to change drastically (Brawn et al. 2001, Chambers
et al. 1999, Harrison and Kilgo 2004), and cause decreased density (Harrison and Kilgo
2004), species richness, (Harrison and Kilgo 2004) and abundance (Chambers et al.
1999). Some birds will use clearcuts for foraging, but do not nest in these areas
(Chambers et al. 1999). Gram et al. (2003) found higher densities of early successional
species after clearcuts, but densities of mature forest species declined. Productivity of
mature forest birds also declined in clearcuts, whereas productivity increased for some
early successional species.
Thinning has the potential to create conditions similar to those caused by natural
disturbance. Thins that retain a greater number of trees will mimic small-scale
disturbance, whereas thins that retain a few trees will mimic large-scale disturbance
(Harrison et al. 2005).
In bottomland hardwood habitat, lower levels of tree retention (0
-20%) result in lower species richness and a lower similarity value between pre- and posttreatment bird communities (Harrison et al. 2005). Birds that are more dependent on
trees and shrubs for nesting or foraging benefit from more residual trees, whereas ground
5
dwellers are less affected by how many trees are retained (Tittler et al. 2001). Over time,
ground nesting and foraging birds increase, but these increases are by open-habitat or
generalist species, and not forest dwelling birds, as were present before thinning
(Harrison et al. 2005, Tittler et al. 2001). Lang et al. (2002) found that thinning and
burning to create red-cockaded woodpecker (Picoides borealis Vieillot) habitat did not
affect the daily movements of adult wood thrushes (Hylociclha mustelina Gmelin),
however fledglings on treated areas moved farther from their nest than those on untreated
areas. The habitat preferences of adults altered slightly after treatment; they preferred to
use riparian areas more than upland areas (Lang et al. 2002).
Group selection cuts are often used to mimic small-scale natural forest
disturbances by creating small gaps in the forest (Jobes et al. 2004). Group selection cuts
increase understory plant production, providing more opportunity for lepidopteron larvae
and creating habitat used by many early successional species (Blake and Hoppes 1986).
Studies report decreased abundance and diversity in group selection cuts compared to
control sites in western coastal forests (Chambers et al. 1999), however in eastern
hardwood forests, densities of early successional species increased (Gram et al. 2003). In
southeastern bottomland forests, worm-eating warbler (Helmitheros vermivorus Gmelin)
densities declined in both group selection cuts and clearcuts, however they rebounded
more quickly in group selection cuts (Harrison and Kilgo 2004). Hooded warblers
(Wilsonia citrina Boddaert) and indigo buntings (Passerina cyanea Linnaeu) had higher
nesting success in group selection cuts than in clearcut hardwood stands (Alterman et al.
2005, Whittam et al. 2002).
6
Prescribed burning is used in many different ecosystems and has different effects
in each. In oak savannas, avian density and species richness were lower in unburned than
in frequently burned sites (Davis et al. 2000), whereas density and species richness in
coastal sage scrub and Florida pine scrub were higher in unburned sites (Greenberg et al.
1995, Stanton 1986). Other studies have found no changes in species richness when
comparing burned areas to unburned areas (Artman et al. 2001, White et al. 1999). White
et al. (1999) found burned pine forests to be preferred nesting habitat over unburned sites
for a variety of species; however, productivity rates in these sites were low. In mixedoak forests in Ohio, densities of ground and low-shrub nesting birds declined in burned
sites compared to unburned sites (Artman et al. 2001, Blake 2005). In burned sites,
hooded warblers shifted their territories and wood thrushes moved their nests from xeric
and intermediate sites to mesic sites (Artman and Downhower 2003, Artman et al. 2001).
Wood thrushes also placed their nests higher off the ground and in areas where burn
intensity was low. Due to this, nest survival rates did not differ between sites (Artman
and Downhower 2003). Apparently, the birds were able to minimize the affects of
burning by altering their nest placement.
The effect of forest management on nest parasitism rates is inconclusive. Some
studies report that the effects of fragmentation on nest parasitism rates in forested
ecosystems are negligible (Alterman et al. 2005, Artman and Downhower 2003, Harrison
and Kilgo 2004, Stuart-Smith and Hayes 2003), whereas other studies found parasitism
rates to be related to forest edges (Evans and Gates 1997, Moorman et al. 2002). Evans
and Gates (1997) found the highest occurrence of cowbirds at forest-brush and foreststream edges, with the probability of parasitism greater nearer to the edge. Cowbird
7
occurrence is higher in open-canopy plots than closed-canopy plots (Evans and Gates
1997, Chambers et al. 1999). Areas with high basal area of snags and high total
vegetation volume are also correlated with high cowbird use because they provide
perches from which cowbirds can search for host nests (Evans and Gates 1997).
Nest predation is the major cause of nest failure for passerines (Alterman et al.
2005, Artman and Downhower 2003, Harrison and Kilgo 2004, Stuart-Smith and Hayes
2003, White et al. 1999). Nest predation rates have not been found to be correlated to any
type of forest management (Brawn et al. 2001, King and DeGraaf 2000, King et al. 1998,
Stuart-Smith and Hayes 2003) however; predation rates are higher closer to edge than the
interior forest (King et al. 1998). Because nest predation is related to edge, silvicultural
treatments that produce more edge (i.e., group selection cuts) have the potential for
greater impact on the avian community (King et al. 1998).
Home Range Size
An animal’s home range, as defined by Burt (1943), is “that area traversed by the
individual in its normal activities of food gathering, mating, and caring for young.” A
birds’ territory is the area that he overtly defends from conspecifics (Wilson 1979). The
home range differs from the territory in that it may contain areas that the bird uses, but
does not defend as his own. For birds, breeding home ranges and territories provide
resources for nesting and fledging young. Historically most research has been able to
report on territory size, but due to recent developments in smaller radio transmitters,
analyzing home range has become a viable option.
Home range size varies in differing habitat quality levels, with poor quality
habitat resulting in the need for a larger home range and subsequently in a lower bird
8
density at a site (Mazarolle and Hobson 2004). Home range size and quality can be
influenced by a variety of factors, including the availability of nest sites, perch sites, and
foraging areas; food abundance; predator abundance; bird density; competition with
conspecifics; and age and size of the bird (Fretwell and Lucas 1970, Mazarolle and
Hobson 2004, Petit and Petit 1996, Smith and Shugart 1987, Wilson 1979). The role of
these factors in home range and habitat selection has long been a focus in avian ecology.
The ideal dominance model of habitat selection states that dominant individuals
occupy optimal habitat and subordinates are forced to occupy lower quality habitat or to
not defend a territory at all (Fretwell and Lucas 1970). Mazarolle and Hobson (2004)
found that older and larger ovenbirds (Seiurus aurocapillus Linnaeus) had larger
territories than smaller and younger males, thus supporting the model. The older males
also had more exclusive territories with fewer overlaps. Older and larger prothonotary
warblers (Protonotaria citrea Boddaert) occupy higher quality habitats than subordinate
males (Petit and Petit 1996). Older males arrive on the breeding grounds earlier than
younger males, so prime habitat is often ‘claimed’ before younger males arrive (Petit and
Petit 1996). Smith and Moore (2005) found that American redstart (Setophaga ruticilla
Linnaeus) males that arrived on the breeding grounds earlier appeared to settle on higher
quality territories and hatched nestlings earlier, while females began their clutches earlier
and produced heavier nestlings than later arrivals. The model assumes that fitness at high
population levels is lowered by some other mechanism, such as resource abundance or
predation risk. Forest management has the potential to alter the mechanisms which affect
the ideal dominance model.
9
The food-value theory states that birds defend territories to have sufficient
resources to raise offspring (Wilson 1979). Many studies concur that territory size is
related to the abundance of prey availability (Marshall and Cooper 2004, Nagy and
Holmes 2005, Petit and Petit 1996, Smith and Shugart 1987). There are three hypotheses
that imply different mechanisms by which territories and prey abundance are related
(Smith and Shugart 1987). The first is the direct-monitoring hypothesis, which states that
individuals directly monitor prey abundance and adjust territory size, thereby insuring
sufficient resources. This hypothesis has been rejected because territory size appears to
be relatively constant during the breeding season, whereas prey abundance is not
(Marshall and Cooper 2004). With the second hypothesis, an individual defends an area
as large as possible, but territory size is limited by intra-specific competition for habitat
(intra-specific competition hypothesis). The third hypothesis, the structural cues
hypothesis, states that territory size is a function of expected prey abundance based on
habitat structure. The latter two hypotheses are consistent with and complement each
other. If the structural cues hypothesis is true, it explains how birds acquire information
about resource abundance within an area, thus explaining the cause of competition
intensity at a site (Smith and Shugart 1987). Marshall and Cooper (2004) found that food
abundance fluctuated unpredictably over the course of the breeding season and among
territories, although territory sizes remained relatively constant. They also found that
territory size of red-eyed vireos (Vireo olivaceus Linnaeus) was correlated with the
foliage density at the time the territory was established. Foliage density was correlated
with food abundance at the nestling stage, demonstrating a link between foliage density,
food abundance, and territory size. This suggests that birds are using foliage density as an
10
indicator of food abundance at the most critical point in the nesting cycle (the nestling
stage). Smith and Shugart (1987) also found an inverse relationship between territory size
and expected prey abundance (based on habitat structure). They suggest a habitat quality
gradient that is defined by relationships between habitat structure and prey abundance
which influence variation in territory size. Since forest management affects foliage
density and habitat structure, it is important to understand how management will alter
resources available to birds.
Importance of the study
Much research of avian responses to forest disturbance has focused on avian
species richness and abundance (Blake 2005, Chambers et al. 1999, Davis et al. 2000,
Harrison and Kilgo 2004, Stanton 1986), but richness and abundance are not necessarily
good indicators of habitat quality since they do not necessarily reflect individual fitness
(Van Horne 1989, Vickery et al. 1992a). To understand how and why forest disturbance
affects avian ecology, the mechanisms that cause these changes must be evaluated
(Marzluff et al. 2000). Results from early studies of avian response are often inconsistent
and too simplified because of difficulties in conducting large-scale ecological research in
“true” experimental and replicated fashion (Marzluff et al. 2000).
My research addresses many of these issues because (1) it was a manipulative
experiment so that the treatment conditions were randomly assigned to selected forest
stands within logistical limitations, (2) I collected data before and after treatments so that
treatment effect could me more accurately estimated, (3) each treatment was replicated
three times so that the treatment effects could be relatively precisely estimated, and (4) I
examined both the general bird community structure and responses as well as parameters
11
such as home range size, habitat selection, and reproductive success index which are
closely related to individual bird fitness.
Bibliography
Alterman, L.E., J.C. Bednarz, and R.E. Thill. 2005. Use of group-selection and seed-tree
cuts by three early-successional migratory species in Arkansas. Wilson Bulletin
117: 353-363.
Artman, V.L. and J.F. Downhower. 2003. Wood Thrush (Hylocichla mustelina) nesting
ecology in relation to prescribed burning of mixed-oak forest in Ohio. The Auk
120: 874-882.
Artman, V.L., E.K. Sutherland, and J.F. Downhower. 2001. Prescribed burning to
restore mixed-oak communities in southern Ohio: Effects on breeding-bird
populations. Conservation Biology 15: 1423-1434.
Askins, R.A. 2001. Sustaining biological diversity in early successional communities: the
challenge of managing unpopular habitats. Wildlife Society Bulletin 29: 407-412.
Askins, R.A., J.F. Lynch, and R. Greenberg. 1990. Population declines in migratory
birds in eastern North America. Current Ornithology 7: 1-57.
Blake, J.G. 2005. Effects of prescribed burning on distribution and abundance of birds in
a closed canopy oak-dominated forest. Biological Conservation 121: 519-531.
Blake, J.G. and W.G. Hoppes. 1986. Influence of resource abundance on use of tree-fall
gaps by birds in an isolated woodlot. The Auk 103: 328-340.
Brawn, J.D., S.K. Robinson, and F.R. Thompson III. 2001. The role of disturbance in
the ecology and conservation of birds. Annu. Rev. Ecol. Syst. 32: 251-276.
Burt, W.H. 1943. Territoriality and home range concepts as applied to mammals. Journal
of Mammalogy 24: 346-352
Chambers, C.L., W.C. McComb and J.C. Tappeiner. 1999. Breeding bird responses to
three silvicultural treatments in the Oregon Coast Range. Ecological Applications
9: 171-185.
Davis, M.A, D.W. Peterson, P.B. Reich, M. Crozier, T. Query, E. Mitchell, J.
Huntington, and P. Bazakas. 2000. Restoring savanna using fire: Impact on the
breeding bird community. Restoration ecology 8: 30-40.
12
Donovan, T.M. and C.H. Flather. 2002. Relationships among North American songbird
trends, habitat fragmentation and landscape occupancy. Ecological Applications
12: 367-374.
Evans, D.R. and J.E. Gates. 1997. Cowbird selection of breeding areas: the role of
habitat and bird species abundance. Wilson Bulletin 109: 470-480.
Fretwell, S.D. and H.L. Lucas. 1970. On territorial behavior and other factors
influencing habitat distribution in birds. I. Theoretical development. Acta
Biotheoretica 19: 16-36.
Gram, W.K., P.A. Porneluzi, R.L. Clawson, J. Faaborg, and S.C. Richter. 2003. Effects
of experimental forest management on density and nesting success of bird species
in Missouri Ozark Forests. Conservation Biology 17: 1324-1337.
Greenberg, C.H., L.D. Harris, and D.G. Neary. 1995. A comparison of bird
communities in burned and salvaged-logged, clearcut, and forested Florida sand
pine scrub. Wilson Bulletin 107: 40-54.
Harrison, C.A. and J.C. Kilgo. 2004. Short-term breeding bird response to two harvest
practices in a bottomland hardwood forest. Wilson Bulletin 116: 314-323.
Harrison, R.B., F.K.A. Schmiegelow and R. Naidoo. 2005. Stand-level response of
breeding forest songbirds to multiple levels of partial-cut harvest in four boreal
forest types. Canadian Journal of Forest Resources 35: 1553-1567.
Jobes, A.P, E. Nol, and D.R. Voigt. 2004. Effects of selection cutting on bird
communities in contiguous eastern hardwood forests. Journal of Wildlife
Management 68: 51-60.
Kilgo, J. C., K.V. Miller and W.P. Smith. 1999. Effects of group-selection timber
harvests in bottomland hardwoods on fall migrant birds. Journal of Field
Ornithology 70: 404-413.
King, D.I. and R.M. DeGraaf. 2000. Bird species diversity and nesting success in
mature, clearcut, and shelterwood forests in northern New Hampshire, USA.
Forest Ecology and Management 129: 227-235.
King, D.I., R.M. DeGraaf and C.R. Griffin. 1998. Edge-related nest predation in
clearcut and groupcut stands. Conservation Biology 12: 1412-1415.
Lang, J.D., L.A. Powell, D.G. Krementz and M.J. Conroy. 2002. Wood thrush
movements and habitat use: effects of forest management for red-cockaded
woodpeckers. The Auk 119: 109-124.
13
Marshall, M.R., J.A. DeCecco, A.B. Williams, G.A. Gale, and R.J. Cooper. 2003. Use
of regenerating clearcuts by late-successional bird species and their young during
the post-fledgling period. Forest Ecology and Management 183: 127-135.
Marshall, M.R. and R.J. Cooper. 2004. Territory size of a migratory songbird in response
to caterpillar density and foliage structure. Ecology 85: 432-445.
Marzluff, J.M., M.G. Raphael and R. Sallabanks. 2000. Understanding the effects of
forest management on avian species. Wildlife Society Bulletin 28: 1132-1143.
Mazerolle, D.F. and K.A. Hobson. 2004. Territory size and overlap in male Ovenbirds:
contrasting a fragmented and contiguous boreal forest. Canadian Journal of
Zoology 82: 1774-1781.
Moorman, C.E., D.C. Guynn Jr, and J.C. Kilgo. 2002. Hooded Warbler nesting success
adjacent to group-selection and clearcut edges in a southeastern bottomland
forest. The Condor 104: 366-377.
Nagy, L.R. and R.T. Holmes. 2005. Food limits annual fecundity of a migratory
songbird: an experimental study. Ecology 86: 675-681.
Petit, L.J. and D.R Petit. 1996. Factors governing habitat selection by prothonotary
warblers: Field tests of the Fretwell-Lucas models. Ecological Monographs 66:
367-387.
Rappole, J.H. and M.V. McDonald. 1994. Cause and effect in population declines of
migratory birds. The Auk 111: 652-660.
Smith, D.M., B.C. Larson, M.J. Kelty, and P.M.S. Ashton. 1997. The practice of
silviculture: applied forest ecology. 9th edition. John Wiley & Sons, Inc., New
York, New York, USA
Smith, R. and F.R. Moore. 2005. Arrival timing and seasonal reproductive performance
in a long-distance migratory landbird. Behavioral Ecology and Sociobiology 57:
231-239.
Smith, T.M. and H.H. Shugart. 1987. Territory size variation in the Ovenbird: The role
of habitat structure. Ecology 68: 695-704.
Stanton, P.A. 1986. Comparison of avian community dynamics of burned and unburned
coastal sage scrub. The Condor 88: 285-289.
Stuart-Smith, A.K. and J.P. Hayes. 2003. Influence of residual tree density on predation
of artificial and natural songbird nests. Forest Ecology and Management 183:
159-176.
14
Tittler, R., S.J. Hannon and M.R. Norton. 2001. Residual tree retention ameliorates shortterm effects of clear-cutting on some boreal songbirds. Ecological Applications
11: 1656-1666.
Trani, M.K., R.T. Brooks, T.L. Schmidt, V.A. Rudis, C.M. Gabbard. 2001. Patterns and
trends of early successional forests in the eastern United States. Wildlife Society
Bulletin 29: 413-424.
Van Horne, B. 1989. Density is a misleading indicator of habitat quality. Journal of
Wildlife Management 47: 893-901.
Vickery, P.D., M.L. Hunter and J.V. Wells. 1992. Is density an indicator of breeding
success? The Auk 109: 706-710.
Wiens, J. A. 1989. The ecology of bird communities. Volume 1: Foundations and
Patterns. Cambridge University Press, Cambridge, UK
White, D.H., B.R. Chapman, J.H. Brunjes IV, R.V. Raftovich, Jr., J.T. Seginak. 1999.
Abundance and reproduction of songbirds in burned and unburned pine forests of
the Georgia piedmont. Journal of Field Ornithology 70: 414-424.
Whittam R. M, J.D. McCracken, C.M. Francis, and M.E. Gartshore. 2002. The effects of
selective logging on nest-site selection and productivity of hooded warblers
(Wilsonia citrina) in Canada. Canadian Journal of Zoology 80: 644-654.
Wilson, E.O. 1979. Sociobiology. Belknap, Cambridge, Massachusetts, USA.
15
CHAPTER 2
MICROCLIMATE, MICROHABITAT, AND BIRD COMMUNITY CONDITIONS
BEFORE FOREST STAND TREATMENT
Introduction
Forest ecosystems on the Cumberland Plateau provide habitat for a vast number
bird species, some of which have experienced population declines over the last four
decades (Askins et al. 1990, James et al. 1996). The US Department of Agriculture
Forest Service focus has shifted from primarily timber production in the early 1900s to
more focus on ecosystem management, biodiversity, and species viability from the late
1900s to the present (Quigley 2005). More emphasis on managing for healthy
ecosystems has heightened interest in identifying causal relationships between wildlife
and habitat. Understanding these relationships is imperative to understand declines in
bird populations. Recent research on effects of timber harvest on bird populations have
generally been short-term, unreplicated, and correlative (Sallabanks et al. 2000,
Thompson et al. 2000). It has been recommended that to be more useful, future research
should incorporate hypothesis testing, manipulative experiments, data collection on both
pre-and post-treatment conditions, and evaluate the causal relationships underlying
changes in bird communities (Anderson et al. 2001, Brawn et al. 2001, Marzluff et al.
2000, Sallabanks et al. 2000, Thompson et al . 2000). Pretreatment data is especially
16
helpful because it allows the researcher to evaluate the magnitude of change in addition
to differences among treatments.
The objective of this portion of the study was to quantify the bird community,
microhabitat characteristics, and microclimate in eighteen study plots scheduled for
silviculture treatment (treatments scheduled within the following two years). I explored
the relationships between the bird community and microhabitat, and tested the null
hypothesis that there is no difference in any of the collected variables among the eighteen
stands prior to treatment.
Study Area and Methods
Study Area
The study was located in the northern third of William B. Bankhead National
Forest (Fig. 2.1), located in Lawrence and Winston counties, northwestern Alabama.
Bankhead National Forest (BNF) is a 72,800 ha multi-use forest located in the strongly
dissected plateau subregion of the southern Cumberland Plateau (Smalley 1979). Soils
are dominated by Hartsells, Linker, Nectar, Wynnville, Albertville, and Enders soil types.
Slopes are gentle and drainage is good (Smalley 1979). The forests in this region have a
diverse species composition due to a variety of past disturbances – agriculture in the
1800s, heavy cutting and wildfire in the early 1900s, fire suppression in the last decade
and the recent large scale infestations of the southern pine beetle (Dendroctonus frontalis
Zimmerman) (Gaines and Creed 2003). In the 1930’s, abandoned farm land and other
open lands were reestablished with loblolly pine, Pinus taeda Linnaeus (Gaines and
Creed 2003). This has resulted in 31,600 ha of loblolly pine throughout BNF. Once
17
established, intensive pine plantation management was not implemented, and
subsequently, a variety of hardwood species voluntarily invaded the sites. Over the past
decade, southern pine beetle infestations have killed a major portion of loblolly pine,
increasing fuel loads and the risk of wildfires (Gaines and Creed 2003). Bankhead
National Forest has initiated a Forest Health and Restoration Project to promote healthy
forest growth via thinning and fire disturbance. The thinning and fire prescriptions were
administered to return the forest to a mixed oak-pine upland ecosystem. My research was
conducted in conjunction with BNF’s restoration project.
Pretreatment data were collected from eighteen research stands located on upland
sites composed of 20 to 35 year old loblolly pine. Stands were comprised of a minimum
of sixty percent pine (loblolly pine or Virginia pine, P. virginiana Mill.), with the
remainder mainly oak species. Average stand size was 12 ha and stands had similar age
and stand density. Pre-treatment data were collected between April and August 2005.
Figure 2.1. Location of study plots in Bankhead National Forest, AL.
18
Sampling
Microclimate. I used Hobo dataloggers (Onset Corp., Bourne, MA) to collect
microclimate data. One data logger was placed in each stand and recorded ambient
temperature (oFahrenheit) and relative humidity every four hours from May 15 – July 11,
2005. Each data logger was attached to the top of a wooden stake and protected by a 1
liter plastic container with the bottom removed to allow for access and ventilation.
Microhabitat. I performed line transect habitat surveys at the end of the breeding
season (July – August 2005) to assess the microhabitat within each stand. Placement of
three habitat plots was determined by a random compass bearing and distance (30 – 50
m) from a central point in the stand. Two 20 m perpendicular transects placed northsouth and east-west from the center of the habitat plot formed the structure for the survey.
I recorded presence or absence of the following parameters at 0.5 m intervals along each
transect: litter, bare ground, herbaceous cover, and woody cover. I measured litter depth
(to nearest mm) at the center point and at 2 m intervals along each transect. At 5 m
intervals, I recorded percent canopy cover (using a convex spherical densitometer, to
nearest percent) and the presence of each vertical forest layer. Vertical forest layers were
assigned a value of 1-4, with the following designations: 1) ground cover (< 2 m); 2)
understory (> 2 m - < 4 m); 3) mid-story (> 4 m - < 6 m); and 4) overstory (>6 m) (FIA,
1998). Forest level 4 was approximately 14 m high.
Additional forest characteristics (basal area, tree species richness, and tree
abundance) were calculated from data collected by the USDA Forest Service Southern
Research Station (provided by Callie Schweitzer). These data were collected at five 0.08
ha circular plots systematically arranged within each stand. The species and diameter at
19
breast height (DBH) of all trees greater than 14.2 cm DBH were recorded to the nearest
tenth of an inch using a diameter tape.
Bird Sampling. I sampled the bird community using line-transect surveys and
distance sampling methods (Buckland et al. 2001). Line transects were established on
each of the stands and flagged every 25 m. Each transect was 50 m from the edge of the
stand and 100 m wide; the observer slowly walked down the middle of the transect and
recorded all birds heard or seen within 50 m on either side. The observer recorded the
following: species, sex, age, the location of the bird in relation to the transect.
All stands were surveyed three times during the 2005 breeding season (15 May –
30 June) between 530 and 1030 Central Standard Time. Surveys were done in random
order and the transects walked in a different order at each visit. I conducted all surveys to
avoid observer bias.
Data Analysis
Microclimate data collected concurrently from all stands were used for
comparisons. Each 24-hour period was divided into day and night time periods (daytime
= 6:00, 10:00, 14:00; nighttime = 18:00, 22:00, 2:00), and variables included in the
analysis were mean day and night time air temperature and relative humidity. I averaged
microhabitat characteristics for the three habitat plots in each stand for comparison. I
calculated average basal area (BA) for the five tree plots using the equation 2.1. Basal
area was calculated in English measurements and then converted to metric. I inspected all
microclimate and microhabitat variables for normality visually and statistically using a
BA 
d 2
(Eq. 2.1)
(4)(144)
20
Shapiro-Wilks tests. Nighttime May air temperature, nighttime June relative humidity,
canopy cover, forest level four, litter ground cover, herbaceous ground cover, woody
ground cover, and litter depth variables were log transformed to meet normality
assumptions. I used principle components analysis (PCA, SPSS v.15.0) to group the
original variables.
To create a relative bird abundance index, I divided the number of detections by
the transect length for each stand. Stands differed in size and shape and transect lengths
differed among stands as well. I used the greatest number of individuals detected among
the three surveys to estimate the relative abundance of each species. I grouped species
into four guilds based on their migration patterns (Sauer et al. 1996, Imhof 1976), nesting
location (Ehrlich et al. 1986), foraging location (Ehrlich et al. 1986) and habitat
association (Blake and Karr 1987, Freemark and Collins 1992) (Table 2.1). To evaluate
similarity among stands, I calculated Morisita’s similarity index. Morisita’s index is
recommended as the best overall measure of similarity for ecological use (Magurran
1988). The index ranges from 0 to 1, with 0 representing pairs of sites with no species in
common and values of 1 representing complete overlap in species between sites. I used
the Shannon-Weiner diversity index, evenness, and species richness to describe the
community in each stand (Krebs 1998). To standardize species richness because transect
lengths differed among plots, I used rarefaction (Krebs 1998). I inspected all variables
for normality visually and statistically using Shapiro-Wilks tests and log transformed the
following variables to better meet normality assumptions: diversity, cavity nesting guild
abundance, foliage foraging guild abundance, and edge/open guild abundance.
21
Table 2.1. Guild memberships of songbird species detected on eighteen upland pine-hardwood stands in Bankhead National Forest,
AL, classified by: forage guild (A, aerial; F, foliage; G, ground, B, bark) (Ehrlich et al. 1986), nest location (G, ground; S, shrub; T,
tree; C, cavity) (Ehrlich et al. 1986), migratory destination (N, Neotropical migrant; T, temperate migrant; R, resident) (Sauer et al.
1996, Imhof 1976), and habitat association (O/E, open-edge; I/E, interior-edge; I, interior) (Blake and Karr 1987, Freemark and
Collins 1992).
22
Species
Code
ACFL
BAWW
BGGN
BHCB
BHVI
BLGR
BLJA
BRTH
BTGW
CACH
CARW
DOWO
ETTI
GCFL
HAWO
HOWA
INBU
KEWA
LOWA
NOCA
NOMO
NOPA
OVEN
PIWA
PIWO
Common Name
Acadian Flycatcher
Black-and-White Warbler
Blue-gray Gnatcatcher
Brown-headed Cowbird
Blue-headed Vireo
Blue Grosbeak
Blue Jay
Brown Thrasher
Black-throated Green Warbler
Carolina Chickadee
Carolina Wren
Downy Woodpecker
Eastern Tufted Titmouse
Great Crested Flycatcher
Hairy Woodpecker
Hooded Warbler
Indigo Bunting
Kentucky Warbler
Louisiana Waterthrush
Northern Cardinal
Northern Mockingbird
Northern Parula
Ovenbird
Pine Warbler
Pileated Woodpecker
Scientific Name
Empidonax virescens Vieillot
Mniotilta varia Linnaeus
Polioptila caerulea Linnaeus
Mothorous ater
Vireo solitarius Wilson
Guiraca caerulea Linnaeus
Cyanocitta cristata Linnaeus
Toxostoma rufum Linnaeus
Dendroica virens Gmelin
Poecile carolinensis Audubon
Tryothorus ludovicianus Latham
Picoides pubescens Linnaeus
Baeolophus bicolor Linnaeus
Myiarchus crinitus Linnaeus
Picoides villosus Linnaeus
Wilsonia citrina Boddaert
Passerina cyanea Linnaeus
Oporornis formosus Wilson
Seiurus motacilla Vieillot
Cardinalis cardinalis Linnaeus
Minus polygottos Linnaeus
Parula Americana Linnaeus
Seiurus aurocapillus Linnaeus
Dendroica pinus Wilson
Dryocopus pileatus Linnaeus
22
Nesting
Guild
T
G
T
P
T
S
T
S
T
C
C
C
C
C
C
S
S
G
G
S
S
T
G
T
C
Migration
Guild
N
T
T
R
N
N
R
T
N
R
R
R
N
R
N
N
N
N
R
R
T
N
R
R
T
Habitat
Guild
I
I
I/E
O/E
I/E
O/E
I/E
O/E
I
I/E
O/E
I/E
I/E
I/E
I
I
O/E
I/E
I/E
I/E
E
I/E
I
I/E
I
Forage
Guild
A
B
F
G
F
G
G
G
F
F
G
B
F
A
B
F
F
G
G
G
G
F
F
B
B
23
Species
Code
PRWA
RBWO
REVI
SCTA
SUTA
WBNU
WEVI
WEWA
WOTH
YBCH
YBCU
Common Name
Prairie Warbler
Red-bellied Woodpecker
Red-eyed Vireo
Scarlet Tanager
Summer Tanager
White-breasted Nuthatch
White-eyed Vireo
Worm-eating Warbler
Wood Thrush
Yellow-breasted Chat
Yellow-billed Cuckoo
Scientific Name
Dendroica discolor Vieillot
Melanerpes carolinus Linnaeus
Vireo olivaceus Linnaeus
Piranga olivacea Gmelin
Piranga ruba Linnaeus
Sitta carolinensis Latham
Vireo griseus Boddaert
Helmitheros vermivorus Gmelin
Hylocichla mustelina Gmelin
Icteria virens Linnaeus
Coccyzus americanus Linnaeus
23
Nesting
Guild
S
C
S
T
T
C
S
G
T
S
T
Migration
Guild
N
R
N
N
R
T
N
N
N
N
N
Habitat
Guild
O/E
I/E
I/E
I
I/E
I
O/E
I
I/E
O/E
I/E
Forage
Guild
F
B
F
F
F
B
F
F
G
F
F
I used analysis of variance (ANOVA) with treatment and block as main factors to
test for differences among stands in bird community, microclimate, and microhabitat
PCs. Tukey’s multiple comparisons test was performed based on results of the ANOVA.
All analyses were performed in SPSS (v.15.0) using an alpha level of 0.05.
To investigate variation in abundance of species and guilds as they relate to
microhabitat measures, I used canonical correspondence analysis (CCA, CANOCO
v.4.5). I used only species that had greater than five detections in the analysis so that rare
species would not bias the results. This procedure is a direct gradient analysis technique
that compares community composition directly to environmental variables across a
gradient (Palmer 1993). CCA is appropriate to use when there are no differences among
stands because it evaluates gradients on a different scale; it examines the trends and
variability within stands.
Results
Microclimate. There were no differences in microclimate characteristics among
stands (p > 0.05). Temperature and relative humidity increased over the season on all the
plots (Table 2.2)
The original twelve variables showed low multivariate correlation (Kaiser-MeyerOlkin [KMO] Measure of Sampling Adequacy = 0.355) so the three variables with the
lowest multivariate correlation (nighttime July temperature, nighttime July relative
humidity, and nighttime May temperature) were removed from the PCA to increase the
KMO measure of sampling adequacy to 0.55. The remaining 9 variables were grouped
into 3 principle components (PCs), the first representing daytime climate, the second
24
Table 2.2. Mean microclimate and standard error for microclimate characteristics in
upland pine-hardwood stands before treatment in Bankhead National Forest, AL, 2005.
Month and Time
Degrees Celsius
Relative Humidity
May Day
20.6 ± 0.6
65.1 ± 2.1
May Night
17.9 ± 0.5
73.1 ± 2.5
June Day
23.5 ± 0.5
76.1 ± 2.1
June Night
21.2 ± 0.4
84.1 ± 2.4
July Day
24.2 ± 0.9
80.1 ± 2.7
July Night
22.1 ± 0.8
86.7 ± 1.6
25
representing nighttime climate, and the third representing temperature (Table 2.3). The 3
components retained approximately 92 % of the original variation (Bartlett’s Test of
Sphericity χ2 = 195.209, df = 36, p = 0.000).
Microhabitat. There was no difference in microhabitat among the stands (p > 0.05).
Plots were characterized by high litter ground cover, high percent canopy cover, and a
high presence of forest level four (Table 2.4).
The original twelve variables were grouped into 4 principle components (PCs), the
first representing understory cover and forest structure, the second representing canopy
cover, the third representing BA, and the fourth representing litter depth (Table 2.5). The
4 components retained approximately 79% of the original variation (Bartlett’s Test of
Sphericity χ2 = 103.832, df = 66, p = 0.002).
Bird Community. A total of 1185 birds were detected, representing 35 species (Table
2.1). The most abundant species were the red-eyed vireo (Vireo olivaceus Linnaeus),
comprising 20.9% (249 detections) of total individuals, and the pine warbler (Dendroica
pinus Wilson), comprising 11.6% (138 detections) of total individuals. The majority of
the community comprised bids in the following guilds: shrub nesting, Neotropical
migrant, and interior/edge habitat specialist (Table 2.6). There were significantly more
aerial foraging birds on the plots to be burned than those to be thinned and more
interior/edge. Morisita’s similarity indices indicate that species composition on all stands
were similar to one another (Table 2.8).
The bird community included twelve species that are listed in Partners in Flight’s
(PIF) North American Landbird Conservation Plan as Species of Continental Importance
(Rich et al. 2004, Table 2.7). PIF lists species in two categories; WatchList species are
26
Table 2.3. Principle component analysis loadings, eigenvalues, and percent variance
for microclimate variables on uplandpine-hardwood stands in Bankhead National Forest,
AL, 2005.
PC1
PC2
PC3
Daytime May Temperature
0.907
Daytime June Temperature
0.923
Daytime July Temperature
0.604
Daytime May RelHumdity
-0.838
0.372
0.332
Daytime June RelHumdity
-0.662
0.592
0.339
Daytime July RelHumdity
-0.904
Nighttime June Temperature
-0.603
0.754
Nighttime May RelHumdity
0.946
0.727
Nighttime June RelHumdity
0.346
0.921
Eigenvalue
4.18
2.66
1.43
% Variance
46.49
29.56
15.87
27
Table 2.4. Means for microhabitat characteristics on upland pine-hardwood stands in
Bankhead National Forest, AL, 2005.
Microhabitat Characteristic
Mean ± SE
Ground cover: Percent litter
99.4 ± 0.7
Ground cover: Percent herbaceous plants
16.9 ± 13.2
Ground cover: percent woody plants
11.0 ± 7.6
Litter depth (cm)
6.3 ± 1.0
Percent canopy cover
82.5 ± 6.8
Forest level 1 present
20.7 ± 5.8
Forest level 2 present
20.7 ± 3.3
Forest level 3 present
17.2 ± 6.4
Forest level 4 present
24.3 ± 18.5
28
Table 2.5. Principle component analysis loadings, eigenvalues, and percent
variance for microhabitat variables on upland pine-hardwood stands in
Bankheand National Forest, AL, 2005.
PC1
PC2
PC3
PC4
Herb Cover
0.885
Woody cover
0.857
Canopy Cover
0.663
Litter Cover
-0.617
0.788
Litter Depth
-0.379
Forest Level 1
0.819
Forest Level 2
Forest Level 3
0.337
-0.376
Forest Level 4
Basal Area
-0.439
Tree Species Richness
0.760
Tree Abundance
-0.859
-0.381
0.688
0.528
0.539
-0.772
0.336
0.406
0.780
-0.743
0.380
Eigenvalue
4.50
2.25
1.51
1.18
% Variance
37.50
18.73
12.62
9.86
29
Table 2.6. Significance value, mean, and standard error by proposed treatment for bird
community on upland pine-hardwood stands in Bankhead National Forest, AL, 2005.
Superscript letters indicate differences among means in the same row (Tukey test p
<0.05).
Means by treatment
Community
Variable
To be
To be
To be
p
Control
burned
thinned
thinned/burned
0.24
12.45±0.39
12.39±1.66
13.30±0.63
14.54±0.56
abundance
0.24
99.57±3.16
99.08±13.28
106.38±5.05
116.34±4.47
Div. index
0.32
2.45±0.53
2.39±0.23
2.55±0.04
2.64±0.05
Evenness
0.06
0.90±0.02
0.87±0.03
0.95±0.10
0.92±0.01
Tree
0.06
11.29±2.73
11.85±1.64
11.13±1.65
17.41±1.43
Shrub
0.31
10.65±3.19
13.73±2.90
16.04±2.81
17.51±2.34
Ground
0.82
5.91±2.15
5.26±2.44
7.81±1.95
6.30±1.42
Parasite
0.29
0±0
0±0
0.17±0.17
0.50±0.22
Cavity
0.07
6.85±1.33
5.05±1.58
7.32±1.07
12.22±2.49
Foliage
0.05
24.35±3.43
22.78±3.75
27.05±4.83
36.87±5.05
Aerial
0.02
1.82±0.57ab
2.16±1.38a
0.75±0.15b
2.90±0.20ab
Ground
0.45
3.11±0.89
4.38±1.73
4.36±0.98
5.79±0.73
Bark
0.65
6.80±2.17
6.57±1.33
8.17±1.18
8.72±1.38
Neotropical
0.07
19.80±2.60
21.49±4.34
24.78±3.99
31.07±2.78
Temperate
0.95
3.95±0.91
3.10±1.58
3.26±0.71
19.83±2.09
Resident
0.06
12.05±2.64
11.30±2.13
14.41±1.96
19.83±2.09
0.55
12.33±3.15
9.30±3.70
15.03±4.34
16.10±2.28
Interior/edge
0.03
17.74±2.84
a
Edge/open
0.24
2.83±1.26
Species rich.
Relative
Nesting Guild
Foraging Guild
Migration Guild
Habitat Guild
Interior
23.80±3.38
ab
2.84±1.14
30
22.34±2.25
ab
4.47±0.64
30.72±2.44b
6.90±2.21
Table 2.7. Species of Continental Importance, as listed by the Partners in Flight
Landbird Conservation Plan (Rich et al. 2004), detected on eighteen upland pinehardwood stands in Bankhead National Forest, AL, 2005. See Table 2.1 for scientific
names.
Species
PIF Listing* Individuals
Plots
Acadian Flycatcher
S
7
8
Brown Thrasher
S
1
1
Carolina Wren
S
15
11
Hooded Warbler
S
52
13
Indigo Bunting
S
38
16
Kentucky Warbler
W
6
6
Louisiana Waterthrush
S
1
1
Pine Warbler
S
80
18
Prairie Warbler
W
5
5
White-eyed Vireo
S
5
4
Worm-eating Warbler
W
53
15
Wood Thrush
W
2
1
* S = Stewardship Species, W = WatchList Species (Rich et al. 2004)
31
Table 2.8. Morisita’s similarity index for the breeding bird community on eighteen
upland pine-hardwood stands in Bankhead National Forest, AL, 2005.
To be
To be
To be
treatment
control
burned
thinned
thinned/burned
--
0.97
1.04
1.06
To be Burned
0.97
--
1
1
To be thinned
1.04
1
--
1.04
To be thinned/burned
1.06
1
1.04
--
Control
32
species that have multiple reasons (restricted distribution, low population size,
widespread population declines, high threats to habitat, etc.) for conservation concern the
variation in the first three axes. Axis one explained 33.1% (Eigenvalue = 0.023), the
second axis 20.1% (Eigenvalue = 0.014), and the third 14.8% (Eigenvalue = 0.010).
The CCA of microhabitat characteristics and species abundance revealed a gradient
of microhabitat characteristics evident by the position of variables along the axes of the
ordination plots (Fig. 2.2). On one end of the gradient is canopy cover and tree
abundance, on the other is percent woody ground cover, percent herbaceous ground
cover, presence of forest level one, and tree species richness. The Kentucky warbler
(Oporornis formosus Wilson), hooded warbler (Wilsonia citrina Boddaert), worm-eating
warbler (Helmitheros vermivorus Gmelin) were strongly associated with the presence of
woody and herbaceous ground cover (Fig. 2.2). The blue jay (Cyanocitta cristata
Linnaeus) was associated with tree abundance (Fig. 2.2), the white-eyed vireo (Vireo
griseus Boddaert) with the presence of forest level one (Fig. 2.2), and the Carolina wren
(Tryothorus ladovicianus Latham) with litter depth (Fig. 2.2).
The CCA of microhabitat characteristics and nesting guild associations revealed a
similar gradient, ranging from canopy cover and presence of forest level three on one end
of the gradient to herbaceous ground cover, woody ground cover and presence of forest
level one on the other (Fig. 2.3). A second gradient was also evident, one end with basal
area, litter depth, and presence of forest level four and at the other end, presence of forest
level two (Fig. 2.3). The nesting guilds were evenly distributed across the ordination.
The shrub nesting guild was associated with the presence of forest level one and two, and
a high percentage of woody and herbaceous ground cover (Fig. 2.3). Ground nesting
33
.
Figure 2.2. First and second canonical correspondence axes for bird species and
microhabitat variables, Bankhead National Forest, AL. See table 2.1 for species code
explanations
34
.
Figure 2.3. First and second canonical correspondence axes for nesting guild associations and
microhabitat variables, Bankhead National Forest, AL, 2005
35
species were also associated with a high percentage of woody and herbaceous ground
cover, and the presence of forest level two (Fig. 2.3). Tree nesting species were
associated with high canopy cover and the presence of forest level three (Fig. 2.3).
Cavity nesting species appear to be associated with basal area and litter depth (Fig. 2.3).
The CCA of microhabitat characteristics and foraging guild associations revealed a
gradient from understory to overstory cover. On one end of the gradient was the presence
of forest level one, two, and four, tree species richness, and woody and herbaceous
ground cover; on the other end was canopy cover, litter ground cover, litter depth, and the
presence of forest level three (Fig. 2.4). In the center of the plot was foliage foraging
species, indicating that they are more generalist species. On the far edge of the plot was
aerial feeding species, associated with BA (Fig. 2.4). Bark foraging species showed an
association with the presence of forest level three and ground foraging species showed an
association with canopy cover and litter ground cover (Fig. 2.4).
Discussion
The overall structure of the bird and forest community before treatment appears to
be a mid-successional forest. Forest structure was defined by high percentage of canopy
cover and also a high presence of understory vegetation. The bird community consisted
of a majority of shrub nesting species and interior/edge dwelling species. Optimal habitat
for these guilds was created by the presence of wildlife openings, roads, and southern
pine beetle damaged areas within many of the plots, which create small pockets of open
areas and increase the amount of edge. All stands were similar to one another in terms of
microhabitat and microclimate characteristics, although there were differences in some
36
Figure 2.4. First and second canonical correspondence axes for foraging guild associations and
microhabitat variables, Bankhead National Forest, AL, 2005.
37
aspects of the bird community among the stands. Knowing these differences will give
insight when evaluating post-treatment data.
The bird community included no federally listed species; however twelve species
of continental importance were detected on the stands, including four species listed on the
PIF WatchList. These are relatively abundant species that have multiple reasons for
conservation concern (Rich et al. 2004). Many of the species of concern found on the
stamds are disturbance-dependent shrub- or ground-nesting species (brown thrasher
[Toxostoma rufum Linnaeus], hooded warbler, indigo bunting [Passerina cyanea
Linnaeus], Kentucky warbler, prairie warbler [Dendroica virens Vieillot], and white-eyed
vireo), which represents the major habitat type of conservation concern in the eastern US
(Rich et al. 2004).
Bird abundance and nesting guild abundance were related to the gradient of
microhabitat characteristics occurring on the plots. The distribution of birds along this
gradient corresponds with what is known about the natural history of the animals.
Species that forage and nest in the understory (Kentucky warbler, hooded warbler, wormeating warbler) were strongly associated with the presence of woody and herbaceous
ground cover. The presence of ground cover would provide both cover and nesting
habitat for theses species. The white-eyed vireo, an edge-associated species, was
associated with the presence of forest level one. Forest level one was part of the gradient
from high canopy cover percentage to low canopy cover percentage, so an indication of
the presence of forest level one also indicates a lower amount of canopy cover. The
ordination plot shows an association between blue jays and tree abundance, which may
just be a random occurrence due to the low number of detections (n=9).
38
The distribution of nesting guild abundance along the gradients indicates a
distinct separation of habitat suitable for each guild. Ground and shrub nesting species
were associated with high percent of ground cover and the presence of forest level two,
which would be the primary source of cover and nesting habitat for these species. Tree
nesting species were associated with high canopy cover and the presence of forest level
three, again this would be the primary source of cover and nesting habitat for these
species. The association between cavity nesting species and litter depth and basal area
does not make sense. Cavity nesting guild species are made up of both primary (i.e.,
woodpeckers) and secondary (i.e., eastern tufted titmouse [Baeolophous bicolor
Linnaeus], Carolina wren) cavity nesters. There could be a relationship between the
volume of trees (basal area) and the number of snags present or the number of cavities
available for secondary nesting species; however none of the microhabitat characteristics
that we measured appears to explain the variation in the abundance of cavity nesting
species.
Bibliography
Anderson, D.R., W.A. Link, D.H. Johnson, K.P. Burnham. 2001. Suggestions for
presenting the results of data analyses. Journal of Wildlife Management 65: 373378.
Askins, R.A., J.F. Lynch, and R. Greenberg. 1990. Population declines in migratory
birds in eastern North America. Current Ornithology 7:1-57.
Blake, J.G. and J.R. Karr. 1987. Breeding birds of isolated woodlots: area and habitat
relationships. Ecology 68:1724-1734.
Brawn, J.D., S.K. Robinson, and F.R. Thompson III. 2001. The role of disturbance in
the ecology and conservation of birds. Annual Review of Ecological Systems 32:
251-276.
39
Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L.
Thomas. 2001. Introduction to distance sampling. Oxford University Press,
Oxford.
Ehrlich, P.R., D.S. Dobkin, and D. Wheye. 1988. The birder’s handbook: a field guide
to the natural history of North American birds. Simon and Schuster, New York,
NY. 785 pp.
FIA. 1998. Field instructions for southern forest inventory. Remeasurement of prism
plots. Southern Research Station, Forest Service, U.S. Department of Agriculture.
Item 26 version of manual.
Freemark, K. and B. Collins. 1992. Landscape ecology of birds breeding in temperate
forest fragments. Pp. 443-454 in Ecology and Conservation of Neotropical Migrant
Landbirds (J.M. Hagen III and D.W. Johnston, eds.). Smithsonian Institution Press,
Washington, D.C.
Gaines, G.D. and J.W. Creed. 2003. Forest health and restoration project. National
forests in Alabama, Bankhead National Forest Franklin, Lawrence and Winston
Counties, Alabama. Final environmental impact statement. Management Bulletin
R8-MB 110B.
Imhof, T.A. 1976. Alabama birds. The University of Alabama Press, University, AL.
445 pp.
James, F.C., C.E. McCullough, and D.A. Wiedenfeld. 1996. New approaches to the
analysis of population trends in land birds. Ecology 77: 13-27.
Krebs, C.J. 1998. Ecological methodology. Addison Wesley Longman, Menlo Park,
CA. 620 pp.
Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University
Press, Princeton, NJ. 179pp.
Marzluff, J.M., M.G. Raphael, and R. Sallabanks. 2000. Understanding the effects of
forest management on avian species. Wildlife Society Bulletin 28: 1132-1143.
Palmer, M.W. 1993. Putting things in even better order: the advantages of canonical
correspondence analysis. Ecology 74: 2215-2230.
Quigley, T.M. 2005. Evolving views of public land values and management of natural
resources. Rangelands 27: 37-44.
Rich, T. D., C. J. Beardmore, H. Berlanga, P. J. Blancher, M. S. W. Bradstreet, G. S.
Butcher, D. W. Demarest, E. H. Dunn, W. C. Hunter, E. E. Iñigo-Elias, J. A.
40
Kennedy, A. M. Martell, A. O. Panjabi, D. N. Pashley, K. V. Rosenberg, C. M.
Rustay, J. S. Wendt, T. C. Will. 2004. Partners in Flight North American Landbird
Conservation Plan. Cornell Lab of Ornithology. Ithaca, NY.
Sallabanks, R., E.B. Arnett, and J.M. Marzluff. 2000. An evaluation of research on the
effects of timber harvest on bird populations. Wildlife Society Bulletin 28: 11441155.
Sauer, J.R., S. Schwartz, and B. Hoover. 1996. The Christmas Bird Count Home Page.
Version 95.1, http://www.mbr-pwrc.usgs.gov/bbs/cbc.html. USGS Patuxent
Wildlife Research Center, Laurel, MD.
Smalley, G.W. 1979. Classification and evaluation of forest sites on the Southern
Cumberland Plateau: U.S. Department of Agriculture, Forest Service, Southern
Forest Experiment Station, general technical report SO-23.
Thompson, F.R., III, J.D. Brawn, S. Robinson, J. Faaborg, R.L. Clawson. 2000.
Approaches to investigate effects of forest management on birds in eastern
deciduous forests: How reliable is our knowledge? Wildlife Society Bulletin 28:
1111-1122.
41
CHAPTER 3
MICROHABITAT, MICROCLIMATE, ARTHROPOD AVAILABILITY, AND BIRD
COMMUNITY IN FOREST STANDS TREATED WITH BURNING AND THINNING
Introduction
The decline of Neotropical migratory songbirds in eastern North America has
been a subject of much discussion among ornithologists over the past two decades
(Askins et al. 1990, Finch 1991, James et al. 1996, Rappole and McDonald 1994,
Robbins et al. 1989). Although some evidence of declines is conflicting, it is generally
accepted that due to general trends of habitat loss or degradation, and their importance to
the ecosystems, giving priority conservation status to Neotropical songbirds is justified.
In recent studies, the decline of birds associated with early successional breeding habitat
has been noted (Askins et al. 1990, Hunter et al. 2001, Litvaitis et al. 1999). Trani et al.
(2001) reported that, according to Forest Inventory and Analysis (FIA) data, young forest
habitats are declining due to forest maturation and the absence of timber removal on
much public land. Tree removal creates early successional habitat by removing trees to
create an environment favorable for tree growth or regeneration (Smith et al. 1997). As
forest management evolves to employ multiple silvicultural tools to meet a myriad of
objectives, it is important to understand how such management affects the bird
community and if quality early successional wildlife habitat is produced.
42
Prescribed burning has garnered heightened awareness on public lands as a
silvicultural technique since fire suppression in eastern forests has been questioned
(Brose et al. 2001, Van Lear and Waldrop 1989). Although the effect of silviculture and
fire on birds has been studied individually in eastern forests, there is little research
assessing the effect of thinning and prescribed burning (Greenberg et al. 1995, Greenberg
et al. 2007) and only one study reports the effects when tree reduction and burning are
combined (Wilson et al. 1995). It is important to understand how these treatments will
affect the bird community when compared to other silvicultural techniques.
The objective of this portion of the study was to quantify the bird community,
microhabitat characteristics, microclimate, and arthropod availability on six silvicultural
treatments in the William B. Bankhead National Forest. I examined the change in
microhabitat and microclimate features after implementation of the silvicultural
treatments, explored the relationships between the bird community; and the structure and
features of microhabitat, microclimate, and arthropod availability and tested the null
hypothesis that there is no difference in any of the collected variables among the
treatments.
Study Area and Methods
Study Area
The study was located in the northern third of William B. Bankhead National
Forest (Fig 3.1), located in Lawrence and Winston counties, northwestern Alabama.
Bankhead National Forest (BNF) is a 72,800 ha multi-use forest located in the Strongly
Dissected Plateau subregion of the southern Cumberland Plateau (Smalley 1979). Soils
43
are dominated by Hartsells, Linker, Nectar, Wynnville, Albertville, and Enders soil types.
Slopes are gentle and drainage is good (Smalley 1979). The forests in this region have a
diverse species composition due to a variety of past disturbances – agriculture in the
1800s, heavy cutting and wildfire in the early 1900s, fire suppression in the last decade
and the recent infestation of the southern pine beetle (Dendroctonus frontalis
Zimmerman) (Gaines and Creed 2003). In the 1930’s, abandoned farm land and other
open lands were reestablished with loblolly pine, Pinus taeda Linnaeus (Gaines and
Creed 2003). This has resulted in 31,600 ha of loblolly pine throughout BNF. Once
established, intensive pine plantation management was not implemented, and
subsequently, a variety of hardwood species voluntarily invaded these sites. Over the
past decade, southern pine beetle infestations have killed a major portion of loblolly pine,
increasing fuel loads and the risk of wildfires (Gaines and Creed 2003). Bankhead
National Forest has initiated a Forest Health and Restoration Project to promote healthy
forest growth via thinning and fire disturbance. The thinning and fire prescriptions were
administered to return the forest to a more healthy state and to promote regeneration of
native species. My research was conducted in conjunction with BNF’s restoration
project.
The study design consisted of a randomized complete block design with two
factors – three thinning levels (no thin, 11 m2 ha-1 residual basal area, and 17 m2 ha-1
residual basal area) and two burn treatments (no burn and burn). Each treatment was
replicated three times and blocked by year. Treatments were assigned randomly to
44
Figure 3.1. Location of study plots in Bankhead National
Forest, Alabama.
Burning treatment
Burn
No Burn
Control
3
3
Thin
6
6
Figure 3.2. Experimental design: two-factor, randomized
complete block design. Treatments include two burn
treatments and two thinning levels.
45
delineated stands. After the treatments were completed, I collapsed the thinned treatments
together because there was no difference in basal area between the two thinning levels (F
= 0.07, df = 1, p = 0.8). Although this was not the response variable I was studying, the
variability within individual treatment levels was uneven, with some stands within the
same treatment level having greater BA than the target and others having lower BA than
the target. This variation was too large to detect any difference between the thinning
levels. This created three replicates each of the control and burn, and six replicates each
of the thin and the thin/burn (Fig. 3.2). The research stands were located on upland sites
composed of 20 to 35 year old loblolly pine. Stands were comprised of a minimum of
sixty percent pine (loblolly pine or Virginia pine, P. virginiana Mill.), with the remainder
mainly oak species (Quercus spp.). Average stand size was 12 ha and plots had similar
age and stand density. Thinning favored the retention of hardwood species and was done
before fire prescriptions. Prescribed burning was completed in the dormant season
(January – March) with low-burning surface fires. Treatments on block one were
completed between August 2005 and 1 February 2006; blocks two and three were treated
between April 2006 and March 2007. Post- treatment data was collected from block one
between April and August 2006, and from blocks two and three between April and
August 2007. Post treatment data was collected in two different years, but because
treatments were blocked by year, any differences in year would be detected in the block
factor.
Sampling
46
Microclimate. Microclimate data were collected with Hobo dataloggers (Onset
Corp., Bourne, MA). One data logger was placed in each stand and recorded ambient
temperature (to tenth of a degree Centigrade [C]) and relative humidity every four hours
from May 15 – June 13. Each data logger was attached to the top of a wooden stake and
covered by a 1 liter plastic container with the bottom removed to allow for access and
ventilation.
Microhabitat. I performed line transect habitat surveys at the end of the breeding
season (July – August) to assess the microhabitat within each stand. Placement of three
habitat plots was determined during pre-treatment data collection by a random compass
bearing and distance (30 – 50 m) from a central point in the stand. The central point was
marked with a metal stake during pre-treatment data collection and the same distance and
compass bearing were used to locate post-treatment habitat plots. Two 20 m
perpendicular transects placed north-south and east-west from the center of the habitat
plot formed the structure for the survey. I recorded presence or absence of the following
parameters at 0.5 m intervals along each transect: litter, bare ground, herbaceous cover,
and woody cover. I measured litter depth (to nearest mm) at the center point and at 2 m
intervals along each transect. At 5 m intervals, I recorded percent canopy cover (using a
convex spherical densitometer, to nearest percent) and the presence of each vertical forest
layer. I assigned vertical forest layers a value of 1-4, with the following designations: 1)
ground cover (< 2 m); 2) understory (> 2 m - < 4 m); 3) mid-story (> 4 m - < 6 m); and 4)
overstory (>6 m) (FIA, 1998). I also recorded basal area (BA) at the center of each
habitat plot using a 10 factor basal area prism.
47
I calculated additional forest characteristics (basal area, tree species richness, and
tree abundance) from data collected by the USDA Forest Service Southern Research
Station (provided by Callie Schweitzer). These data were collected at five 0.08 ha
circular plots systematically arranged within each stand. The species and diameter at
breast height (DBH) of all trees greater than 14.2 cm DBH were recorded to the nearest
tenth of an inch using a diameter tape. All plots were marked with PVC pipe and trees
number tagged during pre- treatment data collection. The same plots and trees were resampled post-treatment.
Arthropod availability. To sample the arthropod abundance in each stand, I used
the branch clipping method (Johnson 2000). Samples were collected at 50 m intervals
along each bird transect survey. A branch clip (approximately 25 cm), included the
terminal leaf cluster and was collected from either an oak or red maple (Quercus spp. and
Acer rubrum), alternating species at each sampling point. These two species were
selected because they are common hardwood species in the stand and commonly used for
foraging by songbirds. Each branch was randomly clipped from either 0-3 m or 3-6 m.
Each branch clip was collected using pruning sheers and collected in a white plastic
garbage bag. Once inside the bag, the leaves were sprayed with an insecticide to kill all
insects. After a minimum of 5 h, the leaves were removed from the bag, all arthropods
extracted, identified to Order, and lengths were measured to the nearest 0.05 cm. Each
branch was de-leafed, the leaves were air dried in paper bags, and the dry weight
recorded. Each stand was sampled monthly throughout the breeding season (April - July)
between 1200 h and 1600 h.
48
Bird Sampling. I sampled the bird community using line-transect surveys and
distance sampling methods (Buckland et al. 2001). Line transects were established on
each of the stands and flagged every 25 m. Each transect was 50 m from the edge of the
stand and 100 m wide; the observer slowly walked down the middle of the transect and
recorded all birds heard or seen within 50 m on either side. The observer recorded the
following: species, sex, age, the location of the bird in relation to the transect.
All stands were surveyed three times during the breeding season (15 May – 30 June)
between 530 and 1030 Central Daylight Savings Time. Surveys were done in random
order and the transects walked in a different order at each visit. I conducted all surveys to
avoid observer bias.
Data Analysis
Microclimate data collected concurrently from all stands were used for comparisons.
Each 24-hour period was divided into day and night time periods (daytime = 6:00, 10:00,
14:00; nighttime = 18:00, 22:00, 2:00), and variables included in the analysis were mean
day and night time air temperature and relative humidity. Data was lost (due to computer
crash) from three stands in block one. In these cases, the average from the remaining two
blocks was substituted for the lost data. I averaged microhabitat characteristics for the
three habitat plots in each stand for comparison. I calculated an average basal area for the
five tree plots using equation 3.1. Basal area was calculated in English measurements and
then converted to metric. I inspected all microclimate and microhabitat variables for
BA 
d 2
(4)(144)
49
(Eq.3.1)
normality visually and statistically using a Shapiro-Wilks test. Daytime May relative
humidity and bare ground cover were square root transformed, litter cover was arcsine
transformed, and nighttime June temperature and tree species richness were log
transformed to meet normality assumptions. I used principle components analysis (PCA,
SPSS v. 15.0) to group the original variables.
Arthropod biomass was estimated based on length using regression models (SPSS v.
15.0) (Ganihar 1997, Sample et al. 1993). Ganihar (1997) calculated beta coefficients for
each arthropod Order; I used these coefficients as well as the recommended model to
predict total biomass per sample (by stand and month). I calculated relative biomass
index by dividing the total biomass by dry leaf weight.
To create a relative bird abundance index, I divided the number of detections by the
transect length for each stand. Stands differed in size and shape and transect lengths
differed among stands as well. I used the greatest number of individuals detected among
the three surveys to estimate the relative abundance of each species. I grouped species
into four guilds based on their migration patterns (Sauer et al. 1996, Imhof 1976), nesting
location (Ehrlich et al. 1986), foraging location (Ehrlich et al. 1986), and habitat
50
Table 3.1. Guild memberships of all songbird species detected on eighteen upland pine-hardwood plots one year after forest treatment
in Bankhead National Forest, AL, classified by: forage guild (A, aerial; F, foliage; G, ground; N, nectar; B, bark) (Ehrlich et al. 1986),
nest location (G, ground; S, shrub; T, tree; C, cavity) (Ehrlich et al. 1986), migratory destination (N, Neotropical migrant; T, temperate
migrant; R, resident) (Sauer et al. 1996, Imhof 1976), and habitat association (O/E, open-edge; I/E, interior-edge; I, interior) (Blake
and Karr 1987, Freemark and Collins 1992).
51
Species
Code
ACFL
BAWW
BGGN
BHCB
BHNU
BHVI
BLJA
BRTH
BTGW
CACH
CARW
DOWO
EAPH
EATO
EAWP
ETTI
GCFL
HAWO
HOWA
INBU
KEWA
LOWA
Common Name
Acadian Flycatcher
Black-and-White Warbler
Blue-gray Gnatcatcher
Brown-headed Cowbird
Brown-headed Nuthatch
Blue-headed Vireo
Blue Jay
Brown Thrasher
Black-throated Green Warbler
Carolina Chickadee
Carolina Wren
Downy Woodpecker
Eastern Phoebe
Eastern Towhee
Eastern Wood-pewee
Eastern Tufted Titmouse
Great Crested Flycatcher
Hairy Woodpecker
Hooded Warbler
Indigo Bunting
Kentucky Warbler
Louisiana Waterthrush
Scientific Name
Empidonax virescens Vieillot
Mniotilta varia Linnaeus
Polioptila caerulea Linnaeus
Molothrus ater Boddaert
Sitta pusilla Latham
Vireo solitarius Wilson
Cyanocitta cristata Linnaeus
Toxostoma rufum Linnaeus
Dendroica virens Gmelin
Poecile carolinensis Audubon
Tryothorus ludovicianus Latham
Picoides pubescens Linnaeus
Sayornis phoebe Latham
Pipilo erythrophthalmus Linnaeus
Contopus virens Linnaeus
Baeolophus bicolor Linnaeus
Myiarchus crinitus Linnaeus
Picoides villosus Linnaeus
Wilsonia citrina Boddaert
Passerina cyanea Linnaeus
Oporornis formosus Wilson
Seiurus motacilla Vieillot
51
Forage
Guild
A
B
F
G
B
F
F
G
F
F
G
B
C
G
A
F
A
B
F
F
G
G
Nest
Guild
T
G
T
P
C
T
T
S
T
C
C
C
C
G
T
C
C
C
S
S
G
G
Migration
Guild
N
T
T
R
R
N
R
T
N
R
R
R
R
N
R
N
R
N
N
N
NR
R
Habitat
Guild
I
I
I/E
O/E
I
I/E
I/E
O/E
I
I/E
O/E
I/E
I/E
I/E
I/E
I/E
I/E
I
I
O/E
I/E
I/E
52
Species
Code
MODO
NOCA
NOPA
OVEN
PIWA
PIWO
PRWA
REVI
RTHU
SCTA
SUTA
WBNU
WEVI
WEWA
WOTH
YBCH
YBCU
YTVI
Common Name
Mourning Dove
Northern Cardinal
Northern Parula
Ovenbird
Pine Warbler
Pileated Woodpecker
Prairie Warbler
Red-eyed Vireo
Ruby-throated Hummingbird
Scarlet Tanager
Summer Tanager
White-breasted Nuthatch
White-eyed Vireo
Worm-eating Warbler
Wood Thrush
Yellow-breasted Chat
Yellow-billed Cuckoo
Yellow-throated Vireo
Scientific Name
Lenaida macroura Linnaeus
Cardinalis cardinalis Linnaeus
Parula Americana Linnaeus
Seiurus aurocapillus Linnaeus
Dendroica pinus Wilson
Dryocopus pileatus Linnaeus
Dendroica discolor Vieillot
Vireo olivaceus Linnaeus
Archilochus coulbris Linnaeus
Piranga olivacea Gmelin
Piranga ruba Linnaeus
Sitta carolinensis Latham
Vireo griseus Boddaert
Helmitheros vermivorus Gmelin
Hylocichla mustelina Gmelin
Icteria virens Linnaeus
Coccyzus americanus Linnaeus
Vireo flavifrons Vieillot
52
Forage
Guild
G
G
F
F
B
B
F
F
N
F
F
B
F
F
G
F
F
F
Nest
Guild
T
S
T
G
T
C
S
S
T
T
T
C
S
G
T
S
T
T
Migration
Guild
N
R
N
R
R
T
N
N
N
N
R
T
N
N
N
N
N
N
Habitat
Guild
O/E
I/E
I/E
I
I/E
I
O/E
I/E
O/E
I
I/E
I
O/E
I
I/E
O/E
I/E
I/E
association (Blake and Karr 1987, Freemark and Collins 1992) (Table 3.1). To evaluate
similarity among the stands and across years, I calculated Morisita’s similarity index
(Magurran 1988). Morisita’s index is recommended as the best overall measure of
similarity for ecological use (Magurran 1988). The index ranges from 0 to 1, with 0
representing pairs of sites with no species in common and values of 1 representing
complete overlap in sites. I used the Shannon-Weiner diversity index, evenness, and
species richness to describe the community in each stand (Krebs 1998). To standardize
species richness because transect lengths differed among plots, I used rarefaction (Krebs
1998). I inspected all variables for normality visually and statistically using ShapiroWilks tests and all variables met assumptions.
I used two-way analysis of variance (ANOVA) with thin, burn, and block as
factors to test for differences among treatments in the post-treatment bird community,
microclimate, microhabitat, and arthropod availability. I also calculated the differences
between pre- and post-treatment for bird community, microclimate principle components,
and microhabitat principle components and tested these differences using a two-way
ANOVA with thin, burn, and block as main factors. Tukey’s multiple comparisons test
was preformed based on the results of the ANOVA. All analyses were performed in
SPSS (v.15.0) using an alpha level of 0.05.
To investigate variation in abundance of species and guilds as they relate to
microhabitat measures and arthropod availability, I used canonical correspondence
analysis (CCA, CANOCO v. 4.5). I eliminated variables with high correlation (Pearson
correlation > 0.7) to avoid redundancy and over-fitting the model and used only species
that had greater than five detections in the analysis. CCA is a direct gradient analysis
53
technique that compares community composition directly to environmental variables
across a gradient (Palmer 1993). This procedure is a type of ordination, and therefore not
a hypothesis testing technique. CCA is appropriate to use when there are no differences
among stands because it evaluates gradients on a different scale; it examines the trends
and variability within stands.
Results
Microclimate. There were no interactions between burning and thinning for
microclimate one year after treatment (Table 3.2) or for changes following treatment
(Table 3.3). Daytime May relative humidity differed among the treatments (Table 3.4).
Daytime May humidity was higher on burned stands than on burned/thinned stands. The
change in daytime May and June temperature following the treatment was significant
(Table 3.4). The change in daytime temperature was greater on thinned/burned stands
than burned stands in both May and June.
The original eight microhabitat difference variables showed low multivariate
correlation (Kaiser-Meyer-Olkin [KMO] Measure of Sampling Adequacy = 0.423) so I
removed the variable with the lowest multivariate correlation (nighttime May relative
humidity) from the principle component analysis to increase the KMO measure of
sampling adequacy to 0.55. The remaining seven variables were condensed to 2 principle
components (PCs), the first representing daytime climate and the second representing
nighttime climate (Table 3.6). All components with eigen values greater than 1 were
retained. The 2 components retained approximately 84% of the original variation
54
Table 3.2. Results of two-way analysis of variance (ANOVA) on microclimate,
microhabitat, and bird community variables one year after silvicultural treatments in
Bankhead National Forest, 2006-2007.
Burn
Thin
Thin*Burn
Variable
F
P
F
P
F
P
Microclimate
Daytime May Temp
0.03 0.86
7.61 0.02
2.52 0.14
Daytime June Temp
0.02 0.88
9.36 0.01
0.75 0.40
Daytime May RH
0.34 0.57
7.26 0.02
3.48 0.09
Daytime June RH
0.18 0.68
6.10 0.03
1.22 0.29
Nighttime May Temp
0.10 0.76
0.34 0.57
0.08 0.79
Nighttime June Temp
0.12 0.73
1.09 0.32
0.88 0.37
Nighttime May RH
0.24 0.63
1.56 0.24
0.04 0.84
Nighttime June RH
0.12 0.73
1.53 0.24
0.37 0.56
Microhabitat
Herbaceous ground cover
0.10 0.75
3.86 0.07
0.17 0.68
Woody ground cover
0.01 0.92
11.57 0.01
0.50 0.49
Litter ground cover
8.75 0.01
8.81 0.01
0.09 0.77
Bare ground cover
12.32 0.00
1.63 0.23
3.50 0.09
Litter depth
14.73 0.00
4.14 0.06
8.17 0.01
Canopy cover
0.25 0.63
24.66 0.00
0.17 0.69
Forest level 1
0.34 0.57
1.55 0.24
1.17 0.30
Forest level 2
0.65 0.44
30.94 0.00
0.90 0.36
Forest level 3
7.16 0.02
34.21 0.00
9.79 0.01
Forest level 4
5.35 0.04
11.62 0.01
6.22 0.03
Total basal area
0.01 0.92
20.51 0.00
0.09 0.77
Hardwood basal area
0.15 0.71
11.49 0.01
0.15 0.71
Pine basal area
0.00 0.99
12.72 0.00
1.14 0.31
Snag basal area
3.42 0.09
8.79 0.01
2.15 0.17
Bird Community
Species richness
1.25 0.29
3.51 0.09
0.25 0.63
Relative abundance
1.25 0.29
3.51 0.09
0.25 0.63
Shannon-Weiner Diversity
index
0.10 0.75
3.14 0.10
6.67 0.02
Evenness
0.55 0.47
2.01 0.18
0.06 0.80
Tree nesting abundance
1.62 0.23
11.12 0.01
0.49 0.50
Shrub nesting abundance
0.71 0.42
3.70 0.08
0.08 0.78
Cavity nesting abundance
0.76 0.40
2.33 0.15
1.93 0.19
Ground nesting abundance
0.46 0.51
2.34 0.15
0.05 0.83
Parasite nesting abundance
1.41 0.26
9.95 0.01
1.41 0.26
Bark foraging abundance
0.00 0.98
4.11 0.07
0.45 0.52
Aerial foraging abundance
1.99 0.18
0.87 0.37
0.44 0.52
Ground foraging abundance
1.99 0.18
0.87 0.37
0.44 0.52
Foliage foraging abundance
0.32 0.58
4.93 0.05
0.45 0.52
Neotropical migrant
0.41 0.53
1.32 0.27
0.52 0.48
55
Burn
Variable
abundance
Temperate migrant abundance
Resident abundance
Interior species abundance
Interior/edge species
abundance
Edge/open species abundance
Thin
Thin*Burn
F
P
F
P
F
P
0.03
0.01
0.89
0.87
0.93
0.36
2.82
8.76
0.49
0.12
0.01
0.50
1.89
0.65
0.00
0.19
0.44
0.95
0.03
0.14
0.86
0.71
4.56
39.31
0.05
0.00
1.57
2.28
0.23
0.16
56
Table 3.3. Results of two-way analysis of variance (ANOVA) on differences in
microclimate, microhabitat, and bird community variables before and after silvicultural
treatments in Bankhead National Forest, 2005-2007.
Burn
Thin
Thin*Burn
Variable
F
P
F
P
F
P
Microclimate
Daytime May Temp
0.06 0.81
8.14 0.01
3.41 0.09
Daytime June Temp
0.07 0.79
8.04 0.02
0.84 0.38
Daytime May RH
0.29 0.60
3.82 0.07
2.89 0.12
Daytime June RH
0.18 0.68
4.76 0.05
0.70 0.42
Nighttime May Temp
0.05 0.83
0.19 0.67
0.34 0.57
Nighttime June Temp
0.46 0.51
1.37 0.26
0.72 0.41
Nighttime May RH
0.01 0.91
1.44 0.25
0.07 0.80
Nighttime June RH
0.01 0.91
1.40 0.26
1.16 0.30
Microhabitat
Herbaceous ground cover
0.43 0.53
0.02 0.90
0.65 0.44
Woody ground cover
0.04 0.84
0.66 0.43
0.06 0.81
Litter ground cover
5.36 0.04
4.33 0.06
1.78 0.21
Bare ground cover
10.46 0.01
6.79 0.02
0.34 0.57
Litter depth
6.27 0.03
4.73 0.05
1.14 0.31
Canopy cover
0.13 0.72
15.88 0.00
0.04 0.84
Forest level 1
0.87 0.37
0.64 0.44
0.87 0.37
Forest level 2
0.57 0.46
23.98 0.00
0.04 0.85
Forest level 3
0.01 0.94
28.95 0.00
5.76 0.03
Forest level 4
1.30 0.28
12.83 0.00
0.66 0.43
Bird Community
Species richness
0.00 0.99
0.33 0.58
0.95 0.35
Relative abundance
0.00 0.99
0.33 0.58
0.95 0.35
Shannon-Weiner Diversity
index
0.02 0.89
0.03 0.88
5.66 0.03
Evenness
0.22 0.65
1.13 0.31
1.91 0.19
Tree nesting abundance
7.78 0.02
4.17 0.06
1.18 0.30
Shrub nesting abundance
2.70 0.13
0.44 0.52
0.00 0.97
Cavity nesting abundance
0.00 0.97
1.09 0.32
26.37 0.00
Ground nesting abundance
0.01 0.91
2.32 0.15
0.01 0.94
Parasite nesting abundance
0.05 0.83
1.69 0.22
0.05 0.83
Bark foraging abundance
0.00 0.95
0.73 0.41
0.61 0.45
Aerial foraging abundance
0.48 0.50
0.97 0.34
0.63 0.44
Ground foraging abundance
0.96 0.35
0.32 0.58
0.33 0.58
Foliage foraging abundance
2.67 0.13
0.34 0.57
7.12 0.02
Neotropical migrant
abundance
2.43 0.14
1.01 0.34
0.00 0.98
Temperate migrant abundance 0.10 0.76
3.56 0.08
4.00 0.07
Resident abundance
1.37 0.27
1.38 0.26
5.69 0.03
Interior species abundance
0.03 0.86
2.43 0.14
0.25 0.63
57
Burn
Variable
Interior/edge species
abundance
Edge/open species abundance
Thin
F
P
F
P
7.75
0.24
0.02
0.64
0.16
9.88
0.70
0.01
58
Thin*Burn
F
P
2.05
1.75
0.18
0.21
Table 3.4 Results of one-way analysis of variance (ANOVA) for microclimate variables among four silvicultural treatments in
Bankhead National Forest, AL, one year after treatment, 2006-2007. Means within a row with different superscript numbers differ
(Tukey P<0.05).
ANOVA
Climate Variable
Mean ± SE by treatment
P
F
df
Control
Burn
Thin
Thin/Burn
May Temperature
0.053
3.423
3
23.09±0.98
21.25±0.81
24.31±1.02
25.78±0.63
June Temperature
0.052
3.444
3
24.85±1.35
23.94±0.52
27.65±0.80
28.94±1.28
May Relative Humidity
0.047
3.58
3
52.61±8.72ab
60.37±4.93b
49.99±1.54ab
45.93±1.80a
June Relative Humidity
0.115
2.442
3
59.82±6.25
64.77±1.51
5.41±1.92
53.23±2.51
May Temperature
0.894
0.201
3
18.82±1.19
18.79±0.32
19.60±0.90
19.06±0.74
June Temperature
0.507
0.822
3
20.87±0.93
21.42±0.80
22.72±1.09
21.52±0.17
May Relative Humidity
0.626
0.602
3
63.44±7.35
66.58±7.05
58.77±4.73
60.05±1.43
June Relative Humidity
0.545
0.746
3
71.87±4.18
70.89±4.94
64.67±3.73
68.42±2.26
Daytime
59
Nighttime
59
Table 3.5 Results of one-way analysis of variance (ANOVA) of changes in microclimate variables following four silvicultural
treatments in Bankhead National Forest, AL, 2006-2007.
Means within a row with different superscript numbers differ
(Tukey P < 0.05).
ANOVA
Variable
Mean ± SE by treatment
P
F
df
Control
Burn
Thin
Thin/Burn
May Temperature
0.04
3.90
3
2.66±1.07ab
0.72±1.07a
3.59±0.76ab
5.08±0.76b
June Temperature
0.07
3.08
3
1.49±1.49
0.66±1.49
4.0±1.06
5.50±1.06
May Relative Humidity
0.14
2.23
3
-12.82±7.64
-5.27±6.89
-15.64±1.42
-18.29±1.96
June Relative Humidity
0.20
1.83
3
-16.59±3.58
-12.70±3.58
20.77±2.53
-22.06±2.53
May Temperature
0.91
0.18
3
0.61±1.06
1.36±1.06
1.54±0.75
1.21±0.75
June Temperature
0.54
0.76
3
-0.50±0.62
0.45±0.75
1.37±1.10
0.46±0.27
May Relative Humidity
0.68
0.52
3
-8.10±4.87
-8.70±4.87
-14.27±3.44
-12.66±3.44
June Relative Humidity
0.45
0.94
3
-11.36±4.19
-17.82±4.19
-19.56±2.96
-15.22±2.96
PC 1 (daytime climate)
0.05
3.47
3
0.43±0.47
1.13±0.47
-0.16±0.33
-0.62±0.33
PC 2 (nighttime climate)
0.55
0.74
3
0.71±0.63
0.23±0.63
-0.38±0.44
-0.09±0.44
Daytime
60
Nighttime
Principle Components
60
Table 3.6. Principle component analysis loadings, eigenvalues, and percent variance for
the change in microclimate variables following silvicultural treatment in Bankheand
National Forest, AL, 2005-2007.
PC1
PC2
60
Day May Temp
-0.874
-0.241
Day June Temp
-0.637
-0.608
Day June Relative Humdity
0.889
0.406
Night May Temperature
0.737
-0.267
Night May Relative Humdity
-0.380
0.779
Night June Relative Humdity
-0.522
0.807
Night June Temperature
0.556
-0.753
Day May Relative Humidity
0.715
0.523
Eigenvalue
3.741
46.770
% Variance
2.76
34.55
61
(Bartlett’s Test of Sphericity χ2 = 117.765, df = 21, p = 0.000). Neither microclimate PC
was significantly different among treatments (Table 3.5).
Microhabitat. There was an interaction between burning and thinning for litter
depth and presence of forest level three and four one year after treatment (Table 3.2).
Thinning reduces the amount of litter depth and the presence of forest levels three and
four regardless of whether they were burned or not (Fig 3.3, 3.4, 3.5). However, when
burning is not combined with thinning, it results in a less litter depth and lower presence
of forest levels three and four (Fig 3.3, 3.4, 3.5). There was an interaction between
burning and thinning in the change in forest level three presence following treatment
(Table 3.3). Burning resulted in a decreased presence of forest level three when
combined with thinning, but when only burning was performed, the presence of forest
level three increased (Fig. 3.6). One year after treatment, all microhabitat variables
except percent herbaceous cover and presence of forest level one differed among the
treatments (Table 3.7). Thinned and thinned/burned stands had higher percentage of
woody ground cover (17% and 19%, respectively) than the untreated or burned stands
(7% and 9% respectively). Litter cover was highest on untreated stands (99%), whereas
bare ground was highest on thinned/burned stands (6%). As expected, canopy cover and
BA were lowest on stands that had been thinned (63% and 67 ft2/acre).
The BA of pines
and snags was higher on the control (119 ft2/acre and 9 ft2/acre, respectively) than on the
thinned stands (97 ft2/acre and 4 ft2/acre, respectively). Presence of forest level two was
greatest on the control and the burn. Burned stands had the greatest presence of forest
level three (22 %).
62
Figure 3.3. Litter depth interaction between burning
and thinning in the Bankhead National Forest, AL,
2006-2007.
Figure 3.4. Presence of forest level 3 interaction between
burning and thinning in the Bankhead National Forest, AL,
2006-2007.
63
Figure 3.5. Interaction between burning and thinning of
presence of forest level 4 in the Bankhead National Forest,
AL, 2006-2007.
Figure 3.6. Interaction between burning and thinning in the
change in presence of forest level 3 following silvicultural
treatment in the Bankhead National Forest, AL, 2006-2007.
64
Table 3.7 Results of one-way analysis of variance (ANOVA) of habitat variables among four silvicultural treatments in Bankhead
National Forest, AL, one year after treatment, 2006-2007. Means within a row with different superscript numbers differ (Tukey P <
0.05).
ANOVA
Habitat Variable
Mean ± SE by treatment
65
P
F
df
Control
Burn
Thin
Thin/Burn
Ground cover: % herbaceous plants
0.30
1.36
3
11.25±4.83
14.58±5.02
22.01±3.93
21.58±2.83
Ground cover: % woody plants
0.03
4.07
3
6.81±1.14
8.61±3.62
19.10±3.14
16.67±2.63
Ground cover: % litter
0.01
6.04
3
99.44±0.28a
97.5±1.05ab
97.92±0.55ab
91.81±2.06b
Ground cover: % bare ground
0.00
8.12
3
0a
1.53±0.77ab
0.76±0.25a
6.11±1.62b
Litter depth (cm)
0.01
7.23
3
7.7±0.3b
3.7±0.8a
4.8±0.3a
4.2±0.8a
Canopy cover
0.00
8.43
3
91.03±1.01b
90.44±2.22b
63.29±1.63a
60.39±6.88a
Forest level 1 present
0.46
0.92
3
11.33±3.38
8.00±1.15
11.67±1.63
12.67±1.33
Forest level 2 present
0.00
11.08
3
22.00±1.53b
23.00±2.65b
7.33±0.95a
5.67±1.73a
Forest level 3 present
0.00
15.66
3
11.67±1.20a
22.33±2.19b
6.67±0.92a
5.83±2.01a
Forest level 4 present
0.01
6.77
3
25.67±1.33a
20.33±4.26b
18.83±2.71b
19.33±2.76b
Basal Area: Total (ft2/acre)
0.01
6.81
3
66.67±7.79a
77.78±19.71a
Basal Area: Hardwoods (ft2/acre)
0.04
3.98
3
44.44±25.02
25.00±3.15
20.00±17.33
ab
a
57.78±3.44ab
2
162.22±12.81b 156.67±13.88b
44.44±24.44
Basal Area: Pines (ft /acre)
0.02
4.67
3
118.89±11.28
Basal Area: Snags (ft2/acre)
0.03
4.34
3
8.89±1.11b
65
b
96.67±4.84
4.44±2.94ab
39.44±5.76
3.33±2.11a
2.22±1.11a
Table 3.8 Results of analysis of variance (ANOVA) of changes in microhabitat variables following four silvicultural treatments in
Bankhead National Forest, AL, 2006-2007. Means within a row with different superscript numbers differ (Tukey P<0.05).
ANOVA
Variable
Mean ± SE by treatment
66
P
F
df
Control
Burn
Thin
Thin/Burn
Ground cover: % herbaceous plants
0.66
0.54
3
1.94±6.25
1.11±6.25
-1.74±4.42
6.16±4.42
Ground cover: % woody plants
0.84
0.84
3
1.67±4.00
1.53±4.00
3.61±2.82
5.21±2.82
Ground cover: % litter
0.01
0.19
3
0.42±0.64b
-1.95±1.01ab
-1.39±0.65ab
-7.78±1.92a
Ground cover: % bare ground
0.02
6.79
3
0a
1.53±0.77ab
0.76±0.25ab
6.11±1.62b
Litter depth
0.04
0.04
3
3.01±0.85b
0.37±0.85ab
0.62±0.60ab
-0.44±0.60a
Canopy cover
0.01
5.38
3
4.79±6.68b
3.88±6.68ab
-17.04±4.72ab
-20.38±4.72a
Forest level 1 present
0.60
0.65
3
-8.33±2.89
-13.00±2.89
-8.67±2.05
-8.67±2.05
c
bc
-12.33±2.43
-15.17±2.43a
Forest level 2 present
0.00
8.26
3
1.67±3.44
Forest level 3 present
0.00
11.77
3
-2.67±2.33bc
2.33±2.33c
-8.67±1.64ab
-13.33±1.54a
Forest level 4 present
0.02
4.78
3
-1.00±2.36ab
3.00±2.36b
-6.67±1.67a
-6.00±1.67a
PC 1 (Ground cover)
0.00
14.96
3
0.98±0.17b
1.35±0.10b
-0.61+0.20a
-0.56±0.29a
PC 2 (Understory cover)
0.03
4.24
3
0.77±0.20b
-0.16±0.3ab
0.46±0.11ab
-0.76±0.55a
PC 3 (Midstory cover)
0.88
0.22
3
-0.21±0.21
-0.24±0.08
-0.03±0.39
0.26±0.62
PC 4 (Overstory cover)
0.19
1.84
3
0.35±0.34
0.50±0.68
-0.47±0.37
0.05±0.48
66
1.55E-15±3.40
ab
The change in percent litter cover, percent bare ground, litter depth, canopy cover,
and presence of forest levels two and four were different across treatments (Table 3.8).
The change in litter ground cover and bare ground cover was greatest on the
thinned/burned stands. Litter depth decreased on the thinned/burned stands and increased
slightly on burned, thinned, and untreated stands. As expected, the change in canopy
cover was greatest on the stands that had been thinned. The presence of forest level two
decreased on all treated stands; the greatest decrease was on thinned and thinned/burned
stands. The presence of forest level four decreased on the untreated, thinned, and
thinned/burned stands, and increased on the burned stands.
I grouped the original thirteen difference variables into 4 principle components (PCs),
the first representing ground cover, the second representing understory cover, the third
representing midstory cover, and the fourth representing overstory cover (Table 3.9). All
components with an eigen value greater than 1 were retained. The 4 components retained
approximately 85% of the original variation (Bartlett’s Test of Sphericity χ2 = 194.022, df
= 78, p = 0.000). PC1 (ground cover) and PC2 (understory cover) differed among the
treatments (Table 3.8). PC1 (ground cover) decreased on the thinned and thinned/burned
stands and it increased on the untreated and burned stands (Table 3.8). The increase in
PC2 understory cover) was greater on control stands than on the thinned/burned stands
(Table 3.8).
Arthropod availability. There were no differences in arthropod biomass index among
treatments (Table 3.10), nor were there any interactions between the treatments (Table
3.2).
67
Table 3.9. Principle component analysis loadings, eigenvalues, and percent variance for
the change in microhabitat variables following silvicultural treatment in Bankhead
National Forest, AL, 2005-2007.
PC1
PC2
PC3
PC4
Herb Cover
0.099
0.795
-0.062
0.383
Woody Cover
0.003
0.925
0.073
-0.192
Litter depth
0.702
0.013
-0.408
0.504
Canopy Cover
0.561
-0.557
0.371
0.141
Forest Level 1
0.198
0.616
-0.446
0.401
Forest Level 2
0.361
-0.461
0.604
0.233
Forest Level 3
0.233
0.020
0.916
0.163
Forest Level 4
0.060
0.044
0.296
0.914
Bare Cover
-0.939
0.025
-0.195
-0.111
Litter Cover
0.911
0.172
0.262
-0.072
Eigenvalue
3.56
2.70
1.26
1.13
% Variance
35.58
29.96
12.57
11.29
68
Table 3.10. Results of one-way analysis of variance (ANOVA) of arthropod biomass index among four silvicultural treatments in
Bankhead National Forest, AL, one year after treatment, 2006-2007.
ANOVA
Mean ± SE by treatment
Arthropod Variable
P
F
df
Control
April Biomass (g/10g dry leaves)
0.435
0.978
3
49.98±10.93
35.29±6.23
58.61±15.77
120.21±53.85
May Biomass (g/10g dry leaves)
0.069
3.061
3
133.99±29.23
44.94±22.48
60.23±25.18
41.90±9.38
June Biomass (g/10g dry leaves)
0.163
2.031
3
126.69±61.84
33.97±14.69
68.17±22.06
54.76±25.22
69
69
Burn
Thin
Thin/Burn
Bird community. A total of 983 birds were detected one year after treatment,
representing 40 species (Table3.1). The most abundant species were the red-eyed vireo
(Vireo olivaceus Linnaeus), comprising 16.5% (162 detections) of total individuals, and
the pine warbler (Dendroica pinus Wilson), comprising 14.0% (138 detections) of total
individuals. Species detected post treatment that were not detected before treatment were
the brown-headed nuthatch (Sitta pusilla Latham), eastern phoebe (Sayornis phoebe
Latham), eastern towhee (Pipilo erythrophthalmus Linnaeus), eastern wood-pewee
(Contopus virens Linnaeus), mourning dove (Zenaida macroura Linnaeus), ruby-throated
hummingbird (Archilochus coulbris Linnaeus), and yellow-throated vireo (Vireo
flavifrons Vieillot). Two species (blue grosbeak [Guiraca caerulea Linnaeus] and redbellied woodpecker [Melanerpes carolinus Linnaeus]) detected before treatments were
not detected post-treatment.
There was an interaction in Shannon-Weiner diversity index between burning and
thinning one year after treatment (Table 3.2). When combined with thinning, burning
results in lower diversity, but burning alone results in higher bird diversity (Fig. 3.7).
There was an interaction between the two treatments in changes in Shannon-Weiner
diversity index, cavity nesting species abundance, foliage foraging species abundance,
and resident species abundance following the treatment (Table 3.3). For all four
variables, when thinning is combined with burning the result is a smaller change in the
variable (Fig. 3.8, 3.9, 3.10, 3.11). However, when burning is done alone, there is a
larger change in the variable.
Parasite nesting species abundance (i.e. brown-headed cowbirds), and edge/open
habitat species abundance differed among the treatments (Table 3.11). Parasite nesting
70
Figure 3.7. Interaction between burning and thinning in the
Shannon-Weiner Diversity Index following silvicultural treatment
in the Bankhead National Forest, AL, 2006-2007.
Figure 3.8. Interaction between burning and thinning in the
change in Shannon-Weiner Diversity Index following
silvicultural treatment in the Bankhead National Forest, AL,
2006-2007.
71
Figure 3.9. Interaction between burning and thinning in the
change in cavity nesting bird abundance following silvicultural
treatment in the Bankhead National Forest, AL, 2006-2007.
Figure 3.10. Interaction between burning and thinning in the
change in foliage foraging bird abundance following silvicultural
treatment in the Bankhead National Forest, AL, 2006-2007.
72
Figure 3.11. Interaction between burning and thinning in
the change in resident bird abundance following
silvicultural treatment in the Bankhead National Forest,
AL, 2006-2007.
73
74
Table 3.11. Results of analysis of variance (ANOVA) of bird community variables among silvicultural treatments in Bankhead
National Forest, AL, one year after treatment, 2006-2007. Means within a row with different superscript letters differ (Tukey p<0.05).
ANOVA
Mean ± SE by treatment
Community Variable
p
F
df
Control
Burn
Thin
Thin/Burn
Species richness
0.24
1.59
3
9.99±0.22
10.82±0.51
11.21±0.23 11.53±0.57
Relative abund (detections per
100m)
0.24
1.59
3
79.89±1.77
86.58±4.10
89.71±1.84 92.28±4.52
Shannon-Weiner diversity index
0.05
3.38
3
2.36±0.10a
2.6±0.04ab
2.72±0.08b 2.53±0.07ab
Evenness
0.85
0.49
3
0.90±0.01
0.91±0.02
0.93±0.01
0.93±0.02
Nesting Guild Abundance
Tree
0.03
4.28
3
7.62±0.98ab
5.24±0.59a
12.85±1.68b 12.44±2.71ab
Shrub
0.24
1.60
3
9.05±0.84
8.32±1.43
15.85±4.11 13.69±4.25
Cavity
0.27
1.48
3
4.80±1.55
8.71±1.36
9.86±1.44
8.96±1.60
Ground
0.42
1.01
3
5.09±1.49
4.53±1.78
3.50±1.05
2.42±0.84
a
a
ab
Parasite
0.02
4.73
3
0
0
0.62±0.30
1.37±0.44 b
Foraging Guild Abundance
Bark
0.26
1.54
3
3.86±1.10
5.02±1.05
8.41±0.51
7.32±2.40
Aerial
0.28
1.43
3
1.63±1.00
2.04±0.59
1.78±0.38
2.91±0.42
Ground
0.28
1.43
3
13.02±8.03
16.29±4.74
14.20±3.02 23.27±3.39
Foliage
0.16
2.02
3
16.12±1.48
17.10±4.02
25.96±3.18 23.17±5.38
Migratory Guild Abundance
Neotropical
0.58
0.67
3
17.27±2.62
13.99±1.25
19.02±2.98 20.52±5.14
Temperate
0.22
1.71
3
1.17±0.58
3.07±1.58
5.10±1.65
2.97±1.00
Resident
0.06
3.18
3
7.82±0.47
9.76±1.85
18.04±0.89 15.61±3.50
Habitat Guild Abundance
Interior
0.70
0.49
3
10.06±2.37
8.24±2.58
8.68±1.32
7.10±1.29
Interior/edge
0.14
2.18
3
14.85±0.76
17.22±1.77
22.29±1.71 19.15±2.56
a
a
3
Edge/open
0.00
14.16
1.64±0.50
3.00±1.81
12.28±2.84b 8.55±4.85b
74
species abundance was highest on thinned and thinned/burned stands, and edge/open
habitat species abundance was higher on thinned and thinned/burned stands than
untreated or burned stands.
Following silviculture treatment, the change in abundance of tree and cavity nesting
species, interior/edge species, and edge/open species differed among the treatments
(Table 3.12). The greatest decrease in tree nesting species was on burned stands, whereas
there was an increase on thinned stands. Cavity nesting species increased on burned and
thinned stands and decreased on untreated and thinned/burned stands. Foliage foraging
species decreased on all stands, but the greatest change was on thinned/burned stands.
Interior edge species decreased on all stands; the biggest decrease was on burned and
thinned/burned stands. Edge/open species increased the most on thinned and
thinned/burned stands.
The post treatment bird community included fifteen species listed in Partners in
Flight’s (PIF) North American Landbird Conservation Plan as Species of Continental
Importance (Rich et al. 2004, Table 3.13). Three new species of concern were detected
after treatment (Table 3.13). PIF lists species in two categories; WatchList species are
species that have multiple reasons (restricted distribution, low population size,
widespread population declines, high threats to habitat, etc.) for conservation concern
across their entire range (Rich et al. 2004). Stewardship species are species which have a
high percentage of their global population within a single North American biome (Rich et
al. 2004).
Morisita’s similarity indices indicate that species composition on the untreated stands
is most similar to the burned stands and least similar to the thinned stands (Table 3.14).
75
76
Table 3.12. Results of one-way analysis of variance (ANOVA) of changes in bird community variables following four silvicultural
treatments in Bankhead National Forest, AL, 2006-2007. Means within a row with different superscript numbers differ (Tukey
P<0.05).
ANOVA
Mean ±SE by treatment
Variable
p
F
df
Control
Burn
Thin
Thin/Burn
Species richness
0.71 0.47 3
2.46±1.08
-1.45±1.08
-2.08±0.76
-3.01±0.76
Relative abundance
0.71 0.47 3
-19.68±8.64
-12.50±8.64 -16.7±6.11
-24.07±6.11
Shannon-Weiner diversity 0.16 2.05 3
-0.11±0.15
0.21±0.15
0.17±0.10
-0.11±0.10
Evenness
0.42 1.01 3
0.00±0.03
0.04±0.03
0.01±0.02
-0.01±0.02
Nesting Guild Abundance
Tree
0.01 5.51 3
-3.68±1.99ab
-6.61±1.99a
1.72±1.41b
-4.97±1.41ab
Shrub
0.37 1.14 3
-1.60±2.61
-5.40±2.61
-0.02±1.85
-3.83±1.85
ab
c
bc
Cavity
0.00 10.3 3
-2.05±1.30
3.67±1.30
2.54±0.9
-3.26±0.92a
Ground
0.53 0.78 3
-0.82±2.53
-0.73±2.53
-4.32±1.79
-3.89±1.79
Parasite
0.62 0.61 3
0
0
0.47±0.21
0.67±0.68
Foraging Guild Abundance
Bark
0.70 0.49 3
-2.95±2.25
-1.55±2.25
0.24±1.59
-1.40±1.59
Aerial
0.48 0.86 3
-0.19±0.79
-0.12±0.79
1.03±0.56
0.01±0.56
Ground
0.55 0.73 3
9.09±5.67
11.91±5.67
9.84±4.01
17.49±4.01
Foliage
0.02 4.87 3
-8.23±2.85ab
-5.68±2.85ab -3.1±2.02b
-13.70±2.02a
Migratory Guild Abundance
Neotropical
0.34 1.24 3
-2.53±3.60
-7.50±3.60
-5.76±2.56
-10.56±2.56
Temperate
0.10 2.57 3
-2.78±1.38
-0.03±1.38
1.85±0.97
-0.16±0.97
Resident
0.04 3.80 3
-4.23±2.55
-1.55±2.55
3.63±1.80
-4.22±1.80
Habitat Guild Abundance
Interior
0.45 0.94 3
-2.27±4.46
-1.05±4.46
-6.35±3.15
-8.99±3.15
Interior/edge
0.02 4.72 3
-2.89±3.12ab
-6.58±3.12ab -0.1±2.23b
-11.57±2.23a
Edge/open
0.03 4.20 3
-1.19±2.33a
0.49±2.33ab
7.81±1.65b
4.16±1.65ab
76
Table 3.13. Species of Continental Importance, as listed by the Partners in
Flight Landbird Conservation Plan (Rich et al. 2004), detected on upland
pine-hardwood plots before and after silvicultural treatments in Bankhead
National Forest, AL, 2005-2007. See table 2.1 for scientific names.
PIF
Pre-treatment
Post-treatment
Species
Listing*
Individuals Plots
Individuals
Plots
Acadian Flycatcher
S
7
8
5
5
Brown-headed Nuthatch
W
0
0
1
1
Brown Thrasher
S
1
1
1
1
Carolina Wren
S
15
11
21
9
Eastern Towhee
S
0
0
1
1
Hooded Warbler
S
52
13
20
12
Indigo Bunting
S
38
16
47
14
Kentucky Warbler
W
6
6
12
8
Louisiana Waterthrush
S
1
1
2
2
Pine Warbler
S
80
18
68
18
Prairie Warbler
W
5
5
28
10
White-eyed Vireo
S
5
4
4
4
Worm-eating Warbler
W
53
15
28
16
Wood Thrush
W
2
1
4
4
Yellow-throated Vireo
S
0
0
2
2
*S = Stewardship Species; W = WatchList Species (Rich et al. 2004)
77
Table 3.14. Morisita’s similarity index for the breeding
bird community one year after silvicultural treatment in
Bankhead National Forest, AL, 2006-2007.
Control Burn
Thin
Thin/Burn
Control
-
0.99
0.74
0.91
Burn
0.99
-
0.85
0.91
Thin
0.74
0.85
-
1.01
Thin/Burn
0.91
0.91
1.01
-
Pre
treatment
Table 3.15. Morisita’s similarity index for the breeding bird
community before and one year after silvicultural treatment in
Bankhead National Forest, AL, 2005-2007.
Post treatment
Control Burn
Thin
Thin/Burn
Control
Burn
1.01
0.94
1
0.9
0.89
0.84
0.95
0.89
Thin
Thin/Burn
0.98
0.99
0.94
0.99
0.94
0.92
0.97
1.02
78
Species composition was similar before and after treatment on all stands, with a small
change on the burned and thinned stands (Table 3.15).
Canonical correspondence analysis. The CCA of microhabitat characteristics,
arthropod availability, and species abundance explained 48.8 % (total inertia = 0.847) of
the variation in the first three axes. Axis one explained 21.1% of the variation
(Eigenvalue = 0.178), the second axis 15.3% (Eigenvalue = 0.130), and the third axis
12.4% (Eigenvalue = 0.104). Based on the CCA of microhabitat characteristics,
arthropod availability, and nesting guild abundance, the first three axes explained 74.9%
(total inertia = 0.119) of the variation in guild abundance. The first axis explained 39.9%
(Eigenvalue = 0.048) of the variation, the second 21.9% (Eigenvalue = 0.026), and the
third 13.1% (Eigenvalue = 0.016). The CCA of the microhabitat characteristics,
arthropod availability and foraging guild abundance explained 90.9% (total inertia =
0.069) of the variation in the first three axes. Axis one explained 37.2% (Eigenvalue =
0.026), the second axis 31.5% (Eigenvalue = 0.022), and the third 22.2% (Eigenvalue =
0.015).
The CCA of microhabitat characteristics, arthropod availability, and species
abundance revealed a gradient in microhabitat characteristics apparent in the position of
variables along the axes of the ordination plots (Fig. 3.12). On one end of the gradient is
canopy cover and presence of forest level four, on the other is woody and herbaceous
ground cover and presence of forest level one. Another gradient from tree species
richness, basal area, and presence of forest level three to a blank area on the ordination
indicates that open areas were not represented by any of the habitat variables collected.
Open habitat species (yellow-breasted chat [Icteria vierns Linnaeus], eastern wood79
pewee, and morning dove) and generalist species (eastern tufted titmouse [Baeolophous
bicolor Linnaeus], Carolina chickadee [Poecile carolinensis Audubon], and summer
tanager [Piranga rubra Linnaeus]) are found in this area of the ordination plot (Figure
3.12). Early successional species (Kentucky warbler [Oporornis formosus Wilson],
indigo bunting [Passerina cyanea Linnaeus], and prairie warbler[Dendroica discolor
Vieillot]) were associated with herbaceous and woody ground cover and presence of
forest level one, whereas more interior species (worm-eating warbler [Helmitheros
vermivorus Gmelin], acadian flycatcher [Empidonax virescens Vieillot], and blackthroated green warbler [Dendroica virens Gmelin]) were on the other end of the gradient,
associated with the presence of forest level three, basal area, basal area of hardwoods and
snags, and litter depth (Fig. 3.12)
A similar gradient was revealed in the CCA of habitat characteristics, arthropod
availability and nesting guild associations. The ordination plot revealed a gradient from
basal area and presence of forest level four to herbaceous and woody ground cover and
presence of forest level one, and also a gradient from canopy cover and presence of forest
level three to litter depth, litter ground cover, and tree species richness (Fig. 3.13).
Parasite nesting guild was on the edge of the plot, indicating that there were no strong
associations with the environmental variables. Ground nesting species were associated
with high litter ground cover; cavity nesting species were associated with basal
area of snags and presence of forest level four; and tree nesting species were associated
with presence of forest level three and canopy cover (Fig. 3.13).
The CCA of microhabitat characteristics, arthropod availability, and foraging guild
associations revealed a gradient that was different from the first two analyses. On
80
Figure 3.12. First and second canonical correspondence axes for microhabitat
characteristics, arthropod availability, and bird species abundance one year after
silvicultural treatments, Bankhead National Forest, AL, 2006-2007. See table 3.1 for
species codes.
81
Figure 3.13. First and second canonical correspondence axes for microhabitat
characteristics, arthropod availability, and nesting guild abundance one year after
silvicultural treatments, Bankhead National Forest, AL, 2006-2007.
82
one end of the gradient was the presence of forest level one and four, tree species
richness, basal area, and woody and herbaceous ground cover; on the other end was
canopy cover, basal area of snags, litter ground cover, litter depth, and the presence of
species was associated with litter cover, snag BA, and hardwoods BA (Fig 3.13).
forest level three (Fig. 3.14). In the center of the plot was foliage foraging species,
indicating that they are more generalist species, and on the far edges of the plot, opposite
one another, were aerial feeding species, and ground foraging species (Fig. 3.14). Bark
foraging
Discussion
One year after treatment there is a treatment affect on some aspects of the bird
community; it is likely a result of changes in microhabitat among treatments. Thinning
had a greater impact on the bird community than burning, although burning affected the
bird community on a smaller scale. Treated stands had a higher diversity than untreated
stands. This is a common effect of any forest disturbance where a more heterogenic
habitat is created (Roth 1976). Thinning resulted in a decrease in canopy cover, vertical
structure, and the lowest basal area of all stands. However, one year after the treatment
there was much shrubs and herbaceous plant regeneration. As a result of this
regeneration, thinned stands had a high abundance of tree nesting species. This
corroborates other studies that have found that may tree and shrub nesting birds use
regenerating stands shortly after cutting (Gram et al. 2003, Harrison et al. 2005, Heltzel
83
Figure 3.14. First and second canonical correspondence axes for microhabitat
characteristics, arthropod availability, and foraging guild abundance one year after
silvicultural treatments, Bankhead National Forest, AL, 2006-2007.
84
and Leberg 2006, Holmes and Pitt 2007). No difference was found in arthropod biomass
however; leaf samples were taken at every sampling location regardless of tree Arthropod
availability. Since I indexed arthropod abundance by branch clippings that did not reflect
the relative availability of leaves in different treatments (i.e., control and burn only stands
had more trees than thinned and thinned/burned stands), there could be a difference that
was not detected by my methods. Biomass in April appears to have affected the
abundance of foliage foraging species. Canonical correspondence analysis shows a strong
relationship between foliage foraging species and arthropod biomass in April, when they
arrived on the breeding grounds. Thinned stands created habitat for edge/open habitat
species, but still maintained many of the interior species. Many other studies have found
that when some trees are retained, as in shelterwood and selection cuts, edge and open
habitat bird species use the habitat for a short time and many mature forest birds remain
(Campbell et al. 2007, Greenberg et al. 2007, Holmes and Pitt 2007, Lanham et al. 2002,
Vanderwel et al. 2007, Weakland et al. 2002). Silvicultural treatments that leave trees
appear to be viable options for creating habitat for early-successional birds if clear cutting
is not an option or if retaining mature forest birds is also a management goal. However,
Costello et al. (2000) suggested that there may be a minimum opening size requirement
for some species associated with early successional habitat. Treatments that retain some
trees may not create openings large enough to support all species that use early
successional habitat for breeding.
Low intensity burns used in this study resulted in a decrease in the presence of the
shrub layer, a decrease in litter depth, and an increase in bare ground. Diversity increased
85
in the stands, indicating that the increased heterogeneity created by the fire’s patchiness
created more diverse habitat than was available before the treatment. Previous studies
found that patchiness created with low-intensity fire results in diverse habitats which are
attractive to a large variety of birds (Greenberg 2007, Blake 2005, Lanham et al. 2002,
Stribling and Barron 1995).
My results suggest that burning and thinning in combination has a more negative
effect on certain bird groups than either treatment by itself one year after treatment.
Burned/thinned plots had a lower diversity than the other treatments, tree and cavity
nesting guilds decreased and foliage foraging birds decreased, likely a result of a decrease
in foliage in the stands. Edge/open habitat species increased while interior/edge habitat
species decreased. Only one study has looked at the effect of combined burning and
thinning (Wilson et al. 1995), but their results differed from mine. They found higher
densities in stands that had been thinned and burned than stands that had only been
thinned.
Large increase in prairie warblers in thinned stands indicates that treatments are
providing habitat for this species of concern, while the constant number in burned stands
indicates that burning is not detrimental to them. Abundance increased the most (from 1
to 19) on thinned stands, which suggests that thinning may provide their preferred
habitat. Gram et al. (2003) found higher prairie warbler abundance in clearcut stands
when 10-15 percent of the trees were left standing, similar to a seed tree cut. They
preferred this habitat to small group selection cuts, single tree selection cuts, or uncut
stands. It appears that they are attracted to treatments that promote understory growth
and remove a large portion of overstory trees.
86
There is no single management system that will provide habitat for all species of
forest birds. Some species respond positively and other respond negatively to any given
management system. Bird responses are often scale dependent and landscape level
effects are also important. My results suggest that ‘free’ thinning, and to a lesser extent,
low intensity prescribed burning, provides habitat for early successional birds while
retaining many mature forest birds one year after treatment. Regenerating forests provide
only temporary habitat for early successional species, so it would be appropriate to time
rotations and burning so that there is consistent availability of early successional breeding
habitat across the landscape, if that is the goal of the management program.
Bibliography
Aquilani, S.M., D.C. LeBlanc, T.E. Morrell. 2000. Effects of prescribed surface fires on
ground- and shrub-nesting Neotropical migratory birds in a mature Indiana Oak
Forest, USA. Natural Areas Journal 20: 317-324.
Askins, R.A., J.F. Lynch, R. Greenberg. 1990. Population declines in migratory birds in
eastern North America. Current Ornithology 7: 1-15.
Blake, J.G. 2005. Effects of prescribed burning on distribution and abundance of birds in
a closed-canopy oak-dominated forest, Missouri, USA. Biological Conservation
121: 519-531.
Blake, J.G. and J.R. Karr. 1987. Breeding birds of isolated woodlots: area and habitat
relationships. Ecology 68: 1724-1724.
Brawn, J.D., S.K. Robinson, and F.R. Thompson III. 2001. The role of disturbance in
the ecology and conservation of birds. Annual Review of Ecology and
Systematics 32: 251-276.
Brose, P., T. Schuler, D.Van Lear, J. Berst. 2001. Bringing fire back: The changing
regimes of the Appalachian mixed-oak forests. Journal of Forestry 32: 30-35.
87
Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L.
Thomas. 2001. Introduction to distance sampling. Oxford University Press,
Oxford.
Campbell, S.P., J.W. Witham, M.L. Hunter Jr. 2007. Long-term effects of groupselection timber harvesting on abundance of forest birds. Conservation Biology
21: 1218-1229.
Costello, C.A, M. Yamasaki, P.J. Penkins, W.B. Leak, C.D. Neefus. 2000. Songbird
response to group selection harvests and clearcuts in a New Hampshire northern
hardwood forest. Forest Ecology and Management 127: 41-54.
Davis, M.A., D.W. Peterson, P.B. Reich, M. Crozier, T. Query, E. Mitchell, J.
Huntington, P. Bazakas. 2000. Restoring savanna using fire: Impact on the
breeding bird community. Restoration Ecology 8: 30-40.
Ehrlich, P.R., D.S. Dobkin, and D. Wheye. 1988. The birder’s handbook: a field guide
to the natural history of North American birds. Simon and Schuster, New York,
NY. 785 pp.
FIA. 1998. Field instructions for southern forest inventory. Re-measurement of prism
plots. Southern Research Station, Forest Service, U.S. Department of Agriculture.
Item 26 version of manual.
Finch, D.M. 1991. Population ecology, habitat requirements, and conservation of
Neotropical migratory birds. General Technical Report R-205. USDA Forest
Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins,
Colorado.
Freemark, K. and B. Collins. 1992. Landscape ecology of birds breeding in temperate
forest fragments. Pp. 443-454 in Ecology and Conservation of Neotropical
Migrant Landbirds (J.M. Hagen III and D.W. Johnston, eds.). Smithsonian
Institution Press, Washington, D.C.
Gaines, G.D. and J.W. Creed. 2003. Forest health and restoration project. National
forests in Alabama, Bankhead National Forest Franklin, Lawrence and Winston
Counties, Alabama. Final environmental impact statement. Management Bulletin
R8-MB 110B.
Ganihar, S.R. 1997. Biomass estimates of terrestrial arthropods based on body length.
Journal of Bioscience 22: 219-224.
Gram, W.K, P.A. Porneluzi, R.L. Clawson, J. Faaborg, S.C. Richter. 2003. Effects of
experimental forest management on density and nesting success of bird species in
Missouri Ozark Forests. Conservation Biology 17: 1324-1337.
88
Greenberg, C.H., L.D. Harris, and D.G. Neary. 1995. A comparison of bird
communities in burned and salvage-logged, clearcut, and forested Florida sand
pine scrub. Wilson Bulletin 107: 40-54.
Greenberg, C.H., A.L. Tomcho, J.D. Lanham, T.A. Waldrop, J. Tomcho, R.J. Phillips, D.
Simon. 2007. Short term effects of fire and other fuel reduction treatments on
breeding birds in a Southern Appalachian upland hardwood forest. Journal of
Wildlife Management 71: 1906-1916.
Harrison, R.B., F.K.A. Schmiegelow, R. Naidoo. 2005. Stand-level response of breeding
forest songbirds to multiple levels of partial-cut harvest in four boreal forest
types. Canadian Journal of Forest Resources 35: 1553-1567.
Heltzel, J.M. and P.L. Leberg. 2006. Effects of selective logging on breeding bird
communities in bottomland hardwood forests in Louisana. Journal of Wildlife
Management 70: 1416-1424.
Holmes, S.B and D.G. Pitt. 2007. Response of bird communities to selection harvesting
in a northern tolerant hardwood forest. Forest Ecology and Management 238:
280-292.
Hunter, W.C., D.A. Buehler, R.A. Canterbury, J.L. Confer, P.B. Hamel. 2001.
Conservation of disturbance-dependent birds in eastern North America. Wildlife
Society Bulletin 29: 440-455.
Imhof, T.A. 1976. Alabama birds. The University of Alabama Press, University, AL.
445 pp.
James, F.C., C.E. McCullouch, D.A. Wiedenfeld. 1996. New approaches to the analysis
of population trends in landbirds. Ecology 77: 13-27.
Johnson, M.D. 2000. Evaluation of an arthropod sampling technique for measuring food
availability for forest insectivorous birds. Journal of Field Ornithology 71: 88109.
Krebs, C.J. 1998. Ecological methodology. Addison Wesley Longman, Menlo Park,
CA. 620 pp.
Lanham, J.D., P.D. Keyser, P.H. Brose, D.H. Van Lear. 2002. Oak regeneration using
the shelterwood-burn technique: Management options and implications for
songbird conservation in the southeastern United States. Forest Ecology and
Management 155: 143-152.
Litvaitis, J.A., D.L. Wagner, M.D. Tarr, E.J. Snyder. 1999. Early successional forests and
shrub-dominated habitats: land use artifact or critical community in the
northeastern United States. Northeast Wildlife 54: 101-118.
89
Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University
Press, Princeton, NJ. 179pp.
Palmer, M.W. 1993. Putting things in even better order: the advantages of canonical
correspondence analysis. Ecology 74: 2215-2230.
Rappole, J.H. and M.V. McDonald. 1994. Cause and effect in population declines of
migratory birds. The Auk 11: 652-660.
Rich, T. D., C. J. Beardmore, H. Berlanga, P. J. Blancher, M. S. W. Bradstreet, G. S.
Butcher, D. W. Demarest, E. H. Dunn, W. C. Hunter, E. E. Iñigo-Elias, J. A.
Kennedy, A. M. Martell, A. O. Panjabi, D. N. Pashley, K. V. Rosenberg, C. M.
Rustay, J. S. Wendt, T. C. Will. 2004. Partners in Flight North American
Landbird Conservation Plan. Cornell Lab of Ornithology. Ithaca, NY. Partners in
Flight website. http://www.partnersinflight.orb/cont_plan/ (VERSION: March
2005).
Robbins, C.S., J.R. Sauer, R.S. Greenberg. S. Droege. 1989. Population declines in North
American birds that migrate to the Neotropics. Proceedings of the National
Academy of Science 86: 7658-7662.
Roth, R.R. 1976. Spatial heterogeneity and bird species diversity. Ecology 57: 773-782.
Sample, B.E., R.J. Cooper, R.D., Greer, and R.C. Whitmore. 1993. Estimation of insect
biomass by length and width. American Midland Naturalist 129: 234-240.
Sauer, J.R., S. Schwartz, and B. Hoover. 1996. The Christmas Bird Count Home Page.
Version 95.1, http://www.mbr-pwrc.usgs.gov/bbs/cbc.html. USGS Patuxent
Wildlife Research Center, Laurel, MD.
Smalley, G.W. 1979. Classification and evaluation of forest sites on the Southern
Cumberland Plateau: U.S. Department of Agriculture, Forest Service, Southern
Forest Experiment Station, general technical report SO-23.
Smith, D.M., B.C. Larson, M.J. Kelty, P.M.S. Ashton. 1997. The practice of silviculture:
Applied forest ecology, ninth edition. John Wiley & Sons, Inc., USA.
Stribling, H.L., and M.B. Barron. 1995. Short-term effects of cool and hot prescribed
burning on breeding songbird populations in the Alabama piedmont. Southern
Journal of Applied Forestry 19: 18-22.
Thompson III, F.R., R.M. DeGraaf. 2001. Conservation approaches for woody, early
successional communities in the eastern United States. Wildlife Society Bulletin
29: 483-494.
90
Trani, M.K., R.T. Brooks, T.L. Schmidt, V.A. Rudis, C.M. Gabbard. 2001. Patterns and
trends of early successional forests in the eastern United States. Wildlife Society
Bulletin 29: 413-424.
Vanderwel, M.C., J.R. Malcom, S.C. Mills. 2007. A meta-analysis of bird responses to
uniform partial harvesting across North America. Conservation Biology 21: 12301240.
Van Lear, D.H., and T.A. Waldrop. 1989. History, uses, and effects of fire in the southern
Appalachians. General Technical Report SF-54. Asheville, NC: USDA Forest
Service.
Weakland, C.A., P.B. Wood, W.M. Ford. 2002. Responses of songbirds to diameter-limit
cutting in the central Appalachians of West Virginia, USA. Forest Ecology and
Management 155: 115-129.
Wilson, C.W., R.E. Masters, G.A. Bukenhofer. 1995. Breeding bird response to pinegrassland community restoration for red-cockaded woodpeckers. Journal of
Wildlife Management 59: 56-67.
Woinarski, J.C.Z. 1990. Effects of fire on the bird communities of tropical woodlands
and open forests in northern Australia. Australian Journal of Ecology 15: 1-22.
91
CHAPTER 4
HOME RANGE SIZE, HABITAT USE, AND REPRODUCTIVE SUCCESS OF
HOODED WARBLERS AND WORM-EATING WARBLERS IN THINNED AND
BURNED FOREST STANDS
Introduction
Avian community parameters such as abundance and species richness provide
information about how silvicultural treatments affect bird community composition and
structure; however these measurements are not necessarily indications of reproductive
success and fitness (Vickery et al. 1992a). For a more thorough understanding of how
the silvicultural treatments affect the fitness of individual birds, a more detailed
examination of bird behavior and reproduction has been recommended (Thompson et al.
2000). Observations of space use by monitoring and estimating home range and core
area are indirect approaches to examine factors that affect individual fitness. Observing
nesting behavior and estimating nest success provides a more direct measure of
individual bird fitness.
Home range has been defined as “that area traversed by the individual in its normal
activities of food gathering, mating, and caring for young” (Burt 1943). For birds, as
most species, home range size is a function of many factors including food availability,
cover and protection from predators, nest site availability, availability of singing perches,
bird density, competition, and the age and size of the bird (Fretwell and Lucas 1970,
92
Mazarolle and Hobson 2004, Petit and Petit 1996, Smith and Shugart 1987, Wilson
1979). Home range size varies in differing quality habitats, with poor quality habitat
resulting in the need for larger territory size and thus, lower bird density at a site
(Mazarolle and Hobson 2004).
The objectives of this portion of the study were to test for differences in home
range, habitat use, and nesting success of two focal songbird species among different
silvicultural treatments used to reduce tree basal area in Bankhead National Forest,
Alabama.
Study Area and Methods
Study Area
The study was located in the northern third of William B. Bankhead National
Forest (Fig. 4.1), located in Lawrence and Winston counties, northwestern Alabama.
Bankhead National Forest (BNF) is a 72,800 ha multi-use forest located in the southern
Cumberland Plateau (Gaines and Creed 2003). The physiography of this region consists
of flat plateaus dissected by small valleys. The forests in this region have a diverse
species composition due to a variety of past disturbances – agriculture in the 1800s,
heavy cutting and wildfire in the early 1900s, fire suppression in the last decade and the
recent infestation of the southern pine beetle (Dendroctonus frontalis Zimmerman)
(Gaines and Creed 2003). In the 1930’s, abandoned farm land and other open lands were
reestablished with loblolly pine, Pinus taeda Linnaeus (Gaines and Creed 2003). This
93
Figure 4.1. Location of study plots in Bankhead National Forest, AL.
Burning treatment
Burn
No Burn
Control
3
3
Thin
6
6
Figure 4.2. Experimental design: two-factor, randomized
complete block design. Treatments include two burn
treatments and two thinning levels.
has resulted in 31,600 ha of loblolly pine throughout BNF. Once established, intensive
pine plantation management was not implemented, and subsequently, a variety of
hardwood species voluntarily invaded these sites. Over the past decade, southern pine
beetle infestations have killed a major portion of loblolly pine, increasing fuel loads and
the risk of wildfires (Gaines and Creed 2003). Bankhead National Forest has initiated a
Forest Health and Restoration Project to promote healthy forest growth via thinning and
fire disturbance. The thinning and fire prescriptions were administered to return the
forest to a mixed oak-pine upland ecosystem. My research was conducted in conjunction
with Bankhead National Forest’s restoration project.
94
The study design consisted of a randomized complete block design with two
factors – three thinning levels (no thin, 11 m2 ha-1 residual basal area, and 17 m2 ha-1
residual basal area) and two burn treatments (no burn and burn). Each treatment was
replicated three times and were blocked by year. Treatments were assigned randomly to
delineated stands. After the treatments were completed, I decided to collapse the thinned
treatments together because there was no difference in basal area between the two
thinning levels (analysis of variance, F = 0.066, df = 1, p = 0.8). Although this was not
the response variable I was studying, the variability within the individual treatment levels
was uneven, with some stands within the same treatment level having greater BA than the
target and others having lower BA than the target. This variation was too large to detect
any difference between the thinning levels. This resulted in three replicates each of the
control and burn, and six replicates each of the thin and the thin/burn (Fig. 4.2). The
research stands were located on upland sites composed of 20 to 35 year old loblolly pine.
Stands were comprised of a minimum of sixty percent pine (loblolly pine or Virginia
pine, P. virginiana Mill.), with the remainder mainly oak species (Quercus spp.).
Average stand size was 12 ha and plots had similar age and stand density. Thinning
favored the retention of hardwood species and was done before fire prescriptions.
Prescribed burning was completed in the dormant season (January – March) with lowburning surface fires.
Treatments on block one were completed between August 2005 and February
2006; blocks two and three were treated between April 2006 and March 2007. Posttreatment data were collected from block one between April and August 2006, and from
blocks two and three between April and August 2007.
95
Target Species
I selected the hooded warbler (Wilsonia citrina Boddaert) and the worm-eating
warbler (Helmitheros vermivorus Gmelin) as target species. These two species were
relatively common in the stands before treatment (see Chapter 2), are insectivorous, and
their nests are relatively easy to find. Hooded warblers are small (11 g) forest-interior
birds that inhabit mixed hardwood forests and cypress-gum swamps (Evans-Ogden and
Stutchbury 1994). They use small clearings and a shrub understory for nesting and often
place nests at the forest edge or in tree fall gaps (Evans-Ogden and Stutchbury 1994).
Worm-eating warblers are also small (12-14 g) forest-interior birds that inhabit
deciduous and mixed forests (Hanners and Patton 1998). The primary factor in their
presence in an area is the availability of slopes for nesting (Hanners and Patton 1998).
Sampling
Radiotelemetry. I captured males of each target species using song playback to
attract them into mistnets. I banded each captured bird with a U.S. Fish and Wildlife
Service aluminum band and plastic color bands to aid in individual identification. I also
attached a radio transmitter (model BD-2, 0.065 g [4-5% of body mass], battery life:
approx. 21 days, Holohil Systems, ltd.) to the back of selected males using a figure-8
harness made with cotton thread (Rappole and Tipton 1991) and secured with glue
(Krazy Glue, Columbus, OH). I released all birds immediately after processing and
waited 24 h to track them to allow for adjustment to leg bands and radio transmitters.
I used burst sampling methods (Barg et al. 2005) when tracking birds, recording
bird locations at 60 second intervals for a total of 30 points per session. If the bird was
lost during the session, I temporarily stopped recording until the bird could be relocated.
96
Each session lasted between 30 and 80 minutes. I tracked each bird every three to four
days, performed as many sessions as possible before the transmitter battery died. Each
location was recorded using a handheld global positioning system (GPS) (eTrex Vista,
Garmin Ltd.) and was downloaded into ArcGIS v. 9.1 (ESRI) for analysis.
Reproductive success. I searched for nests and watched for indications of
reproductive success while radio tracking. When a nest was found, I recorded its location
using GPS and marked the nest with flagging at least 3m away. I monitored nests based
on the standard protocol of Ralph et al. (1993). Nests were checked every 3-4 days and
more often when fledging was expected. At each nest check, I recorded the activity of
the adult, contents of nest, and age of young, as well as any indications of predation or
nest parasitism. To reduce the risk of increasing predation and nest parasitism, I took
precautions detailed in Martin and Geupel (1993) and Ralph et al. (1993).
Microhabitat. I measured microhabitat variables at the end of each breeding
season at 11.3 m radius (0.04 ha) circular plots, using methods based on James and
Shugart (1970). I completed three habitat plots at random points within each territory
and outside of each territory (Smith and Shugart 1987). Random points were determined
using Hawth’s tools random point generator. I also completed a habitat plot at each nest
location. At each plot I recorded the species, size class (7.6-33 cm, 33.1-58.4 cm, 58.583.8, >83.9 cm diameter at breast height [DBH]). At a smaller plot (5 m radius) within
the larger plot I tallied the number of woody and herbaceous stems. Within a 1 m radius
circle within the 5 m circle I recorded percent ground cover. I recorded aspect, percent
slope (using a percent scale clinometer, to nearest percent), percent canopy cover (using a
convex spherical densitometer, to nearest percent) and litter depth (to nearest mm) at the
97
center point of the plot. At the four cardinal directions 5 m from the plot center I
measured percent canopy cover (using a convex spherical densitometer, to nearest
percent) and percent vertical cover from ground level to 3 m (using a checkered drop
cloth, to nearest percent).
Data analysis
Home range delineation. I only used birds that had greater than three observation
sessions and 70 individual observation points for the home range analyses to ensure an
adequate sample size. For each bird, I selected the home range model that had the most
support from the data based on likelihood cross-validation (Horne and Garton 2006)
using Animal Space Use v.1.2 (Horne and Garton 2007). The “best” model was defined
as the model with the smallest Kullback-Leibler distance (i.e., difference between actual
and estimated distributions). This approach allows information-theoretic model selection
to assist in deciding which approximating model is closest to the underlying distribution
of the individual, using the data collected from that individual (Horne and Garton 2006).
The home ranges models I used in the selection process were the exponential power
model, one mode bivariate normal, two mode bivariate circular, two mode bivariate
normal, fixed kernel density, and adaptive kernel density. The exponential power model
is a simple model that is circular in shape and flat on top. It is the most economical home
range, where the animal uses the habitat within the home range uniformly (Horne and
Garton 2006). For the bivariate circular model, the center of activity is first calculated
using the means of all locations. The radius (r) is then calculated from this center of
activity. This model assumes observations decline in density from the center to the edge
of the distribution (Calhoun and Casby 1958). An improvement on this method is the
98
bivariate normal model which applies an elliptical approximation. This model is based on
the assumption of an underlying normal, or bell-shaped, distribution (Jennrich and Turner
1969). Kernel density estimators are equitable to the histogram, the oldest density
estimator. Kernels or ‘bumps’ are calculated based on number of observations in a given
location; the more observations, the ‘higher’ the bump will be and the greater the
probability of the animal occurring in that location (Worton 1989). Animal Space Use
v.2.1 (Horne and Garton 2007) was used to create the home ranges after I chose the
model. Core areas were defined as the area within the 95% home range boundary with
the probability of use greater than that expected from a uniform distribution of use
(Samuel et al. 1985). This was the location where the difference between the actual
home range and the expected uniform home range was the greatest. To test for home
range and core area size differences among treatments I used a Kruskall-Wallace test
(SPSS v.15.0)
Reproductive success. Sample size was too small to evaluate nesting success
using the Mayfield index (Mayfield 1961). I instead used the method developed by
Vickery et al. (1992b) to create a reproductive index based on behaviors indicative of
different stages of the breeding cycle. Behaviors were ranked as follows: (1) territorial
male present > 4 weeks or a nest parasitized by a brown-headed cowbird, (2) territorial
male and female present > 4 weeks, (3) evidence of a nest, and (4) evidence of fledging
success. Nest success was defined as fledging at least one young. This method requires
that sampling efforts are equal among individuals to avoid bias (Vickery et al. 1992b). I
was unable to do so because of manpower limitation, so the presented results should be
99
regarded as coarse estimates. I tested for differences in reproductive index among
treatments with a Kruskall-Wallace test (SPSS v.15.0).
Microhabitat. I averaged all variables across three habitat plots for each bird
territory and paired random locations. To reduce the number of variables in the model
and avoid multicollinearity, I use principle components analysis (PCA, SPSS v.15.0).
From each component, I chose one variable to represent that component in the analysis. I
used these variables instead of the principle components so that the results would be
easier to interpret. I inspected variables for multicollinearity using Pearson’s correlation
matrix and for multivariate outliers using Mahalanobis distance. I then used pairwise
logistic regression (PROC LOGISTIC, SAS v. 9.1) to create a set of habitat models for
each species. To assess the relative degree of fit for each model, I used Akaike’s
Information Criteria (Burnham and Anderson 1998). The best model was selected by
judging the degree of support as indicated by delta AIC and normalized Akaike weights.
Models with delta AIC < 2 were considered to have substantial support (Burnham and
Anderson 1998).
Results
Home range. Over the two years, 23 birds were tracked (13 hooded warblers and
10 worm-eating warblers). The treatment with the most individuals of either species was
the thinned stands, and I was not able to track any birds on the burned stands because of
the low occurrence of the species on the burned stands (Table 4.1). The average number
100
Table 4.1. Distribution of radio-tracked hooded warblers (HOWA) and
worm-eating warblers (WEWA) among silvicultural treatments in
Bankhead National Forest, AL, one year post-treatment, 2006-2007.
Treatment
HOWA
WEWA
Total
Control
2
2
4
Burn
0
0
0
Thin
8
4
12
Thin/Burn
3
4
7
Total
13
10
23
101
of observation points of each bird was 128, from an average of 5.6 observation sessions
per bird. The exponential power model was the ‘best’ model for 15 birds (nine hooded
warblers and six worm-eating warblers), bivariate normal model for seven birds (three
hooded warblers and four worm-eating warblers), and the two mode bivariate circular
model for one bird (hooded warbler). Neither home range nor core area differed among
treatments for either species (Table 4.2). Home ranges for hooded warblers ranged from
3.4 ha to 13.1 ha and from 4.3 ha to 8.8 ha for worm-eating warblers (Table 4.2). Home
ranges overlapped; however core areas did not overlap in the same species. All core
areas were located off the treated stands but in some cases, parts of the home range were
located on the stand (Fig. 4.3).
Microhabitat. The original 18 variables showed low multivariate correlation
(Kaiser-Meyer-Olkin [KMO] Measure of Sampling Adequacy = 0.226) so I removed the
two variables with the lowest multivariate correlation (log cover and rock cover) from the
principle component analysis to increase the KMO measure of sampling adequacy to
0.552. The principle component analysis on the remaining sixteen habitat variables
generated seven principle components (PCs) with eigen values greater than one. The first
was represented by litter cover, the second by number of woody stems, the third by
herbaceous cover, the fourth by litter depth, the fifth by bare ground cover, the sixth by
vertical cover, and the seventh by tree abundance (Table 4.3). All retained components
had an eigen value greater than 1. The seven components retained approximately 73% of
the original variation (Bartlett’s Test of Sphericity χ2 = 405.657, df = 120, p < 0.001).
Log cover showed high multicollinearity, so it was eliminated from the analysis and rock
cover was used to represent ground debris. The variables included in the analyses were
102
Table 4.2. Significance values, mean, and standard error among silvicultural treatment
for home range and core area (ha) of hooded warblers (HOWA) and worm-eating
warblers (WEWA) in Bankhead National Forest, AL, 2006-2007.
Kruskall-Wallace test
Mean by treatment
χ2
df
p
Control
Thin
Thin/Burn
Home Range
0.92
2
0.63
5.44±.5.44
13.05±6.16
3.41±3.01
Core Area
0.66
2
0.72
2.19±2.08
5.84±2.75
2.60±2.41
Home Range
1.32
2
0.52
4.38±0.53
4.53±2.61
8.81±8.81
Core Area
0.94
2
0.62
1.88±0.34
2.15±1.19
3.01±1.21
Variable
Hooded Warbler
Worm-eating Warbler
103
Figure 4.3. Example of 95% probability home range locations in relation to treatment plot
for hooded warbler (HOWA) and worm-eating warbler (WEWA). Inner darker portion
of home range is the bird’s core area. Small circles are individual bird locations.
104
Table 4.3. Principle component analysis loadings, eigenvalues, and percent
variance for microclimate variables following silvicultural treatment in
Bankheand National Forest, AL, 2006-2007.
Tree abundance
Tree spp. Richness
# snags
# woody stems
# herb stems
Bare ground
Litter cover
Rock cover
Log cover
Herb cover
Fern cover
Woody cover
Tree cover
Slope
litter depth
Basal area
Canopy cover
Vertical cover
Eigenvalue
% Variance
PC1
0.205
0.643
0.166
0.011
-0.398
-0.464
0.818
0.082
-0.373
-0.392
-0.039
-0.034
0.301
0.399
0.468
0.684
0.755
-0.115
3.35
18.60
PC2
0.669
-0.450
-0.054
0.645
0.220
0.137
-0.026
-0.148
-0.396
0.002
0.048
0.645
-0.089
0.071
-0.277
0.203
0.194
-0.497
2.15
11.94
PC3
0.153
0.148
-0.223
-0.494
0.686
-0.225
-0.138
0.034
-0.232
0.750
0.179
-0.093
0.035
0.080
0.062
0.143
0.303
-0.074
1.67
9.26
105
PC4
0.160
0.057
0.172
0.025
-0.028
0.333
0.066
0.407
-0.192
-0.071
0.307
-0.467
-0.458
0.223
-0.609
0.111
0.193
0.283
1.46
8.09
PC5
0.057
0.280
-0.285
-0.019
-0.065
0.629
-0.315
0.037
-0.246
-0.018
-0.418
-0.057
0.347
-0.256
-0.008
0.360
0.038
0.128
1.24
6.86
PC6
0.353
-0.090
-0.672
-0.032
-0.017
-0.056
0.033
-0.255
0.308
-0.249
0.340
0.002
0.027
0.052
0.043
0.063
0.067
0.460
1.15
6.41
PC7
0.165
-0.026
0.388
-0.201
-0.072
0.211
-0.043
-0.440
0.024
-0.082
0.519
-0.242
0.297
-0.330
-0.015
0.049
0.052
-0.233
1.06
5.87
the above seven variables (which represented the corresponding principle component)
and rock cover. I used the following six logistic regression models in the AIC analysis:
(1) full model (all variables); (2) all variables except rock cover; (3) foliage ground
cover: woody stems and herbaceous cover; (4) ground cover: litter depth, litter cover,
bare ground; and (5) structure: vertical cover and tree abundance. I included a sixth
model (slope) for worm-eating warblers because slope is known to be an important
component for nesting (Evans-Ogden and Stutchbury 1994). When rock cover was
included in the model with all the variables, there appeared to be high multicollinearity
with the other variables (indicated by high confidence intervals). The best fitting models
for the worm-eating warbler was the structure model (model # 1), containing vertical
cover and tree abundance variables and the slope model (model # 2) (Table 4.4). The
models containing (1) all variables and (2) all variables except rock cover were the best
fitting models for the hooded warbler (Table 4.4). The model without rock cover was
slightly better than the model with all variables (Table 4.4).
Reproductive success. Fourteen nests were found over the two years; 11 wormeating warbler nests and three hooded warbler nests. Only five of the nests were located
in treated stands, the others were located adjacent to stands, but in areas that were not
treated. The two worm-eating warbler nests on the burned stands were successful and the
two nests on the thinned stands fledged only brown-headed cowbirds. The one nest
found on the control failed. Of the nests that were located adjacent to the stands, three
fledged young, five failed, and the outcome of one was unknown. For the 14 nests we
monitored, nest success rate was 36%.
106
Table 4.4. AIC scores for five logistic regression models of habitat preferences of hooded warblers and worm-eating warblers in
Bankhead National Forest, AL, 2006-2007.
Model
Delta
Akaike
P-value
AIC
AIC
weight
Worm-eating Warbler
107
1
Structure (tree abundance, vertical cover)
0.01
15.65
0.00
0.90
2
Slope (slope)
0.01
16.45
0.80
0.38
3
All variables (tree abundance, woody stems, bare ground cover, litter ground
cover, rock ground cover, herbaceous ground cover, litter depth, vertical cover)
0.03
20.94
5.29
0.06
4
All variables except rocks
0.09
23.73
8.08
0.02
5
Foliage ground cover (woody stems, herbaceous ground cover)
0.32
23.90
8.25
0.01
6
Ground cover (bare ground cover, litter ground cover, litter depth)
0.65
26.53
10.88
0.00
<0.01
14.07
0.00
0.68
cover, rock ground cover, herbaceous ground cover, litter depth, vertical cover)
<0.01
16.05
1.98
0.25
3
Structure (tree abundance, vertical cover)
0.03
19.11
5.05
0.05
4
Foliage ground cover (woody stems, herbaceous ground cover)
0.12
21.93
7.86
0.01
5
Ground cover (bare ground cover, litter ground cover, litter depth)
0.99
28.10
14.03
0.00
Hooded Warbler
1
All variables except rocks
2
All variables(tree abundance, woody stems, bare ground cover, litter ground
107
I used the behavior of thirty-nine birds (21 worm-eating warblers and 18 hooded
warblers) to calculate reproductive indices. Reproductive index was not different among
treatments for either species (χ2=0.67, df =2, P = 0.72 for worm eating warblers; χ2=5.10,
df =2, P = 0.08 for hooded warblers).
Discussion
Hooded and worm-eating warblers appear to prefer habitat that is adjacent to the
treated stands in this study. No bird’s home range was entirely located in any of the
treated stands, although some birds had territories that were partially located on the
treated stands. None of the birds’ core areas were located in the stands, indicating that
the stands do not provide habitat crucial for either species. The majority of the home
ranges for both species were located in the small ravines and gullies that surround all of
the stands. These were areas within the boundaries of the treatment that were not treated,
or areas adjacent to the treatment. There was no difference in home range size or core
area among treatments for either species; however it may not be appropriate to make
comparisons among the treatments since the majority of the home ranges were not
located in treated stands.
The three probability distributions (exponential power, one mode bivariate
normal, and one mode bivariate circular) that were the best fit for the two warbler species
describe distributions of animals inhabitating homogeneous habitats (Van Winkle 1975).
This indicates birds are using habitat within their home range in a spatially uniform
manner. The bivariate normal model is characterized by a distribution that is elliptical,
where occurrence probability varies in direction from the home range center (Van Winkle
108
1975). The exponential power and bivariate circular models are described by the
probability of occurrence at distances from the home range center, without regard to
direction (Horne and Garton 2006, Koeppl et al. 1975). Thus, habitat use by these two
species is relatively uniform within the home range and there are not multiple areas of
concentrated use.
Our home range size estimations for hooded warblers were larger than territory
sizes previously reported. There are no studies that report home range size estimates for
hooded warblers. In Pennsylvania, average territory size was 0.88 ha (Howlett and
Stutchbury 1997). They estimated territory size by recording locations of singing males
but failed to describe the method used for delineating territories. Since they estimated
territories based on singing birds, it is expected that their estimates would be smaller than
the home range estimates in this study. Norris and Stutchbury (2001) recorded extraterritorial movements by male hooded warblers and found that they will travel up to 465
m, usually to attain extra-pair copulations with neighboring females. Roughly combining
this distance with their previous estimates of singing territory size (the two studies were
in the same location) provides insight into the home range size of birds in their study
area. These approximations are close to the home range size estimates in this study (3–13
ha).
Hooded warblers are socially monogamous birds, however extra-pair matings
(revealed via DNA fingerprinting) are a common and important component of their
mating system (Evans-Ogden and Stutchbury 1994). When hooded warbler densities are
low, as they are in our study area, it is expected that home range size would increase due
to males traveling farther in search of extra pair copulations.
109
Worm-eating warbler home ranges were larger than territory size previously
reported in the literature. Hanners and Patton (1998) reported a mean territory size of
1.72 ha in Connecticut, although they did not indicate how the data were collected or how
territories were delineated. If it was based on the bird’s singing territory, I would expect
it to be slightly smaller than the bird’s home range. Their measure may coincide more
appropriately with the core areas delineated in my study.
Home range size and quality can be influenced by a variety of factors, including
food abundance; predator abundance; bird density; and availability of nest sites
(Mazarolle and Hobson 2004, Marshall and Cooper 2004, Petit and Petit 1996, Smith and
Shugart 1987). Generally, poor quality habitat results in the need for a larger home range
(Mazarolle and Hobson 2004). Since there is little information available about home
range size for these two species, it is difficult to make assumptions about the quality of
their habitat. Based on rough estimates, it appears that home ranges coincide with other
studies.
Hooded warblers are considered forest interior species, but also have
requirements associated with forest openings and canopy gaps (Evans-Ogden and
Stutchbury 1994, Kilgo et al. 1996, Moorman et al. 2002, Whittam et al. 2002). It
appears that hooded warblers are choosing habitat based on many variables in this study.
Their probability of occurrence increases when herbaceous ground cover and vertical
cover increases. Other studies have reported an association of hooded warblers with
dense understory vegetation (Mooreman et al. 2002, Whittam et al. 2002). Kilgo et al.
(1996) reported that hooded warblers in South Carolina preferred shrubs with thicketforming properties that provide protection from weather and predators. It is likely that as
110
shrub recovery advances and provides more resources, hooded warbler densities will
increase.
Habitat structure in the forest understory and slope appear to be important in
predicting the presence of worm-eating warblers. In Missouri, worm-eating warblers
preferred large forests with dense understory (Wenny et al. 1993). Watts and Wilson
(2005) concluded that dense shrub cover was more important for worm-eating warbler
habitat than plant composition or any other factor, including slope. Slope is considered to
be an important factor in the presence of worm-eating warblers because they only place
their nests on slopes (Gale et al. 1997, Hanners and Patton 1998, Wenny et al. 1993).
The nests I found were all located on some sort of slope, but not necessarily a long slope.
We found nests on the slopes along side logging roads and on slopes created by up-rooted
trees, as well as hillside slopes. In coastal North Carolina, a very flat terrain, wormeating warblers are present in very high densities (Watts and Wilson 2005). It appears
that dense shrub cover is an important habitat feature for worm-eating warblers and the
presence of slope is secondarily important.
Reproductive estimates for both species were low; however, small sample size
and deviation from standard protocols (i.e., not spending equal amounts of time
observing each bird) may have biased these results. For the 14 nests I found, nest success
rate was 36%, considerably lower than nest success reported in other studies. Reported
nest success rates are around 50% for hooded warblers (Howlett and Stutchbury 1997,
Moorman et al. 2002, Whittam et al. 2002) and 76% for worm-eating warblers (Gale et
al. 1997). Of the five nests found in the stands (all worm-eating warbler nests), only two
nests fledged young. Both of these nests were in burned stands. The nests on the thinned
111
stands fledged only cowbirds and the nest found on the control failed due to predation.
After the nest failed, the pair could not be located in the stand and appeared to have left
the area completely.
Average reproductive index shows that there were limited observations of
reproductive behavior, especially in hooded warblers. Evidence of hooded warbler
reproduction was seen only on and adjacent to thinned stands. I never observed any
reproductive evidence other than the presence of territories for birds on or adjacent to the
untreated or thinned/burned stands. Evidence of worm-eating warbler reproduction was
seen on or adjacent to all treated stands, indicating that they may be more flexible in their
choice of nesting habitat than hooded warblers. Neither species were seen with fledglings
in the treated stands, only adjacent to them, so again, it may not be appropriate to make
comparisons among the treatments.
Stuart-Smith and Hayes (2003) found that increased density of residual trees
increased the odds of predation of artificial nests. Other studies have also suggested that
greater structural diversity in regenerating plots may provide habitat for a greater number
of avian nest predators (Barber et al. 2001). There was greater structural diversity in
thinned stands in our study as well, so this may have resulted in the same effect. More
nest predators would naturally lead to lower nest success rates. High structural diversity
also creates more perching sites for use by brown-headed cowbirds (Molothrus ater
Boddaert). Cowbirds use perches in above an open forest canopy to watch and locate
nests to parasitize (Brittingham and Temple 1996, Evans and Gates 1997). Most female
cowbirds travel less than three km between their agricultural feeding grounds and
breeding areas, although some will travel farther (Goguen and Mathews 1999). This
112
results in higher cowbird densities in fragmented landscapes. Although cowbird
abundance increased slightly in the stands after thinning, density is still relatively low. In
this study, stands are located within a large contiguous forest which may minimize the
presence of cowbirds, however if more areas are thinned and more edges are created, it is
likely the cowbird density will increase (Evans and Gates 1997).
Bankhead National Forest is a large contiguous forest; the treatments examined in
this study created relatively small openings and were not the dominant landscape feature.
The landscape mosaic created by cutting small portions of forest may provide hooded
warblers and worm-eating warblers with the types of habitat they require in the short
term. By staggering treatments over time and space, continuous habitat would be
provided for these species.
Bibliography
Barber, D.R., T.E. Martin, M.A. Melchiors, R.E. Thill, T.B. Wigley. 2001. Nesting
success of birds in different silvicultural treatments in Southeastern US pine
forests. Conservation Biology 15: 196-207.
Barg, J. J., J. Jones, and R. J. Robertson. 2005. Describing breeding territories of
migratory passerines: suggestions for sampling, choice of estimator, and
delineation of core areas. Journal of Animal Ecology 74: 139-149.
Brittingham, M.C. and S.A. Temple. 1996. Vegetation around parasitized and nonparasitized nests within deciduous forest. Journal of Field Ornithology 67: 406413.
Burnham, K.P., and D.R. Anderson. 1998. Model selection and inference: A practical
information-theoretic approach. Springer-Verlag New York, Inc., New York,
NY.
Burt, W.H. 1943. Territoriality and home range concepts as applied to mammals. Journal
of Mammalogy 24: 346-352
113
Calhoun, J.B. and J.U. Casby. 1958. Calculation of home range and density of small
mammals. Public Health Monograph No. 55. U.S. Public Health Service,
Washington, D.C.
Evans, D.R. and J.E. Gates. 1997. Cowbird selection of breeding areas: The role of
habitat and bird species abundance. Wilson Bulletin 109: 470-480.
Evans-Ogden, L.J. and B.J. Stutchbury. 1994. Hooded warbler (Wilsonia citrina). In: The
Birds of North America, No. 110 (A. Poole and F. Gills, eds.). The Birds of North
America, Inc., Philadelphia, PA.
Fretwell, S.D. and H.L. Lucas. 1970. On territorial behavior and other factors
influencing habitat distribution in birds. I. Theoretical development. Acta
Biotheoretica 19: 16-36.
Gaines, G.D. and J.W. Creed. 2003. Forest health and restoration project. National
forests in Alabama, Bankhead National Forest Franklin, Lawrence and Winston
Counties, Alabama. Final environmental impact statement. Management Bulletin
R8-MB 110B.
Gale, G.A., L.A. Hanners, S.R. Patton. 1997. Reproductive success of worm-eating
warblers in a forested landscape. Conservation Biology 11: 246-250.
Goguen, C.B., and N.E. Mathews. 1999. Brown-headed cowbird behavior and
movements in relation to livestock grazing. Ecological Applications 11: 15331544.
Hanners, L.A. and S.R. Patton. 1998. Worm-eating Warbler (Helmitheros vermivorus).
In: The Birds of North America, No. 367 (A. Poole and F. Gill, eds.). The Birds
of North America, Inc., Philadelphia, PA.
Howlett, J.S. and B.J.M. Stutchbury. 1997. Within-season dispersal, nest-site
modification, and predation in renesting hooded warblers. Wilson Bulletin 109:
643-649.
Horne, J.S. and E.O. Garton. 2006. Selecting the best home range model: An
information-theoretic approach. Ecology 87: 1146-1152.
Horne, J.S. and E.O. Garton. 2007. Animal Space Use 1.2.
http://www.cnr.uidaho.edu/population_ecology/animal_space_use.htm
James, F.C. and H.H. Shugart. 1970. A quantitative method of habitat description.
Audubon Field Notes 24: 727-736.
114
Jennrich, R.I. and F.B. Turner. 1969. Measurement of non-circular home range. Journal
of Theoretical Biology 22: 227-237.
Kernohan, B.J., R.A. Gitzen, J.J. Milspaugh. 2001. Analysis of animal space use and
movements. In: Radio Tracking and Animal Populations (J.J. Milspaugh and J.M.
Marzluff, eds.). Academic Press, New York, New York, USA. Pp 125-166.
Kilgo, J.C., R.A. Sargemt. B.R. Chapman, K.V. Miller. 1996. Nest-site selection by
hooded warblers in bottomland hardwoods of South Carolina. Wilson Bulletin
108: 53-60.
Koeppl, J.W., N.A. Slade, R.S. Hoffmann. 1975. A bivariate home range model with
possible application to ethological data analysis. Journal of Mammology 56: 8190.
Marshall, M.R., and R.J. Cooper. 2004. Territory size of a migratory songbird in response
to caterpillar density and foliage structure. Ecology 85: 432-445.
Martin, T.E. and G.R. Guepel. 1993. Nest-monitoring plots: Methods for locating nests
and monitoring success. Journal of Field Ornithology 64: 507-519.
Mayfield, H. F. 1961. Nesting success calculated from exposure. Wilson Bulletin
73:255-261.
Mazerolle, D.F. and K.A. Hobson. 2004. Territory size and overlap in male Ovenbirds:
contrasting a fragmented and contiguous boreal forest. Canadian Journal of
Zoology 82: 1774-1781.
Moorman, C.E., D.C. Guynn, Jr., J.C. Kilgo. 2002. Hooded warbler nesting success
adjacent to group-selection and clear cut edges in a southeastern bottomland
forest. The Condor 104: 366-377.
Norris, R.D. and B.J.M. Stutchbury. 2001. Extra-terrtiorial movements of a forest
songbird in a fragmented landscape. Conservation Biology 15: 729-736.
Petit, L.J. and D.R Petit. 1996. Factors governing habitat selection by prothonotary
warblers: Field tests of the Fretwell-Lucas models. Ecological Monographs 66:
367-387.
Rodgers, A.R., A.P. Carr, L. Smith, and J.G. Kie. 2005. HRT: Home Range Tools for
ArcGIS. Ontario Ministry of Natural Resources, Centre for Northern Forest
Ecosystem Research, Thunder Bay, Ontario, Canada. [Online.] Available at
http://blue.lakeheadu.ca/hre.
Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. DeSante. 1993. Handbook of
field methods for monitoring landbirds. General Technical Report PSW-GTR115
144, Albany, CA: Pacific Southwest Research Station, Forest Service, U. S.
Department of Agriculture.
Rappole, J.H. and A.R. Tipton. 1991. New harness design for attachment of radio
transmitters to small passerines. Journal of Field Ornithology 62: 335-337.
Samuel, M.D., D.J. Pierce, E.O. Garton. 1985. Identifying areas of concentrated use
within the home range. Journal of Animal Ecology 54: 711-719.
Smith, T.M. and H.H. Shugart. 1987. Territory size variation in the Ovenbird: The role
of habitat structure. Ecology 68: 695-704.
Stuart-Smith, A.K. and J.P. Hayes. 2003. Influence of residual tree density on predation
of artificial and natural songbird nests. Forest Ecology and Management 183:
159-176.
Van Winkle, W. 1975. Comparison of several probabilistic home-range models. Journal
of Wildlife Management 39: 118-123.
Vickery, P.D., M.L. Hunter and J.V. Wells. 1992a. Is density an indicator of breeding
success? The Auk 109: 706-710.
Vickery, P.D., M.L. Hunter and J.V. Wells. 1992b. Use of a new reproductive index to
evaluate relationship between habitat quality and breeding success. The Auk 109:
697-705.
Watts, B.D. and M.D. Wilson. 2005. The use of pine plantations by worm-eating
warblers in coastal North Carolina. Southeastern Naturalist 4: 177-187.
Wenny, D.G., R.L. Clawson, J. Faaborg, S.L. Sherrif. 1993. Population density, habitat
selection, and minimum area requirements of three forest-interior warblers in
Central Missouri. The Condor 95: 968-979.
Whittam, R.M., J.D. McCracken, C.M. Francis, M.E. Gartshore. 2002. The effects of
selective logging on nest-site selection and productivity of hooded warblers in
Canada. Canadian Journal of Zoology 80: 644-654.
Wilson, E.O. 1979. Sociobiology. Belknap, Cambridge, Massachusetts, USA.
Worton, B.J. 1989. Kernel methods for estimating the utilization distribution in home
range studies. Ecology 70: 164-168.
116
CHAPTER 5
CONCLUSION AND MANAGEMENT RECOMMENDATIONS
No forest management practice can be generalized as either detrimental or
beneficial; some species respond positively and other respond negatively. Change in the
bird community will continue as forest succession progresses. The results from my
study suggest that early-successional, interior/edge habitat specialist responded positively
to thinning treatments in the short term. To a lesser extent, burning treatments also
resulted in a positive response by early successional, interior/edge species. While
interior/edge species responded positively, neither interior nor open/edge species
responded negatively. The change in habitat available to the bird community was not so
drastic that interior species were no longer able to use the habitat.
Pre-treatment characteristics consisted of thick pine stands with gaps and edges
created by SPB damage, power line throughways, roads, and fields. These conditions
provided habitat for many forest interior and interior/edge bird species. One year after
thinning, stands were more open and provided habitat for interior/edge and edge/open
species. Thinned stands were characterized by an understory that provided habitat for
many shrub nesting species. Prescribed burning had a smaller effect on the habitat and
bird community. Burned stands were similar in habitat and bird community to the
untreated plots. Because all trees were retained in these stands, the bird community
remained fairly consistent between pre and post treatment. Thinning and burning in
117
combination created a habitat that was quite similar to the thinned stands, except that
burned/thinned stands had a greater loss of litter cover. There was an increase in edge
and open habitat species in thinned and burned stands and a decrease in foliage foraging
species. Thinned/burned stands were similar to thinned stands in habitat and bird
community. Burning only affected the litter layer, decreasing the amount of litter ground
cover and the litter depth. Neither of theses habitat characteristics were highly correlated
with the bird community in this study, so there was a limited impact on the bird.
Greenberg et al. (2007) and Stribling and Barron (1995) found that understory fuel
reduction treatments (low intensity fire and understory removal) have few detectable
effects on breeding birds. Shrub nesters were most common on light intensity burns in
other studies (Greenberg et al. 1995, Stribling and Barron 1995, Wilson et al. 1995). It
appears that thinning has a greater impact on the bird community than does low-intensity
burning. My results suggest that, one year after treatment, thinning creates habitat for
interior/edge species (species that use early-successional habitat) without displacing
many of the interior forest birds. Other studies have also come to this conclusion
(Campbell et al. 2007, Costello et al. 2000, Holmes and Pitt 2007, Vanderwel et al.
2007).
Individual use of the plots by hooded warblers and worm-eating warblers indicate
that, one year after treatment, the thinned stands create habitat that is used marginally by
these species, but is not causing either species to leave. Most of the home ranges were
located in the small ravines and gullies that surround the stands with only a small portion
of the bird’s home range in the treated stand. It appears that the habitat available in these
areas provide more suitable habitat than what is available in the treated stands. Habitat
118
that hooded warblers are using is characterized by high herbaceous ground cover and
vertical cover. Worm-eating warbler habitat is characterized by dense understory
vegetation and the presence of slopes. Since there are no previous studies examining
home range size of either species, I cannot make comparisons.
Management Recommendations
Management recommendations made as a result of this research are limited in
their scope. This research was conducted in a large contiguous forest on the southern
Cumberland Plateau; the treatments examined in this study created relatively small
openings and were not the dominant landscape feature. It is uncertain what effect the
treatments would have in a smaller, fragmented landscape; if the treatments were applied
over larger or smaller stands; or of the treatments were used outside of this ecoregion.
Although this is a well designed experimental study, the nature of wildlife studies
limits replication. Limited replication in this study may be a limiting factor in detecting
response to ecological processes (Marzluff et al. 2000).
I collected one year post treatment data for this study, and although I detected
some changes in the bird community, it will change over time as shrubs recover and other
habitat attributes continue to change. Long term data needs to be collected to examine
trends in the bird community as the plots recover from disturbance. There are few longterm studies of the effects of forest management on birds, and understanding long term
trends is as important as understanding short term trends (Marzluff et al. 2000, Thompson
et al. 2000). As succession progresses, the resources available to birds will differ, and
this will alter the bird community (Greenberg et al. 2007, Harrison et al. 2005, Holmes
119
and Pitt 2007, Jobes et al. 2004, Yahner 1997). Some studies suggest there is a lag time
between treatment implementation and changes in the bird community, so one year of
data can sometimes be misleading (Greenberg et al. 2007, Jobes et al. 2004).
With those caveats in mind, I recommend tree thinning as a viable option for
creating habitat for early-successional bird species. It is a particularly attractive method
if clear cutting is not an option and/or retaining interior forest birds is included a
management goal. However, thinned forests create early-successional habitat that is
ephemeral, so I recommend that thinning be used at different intervals across the
landscape to provide early successional habitat throughout the forest. The landscape
mosaic created by cutting small portions of forest may provide early successional species
with the types of habitat they require in the short term while also retaining many of the
interior forest birds. By staggering treatments over time and space, continuous habitat
would be provided for these species.
Bibiography
Campbell, S.P., J.W. Witham, M.L. Hunter, Jr. 2007. Long-term effects of groupselection timber harvesting on abundance of forest birds. Conservation Biology 21:
1218-1229.
Costello, C.A., M. Yamaski, P.J. Pekins, W.B. Leak, C.D. Neefus. 2000. Songbird
response to group selection harvests and clearcuts in a New Hampshire northern
hardwood forest. Forest Ecology and Management 127: 41-54.
Greenberg, C.H., L.D. Harris, D.G. Neary. 1995. A comparison of bird communities in
burned and salvaged-logged, clearcut, and forested Florida sand pine scrub. Wilson
Bulletin 107: 40-45.
Greenberg, C.H., A.L. Tomcho, J.D. Lanham, T.A. Waldrop, J. Tomcho, R.J. Phillips, D.
Simon. 2007. Short term effects of fire and other fuel reduction treatments on
120
breeding birds in a Southern Appalachian upland hardwood forest. Journal of Wildlife
Management 71: 1906-1916.
Harrison, R.B., F.K.A. Schmiegelow, R. Naidoo. 2005. Stand-level response of breeding
forest songbirds to multiple levels of partial-cut harvest in four boreal forest types.
Canadian Journal of Forest Resources 35: 1553-1567.
Holmes, S.B. and D.G. Pitt. 2007. Response of bird communities to selection harvesting
in a northern tolerant hardwood forest. Forest Ecology and Management 238: 280292.
Jobes, A.P., E. Nol, D.R. Voigt. 2004. Effects of selection cutting on bird communities in
contiguous eastern hardwood forests. Journal of Wildlife Management 68: 57-60.
Marzluff, J.M., M.G. Raphael, R. Sallabanks. 2000. Understanding the effects of forest
management practices on avian species. Wildlife Society Bulletin 28: 1132-1143.
Stribling, H.L., M.G. Barron. 1995. Short-term effects of cool and hot prescribed burning
on breeding songbird populations in the Alabama Piedmont. Southern Journal of
Applied Forestry 19: 18-22.
Thompson, F.R. III, J.D. Brawn, S. Robinson, J. Faaborg, R.L. Clawson. 2000.
Approaches to investigate effects of forest management on birds in eastern deciduous
forests: How reliable is our knowledge? Wildlife Society Bulletin 28: 1111-1122.
Vanderwel, M.C., J.R. Malcom, S.C. Mills. 2007. A meta-analysis of bird responses to
uniform partial harvesting across North America. Conservation Biology 21: 12301240.
Wilson, C.W., R.E. Masters, G.A. Bukenhofer. 1995. Breeding bird response to pinegrassland community restoration for red-cockaded woodpeckers. Journal of Wildlife
Management 59: 56-67.
Yahner, R.H. 1997. Long term dynamics of bird communities in a managed forested
landscape. Wilson Bulletin 109: 595-613.
121
VITA
Jill Wick, daughter of Norm and Diane Wick, was born May 11th 1979 in Two
Rivers, Wisconsin. She attended University of Wisconsin at Stevens Point for her
undergraduate education in the Wildlife Biology department and earned her BS in 2002.
She moved to Alabama in 2005 to begin her avian research and will graduate in May of
2008. She has since secured a position in Albuquerque New Mexico at an environmental
consulting firm.
122