THE EFFECTS OF SHELTERWOOD HARVESTING ON OAK
REGENERATION ONE AND TWO YEARS AFTER HARVEST
IN SOUTHERN OHIO
A Thesis
Presented in Partial Fulfillment of the Requirements
for the Degree Master of Science in the
Graduate School of The Ohio State University
By
James Daniel Downs, B.S.
*****
The Ohio State University
2008
Master’s Examination Committee:
Approved by:
Dr. Roger A. Williams, Adviser
Dr. David M. Hix
Dr. P. Charles Goebel
_____________________________
Adviser
Natural Resources Graduate Program
Copyright by
James D. Downs
ABSTRACT
Oak ecosystems in the Central Hardwoods region are diverse,
supporting thousands of species of plants, numerous insects and
invertebrates, and hundreds of wildlife species. Oak ecosystems are not only
important for wildlife, but are important to the forest product industry, as well
as for many recreational activities. This study examines the effects of two
intensities of shelterwood harvesting (50 and 70 percent of full stocking). The
main goal of this study is to develop and understand the relationships
between disturbance levels and the successful release of oak regeneration.
The silvicultural treatments were conducted during the fall of 2005 at two Ohio
State forests (Richland Furnace and Zaleski), which are located in the heavily
forested southeastern portion of the state. Multiple Response Permutation
Procedure (MRPP), Analysis of Variance (ANOVA) and Canonical
Correspondence Analysis (CCA) procedures were used to analyze the data.
Patterns in species densities were used to suggest that the silvicultural
treatments were successful in establishing the current regeneration. The
results from these silvicultural treatments should remain consistent across
similar oak-hickory forests of the central hardwood region.
ii
ACKNOWLEDGMENTS
Many thanks to my advisor, Dr. Roger Williams, for his support,
patience, and encouragement, which made this project possible. I am also
grateful to have had Dr. David Hix and Dr. P. Charles Goebel as my
committee members. I would also like to thank Dr. Randall Heiligmann for his
advise and support.
I would also like to thank the Ohio Division of Forestry for their
cooperation and assistance during this project, without which this study would
not have been possible. Additionally, I owe a debt of gratitude to Marne
Titchenell and Stephen Rist for their assistance in data collection.
iii
VITA
1984..............……………………………....Born: Logan, Ohio
2004…………………………………………A.A.S., Forestry Management,
Hocking College
2006..............………………………………B.S., Enivironment and
Natural Resources,
The Ohio State University
FIELDS OF STUDY
Major Field: Natural Resources
iv
TABLE OF CONTENTS
Page
Abstract……………………………………………………………...……………….ii
Acknowledgments……………………………………………..……………………iii
Vita…………………………………………………………...…………………...….iv
List of Tables…………………………………………………...…………………..vii
List of Figures…………………………………………………...………………..…ix
Chapters
1.
Introduction…………..………………………………………………………1
1.1 Study Context……………………………………..………………...…..1
1.2 Purpose and Objectives……..…………………………………………3
2.
Literature Review…………………………………………………………...4
3.
Methods…….………………………………………………………………10
3.1 Description of Site and Treatment………………….……………….10
3.2 Vegetation Plot Establishment and Measurement…..…………….16
3.3 Statistical Analysis…………………………………………………….17
4.
Results and Discussion………...…………………………………………22
4.1 Overstory Attributes…………………………………………………...22
4.2 Understory Composition:
Multiple Response Permutation Procedure………………………...25
4.3 Pre-harvest Data………………………………………………………31
4.4 One Growing Season After Harvest……………………………...…33
4.5 Two Growing Seasons After Harvest……………………………….35
4.6 Deer Herbivory Two Years After Harvest…………………………..38
4.7 Deer Herbivory on Stump Sprouts Two Years After Treatment.…41
v
4.8 Summary of Comparisons Between
Red Maple, White Oak, and Red Oak………………….………...…47
4.9 Canonical Correspondence Analysis………………………………..57
5. Conclusions and Management Implications…………………………………68
Literature Cited…………………………………………………………………….71
vi
LIST OF TABLES
Table
Page
1. Key to species included in species groups…………………………….…….13
2. Pre-harvest overstory attributes……………………………………………….14
3. Pre-harvest and post-harvest overstory attributes…………………………..24
4. Results from Multiple Response Permutation Procedure for Richland
Furnace State Forest and Zaleski State Forest for pre-harvest, 1 year
post-harvest, and 2 years post harvest data………………………..……….27
5. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest small seedling count data between treatments within
Richland Furnace State Forest and Zaleski State Forest……………….….28
6. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest large seedling count data between treatments within
Richland Furnace State Forest and Zaleski State Forest……………..…...29
7. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest sapling count data between treatments within Richland
Furnace State Forest and Zaleski State Forest……………………….….…30
8. Summary ANOVA of pre-harvest small seedling stems per acre…….…...43
9. Summary ANOVA of 1 year post-harvest small seedling stems
per acre…………………………………………………………………………..43
10. Summary ANOVA of 2 years post-harvest small seedling stems
per acre……………………………………………………………….….…….43
11. Summary ANOVA of pre-harvest large seedling stems per acre……..…44
vii
12. Summary ANOVA of 1 year post-harvest large seedling stems
per acre……………………………………………………………………..….44
13. Summary ANOVA of 2 years post-harvest large seedling stems
per acre……………………………………………………………………..….44
14. Summary ANOVA of pre-harvest sapling stems per acre…………….…..45
15. Summary ANOVA of 1 year post-harvest sapling stems per acre…….…45
16. Summary ANOVA of 2 years post-harvest sapling stems per acre…...…45
17. Summary ANOVA of 2 years post-harvest sapling and sprout stems
per acre……………………………………………………………………...…46
18. Summary ANOVA of 2 years post-harvest sapling versus sprout origin..46
19. Summary of pre-harvest Canonical Correspondence Analysis Monte
Carlo test results……………………………………………………………….61
20. Summary of 1 year post-harvest Canonical Correspondence Analysis
Monte Carlo test results………………………………………………………61
21. Summary of 2 years post-harvest Canonical Correspondence Analysis
Monte Carlo test results……………………………………………………….61
viii
LIST OF FIGURES
Figures
Page
1. Location of oak-hickory forest sites…………………………….…..…………15
2. Treatment layout for study………………………………………….…….……16
3. Plot layout used for this study………………………………………….……...21
4. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 50 percent treatment small seedling
Plots………………………………………………………………………………51
5. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 70 percent treatment small seedling
plots………………………………………………………………………………51
6. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest control small seedling plots……………….51
7. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 50 percent treatment small seedling plots…….……..52
8. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 70 percent treatment small seedling plots….….….…52
9. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest control small seedling plots……………………………52
10. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 50 percent treatment large seedling
plots………………………………………………………………….………….53
11. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 70 percent treatment large seedling
plots……………………………………………………………………………..53
ix
12. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest control large seedling plots……….……53
13. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 50 percent treatment large seedling plots….……..54
14. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 50 percent treatment large seedling plots….……..54
15. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest control large seedling plots……………….…………54
16. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 50 percent treatment sapling plots….….55
17. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest 70 percent treatment sapling plots….….55
18. Summary of stems per acre of ACRU, QUAL, and QURU species in
Richland Furnace State Forest control sapling plots………………….…..55
19. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 50 percent treatment sapling plots………………….56
20. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest 70 percent treatment sapling plots………………….56
21. Summary of stems per acre of ACRU, QUAL, and QURU species in
Zaleski State Forest control sapling plots…………………………………..56
22. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest small seedling plots………………………….………62
23. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest large seedling plots……………………………...…..63
24. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest sapling plots………………………………………….64
25. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest small seedling plots……………………………………………65
26. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest large seedling plots…………………………………….………66
x
27. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest sapling plots…………………………………………………….67
xi
CHAPTER 1
INTRODUCTION
1.1 Study Context
Oak (Quercus spp.) have been the dominant component of the deciduous
forests of the eastern United States for at least the past 5,000 years (Fralish
2004), and have been an important component for 10,000 years (Dolan and
Parker 2004). During the past century, these forests have been changing from
oak dominance to increasing proportions of red maple (Acer rubrum) and other
hardwood species (Griffith et al. 1993, Van Lear 2004). This change in species
composition is largely attributable to alterations of the local natural disturbance
regimes and changes in land use. Periodic fires were once common across the
forested landscape, which favored the regeneration of oaks over red maple and
other hardwood species; but the successful implementation of an aggressive
national fire suppression policy has shifted forest stand composition toward fireintolerant species (Abrams 1998). Forests in southern Ohio have not escaped
the effects of this fire suppression policy.
The major changes from the historical natural disturbance regimes are
consequences of the anthropogenic disturbances associated with
industrialization. Most natural disturbances in forests, which disturb and reduce
1
the amount of tree canopy, increase the available light for vegetative
regeneration. Without canopy disturbance, shade-intolerant species cannot
successfully regenerate. While oaks are intermediate in shade tolerance,
prolonged shading can lead to seedling mortality (Crow 1988). However, when
small canopy-gap disturbances occur, oaks are often out-competed by more
shade-tolerant species, and when major canopy disturbances occur, they are
out-competed by faster-growing, more shade-intolerant species (Van Lear 2004).
This suggests that oaks require intermediate levels of disturbance. While studies
by Sander (1979) and Loftis (1983) demonstrate that shelterwood cutting
techniques, which produce intermediate levels of disturbance, often regenerate
oaks on drier uplands, oaks are still often out-competed by other woody
vegetation on higher-quality sites. These findings suggest that to successfully
regenerate oaks some form of disturbance is needed that gives them a
competitive edge over other competing vegetation to successfully regenerate, as
they had before the fire suppression era.
In the eastern deciduous forests, white-tailed deer (Odocoileus
virginianus) browsing is considered a contributing factor to the inability of oak to
successfully regenerate (Lorimer 1993). Campbell et al. (2006) documented that
white-tailed deer in West Virginia can reduce seedling heights, cause shifts in
species composition, and reduce the survival of stump sprouts. Since oak stump
sprouts are considered to be an important source of regeneration when oak
seedlings and saplings are not advancing into larger size classes in oak stands
(Buckley and Evans 2004), it is important to consider the effects of white-tailed
2
deer herbivory on stump sprouts. However, few studies have been conducted to
determine the extent to which deer browse effects oak regeneration and stump
sprouts in Ohio.
1.2 Purpose and Objectives
The main goal of this study is to develop and understand the relationships
among disturbance levels and white-tailed deer herbivory and the releasing of
oak regeneration. Accordingly, the objectives of this research project are: (1) to
evaluate the effects of two intensities of shelterwood harvesting, specifically
reductions to 50 and 70 percent stocking, respectively on woody vegetation, and
(2) to determine the effects of deer herbivory on hardwood regeneration.
3
CHAPTER 2
LITERATURE REVIEW
Oak ecosystems in the Central Hardwoods region are diverse, supporting
thousands of species of plants, numerous insects and invertebrates, and
hundreds of wildlife species (Dickson 2004). White-tailed deer, black bear, wild
turkey, blue jay, squirrel, and mice depend upon acorns for survival (Hicks 1998,
Dickson 2004). Many wildlife populations often follow fluctuations in acorn crops,
and major declines in oak ecosystems can lead to population crashes of
dependent species if alternate food resources are unavailable (Hicks 1998).
Although oak ecosystems are mostly considered xeric, these forests support an
array of reptile and amphibian species, several endangered species of bats, and
maintain a required habitat for neotropical migratory birds, as well as for early
successional grass-forb associated birds (Dickson 2004).
Oak ecosystems are not only important for wildlife, but are important to the
forest product and recreation-tourism industries (Jackson and Buckley 2004). In
Ohio, 8 million board feet of oak timber is removed annually; this amount equals
or exceeds the total amount of oak board feet grown in Ohio forests each year
(Griffith et al. 1993). However, for species such as beech (Fagus grandifolia),
and yellow-poplar (Liriodendron tulipifera), the annual board footage growth rate
exceeds the total amount of board feet removed per year (Griffith et al. 1993).
4
Additionally, the net economic impact of losing oak ecosystems from the
landscape is impossible to predict, but it greatly exceeds the losses generated by
oak timber alone (Jackson and Buckley 2004).
Currently, oaks are a dominant component of approximately 46% of the
forestland in the eastern United States (McWilliams et al. 2002). In Ohio, oak
forests occupy 59% of the total forestland and 77% in southeastern Ohio (Griffith
et al. 1993). Most of the mature second-growth forest in southern Ohio
regenerated following severe anthropogenic disturbances from the charcoal
industry, and agricultural usage within the past recent centuries (Hutchinson et
al. 2003). Preceding the anthropogenic disturbances, these southern Ohio
forests were maintained by periodic fires ranging from every 5 to 15 years
(Sutherland et al. 2003). The frequent fires occurred across the landscape until
coordinated fire suppression efforts began during 1923 (Sutherland 1997).
At present, the advance regeneration layer varies from dominance of red
maple and other shade-tolerant species on sites with minimal canopy
disturbance to the dominance of yellow-poplar on sites with major canopy
alterations (Lorimer 1993). Numerous studies have found that oak is nearly
absent from the regeneration layers except on the driest upland sites (Clark
1993, Rebbeck et al. 2004). Many studies report similar trends, and with the
current low levels of advance oak regeneration, there will be an inadequate
amount to comprise a significant proportion in these forest following any canopy
disturbance (McCarthy et al. 1987, Goebel and Hix 1997). The reasons most
frequently cited for the sparse advance oak regeneration are: 1) single-tree
5
harvesting techniques which do not to provide adequate light conditions because
of insufficient canopy disturbance (Burns and Honkala 1990), and 2) the lack of
repeated disturbance from fire since oaks are less susceptible to injury and are
able to repeatedly produce sprouts (Dey 2002, Brose and VanLear 1998).
While oaks display many adaptations over other species in a fire regime
(Niering et al. 1971, Swan 1970), fire alone is not all that is necessary for the
successful regeneration of oak. A combination of an ample reduction in canopy
cover, to provide a suitable amount of sunlight to the forest floor, and reductions
in the lower forest strata to minimize or eliminate competition is required.
Shelterwood harvests have often been used to create the desired conditions by
successfully regulating the light conditions on the forest floor (Schlesinger 1993).
Shelterwood treatments typically require a reduction in canopy cover
through the use of a low thinning (Sander 1979, Jacobs and Wray 1992, Dey and
Parker 1996). Although crown cover is the best indicator of understory light
conditions, basal area or stocking level is used most often to regulate stand
density because it is easily quantified (Dey and Parker 1996). However, the
amount of sunlight that reaches the forest floor is not always directly related to
basal area or stocking level (Sloan and Zastrow 1985).
When reducing the overstory density through use of a shelterwood cut,
most prescriptions call for a uniform reduction in canopy by thinning from below
(Sander 1979, Jacobs and Wray 1992). Correspondingly, the initial shelterwood
harvests should emphasize the removal of overtopped trees in combination with
the prescribed overstory reductions (Sander 1979; Loftis 1983, 1990; Lorimer et
6
al. 1994). As a guide to these cuts, codominant or dominant oaks should be
maintained to provide enough shade to inhibit aggressive understory competition,
and to enhance the potential of acorn production in years of good seed crops
(Dey and Parker 1996).
Since the amount of existing understory vegetation is also an important
factor in determining the success or failure of oak regeneration establishment
and growth, it should be considered when prescribing a shelterwood harvest.
When there is an abundant amount of vegetation present in the understory, a
shelterwood cut that retains 30-50 % of the canopy may result in severe
understory competition (Sander 1979, Loftis 1983). In such cases, a canopy
reduction of only 30% is recommended to reduce competition with oak
regeneration (Lorimer 1989). The overstory canopy can be reduced to 50% in the
initial harvest, if understory vegetation is sparse, or if it has been reduced
through vegetation control management (Wolf 1988). However, there are some
instances where a lighter harvest that leaves 85% of the original canopy, when
combined with understory vegetation removal, can improve the survival and
growth of oak regeneration (Pubanz and Lorimer 1992, Lorimer et al. 1994). In all
situations, as site quality increases, a denser residual overstory is required to
prevent the intense release of undesirable vegetation (Dey and Parker 1996).
The major issue when applying the shelterwood method has consequently
revolved around the amount of the residual overstory to maintain. While retaining
a 70% stocking level has been recommended when advance oak regeneration is
absent (Jacobs and Wray 1992), reducing overstory density to 60% of full
7
stocking has been recommended when advance oak regeneration is present
(Graney and Rogerson 1985, Sander 1988, Jacobs and Wray 1992). There are
obviously trade-offs between limiting the release of understory competition and
providing enough light for the successful establishment and growth of oak
regeneration. Sander (1979) suggests that the first shelterwood harvest leave an
overstory stocking of 60%, the minimum stocking for full site utilization, to help
control understory development. Schlesinger et al. (1993) found that removing all
non-oak stems <4 cm DBH and reducing the overstory to 60% stocking
minimized vegetation competition on good quality sites. On average quality sites,
they found that the reduction of understory competition had little effect on
advance oak regeneration, and that reducing of the overstory to 40% stocking
resulted in the highest amount of large oak advance reproduction.
The control of understory vegetation is warranted when intense
competition is expected to be problematic to oak regeneration establishment and
growth following a shelterwood harvest (Ehrenfeld 1980, McGee 1984). Farmer
(1975) found that the establishment and growth of oak regeneration will be
enhanced when understory competition is reduced simply because seedlings and
new sprouts are capable of rapid height growth when both moisture and light are
not limiting. Many different methods of reducing understory competition have
been recommended, including the use of herbicides, mechanical methods, and
prescribed burning (Johnson and Jacobs 1981, Nyland et al. 1982, Lorimer 1985,
Van Lear and Waldrop 1988).
8
In regard to the effects of white-tailed deer herbivory on woody vegetation,
Jacobs (1969) found that high densities of white-tailed deer can significantly
influence the morphology and growth rates of hardwood seedlings. Numerous
other studies have indicated that white-tailed deer prevent the progression from
tree seedling into sapling classes, which are most often unaffected by whitetailed deer browse (Apsley and McCarthy 2004, Castleberry et al. 1999, Oswalt
et al. 2006). Furthermore, Healy (1997) and Hughes and Fahey (1999) found
that because white-tailed deer browse selectively, species diversity could be
significantly altered. Although white-tailed deer are presumed to be a
contributing factor with the problem of oak regeneration survival, little research
has been conducted in oak-dominated systems in Ohio to verify this assumption
(Apsley and McCarthy 2004), and even less research has been conducted to
ascertain the effect of white-tailed deer browse on the survival of oak stump
sprouts.
9
CHAPTER 3
METHODS
3.1 Description of Forests and Treatments
Mixed oak-hardwood forests located on upland sites in two of southern
Ohio’s state forests served as the locations for this study: (1) Richland Furnace
State Forest, Jackson County, Ohio (39o10’ N latitude, 82o36’ W longitude) and
(2) Zaleski State Forest, Vinton County, Ohio (39o15’ N latitude, 82o23’ longitude)
(Figure 1). Both state forests contain large contiguous tracts of upland oakdominated forests, but also contained many other hardwood species (Table 1).
The upland oak forests selected for this study both had stocking of approximately
100%, or a stand density index values (SDI) of 215 (Williams 2003), and are
located on medium sites (black oak site index 60-75 feet, base age 50 years).
At Richland Furnace, 80% of the overstory basal area is comprised of
upland oak, with the majority of that from the white oak group (Table 2). Similarly,
at Zaleski, 86% of the overstory basal area consists of oaks, with the majority of
those species in the white oak group. In terms of trees per acre, white oak is the
dominant tree species within both forests at 44% of the total trees per acre
(Table 2). While red maple is the second most common species in terms of trees
per acre, at Richland Furnace 18% and at Zaleski 27%. The red oak group is the
10
third most common tree species in terms of trees per acre, with 16% of the
overstory trees at both study locations.
Both forests are located within the unglaciated hill country of southeastern
Ohio where the topography is characterized by deeply dissected terrain of the
Allegheny Plateau (Kerr 1985). These topographic features create a gradient of
moisture regimes and microclimates across the landscape (Kerr 1985). The plots
slopes range from 10-24% with an average of 16%. The plots aspect ranges from
northeast (43o azimuth) to southwest (268o azimuth) with an average aspect of
southeast (157o azimuth).
Two different commercial harvest levels were implemented at each forest
during the fall of 2005, reducing the stocking levels to 50 and 70 percent of full
stocking (hence forth referred to as 50 percent treatment and 70 percent
treatment, respectively), which represent the upper and lower bounds of previous
management reccomendations. These are equivalent to SDI values of 150 and
110, respectively (Williams 2003). SDI values were used as the residual target
for marking trees because it is independent of site quality and is easier to
quantify (Williams 2003).
The silvicultural system was designed as a one-cut shelterwood method.
A combined crown thinning and low thinning was implemented to reduce the
stocking. Dominant and codominant oaks were the favored retention trees while
all other species were discriminated against where possible. However, since this
was a commercial harvesting operation, where there was little deliberate marking
of non-commercial species. The Zaleski treatment areas were originally marked
11
for a gypsy moth pre-salvage cut, but were later remarked to meet the residual
stand criteria for this study. This remarking process, may provide the explanation
as to the over marking and harvesting in the Zaleski 70 percent treatment areas.
Zaleski treatment areas were harvested using cable skidders, the harvesting in
Richland Furnace was carried out through using a cut to length logging system.
Each treatment area was 25 acres in, and since there are five treatments
(two reduced to 50% stocking, two reduced to 70% stocking, and one control
area) per state forest, the treatment layout was arranged so that harvesting could
be more easily expedited (Figure 2). Accordingly, treatments involving the same
reductions in stocking levels were placed side-by-side to ensure that the marking
and harvesting was conducted in a consistent manner. The rationale for two of
each treatment type at each location compared to only one control area was this
study is preliminary to a longer-term project, which will be a growing season
prescribed fire and a dormant season prescribed fire, respectively, into one of
each of the treatment areas within both study locations.
12
Species Group Species Included
Acer rubrum
ACRU
CASP
Carya glabra
Carya ovata
Carya tomentosa
LITU
Lirodendron tulipiifera
NONCOM
Amelanchier spp.
Carpinus caroliniana
Castanea pumila
Cercis canadensis
Cornus florida
Corylus americana
Crategus spp.
Hamamelis virginiana
Lindera benzoin
Rhus copallinum
Rhus glabra
Viburnum acerfolium
Viburnum dentata
Viburnum prunifolium
Species Group Species Included
Nyssa sylvatica
NYSY
OTHER
Acer nigrum
Acer saccharum
Castanea dentata
Fagus grandifolia
Juglans nigra
Oxydendron arboreum
Pinus rigida
Pinus virginiani
Populus grandidentata
Prunus serotina
Ulmus americana
QUAL
Quercus alba
Quercus prinus
QURU
Quercus rubra
Quercus coccinea
Quercus velutina
Quercus imbricaria
SAAL
Sassafras albidum
Table 1. Key to species included within species group used in this study to
examine the effects of shelterwood harvests on oak regeneration one and two
years after harvest in southern Ohio.
13
Basal
Quadratic mean
2
Species Group area (ft /ac) Trees/acre
diameter (in.)
-----------------------Richland------------------------ACRU
6.8
23.1
7.3
CASP
5.0
8.7
10.2
LITU
5.9
4.5
15.5
NONCOM
0.0
0.3
4.1
NYSY
1.1
5.5
6.1
OTHER
3.2
8.1
8.5
QUAL
61.3
57.3
14.0
QURU
26.1
21.0
15.1
SAAL
0.2
0.6
7.0
TOTAL
109.5
129.1
9.8 (average)
------------------------Zaleski-------------------------ACRU
9.6
36.6
6.9
CASP
1.4
4.9
7.2
LITU
0.7
1.5
9.2
NONCOM
0.1
0.5
5.7
NYSY
0.2
1.1
6.0
OTHER
2.7
10.3
6.9
QUAL
53.3
59.9
12.8
QURU
36.1
21.3
17.6
SAAL
0.0
0.3
5.3
TOTAL
104.1
136.3
8.6 (average)
Table 2. Pre-harvest overstory attributes for Richland Furnace and Zaleski State
Forests by species group. For key to species groups see Table 2.
14
Figure 1. Location of oak-hickory forests located at Richland Furnace State
Forest, Jackson County, Ohio and Zaleski State Forest, Vinton County, Ohio.
15
70
25 ac
70
25 ac
50
25 ac
50
25 ac
Control
25 ac
Figure 2. Generalized treatment layout for study. Treatments labeled with 70
indicate reduction of the overstory to 70% of full stocking, and those labeled with
50 indicate reduction of the overstory to 50% of full stocking. No cutting was
performed in the control.
3.2 Vegetation Plots Establishment and Measurements
Eight overstory plots were located within each treatment block using a
systematic scheme for plot location rather than a random approach (Figure 3). It
was desirous to have plots evenly distributed over the treatment block rather than
a possible cluster of plots, which could have occurred in a random approach. All
transects were installed parallel to the slope where possible.
During the summer of 2005, 40 vegetation plots were established before
the silvicultural treatments were initiated to describe pre-harvest overstory and
understory conditions. During the summers of 2006 and 2007, one and two
growing seasons after the silvicultural treatment the plots were remeasured.
Each sample plot was stratified into four sample plots based on vegetation size
category (Figure 3), and the following vegetation information was measured and
recorded: (1) overstory plot (0.2 acre) species, diameter at breast height (dbh),
and total height were recorded for trees greater than 4 inches dbh, measured; (2)
16
sapling plot (0.05 acre) species, dbh, and total height was recorded for all trees
between 4.5 feet tall and 3.9 inches dbh, (3) large seedling plot (0.025 acre) a
tally by species for all woody stems 1-4.5 feet tall; (4) small seedling plot (0.01
acre) a tally by species for all woody stems less than 1 feet tall. In addition to the
vegetation data, information on aspect, slope, slope shape, and slope position
were assessed at each overstory plot.
During the 2007 data collection period, additional information was
recorded for the small seedling, large seedling, and sapling plots. The origin of
each stem whether growing from a cut stump or not, was recorded. In addition,
each stem was observed for evidence of deer browsing.
3.3 Statistical Analysis
Multiple Response Permutation Procedure (MRPP), a non-parametric
procedure that tests the hypothesis of no difference between groups, was used.
MRPP does not assume multivariate normality and homogeneity of variances,
which makes it advantageous over multivariate analysis of variance (MANOVA).
MRPP was used to provide a test of whether there is a significant difference
between the treatment groups and count data. The Sorensen (Bray-Curtis)
distance measure was utilized to measure the distance matrices associated with
MRPP. The A value equal (1 - (observed delta/expected delta)), which is a
chance-corrected estimate of the proportion of the distances explained by group
identity.
17
As a result of 38 woody species being identified in the sampling of the
plots, individual species were lumped into similar species groups (Table 1) for
simplification in data analysis. Red maple was a group (ACRU) by itself give the
large number of it’s stems per acre, as well as because the competition between
red maple and oak has been well documented (Abrams 1998, Fei and Steiner
2007, Nowacki and Abrams 2008). Three hickory species, shagbark, pignut, and
mockernut, were placed into the same species group (CASP) because they
belong to the same species guild (Sutherland et al. 2000). Yellow-poplar was a
group (LITU) by itself because there were few other shade-intolerant species
found in the understory and it was significantly affected by the silvicultural
treatment. The non-commercial group (NONCOM) consists of all woody shrub
species. Black gum was a group (NYSY) because there is evidence to suggest
that deer herbivory significantly affects black gum seedlings. The other group
(OTHER) consists of tree species where there were either few stems per acre or
no significant affects of treatment or deer herbivory on the species to justify the
placement into a separate group. The white oak group (QUAL) consists of both
white oak and chestnut oak, which are both members of the white oak species
guild and display similar seed dispersion and growth patterns (Sutherland et al.
2000). The red oak group (QURU) consists of red oak, black oak, scarlet oak,
and shingle oak. These species were a separate group as a result of them being
in the same species guild as well as displaying similar life strategies (Sutherland
et al 2000). Finally, sassafras was left in a group by itself not only because of the
18
quantity of stems per acre in the regeneration plots, but also because of the
effect that the treatment had upon the release of sassafras.
Following this classification, analysis of variance (ANOVA) was performed
to test for both significant differences within species group densities and among
treatment intensities and to test for significant differences between species group
densities within each treatment. ANOVA tests were performed using pre-harvest,
one growing season after harvest, and two growing seasons after harvest
species data sets, at p=0.05 level. In addition, ANOVA was conducted to
compare sapling species stem origin and means on the two growing seasons
after harvest sapling data.
Following MRPP and ANOVA Canonical Correspondence Analysis (CCA)
was conducted to examine patterns and variation along gradients and among
sampling entities. CCA ordination diagrams are presented for Richland Furnace
and Zaleski small seedlings, large seedlings, and saplings two growing seasons
after harvest data sets. As implied, CCA provides a direct ordination of multiple
explanatory variables. CCA uses a Monte Carlo permutation procedure, which is
robust to non-normal distributions. The Monte Carlo permutation procedure
yields two important values with associated p-values. The first is an Eigen value,
which represents the variance in the matrix axis (McCune 1997). The Pearson
Correlation, species-environment correlation value, is the second important
value; it is a measure of the strength of the species environment relationship in
proportion with the variance explained by the environmental matrix (McCune
1997). All multivariate data analyses were conducted using PC-ORD version 5.0
19
(McCune and Mefford 1999) a program designed for multivariate analysis of
ecological data.
Ideally the silvicultural treatment markings would be carried out by the
same crew to provide consistency across both forests. However, as a result of
different crews selecting and marking the residual trees in each state forest and
the remarking of the Zaleski gypsy moth treatments into a shelterwood treatment,
there are differences in the final residual stocking of treatments between the two
state forests. These issues make direct comparisons between the two locations
through standard univariate statistical methods inappropriate. Univariate
statistical methods can still be used to compare regeneration data within the
same forest, but multivariate statistical methods must be used to compare
inferences between Richland Furnace and Zaleski State Forests.
20
3 chains
4 chains
Slope
25 acre Treatment Block
3 chains
8 chains
Overstory Plot (0.2 acre)
Sapling plot (0.05 acre)
Large seedling plot (0.025 acre)
Small Seedling plot (0.01 acre)
Figure 3. Plot layout used for this study, including the overstory plot (0.2 acre),
sapling plot (0.05 acre), large seedling plot (0.025 acre) and small seedling plot
(0.01 acre).
21
CHAPTER 4
RESULTS AND DISCUSSION
4.1 Overstory Attributes
The pre-harvest overstory data for Richland Furnace and Zaleski state
forests are very similar with respect to basal area ft2/ac, trees per acre, quadratic
mean diameter, and percent stocking (Table 3). Accordingly, the effect of
treatment on the regeneration layer between the two locations should be
comparable. However, as indicated in Table 3 there are obvious differences in
treatment intensity in the overstory post-harvest stand data, which provide a
unique challenge in the interpretation of the effect of treatment on the
regeneration layer.
The problem, which makes comparisons between Richland Furnace and
Zaleski State Forests inappropriate, is that the silvicultural marking guidelines
were not followed correctly in Zaleski State Forest in the 70 percent treatment
areas. The basal area (ft2/ac), density, and the percent stocking were all lower in
the 70 percent treatment compared to the 50 percent treatment, which were
much lower than the treatment called for. As a result, the 70 percent treatment is
similar to a 50 percent canopy reduction treatment with a residual stocking of 45
22
percent. This problem must be recognized before interpretations of the effect of
treatment on the regeneration layer can be made in Zaleski State Forest.
Even though there are problems with the treatments in Zaleski State
Forest, the post-harvest overstory data indicates that the treatments conducted in
Richland Furnace State Forest were done properly. The basal area, density,
quadratic mean diameter, and percent stocking within the 50 and 70 percent
treatments, respectively, all fall within acceptable levels of variability. Thus, in
the Richland Furnace State Forest inferences between the 50 and 70 percent
treatments were made using univariate statistical methods.
23
Forest
Richland
Zaleski
Basal
Density
Quadratic mean
2
Treatment area (ft /ac) (Trees/Acre)
diameter (in.)
Stocking %
------------------------Pre-harvest------------------------50%
114.3
111
13.8
95
70%
114.1
143
12.1
96
Control
92.2
144
10.8
79
------------------------Post-harvest------------------------50%
54.3
22
21.5
43
70%
77.1
52
16.6
63
------------------------Pre-harvest------------------------50%
105.7
127
12.3
70%
100.2
138
11.5
Control
99.5
151
11.0
------------------------Post-harvest------------------------50%
62.2
45
16.0
70%
55.9
38
16.3
89
85
85
51
45
Table 3. Pre-harvest and post-harvest overstory attributes for Richland Furnace
State Forest and Zaleski State Forest by treatment.
24
4.2 Understory Composition: Multiple Response Permutation Procedure
As indicated by the test statistic T and associated p-values, the species
distributions are significantly different than random. All of the pairwise
comparisons between treatments and control for pre-harvest data have
statistically significant test statistic-T values and p-values (Tables 5-7). All of the
pairwise comparisons in Zaleski are significant with the exceptions of small
seedling plots 50 percent vs. 70 percent and sapling plots 50 percent vs. 70
percent (Table 5). The pairwise comparisons for the data collected in Richland
Furnace one growing season after treatment are significantly different with the
exception of large seedling plots comparing 50 percent treatments with 70
percent treatments and with control treatment (Table 6). The pairwise
comparisons between the 50 percent vs. 70 percent sapling plots is the only non
significant difference observed in the regeneration data two years after treatment
data for Richland Furnace (Table 7). All of the pairwise comparisons for Zaleski
two years after treatment data were statistically significant with the exception of
the 50 vs. 70 plots for all three plot sizes: (1) small seedling, (2) large seedling,
and (3) sapling (Tables 5-7). However, in the case of the lack of significant
difference between 50 percent vs. 70 percent treatments at Zaleski, the most
likely cause of the lack of significance is due to the problems associated with the
improper marking and harvesting of the 70 percent treatment.
Results from the MRPP and the associated pairwise comparisons indicate
that the silvicultural treatments have a statistically important effect on the
understory regeneration. MRPP suggests that species densties are not randomly
25
distributed across locations or treatments. Therfore, further investigation into the
data through Analysis of Variance, a univariate method of statistical analysis, and
Canonical Correspondence Analysis, a multivariate method of statistical analysis,
were warranted.
26
Forest
Richland
Zaleski
Test
Plot Size
Statistic T
A value
p value
-------------------------Pre-harvest-------------------------Small Seedling
-9.950
0.155
<0.001
Large Seedling
-11.283
0.146
<0.001
Sapling
-19.985
0.358
<0.001
--------------------1 year post-harvest--------------------Small Seedling
-8.340
0.112
<0.001
Large Seedling
-3.077
0.035
0.006
Sapling
-2.776
0.038
0.01
-------------------2 years post-harvest-------------------Small Seedling
-9.919
0.123
<0.001
Large Seedling
-4.291
0.050
<0.001
Sapling
-6.654
0.075
<0.001
-------------------------Pre-harvest-------------------------Small Seedling
-9.157
0.119
<0.001
Large Seedling
-10.945
0.157
<0.001
Sapling
-22.535
0.399
<0.001
--------------------1 year post-harvest--------------------Small Seedling
-6.574
0.079
<0.001
Large Seedling
-6.207
0.078
<0.001
Sapling
-4.846
0.072
<0.001
-------------------2 years post-harvest-------------------Small Seedling
-5.360
0.059
<0.001
Large Seedling
-5.475
0.066
<0.001
Sapling
-4.493
0.057
<0.001
Table 4. Results from Multiple Response Permutation Procedure for both the
Richland Furnace State Forest and the Zaleski State Forest for pre-harvest, 1
year post-harvest, and 2 years post-harvest data.
27
Forest
Richland
Zaleski
Group Codes
Test
A value
p-value
Compared
Statistic T
-------------------------Pre-harvest-------------------------50 vs. 70
-6.885
0.056
<0.001
50 vs. 100
-10.184
0.195
<0.001
70 vs. 100
-4.228
0.084
0.004
--------------------1 year post-harvest--------------------50 vs. 70
-3.963
0.046
0.004
50 vs. 100
-9.154
0.153
<0.001
70 vs. 100
-5.187
0.084
<0.001
-------------------2 years post-harvest-------------------50 vs. 70
-5.369
0.056
<0.001
50 vs. 100
-10.037
0.163
<0.001
70 vs. 100
-5.690
0.087
<0.001
-------------------------Pre-harvest-------------------------50 vs. 70
-3.703
0.041
0.003
50 vs. 100
-8.817
0.153
<0.001
70 vs. 100
-6.493
0.114
<0.001
--------------------1 year post-harvest--------------------50 vs. 70
-1.430
0.014
0.09
50 vs. 100
-8.818
0.136
<0.001
70 vs. 100
-4.113
0.063
0.002
-------------------2 years post-harvest-------------------50 vs. 70
-0.981
0.009
0.155
50 vs. 100
-4.051
0.055
0.002
70 vs. 100
-6.473
0.096
<0.001
Table 5. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest small seedling count data between treatments within the Richland
Furnace State Forest and the Zaleski State Forest.
28
Forest
Richland
Zaleski
Group Codes
Test
Compared
Statistic T
A value
p-value
-------------------------Pre-harvest-------------------------50 vs. 70
-7.840
0.094
<0.001
50 vs. 100
-8.719
0.144
<0.001
70 vs. 100
-6.979
0.110
<0.001
--------------------1 year post-harvest--------------------50 vs. 70
-0.430
0.005
0.3
50 vs. 100
-2.714
0.041
0.2
70 vs. 100
-3.986
-0.051
0.002
-------------------2 years post-harvest-------------------50 vs. 70
-4.692
0.046
<0.001
50 vs. 100
-1.921
0.029
0.05
70 vs. 100
-2.073
0.030
0.04
-------------------------Pre-harvest-------------------------50 vs. 70
-7.898
0.108
<0.001
50 vs. 100
-7.768
0.131
<0.001
70 vs. 100
-7.292
0.136
<0.001
--------------------1 year post-harvest--------------------50 vs. 70
-2.100
0.023
0.04
50 vs. 100
-5.413
0.085
<0.001
70 vs. 100
-6.377
0.095
<0.001
-------------------2 years post-harvest-------------------50 vs. 70
-1.318
0.014
0.10
50 vs. 100
-6.587
0.099
<0.001
70 vs. 100
-4.179
0.061
0.001
Table 6. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest large seedling count data between treatments within the Richland
Furnace State Forest and the Zaleski State Forest.
29
Forest
Richland
Zaleski
Group Codes
Test
Compared
Statistic T
A value
p-value
-------------------------Pre-harvest-------------------------50 vs. 70
-12.981
0.196
<0.001
50 vs. 100
-13.714
0.349
<0.001
70 vs. 100
-14.210
0.420
<0.001
--------------------1 year post-harvest--------------------50 vs. 70
-2.305
0.029
0.03
50 vs. 100
-1.914
0.035
0.05
70 vs. 100
-1.204
0.020
0.01
-------------------2 years post-harvest-------------------50 vs. 70
-0.832
0.008
0.18
50 vs. 100
-5.851
0.076
<0.001
70 vs. 100
-8.670
0.119
<0.001
-------------------------Pre-harvest-------------------------50 vs. 70
-17.441
0.264
<0.001
50 vs. 100
-14.672
0.396
<0.001
70 vs. 100
-14.538
0.359
<0.001
--------------------1 year post-harvest--------------------50 vs. 70
0.764
-0.012
0.76
50 vs. 100
-6.495
0.113
<0.001
70 vs. 100
-5.994
0.099
<0.001
-------------------2 years post-harvest-------------------50 vs. 70
-0.798
0.008
0.18
50 vs. 100
-5.643
0.100
<0.001
70 vs. 100
-3.097
0.046
0.009
Table 7. Pairwise comparisons of pre-harvest, 1 year post-harvest, and 2 years
post-harvest sapling count data between treatments within the Richland Furnace
State Forest and the Zaleski State Forest.
30
4.3 Pre-harvest Data
Analysis of Variance of the pre-harvest data indicates that in the majority
of cases the density of a species group are similar across all treatments and plot
sizes (Tables 8, 11, and 14). As a result, changes in the regeneration, through
time may be partially attributed to the effects of the silvicultural treatment. Red
maple, nearly without exception, had the highest densities across all three of the
regeneration plot sizes, both study sites, and all treatment locations (Tables 8,
11, and 14). This phenomenon of abundant red maple throughout the Central
Hardwoods region has been well documented (Abrams 1998, Fei and Steiner
2007, Nowacki and Abrams 2008). However, large sassafras seedlings in the
Zaleski control large is the only example of a situation where a particular species
had higher densities compared to red maple pre-harvest.
As indicated previously, red maple had significantly more small seedlings
per acre compared to any other species across both study locations and within
all treatments (Table 8). At Richland Furnace there were significantly more small
white oak seedlings compared to all other species, with the exception of red
maple. There were no other consistent statistical differences observed in the
small seedling densities for the Richland Furnace small seedling layer. Within
Zaleski, sassafras and the non-commercial group both had significantly more
small seedlings per acre compared all the other species groups, with the
exception of red maple (Table 8). There were no other significance differences
between densities of small seedlings by species groups at Zaleski.
31
There were significantly more large red maple and sassafras seedlings
compared to all other species groups at Richland Furnace (Table 11). While
there were statistically less large white oak seedlings compared to red maple and
sassafras in Richland Furnace, there were significantly more compared to all
other species. At Zaleski, sassafras had statistically the highest density of large
seedlings compared to all other species groups. Red maple and the noncommercial group had significantly more large seedlings per acre compared to all
the other species groups.
There were significantly more red maple saplings per acre compared to all
other species in the pre-harvest data (Table 14). There was no other statistically
significant differences between the number of saplings within each species group
with the exception of black gum at Richland Furnace, which had more saplings
per acre compared to all other species groups besides red maple. However,
there was a relative absence of both white oak and red oak saplings at both
Richland Furnace and Zaleski State Forests in this size class (Table 14).
While there are a relatively high number of oak small and large seedlings
per acre there is almost an absence of oak saplings throughout the study area.
Forest managers usually depend upon stump sprouts and the existing advance
regeneration, i.e., saplings, to comprise the future forest stand following canopy
disturbance (Buckley and Evans 2004). Since there are very few oak saplings
per acre, these forests are primarily dependent upon the ability of oak stumps to
vigorously resprout following harvest, in order for oaks to maintain themselves in
the future overstory. The analysis of the pre-harvest data indicate that the two
32
study locations are prime examples of this classic problem that exists with the
Central Hardwoods region and the insufficient amount of oak regeneration to
successfully maintain a significant proportion of oaks in these forests following a
natural canopy disturbance (Van Lear 2004).
4.4 One Growing Season After Harvest
While the results from the one growing season after harvest data were
preliminary, there are still general trends in the data that can be identified in order
to properly understand the development of the regeneration layers through time.
There was a decrease in the densities of all species and size classes one
growing season after harvest compared with the pre-harvest data with relatively
few exceptions (Tables 9, 12, and 15). While there remained a relatively high
number of small and large seedlings one growing season after harvest, there
were very few saplings per acre compared with the pre-harvest conditions. There
are relatively few exceptions to this observed decrease in the densities of small
seedlings, large seedlings, and saplings. The one noticeable exception to these
trends is the increase, by approximately 31 percent, in the number of white oak
small seedlings at Richland Furnace. Based on field observations, this increase
in the number of white oak small seedlings is most likely due to the excellent crop
of white oak acorns during 2005. In addition, yellow-poplar and sassafras both
experienced increases in the number of small seedling densities one growing
season after harvest.
33
There were significantly more small red maple seedlings per acre at
Richland Furnace compared with all other species groups one growing season
after harvest, with the exception of yellow-poplar in the 50 percent treatment
(Table 9). Additionally, there were significantly more small white oak and
sassafras seedlings at Richland Furnace compared with the other species
groups. At Zaleski, there were significantly more small red maple seedlings per
acre compared with all other species groups with the exception of sassafras in
the 50 percent treatment. While there were more small red oak and yellow-poplar
seedlings compared with the other species groups, treatment had no significantly
important effects upon the small seedling densities one growing season after
harvest. However, there were significantly more small yellow-poplar and
sassafras seedlings in the treatment areas relative to the control (Table 9).
Similar to the pre-harvest conditions, there were significantly more large
red maple seedlings at Richland Furnace compared with all the other species
groups (Table 12). While sassafras was the second most abundant species at
Richland Furnace, in which there were significantly more large seedlings
compared with all the other species groups. In Zaleski, there were more large red
maple seedlings compared to all other species with the exception of sassafras in
the 50 percent treatment. Similar to Richland Furnace, sassafras was the second
most abundant species in the large seedling plots one growing season after
harvest (Table 12). Overall, the density of large seedlings was reduced from the
pre-harvest to the one growing season after harvest in the treatment areas.
34
There is no significant difference in the one growing season after harvest to
support that the treatment increased the number of large seedlings per acre.
There was also a dramatic decrease between the density of saplings preharvest (Table 14) and the density of saplings one growing season after harvest
(Table 15). However, similar to the pre-conditions, there were more red maple
saplings per acre compared with all other species groups (Table 15). There was
also nearly a complete absence of hickories, yellow-poplar, the non-commercial
group, white oak group, red oak group, and sassafras saplings one growing
season after harvest.
4.5 Two Growing Seasons After Harvest
There was an overall decrease across nearly all species in the densities of
small seedlings two growing seasons after harvest (Table 10) compared with
both pre-harvest (Table 8) and one growing season after harvest (Table 9) for
both Richland Furnace and Zaleski. The only exception is there were significantly
more small yellow-poplar seedlings two growing seasons after harvest in the
treatment areas compared to the pre-harvest data. There was however, an
increase in the total density of large seedlings per acre across both study
locations two growing seasons after harvest (Table 13) compared with preharvest (Table 11) and one growing season after harvest (Table 12). However,
there was a decrease in the number of large sassafras seedlings per acre two
years after treatment compared to the pre-harvest data. There was also an
increase in the number of saplings per acre for all species and all species groups
35
two growing seasons after harvest (Table 16) compared with one growing
season after harvest data (Table 15). However, only red maple, white oak group,
and red oak group had more saplings per acre two growing seasons after harvest
(Table 16) compared with pre-harvest (Table 14). While all the other species
groups had approximately the same densities of saplings two growing seasons
after treatment (Table 16) compared to pre-harvest (Table 14).
There were significantly more small yellow-poplar seedlings in the
Richland Furnace 50 percent treatment compared with any other species groups
(Table 10). There were significantly more small red maple, yellow-poplar and
sassafras seedlings at Richland Furnace treatment areas compared to all other
species groups. At Zaleski, there were significantly more small red maple
seedlings per acre compared with all other species groups across both
treatments (Table 10). There were no other significant differences between
species groups within the small seedling layer two growing seasons after
treatment. There were significantly fewer small red maple, hickories, the noncommercial group, and white oak group seedlings in the treatment areas
compared with the control in Richland Furnace. While there were significantly
more small yellow-poplar seedlings in the treatment areas relative to the control
at Richland Furnace. Similar trends were observed at Zaleski, where there were
more small yellow-poplar seedlings per acre in the treatments compared with the
control. There were significantly fewer small red maple, the non-commercial
group, the other group, white oak group, and sassafras seedlings in the
treatments relative to the control.
36
There were significantly more large red maple seedlings compared with all
other species, with the exception of yellow-poplar in the 50 percent treatment at
Richland Furnace two growing season after harvest (Table 13). At Zaleski, there
were significantly more large sassafras seedlings compared with all other
species groups. Following sassafras, there were more large red maple seedlings
per acre at Zaleski compared with all other species groups. There were more
large yellow-poplar seedlings in the 50 percent treatment at Richland Furnace
compared with the control and the 70 percent treatment (Table 13). There were
significantly more large black gum seedlings per acre in the treatment areas
compared with the control at Richland Furnace. However, there were significantly
fewer large red oak seedlings in the Richland Furnace treatment areas relative to
the control. At Zaleski, large red maple and the other group seedlings exhibited
similar trends from the effects of treatments. Both species groups had
significantly more large seedlings in the 50 percent treatment compared with the
70 percent treatment were there was significantly more compared with the
control.
There were significantly more red maple saplings at both Richland
Furnace and Zaleski compared with all other species groups following two
growing seasons after treatment (Table 16). The other species group at Zaleski
is the only other example where there are more saplings per acre compared with
all the other species groups with the exception of red maple. Both treatments at
Richland and Zaleski significantly increased the number of red maple and white
oak saplings per acre relative to the control. While the 50 percent treatment at
37
Richland Furnace significantly increased the number of red oak saplings
compared with the control and 70 percent treatment (Table 16).
Of the total number of red maple, white oak group, and red oak group
sapling stems, the overwhelming majority were of sprout origin versus seed
origin (Table 18). There is a significant difference between the density of red
maple, white oak group and red oak group saplings grown from sprouts versus
seed following two growing seasons after harvest. However, there is no statistical
difference between the 50 and 70 percent treatments with respect to the density
of red maple, white oak group, or red oak group saplings originating from cut
stumps versus originating from seed (Table 17).
There is significant evidence to suggest that there were more sassafras,
non-commercial, and black gum saplings at Zaleski, originating from seed versus
originating from cut stumps (Table 18). However, there is no significant difference
between the 50 and 70 percent treatments with respect to the density of
sassafras, non-commercial, or black gum saplings originating from seed versus
originating from cut stumps (Table 17). There are also no significant differences
between the densities of saplings originating from seed versus from a cut stump
for hickories, yellow-poplar, or the other species group.
4.6 Deer Herbivory Two Years After Harvest Data
At Richland Furnace, there were a significantly higher densities of
sassafras seedlings browsed in the control plots (13 stems per acre browsed)
compared with either of the canopy reduction plots (0 stems per acre browsed).
38
The non-commercial group had a significantly higher densities of small seedlings
browsed in the Richland Furnace (75 stems per acre) compared with the canopy
reduction treatment plots (0 stems per acre). There were no other significant
differences in small seedling densities among treatment types within species
groups at Richland Furnace. Within the 50 percent treatment, small black gum
seedlings were browsed significantly more often (17 stems per acre) compared
to the other species groups. In addition, in the 70 percent treatment small black
gum seedlings were browsed significantly more often (6 stems per acre)
compared to any other species groups. The non-commercial group small
seedlings were browsed significantly more often (75 stems per acre) in the
control treatment compared with either canopy reduction treatment.
At Zaleski, there were no significant differences in browsing observed
between small seedling densities among treatments areas within the species
groups. There were also no significant differences between the density browsed
between species groups in the 50 percent treatment. However, within the 70
percent treatment and control treatments at Zaleski, the other group had a
significantly higher number of stems per acre browsed compared to all other
species groups, 63 and 12 stems per acre, respectively.
At Richland Furnace, there were significantly more large red maple
seedlings per acre browsed (145 stems per acre) in the 50 percent treatment
compared with the 70 percent (30 stems per acre) and control treatment (40
stems per acre). There was also a significant difference between the density of
hickories browsed in the 50 percent treatment (18 stems per acre) compared with
39
the two other treatment areas, where no hickories stems were browsed. Within
the 50 percent treatment, large red maple seedlings were browsed at a
significantly higher frequency (145 stems per acre) compared with all the other
species groups. In the 70 percent treatment, black gum (115 stems per acre) and
the other group (143 stems per acre) were browsed at higher frequencies than
any of the other species groups. The non-commercial group (95 stems per acre)
was browsed significantly more often than any other species in the control
treatment.
At Zaleski, there were significantly more large red maple seedlings per
acre browsed within the 50 percent (218 stems per acre) and 70 percent
treatments (335 stems per acre) compared with the control treatment where no
stems were browsed. There were significantly more large non-commercial group
seedlings browsed in the 70 percent treatment (250 stems per acre) compared
with the 50 percent (3 stems per acre) or the control treatment where no stems
were browsed. There were also more large black gum seedlings browsed in the
50 percent (115 stems per acre) and 70 percent treatments (220 stems per acre)
compared to the control, where no black gum stems were browsed. There were
significantly more sassafras stems browsed in the 50 and 70 percent treatments
(respectively, 218 and 255 stems per acre) relative to the control treatment (0
stems per acre). Red maple (218 stems per acre) and sassafras (218 stems per
acre) were browsed at higher frequencies in the 50 percent treatment compared
with all the other species groups. Red maple (335 stems per acre), the noncommercial group (250 stems per acre), black gum (220 stems per acre), and
40
sassafras (255 stems per acre) were browsed at significantly higher frequencies
compared with the other species groups within the 70 percent treatment.
4.7 Deer Herbivory on Stump Sprouts Two Years After Treatment
Although in the field deer herbivory was recorded on both small and large
seedling sprouts, very few small seedling sprouts were browsed. This is because
the larger sprouts growing from the stumps hid the majority of the smaller
sprouts. As a result only data from large stump sprouts are presented.
There were significantly more large red maple seedlings per acre browsed
in the 50 percent treatment (110 stems per acre) compared with the 70 percent
treatment (9 stems per acre) at Richland Furnace. There were no other
significant differences observed between the densities browsed between
treatments within species groups. Within the 50 percent treatment, red maple
(110 stems per acre) and black gum (84 stems per acre) were browsed at
significantly higher frequencies compared with any other species groups. Within
the 70 percent treatment, black gum (55 stems per acre) and the other species
group (64 stems per acre) had significantly more stems per acre browsed
compared with any other species groups.
Similar to Richland Furnace, there were significantly more large red maple
seedlings browsed in the 50 percent treatment (417 stems per acre) compared
with the 70 percent treatment (198 stems per acre) at Zaleski. There were no
other significant differences among the densities browsed between treatments
within species groups at Zaleski. There were significantly more large red maple
41
seedlings browsed within the 50 percent and 70 percent treatments (respectively
417 and 198 stems per acre) compared with any other species groups.
While it is clear that high densities of white-tailed deer may significantly
influence the morphology and growth rates of seedlings (Jacobs 1969), the deer
densities at Richland Furnace and Zaleski do not appear to be altering the woody
regeneration layers. Even though it is known that white-tailed deer browse
selectively (Healy 1997, Hughes and Fahey 1999), the species diversity should
remain unchanged from the minimal effects from deer herbivory.
Although white-tailed deer were presumed to be a contributing factor with the
problem of oak regeneration survival there is no statistical evidence to verify that
assumption. While the data does suggest that white-tailed deer browse more on
stump sprouts, the effected stump sprouts seemed to be little affected because
of the large number of sprouts per stump and as a result of their rapid growth
rate.
42
Treatment
ACRU
50
70
Control
5912 Ac
11410 Bc
13840 Bc
50
70
Control
13131 Ad
17910 Ac
8700 Ac
CASP
LITU NONCOM NYSY
OTHER
QUAL
QURU
-------------------------Richland Small Seedling-------------------------431 Aa 475 Aa 1656 Ab 506 Aa 1263 Aab 1481 Aab 638 Aa
233 Aa 460 Aa
727 Aa
113 Aa
213 Ba
3513 Bb 1300 Ba
213 Aa 700 Aab 2020 Ab 100 Aa
175 Ba 2850 ABb 738 Aa
-------------------------Zaleski Small Seedling-------------------------112 Aa 331 Aa 2894 ABb 363 Aa 906 ABa 1144 Aa 1750 Aab
121 Aa 556 Aab 3131 Bb 206 Ba 606 Aab 1450 Aab 1356 Aab
50 Aa 25 Aa
2487 Ab
13 Aa 1975 Bab 2863 Aab 1113 Aab
SAAL
838 Ab
1100 Aa
1100 Aa
4530 Ac
3620 Ab
2900 Ab
Table 8. Summary ANOVA of pre-harvest small seedling density, by species
grouping, for Richland Furnace State Forest and Zaleski State Forest. Means
followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
Treatment
ACRU
50
70
Control
3106 Ac
3250 Ac
10125 Bc
50
70
Control
3413 Ac
3275 Ad
3813 Ab
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
-------------------------Richland Small Seedling-------------------------238 Aa 6750 Bd 1075 ABab 612 Aa 800 Bab 2075 Ab 244 Aa
1500 Aa 2300 Ab
575 Aa
363 Aba 300 Aa 3756 Bc 938 Ba
263 Aa 475 Aa
1575 Ba
0 Ba
513 Aa 4438 Bb 963 Ba
-------------------------Zaleski Small Seedling-------------------------100 Aa 1913 Ab
519 Aa
368 Aa 975 Aa 644 Aa 1431 Aab
106 Aa 1231 Bb 600 Aab
356 Aa 713 Aab 1306 Ab 806 Bab
75 Aa
50 Ca
663 Aa
88 Aa 775 Aa 4313 Bb 888 Ba
SAAL
1644 Aab
2119 Ab
625 Ba
5438 Ad
2075 Bc
538 Ca
Table 9. Summary ANOVA of 1 year post-harvest small seedling density, by
species grouping, for Richland Furnace State Forest and Zaleski State Forest.
Means followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
Treatment
50
70
Control
50
70
Control
ACRU
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
SAAL
-------------------------Richland Small Seedling-------------------------1213 Ab 194 Aa 4931 Bc 381 Aab
88 Aa 318 Aab 881 Aab 113 Aa 706 Aab
2700 Ab 125 Aa 1331 Ab
425 Aa
125 Aa 150 Aa 1813 Ab 634 Ba 1569 Bb
12412 Bc 538 Ba 700 Aa 2087 Bab 38 Aa
300 Aa 3413 Bb 938 Ba 1175 ABa
-------------------------Zaleski Small Seedling-------------------------2463 Ac 113 Aa 1081 Aa 1081 Aab 156 Aa 350 Aa 519 Ab 706 Aa 1325 Aab
1475 Ac 163 Aa 919 Ab
606 Aab 188 Aa 363 Aa 575 Aab 388 Ba 1313 Ac
4125 Bc 100 Aa
25 Ba
1875 Bb 100 Aa 613 Ba 3288 Ba 650 Aa 2488 Bbc
Table 10. Summary ANOVA of 2 years post-harvest small seedling st density, by
species grouping, for Richland Furnace State Forest and Zaleski State Forest.
Means followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
43
Treatment
ACRU
50
70
Control
1120 Ac
1067 Ac
1244 Ab
50
70
Control
898 Aab
590 Aab
720 Aab
CASP
LITU NONCOM NYSY
OTHER
QUAL
QURU
-------------------------Richland Large Seedling-------------------------182 Aa 192 Aa 1010 Bc 212 ABa 140 Aa
510 Aab
122 Aa
101 Aa 245 Aa 325 Aab 248 Ba
613 Ba
552 Abc
160 Aa
125 Aa 190 Aa
350 Aa
85 Aa
180 Aa
370 Aa
155 Aa
-------------------------Zaleski Large Seedling-------------------------185 Ba 42 Ba
1082 Ab
65 Aa
320 Aa
280 Aa
422 Aa
68 Aa 10 Aa
910 Ab
70 Aa
102 Ba
288 Aa
290 Aa
170 Ba 0 Aa
1140 Ab
40 Aa
320 Aa
1130 Bb 670 Bab
SAAL
720 Abc
835 Ac
1104 Ab
3415 Bc
1680 Ac
1795 Ac
Table 11. Summary ANOVA of pre-harvest large seedling density, by species
grouping, for Richland Furnace State Forest and Zaleski State Forest. Means
followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
Treatment
ACRU
50
70
Control
532 Ac
728 Ac
1244 Ab
50
70
Control
502 Ab
455 Ab
720 Aab
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
-------------------------Richland Large Seedling-------------------------52 Aa
62 Aa
128 Aab
238 Bb 25 Aa 167 Aab 40 Aa
108 Ba 85 Aa
85 Aa
190 Aab 102 Ba 212 Aab 60 Aa
125 Aa 190 Aa
350 Aa
85 Aa 180 Aa 370 Aa 155 Aa
-------------------------Zaleski Large Seedling-------------------------70 Aa
12 Aa
158 Aa
78 Aa 148 Aa 50 Aa
110 Aa
55 Aa
7 Aa
402 Ab
55 Aab 128 Aa 125 Aa 107 Aa
170 Ba
0 Aa
1140 Ab
40 Aa 320 Aa 1130 Bb 670 Bab
SAAL
252 Ab
265 Ab
1104 Ab
1275 Bc
460 Ab
1795 Ac
Table 12. Summary ANOVA of 1 year post-harvest large seedling density, by
species grouping, for Richland Furnace State Forest and Zaleski State Forest.
Means followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
Treatment
50
70
Control
50
70
Control
ACRU
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
-------------------------Richland Large Seedling-------------------------1645 Ac 262 Aa 1218 Abc 1772 Ac 825 Aab 310 Aa 795 Aab 210 Aa
1478 Ac 120 Ba 335 Bab 315 Bab 582 Ab 382 Aab 955 Ac 168 Aa
1325 Abc 100 Ba 165.2 Ba 685 Bb
30 Ba 225 Aab 530 Aab 300 Bab
-------------------------Zaleski Large Seedling-------------------------1528 Ac 185 Aa
50 Aa
710 Ab 302 Aab 638 Ab 375 Ab 388 Aab
1198 Bc 145 Aa
92 Aa
1112 Ac 370 Aab 368 Bab 548 Ab 398 Aab
310 Ca 115 Aa
0 Aa
885 Ab
50 Ba
110 Ca 655 Aa 440 Aab
SAAL
885 Aab
442 Bab
1000 Ab
2025 Ad
1068 Bc
930 Bb
Table 13. Summary ANOVA of 2 years post-harvest large seedling density, by
species grouping, for Richland Furnace State Forest and Zaleski State Forest.
Means followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
44
Treatment
ACRU
50
70
Control
479 Ac
340 Ac
53 Bb
50
70
Control
290 Ac
219 Ad
175 Ac
CASP
LITU NONCOM NYSY
OTHER
QUAL
QURU
------------------------------Richland Sapling-----------------------------10 Aa 30 Aa
251 Ab 143 Aab
80 Aa
4 Aa
3 Aa
12 Aa 15 ABa
37 Ba
120 Ab
83 Ab
4 Aa
0 Aa
8 Aa
0 Ba
8 Ba
55 Bb
55 Ab
3 Aa
3 Aa
-------------------------------Zaleski Sapling-------------------------------40 ABa 0 Aa
50 Abab 111 Ab
60 Aab
20 Aa
1 Ba
30 Aab 1 Aa
43 Aab
90 Ac
60 Abc
9 Aa
1 Ba
65 Bab 0 Aa
103 Bb
23 Ba
215 Bc
13 Aa
23 Aa
SAAL
41 Aa
24 Aab
18 Aa
3 Ba
13 Ba
70 Aab
Table 14. Summary ANOVA of pre-harvest sapling density, by species grouping,
for Richland Furnace State Forest and Zaleski State Forest. Means followed by
the same capital letter indicate no significant difference within species group
between treatments. Means followed by the same small case letter indicate no
significant difference between species groups within each treatment For all
Tukey’s tests p =0.05.
Treatment
ACRU
50
70
Control
20 Ab
75 Bb
53 Bb
50
70
Control
35 Ac
33 Ab
175 Ac
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
------------------------------Richland Sapling------------------------------0 Ba
0 Aa
8 Aa
19 Ab
1 Ba
6 Aa
6 Aa
5 Aa
0 Aa
5 Aa
16 Aa
6 Ba
1 Aa
0 Aa
8 Aa
0 Ba
8 Ba
55 Bb
55 Ab
3 Aa
3 Aa
-------------------------------Zaleski Sapling--------------------------------5 Aa
0 Aa
2 Aa
20 Ab 10 Aab
0 Aa
0 Aa
8 Aa
0 Aa
9 Aa
11 Aa
30 Ab
0 Aa
1 Aa
65 Bab
0 Aa
103 Bb
23 Ba 215 Bc 13 Aa
23 Aa
SAAL
6 ABa
1 Ba
18 Aa
1 Aa
4 Aa
70 Aab
Table 15. Summary ANOVA of 1 year post-harvest sapling density, by species
grouping, for Richland Furnace State Forest and Zaleski State Forest. Means
followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
Treatment
ACRU
50
70
Control
634 Bb
508 Bc
88 Ab
50
70
Control
600 Cc
355 Bc
143 Ab
CASP
LITU
NONCOM NYSY OTHER QUAL
QURU
-----------------------------Richland Sapling-----------------------------9 Aa
64 Aa
123 Ba
110 Aa
43 Aa 134 Ba 48 Ba
43 Ba
28 Aa
68 Abab 84 Aab
48 Aa 140 Bb 29 ABa
3 Aa
15 Aa
8 Aa
60 Ab
83 Ab
0 Aa
0 Aa
-------------------------------Zaleski Sapling-------------------------------28 Ba
10 Aa
20 Aa
68 Ba
161 Ab 48 Ba
9 Aa
9 Aa
14 Aa
57 Aa
29 Aa
128 Ab 44 Ba
38 Aa
40 Ba
0 Aa
260 Ba
45 Ac
60 Aa
3 Aa
33 Aa
SAAL
39 Aa
6 Aa
65 Ab
56 Aa
33 Aa
45 Aa
Table 16. Summary ANOVA of 2 years post-harvest sapling density, by species
grouping, for Richland Furnace State Forest and Zaleski State Forest. Means
followed by the same capital letter indicate no significant difference within
species group between treatments. Means followed by the same small case
letter indicate no significant difference between species groups within each
treatment. For all Tukey’s tests p =0.05.
45
Treatment ACRU
50
70
Control
81 Ab
116 Ac
88 Ab
50
70
Control
94 Ab
133 Ac
143 Ab
50
70
553 Ab
391 Ac
50
70
506 Bc
222 Ac
CASP LITU NONCOM NYSY OTHER QUAL QURU
SAAL
-----------------------------Richland Sapling-----------------------------1 Aa
9 Aa 119 Bbc 53 Aab
5 Aa
9 Ba
0 Aa 39 ABab
11 Ba 4 Aa
68 Ab
48 Aab
20 Aa
7 ABa
0 Aa
6 Aa
3 Aa 15 Aa
8 Aa
60 Ab
83 Bb
0 Aa
0 Aa
65 Bb
-------------------------------Zaleski Sapling-------------------------------20 Ba 10 Aa
20 Aa
50 Aab
70 Ab
5 Aa
0 Aa
56 Ab
0 Aa
0 Aa
56 Ab
26 Aab
69 Ab
0 Aa
4 Aa
33 Aab
40 Ca 0 Aa
260 Bc
45 Aa
60 Aa
3 Aa
33 Ba
45 Aa
------------------------------Richland Sprouts------------------------------8 Aa 55 Aa
75 Aa
58 Aa
38 Aa 125 Aa 48 Aa
0 Aa
31 Aa 24 Aa
0 Aa
36 Aa
28 Aa 138 Ab 29 Aa
0 Aa
--------------------------------Zaleski Sprouts--------------------------------8 Aa
0 Aa
0 Aa
18 Ba
91 Ab 43 Aab
9 Aa
0 Aa
9 Aa 14 Ba
1 Aa
3 Aa
59 Ab 44 Aab 34 Aab
0 Aa
Table 17. Summary ANOVA of 2 years post-harvest sapling and sprout density
by species grouping for the Richland Furnace State Forest and the Zaleski State
Forest. Means followed by the same capital letter indicate no significant
difference within species group between treatments. Means followed by the
same small case letter indicate no significant difference between species groups
within each treatment. For all Tukey’s tests p =0.05.
Class
ACRU
Sapling
Sprout
81 A
553 B
Sapling
Sprout
116 A
391 B
Sapling
Sprout
94 A
506 B
Sapling
Sprout
133 A
222 A
CASP LITU NONCOM NYSY OTHER QUAL QURU
-----------------------------Richland 50 percent-----------------------------1A
9A
119 B
53 A
5A
9A
0A
8A
55 A
75 A
58 A
38 A
125 B
48 B
-----------------------------Richland 70 percent-----------------------------11 A
4A
68 A
48 A
20 A
7A
0A
31 A
24 A
0A
36 A
28 A
138 B
29 B
------------------------------Zaleski 50 percent------------------------------20 B
10 A
20 B
50 B
70 A
5A
0A
8A
0A
0A
18 A
91 A
43 B
9A
------------------------------Zaleski 70 percent------------------------------0A
0A
56 B
26 B
69 A
0A
4A
9A
14 A
1A
3A
59 A
44 B
34 B
SAAL
39 B
0A
6B
0A
56 B
0A
33 B
0A
Table 18. Summary ANOVA of 2 years post-harvest sapling versus sprout origin,
density, by species grouping for Richland Furnace State Forest and Zaleski State
Forest. Means followed by the same capitol letter indicate no significant
difference between stem origin within species grouping. For all Tukey’s tests p
=0.05.
46
4.8 Summary of Comparisons Between Red Maple, White Oak Group, and
Red Oak Group
At both Richland Furnace and Zaleski, there were decreases in the
number of red maple, white oak group, and red oak group small seedling
densities in both the 50 and 70 percent treatments compared with the preharvest data through one and two growing season after harvest (Figures 4-9). At
both Richland Furnace and Zaleski, the number of red maple, white oak, and red
oak small seedlings remained essentially unchanged in the control plots through
this same time period (Figures 4 and 9). However, there was a considerable
decrease, over 50 percent, in the number of red maple small seedlings between
the pre-harvest and the one growing season after harvest data in the Zaleski
control plots. Suggesting, that the decrease in the number of red maple small
seedlings observed from pre-harvest to one growing season after harvest in the
canopy reduction treatments was due to natural seedling mortality. However, the
densities of red maple small seedlings in the Zaleski control plots one growing
season after harvest and two growing seasons after harvest were very similar
(Figure 9).
In the 50 percent treatment at Richland Furnace, there was a slight
decrease in the density of large seedling red maple stems per acre between the
pre-harvest and the one year after treatment data, followed by an increase two
years subsequent to harvest (Figure 10). Within the 70 percent treatment at
Richland Furnace there was an increase in the density of red maple large
seedlings between the pre-harvest and one year after harvest data, followed by a
47
slight decrease during the second year (Figure 11). However, there was an
overall increase in the density of red maple large seedlings per acre in both the
50 and 70 percent treatments. The number of large red maple seedlings in the
control remained relatively constant over the three-year period (Figure 12). The
density of large red maple seedlings increased over the two years after harvest in
both the 50 and 70 percent treatments (Figures 13 and 14). The density of large
red maple seedlings in the Zaleski control remained almost constant over the
three years (Figure 15).
The density of large white oak group seedlings in both the 50 and 70
percent treatments remained relatively constant between the pre-harvest and the
one growing season after harvest at Richland Furnace. However, their denstiy
nearly doubled two years after treatment (Figures 10 and 11). Similar trends
were observed at Zaleski for large seedlings. The density of large seedlings
between the pre-harvest and one growing season after harvest remained fairly
constant, only dropping slightly in the 50 percent treatment, but nearly
quadrupling after two growing seasons (Figures 13 and 14). The density of large
white oak group seedlings remained fairly constant in the control in Richland
Furnace and Zaleski during the first year after treatment harvests, but increased
slightly two growing seasons after harvest (Figure 15).
The density of large red oak group seedlings remained almost constant
from the pre-harvest to one growing season after harvest in both the 50 and 70
percent treatments, while two growing seasons after harvest there was only a
slight increase in the number (Figures 10 and 11). There was an overall increase
48
in the density of their large seedlings at Richland Furnace harvested areas
however, this trend was also observed in the control treatment area (Figure 12).
In the 50 and 70 percent treatments at Zaleski, the density of large seedling red
oak group stems remained fairly constant between pre-harvest and the one year
after harvest data (Figures 13 and 14). During the second growing season after
harvest the density of large red oak group seedlings nearly tripled compared to
the previous years data. Although, there was an increase in the density of large
red oak group seedlings in the Zaleski control two years after harvest, the
increase was much less than the increase in the harvested areas.
There was a substantial decrease in the density of red maple saplings per
acre between the one growing season after harvest and the pre-harvest at both
Richland Furnace and Zaleski in both the 50 and 70 percent treatments (Figures
16, 17, 19, and 20). However, there was also an increase in the density of red
maple saplings at Richland Furnace and Zaleski in both 50 and 70 percent
treatments following two years after the harvests. In each of these four cases: (1)
Richland 50 percent treatment, (2) Richland 70 percent treatment, (3) Zaleski 50
percent treatment, and (4) Zaleski 70 percent treatment, there was an increase in
the density of red maple saplings from pre-harvest to two growing seasons after
treatment. The density of red maple in Richland Furnace and Zaleski control
plots remained stable throughout the three years sampled (Figures 18 and 21).
The density of white oak group saplings followed similar trends across
both Richland Furnace and Zaleski. There was nearly a complete absence of
white oak saplings pre-harvest and one growing season after harvest (Figures
49
16-21). However, two growing seasons following harvest there was an increase
in the density of white oak saplings in both the 50 and 70 percent treatments
(Figures 16, 17, 19, and 20). However, there are nearly double the density of
white oak group saplings two growing seasons after harvest in the Richland
Furnace harvested areas compared with the Zaleski harvested areas. The
control plots at Richland Furnace and Zaleski remained fairly constant with
respect to the density of white oak group saplings per acre (Figures 18 and 21).
The density of red oak group saplings remained very consistent, i.e., very
few stems per acre, across Richland Furnace and Zaleski 50 and 70 percent
treatments between the pre-harvest and one growing season after harvest
(Figures 16, 17, 19, and 20). However, there was a noticeable increase in the
density of red oak group saplings between two growing seasons after harvest
and pre-harvest across both harvest intensities and both study locations. Similar
to the trends observed for white oak, there were more red oak sapling stems per
acre at Richland Furnace compared with Zaleski. The control plots at Richland
Furnace and Zaleski had a relative absence of red oak saplings from the preharvest to the two growing seasons after harvest data (Figures 18 and 21).
50
20000
18000
16000
Stems per acre
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Species
Figure 4. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 50 percent treatment small seedling plots.
20000
18000
16000
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Sp ecies
Figure 5. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 70 percent treatment small seedling plots.
20000
18000
16000
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Sp ecies
Figure 6. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest control small seedling plots.
51
20000
18000
Stems per acre
16000
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Species
Figure 7. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 50 percent treatment small seedling plots.
20000
18000
16000
Stems per acre
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Species
Figure 8. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 70 percent treatment small seedling plots.
20000
18000
16000
Stems per acre
14000
12000
preharvest
10000
year 1
8000
year 2
6000
4000
2000
0
ACRU
QUAL
QURU
Species
Figure 9. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest control small seedling plots.
52
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 10. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 50 percent treatment large seedling plots.
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 11. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 70 percent treatment large seedling plots.
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 12. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest control large seedling plots.
53
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 13. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 50 percent treatment large seedling plots.
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 14. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 70 percent treatment large seedling plots.
4500
4000
Stems per acre
3500
3000
preharvest
2500
year 1
2000
year 2
1500
1000
500
0
ACRU
QUAL
QURU
Species
Figure 15. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest control large seedling plots.
54
700
600
Stems per acre
500
preharvest
400
year 1
300
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 16. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 50 percent treatment sapling plots.
600
Stems per acre
500
400
preharvest
300
year 1
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 17. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest 70 percent treatment sapling plots.
700
Stems per acre
600
500
preharvest
400
year 1
300
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 18. Summary of stems per acre of ACRU, QUAL, and QURU species in Richland Furnace
State Forest control sapling plots.
55
700
600
Stems per acre
500
preharvest
400
year 1
300
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 19. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 50 percent treatment sapling plots.
700
Stems per acre
600
500
preharvest
400
year 1
300
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 20. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest 70 percent treatment sapling plots.
700
600
Stems per acre
500
preharvest
400
year 1
300
year 2
200
100
0
ACRU
QUAL
QURU
Species
Figure 21. Summary of stems per acre of ACRU, QUAL, and QURU species in Zaleski State
Forest control sapling plots.
56
4.9 Canonical Correspondence Analysis
Originally, CCA was performed on eighteen data sets and ordination
diagrams and Monte Carlo test results were developed for each data set (Tables
19-21). However, only one of the six pre-harvest and two of the six one growing
season after harvest p-values for the species-environment correlation were
statistically significant at p= 0.05. As a result the lack of statistical importance for
the pre-harvest and one year after harvest, only the six CCA ordination diagrams
for the two growing seasons after harvest are presented. Five of the six CCA
diagrams for two growing seasons after harvest resulted with statistically
significant p-values and associated Eigen-values. Four of the six p-values
constructed with the species-environment correlation were significant at p=0.05.
The two CCA ordination diagrams that did not result in statistically significant
species-environment correlation are the Zaleski small seedling plots and the
Zaleski sapling plots. Both may be the result of the improper marking of the
Zaleski 70 percent of full stocking treatment area.
CCA ordination diagrams (Figures 22-27) suggested plots by treatment,
which indicate an effect of treatment on the species density for small seedling,
large seedling, and sapling size classes. These diagrams illustrate the trends
associated between the species groups and the treatment. For example, in all of
the ordination diagrams, yellow-poplar is strongly associated with the 50 percent
treatment in comparison to the control and 70 percent treatment. This is largely
because yellow-poplar, a shade-intolerant species, is most strongly associated
57
with the plots with the lowest percent stocking, which is equivalent to the least
amount of shading.
The CCA ordination diagram for the Richland Furnace small seedling plots
effectively separated the plots by treatment (Figure 22). This allows an
interpretation of the effect of treatment on the species groups distinguished on
the diagram. Black gum, hickories, yellow-poplar, and the other group are most
strongly associated with the 50 percent treatment. While sassafras is the only
species strongly associated with the 70 percent treatment. Red maple, white oak
group, red oak group, and the non-commercial group were most strongly
associated with the control treatment.
The ordination diagram for the Richland Furnace large seedling plots also
effectively separated plots by treatment (Figure 23). Black gum, hickories, yellowpoplar and the non-commercial group were all separated towards the 50 percent
treatment. Red maple, white oak grooup, sassafras, and the other group were
most strongly associated with the 70 percent treatment. While red oak group was
the only species grouped with the control treatment.
The Richland Furnace sapling plots ordination diagram separated plots by
treatment relatively effectively (Figure 24). Although, there are three plots, two
from the 50 and one from the 70 percent treatments, which are separated from
the other sapling plots. However, according to the Monte Carlo test results (Table
21) the results for this CCA are significant. Red maple, yellow-poplar, white oak
group, red oak group, sassafras, and the non-commercial group were all
separated out with the 50 percent treatment. Black gum, hickories, and the other
58
group were associated with the 70 percent treatment. No species most strongly
correlated with the control treatment.
While the CCA from the Zaleski small seedling plots (Figure 25) are not
statistically significant with respect to the species-environment correlation, the
CCA still may be used to devise general trends of the effect of treatment on the
species densities (Table 18). One of the issues with this diagram is that the 50
and 70 percent treatment plots are actually the reverse of what should be
expected i.e., the 50 percent treatment plots are located between the control and
the 70 percent treatment plots. However, this is a direct result of the 70 percent
treatment actually having a lower percent stocking than the 50 percent treatment.
Black gum, hickories, yellow-poplar, and the other group are all associated with
the 50 and 70 percent treatments. Red maple, white oak group, red oak group,
sassafras, and the non-commercial group are separated with the control
treatment.
The Zaleski large seedling plots ordination diagram separated plots fairly
effectively (Figure 26). While some of the 50 and 70 percent treatment plots are
intermingled, this result should have been expected because of the problems
associated with the 70 percent treatment. Red maple, black gum, yellow-poplar,
sassafras, and the other group were separated with the 50 percent treatment.
Hickories, white oak group, and red oak group were associated with the 70
percent treatment. While the non-commercial group was the only group
associated with the control.
59
The ordination diagrams for the Zaleski sapling plots separated the
treatments out relatively effectively (Figure 27). Although the speciesenvironment correlation value is not significant, general trends may still be
observed to relate the correlation between the treatments and species (Table
20). Red maple, black gum, yellow-poplar, white oak group, sassafras, and the
other group are all associated with the 50 percent treatment. The red oak group
was the only species correlated with the 70 percent treatment. Hickories and the
non-commercial group were most strongly related to the control.
60
Location
Richland
Zaleski
Richland
Zaleski
Richland
Zaleski
Plot Size
Small Seedling
Small Seedling
Large Seedling
Large Seedling
Sapling
Sapling
Eigen Value
0.033
0.016
0.060
0.017
0.030
0.005
p-value
0.173
0.312
0.023
0.421
0.202
0.884
Spp-Envt Corr.
0.479
0.410
0.612
0.322
0.365
0.282
p-value
0.381
0.320
0.021
0.681
0.307
0.899
Table 19. Summary of pre-harvest Canonical Correspondence Analysis Monte
Carlo test results by location and plot size.
Location
Richland
Zaleski
Richland
Zaleski
Richland
Zaleski
Plot Size
Small Seedling
Small Seedling
Large Seedling
Large Seedling
Sapling
Sapling
Eigen Value
0.090
0.031
0.023
0.022
0.009
0.035
p-value
0.002
0.110
0.195
0.233
0.301
0.007
Spp-Envt Corr.
0.619
0.435
0.461
0.445
0.417
0.579
p-value
0.002
0.257
0.251
0.307
0.353
0.049
Table 20. Summary of 1 year post-harvest Canonical Correspondence Analysis
Monte Carlo test results by location and plot size.
Location
Richland
Zaleski
Richland
Zaleski
Richland
Zaleski
Plot Size
Small Seedling
Small Seedling
Large Seedling
Large Seedling
Sapling
Sapling
Eigen Value
0.098
0.041
0.068
0.033
0.068
0.029
p-value
0.004
0.013
0.030
0.024
0.029
0.157
Spp-Envt Corr.
0.572
0.503
0.549
0.649
0.602
0.419
p-value
0.029
0.124
0.055
0.003
0.018
0.373
Table 21. Summary of 2 years post-harvest Canonical Correspondence Analysis
Monte Carlo test results by location and plot size.
61
Figure 22. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest small seedling plots two growing seasons after harvest.
Control plots represented by blue circles. 50 percent of full stocking plots
represented by red triangles. 70 percent of full stocking plots represented by
green squares.
62
Figure 23. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest large seedling plots two growing seasons after harvest.
Control plots represented by blue circles. 50 percent of full stocking plots
represented by red triangles. 70 percent of full stocking plots represented by
green squares.
63
Figure 24. Canonical Correspondence Analysis ordination diagram for Richland
Furnace State Forest sapling plots two growing seasons after harvest. Control
plots represented by blue circles. 50 percent of full stocking plots represented by
red triangles. 70 percent of full stocking plots represented by green squares.
64
Figure 25. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest small seedling plots two growing seasons after harvest. Control
plots represented by blue circles. 50 percent of full stocking plots represented by
red triangles. 70 percent of full stocking plots represented by green squares.
65
Figure 26. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest large seedling plots two growing seasons after harvest. Control plots
represented by blue circles. 50 percent of full stocking plots represented by red
triangles. 70 percent of full stocking plots represented by green squares.
66
Figure 27. Canonical Correspondence Analysis ordination diagram for Zaleski
State Forest sapling plots two growing seasons after harvest. Control plots
represented by blue circles. 50 percent of full stocking plots represented by red
triangles. 70 percent of full stocking plots represented by green squares.
67
CHAPTER 5
5.1 CONCLUSIONS AND MANAGEMENT IMPLICATIONS
Both the 50 and 70 percent of full stocking treatments successfully
released the oak regeneration, allowing oak seedlings to advance in size, which
was a main objective of this study. The treatments were also successful in
generating substantial ingrowth of red maple and yellow-poplar. These species
could become problematic to the establishment and growth of oak regeneration
in the forests. There is wide recognition that oaks are highly fire-adapted, and
that fire played an important role in the ecology of oak forests in the past,
particularly in promoting the dominance of oak in the regeneration layers.
Therefore, the use of prescribed fire could be warranted as a means to control
the competitive vegetation.
Oak species have multiple adaptations to fire, including thick bark that
protects them from damage by surface fires. In addition, the adaptability of oaks
to fire is related to their ability to sprout from the root collar after the top has been
killed (Liming and Johnson 1944). Oaks also have hypogeal germination, i.e.,
their cotyledons are below the surface of the soil because acorns are often
buried by squirrels, blue jays, or other birds and animals. Because these
cotyledons are protected from the heat of surface fires (soil is a poor conductor of
heat), oaks are better able to successfully sprout following fire than many of their
68
competitors whose seeds germinate on the soil surface, such as yellow-poplar
(Brose and Van Lear 1998, Brose et al. 1999). The seedlings of other species
are more susceptible to root-kill, thus giving oaks an advantage over their
associates (Niering et al. 1971, Swan 1970).
Within an oak forest, the degree to which the upper canopy layers are
reduced will be influenced by fire intensity, which is often determined by fuel
conditions and site characteristics. Schwemlein and Williams (M.S. Thesis 2004)
found that recorded fire temperatures were higher on upper slope positions of
south-facing slopes in oak forests of Ohio. Theoretically, fires in the late fall will
burn more intensely than fires occurring in late winter-early spring due to the fact
that newly fallen litter is less compact and drier in late fall. However, cooler air
temperatures may offset this effect, and due to varying weather conditions
differences between fall and spring burns have not always been consistent
(DeSeben and Clebsch 1991). Fire intensity will also be influenced by soil
moisture regimes and burn periodicity (Artman et al. 2001).
Spring burns work best when combined with shelterwood cuts because
the best fuel and weather conditions occur during this time to achieve the
necessary hot fires (Brose and Van Lear 1998, Brose et al.1999). If a single burn
is performed in conjunction with these cuts, then a medium-high (flame length
0.75 m – 1.2 m) to high (flame length > 1.2 m) intensity burn is necessary to
achieve increased oak reproduction (Brose et al. 1999a, Cooper et al. 1999).
While fall burns can produce hot fires, competition control is not achieved as in
spring burns because of the root carbohydrate reserves that exist in the
69
competing vegetation during this time of year (Hodgkins 1958, Langdon 1981).
For the Piedmont region, it has been recommended that these prescribed burns
should occur 3-5 years after the initial harvest. This waiting period allows the
oaks competitors to respond to the initial harvest, and the competing
regeneration is more vulnerable to subsequent disturbances than oak (Brose et
al. 1999).
Based on observations, the results from these silvicultural treatments
should remain consistent for similar oak-hickory forests across the central
hardwood region. However, it must be understood that before a particular
silvicultural treatment is conducted, advance oak regeneration is vital to the
overall success of the treatments. Managers should time their shelterwood
harvest with either high levels of oak regeneration or with years of good acorn
production.
70
LITERATURE CITED
Abrams, M. D. 1998. The red maple paradox. BioScience. pp. 355-364.
Apsley, D.K. and B.C. McCarthy,. 2004. White-tailed deer herbivory on forest
regeneration following fire and thinning treatments in southern Ohio mixed
oak forests. In: Yaussy, D.A., Hix, D.M., Long, R.P., Goebel, C.P. (eds.),
Proceedings of the 14th Central Hardwoods Forest Conference, General
Technical Report NE-316. Wooster, OH, March 16-19. USDA Forest
Service, Northeastern Research Station, Newtown Square, PA. pp. 461471.
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. pp. 1423-1434.
Brose, P.H. and D.H. Van Lear. 1998. Responses of hardwood advance
regeneration to seasonal prescribed fires in oak-dominated shelterwood
stands. Canadian Journal of Forest Research. 28: 331- 339.
Brose, P. H., D.H. Van Lear, and R. D. Cooper. 1999. Using shelterwood
harvests and prescribed fire to regenerate oak stands on productive
upland sites. Forest Ecology and Management pp. 125-141.
Brose, P.T., D.H. Van Lear, and P.D. Keyser. 1999b. A shelterwood-burn
technique for regenerating productive upland oak sites in the Piedmont
region. Southern Journal of Applied Forestry. 23:158-163.
Buckley, D.S. and R.M. Evans. 2004. Effect of season of cutting on the
production and growth of white and shumard oak stump sprouts in
Tennessee In: Yaussy, Daniel A.; Hix, David M.; Long, Robert P.; Goebel,
P. C. (eds.) Proceedings, 14th Central Hardwood Forest Conference;
2004 March 16 19; Wooster, OH. Gen. Tech. Rep. NE-316. Newtown
Square, PA: U.S. Department of Agriculture, Forest Service, Northeastern
Research Station: 45-51.
Burns, R. M.; Honkala, B. H. 1990. Silvics of North America. Volume 2:
Hardwoods. Agric. Handbook 654. USDA For. Serv. Washington, DC. pp
878.
71
Campbell, T.A., B.R. Laseter, W.M. Ford, R.H. Odom, and K.V. Miller. 2006.
Abiotic factors influencing deer browsing in West Virginia. Northern
Journal of Applied Forestry pp. 20-26
Castleberry, S.B., W.M. Ford, K.V. Miller, and W.P. Smith. 1999. White-tailed
deer browse preferences in a southern bottomland hardwood forest.
Southern Journal of Applied Forestry pp. 78-82
Clark, F.B. 1993. An historical perspective of oak regeneration. In: Loftis, D;
McGee, C.E. (eds.) Oak regeneration: serious problems, practical
recommendations. Symposium Proceedings; 1992 September 8-10;
Knoxville, TN. Gen. Tech. Rep. SE-84. Asheville, NC: U.S. Department of
Agriculture, Forest Service, Southeastern Forest Experiment Station: pp.
3-13.
Cooper, R. D., D. H. Van Lear, and P. H. Brose (1999). Oak advance
regeneration following seasonal prescribed fires in mixed hardwood
shelterwood stands. In: Haywood, J. D., ed. Proceedings of the tenth
biennial southern silvicultural research conference; 1999 February 16-18;
Shreveport, LA. Gen. Tech. Rep. SRS-30. Ashville, NC; US Department of
Agriculture, Forest Service, Southern Research Station. 618 p.
Crow, T.R. 1988. Reproductive mode and mechanisms for self-replacement of
northern red oak (Quercus rubra) – a review. Forest Science. pp. 19-40.
DeSeben, H. R, and E. E. C. Clebsch. 1991. Response types to prescribed fire in
oak forest understory pages 22-33. In: Nodvin, S.C.; Waldrop, T.A., (eds.)
Fire and the environment: ecological and cultural perspectives:
Proceedings of an international symposium; 1990 March 20-24; Knoxville,
TN. Gen. Tech. Rep. SE-69. Asheville, NC: U.S. Department of
Agriculture, Forest Service, Southeastern Forest Experiment Station.
Dey, D.C. and W.C. Parker. 1996. Regeneration of red oak (Quercus rubra L.)
using shelterwood systems: Ecophysiology, silviculture and management
recommendations. Ontario Ministry of Natural Resources, Ontario Forest
Research Institute, Sault Ste. Marie, Ontario, Forest Research Information
Paper No. 126, pp. 59.
Dey, D.C. 2002. The ecological basis for oak silviculture in Eastern North
America. In: McShea, W.J.; Healy, W.M. (eds.) Oak forest ecosystems:
Ecology and management for wildlife. Baltimore, MD: The Johns Hopkins
University Press: pp. 60-79.
72
Dickson, J. G. 2004. Wildlife and upland oak forests. In Spetich, Martin A., (ed.)
2004. Upland oak ecology symposium: history, current conditions, and
sustainability. Gen. Tech. Rep. SRS-73. Asheville, NC: U.S. Department
of Agriculture, Forest Service, Southern Research Station. pp. 311.
Dolan, B. J., G. R. Parker. 2004. Understory response to disturbance: an
investigation of prescribed burning and understory removal treatments. In
Spetich, Martin A., (ed.) 2004. Upland oak ecology symposium: history,
current conditions, and sustainability. Gen. Tech. Rep. SRS-73. Asheville,
NC: U.S. Department of Agriculture, Forest Service, Southern Research
Station. pp. 311.
Ehrenfeld, J.G. 1980. Understory response to canopy gaps of varying size in a
mature oak forest. Bull. Torrey Bot. Club. pp. 29-41.
Farmer, R.E., Jr. 1975. Growth and assimilation rate of juvenile northern red oak:
effects of light and temperature. Forest Science. pp. 373-381.
Fei, S. and K.C. Steiner. 2007. Evidence for increasing red maple abundance in
the eastern United States. Forest Science pp. 473-477.
Fralish, J. S. 2004. The keystone role of oak and hickory in the central hardwood
forest. In Spetich, Martin A., (ed.) 2004. Upland oak ecology symposium:
history, current conditions, and sustainability. Gen. Tech. Rep. SRS-73.
Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern
Research Station. pp. 311.
Goebel, P.C. and D.M. Hix. 1997. Changes in the composition and structure of
mixed-oak, second growth forest ecosystems during the understory
reinitiation stage of stand development. Ecoscience. pp. 327-340.
Graney, D.L., and T.L. Rogerson. 1985. Development of oak, ash and cherry
reproduction following thinning and fertilization of upland hardwood stands
in the Boston Mountains of Arkansas. In Proceedings. 3rd Biennial
southern silvicultural research conference. USDA For. Serv. Gen. Tech.
Rep. SO-54. pp. 171-177.
Griffith, D.M., D.M. Giovanni, T.L. Witzel, and E.H. Warton. 1993. Forest
statistics for Ohio, 1991. Resource Bulletin NE-128. USDA Forest Service,
Northeastern Forest Experiment Station, Radnor, Pennsylvania.
73
Healy, W.M. 1997. Influence of deer on the structure and composition of oak
forests in Central Massachusetts. In: W.J. Mcshea, H.B. Underwood and
J.H. Rappole, (eds.) The Science of Overabundance: deer ecology and
population management. Washington, DC: Smithsonian Institution Press:
pp. 249-266
Hicks, R.R. 1998. Ecology and management of central hardwood forests. John
Wiley and Sons, Inc., New York. 412 pp.
Hodgkins, E. J. 1958. Effect of fire on undergrowth vegetation in upland southern
pine forests. Ecology. pp. 36-46.
Hughes, J.W. and T.J. Fahey. 1991. Availability, quality, and selection of browse
by white-tailed deer after clearcutting. Forest Science, pp. 261-270
Hutchinson, T.F., D. Rubino, B.C. McCarthy, and E.K. Sutherland. 2003. History
of Forests and Land Use. Chapter 2. In: Sutherland, E. K.; Hutchinson, T.
F., eds. Characteristics of mixed oak forest ecosystems in southern Ohio
prior to the reintroduction of fire. Gen. Tech. Rep. NE-299. Newtown
Square, PA: U.S. Department of Agriculture, Forest Service, Northeastern
Research Station. pp. 17-27.
Jackson, S. W.; Buckley, D. S. 2004. First-year effects of shelterwood cutting,
wildlife thinning, and prescribed burning on oak regeneration and
competitors in tennessee oak hickory forests. Connor, Kristina F., (ed.)
2004. Proceedings of the 12th biennial southern silvicultural research
conference. Gen. Tech. Rep. SRS-71. Ashville, NC: U.S. Department of
Agriculture, Forest Service, Southern Research Station. pp. 594.
Jacobs, R.D., and R.D. Wray. 1992. Managing oak in the driftless Area. Univ.
Minnesota Ext. Serv. Publ. NR-BU-5900-S. pp. 32.
Jacobs, R.D. 1969. Growth and development of deer-browsed sugar maple
seedlings. Journal of Forestry. pp. 870-874
Johnson, E.S. and R.D. Jacobs. 1981. Northern red oak regeneration after
preherbicided clearcutting and shelterwood removal cutting. USDA For.
Serv. Res. Pap. NC-202. pp. 5.
Kerr, J.W. 1985. Soil Survey of Jackson County, Ohio. Department of Agriculture,
Soil and Conservation Service, Ohio Department of Natural Resources,
Division of Soil and Water Conservation, Columbus, Ohio, USA.
74
Langdon, O. G. 1981. Some effects of prescribed fire on understory vegetation in
loblolly pine stands. pp. 143-153. In Proc: Prescribed fire and wildlife in
southern forests. Wood, G. W. (ed.). Clemson University, Georgetown,
SC.
Liming, F. G., and J. P. Johnson. 1944. Reproduction in oak-hickory forest
stands of the Missouri Ozarks. Journal of Forestry. pp. 175-180.
Loftis, D.L. 1983. Regenerating southern appalachian hardwoods with the
shelterwood method. Southern Journal of Applied Forestry. pp. 212-217.
Loftis, D.L 1990. A shelterwood method for regenerating red oak in the southern
Appalachians. Forest Science. pp. 917-929.
Lorimer, C.G. 1985. The role of fire in the perpetuation of oak forests. In Proc.
Challenges in oak management and utilization conf. Univ. Wisconsin
Madison. pp. 8-25.
Lorimer, C.G. 1989. The oak regeneration problem: new evidence on causes and
possible solutions. Univ. Wisconsin-Madison For. Resour. Anal. No. 8.
Publ. R3484. pp. 31.
Lorimer, C.G. 1993. Causes of the oak regeneration problem. In: Loftis, D,
McGee, C. (eds.) Oak regeneration: Serious problems, practical
recommendations; 1992 September 8-10; Knoxville, TN. Gen. Tech. Rep.
SE-84. U.S. Department of Agriculture, Forest Service: pp. 14-39.
Lorimer, C.G., J.W. Chapman, and W.D. Lambert. 1994. Tall understorey
vegetation as a factor in the poor development of oak seedlings beneath
mature stands. Journal of Ecology. pp. 227-237.
McCarthy, B.C., C.A Hammer, G.L. Kauffman, and P.D. Cantino. 1987.
Vegetation patterns and structure of an old-growth forest in southeastern
Ohio. Bulletin of the Torrey Botanical Club. pp. 33-45.
McCune, B. 1997.Influence of noisy environmental data on canonical
correspondence analysis. Ecology. pp. 2617-2623.
McCune, B., and M.J. Mefford. 1999. PC-ORD. Multivariate analysis of ecological
data. Version 5.0. MjM Software, Gleneden Beach, Oregon, U.S.A.
McGee, C.E. 1984. Heavy mortality and succession in a virgin mixed mesophytic
forest. USDA For. Serv. Res. Pap. SO-2099 pp.
75
McWilliams, W.H., R.A. O’Brien, G.C. Reese, and Waddell. 2002. Distribution
and abundance of oaks in North America. In: McShea, W.J.; Healy, W.M.
eds. Oak forest ecosystems: Ecology and management for wildlife.
Baltimore, MD: The Johns Hopkins University Press: pp. 13-33.
Niering, W.A.; R.H. Goodwin: and S. Taylor. 1971. Prescribed burning in
southern New England: introduction to long-range studies. In:
Proceedings, annual Tall Timbers fire ecology conference; 1970 August
20-21; Fredericton, NB. No. 10. Tallahassee, FL: Tall
Timbers Research Station: pp. 267-286.
Nowacki, G.J. and M.D. Abrams. 2008. The demise of fire and “mesophication”
of forests in the Eastern United States. BioScience pp. 123-138.
Nyland, R.D., L.P.Abrahamson, and K.B. Adams. 1982. Use of prescribed fire for
regenerating red and white oak in New York. In Proc. 1982 SAF national
convention, America's hardwood forests - opportunities unlimited. Soc.
Amer. For. Publ. 83-04. pp. 163-167.
Oswalt, C.M., W.K. Clatterbuck, A.E. Houston, S.E. Schlarbaum. 2006.
Implications of large oak seedlings on problematic deer herbivory. Gen.
Tech. Rep. SRS-92. Asheville, NC: USDA, Forest Service, Southern
Research Station: pp. 26-29.
Pubanz, D.M., and C.G. Lorimer. 1992. Oak regeneration experiments in
southwestern Wisconsin: two-year results. Univ. Wisconsin-Madison Publ.
R3552. 16 pp.
Rebbeck, J., R. Long, and D. Yaussy. 2004. Survival of hardwood seedlings and
saplings following overstory thinning and prescribed fires in mixed-oak
forests of southern Ohio. Spetich, Martin A., ed. 2004. Upland oak ecology
symposium: history, current conditions, and sustainability. Gen. Tech.
Rep. SRS-73. Ashville, NC: U.S. Department of Agriculture, Forest
Service, Southern Research Station. pp. 275-284.
Sander, I.L. 1979. Regenerating oaks with the shelterwood system. In: Holt,
Harvey; Fischer, Burnell, eds. Proceedings of the 1979 John S. Wright
forestry conference, regenerating oaks in upland hardwood forests; 1979
February 22-23;West Lafayette, IN. Purdue University: pp. 54-60.
Sander, I.L. 1988. Guidelines for regenerating Appalachian oak stands. In Proc.
Guidelines for regenerating Appalachian hardwood stands. Society of
American Foresters Publication. pp. 189-198
76
Schlesinger, R.C., I.L. Sander, and K.R. Davidson. 1993. Oak regeneration
potential increased by shelterwood treatments. Northern Journal of
Applied Forestry pp. 149-153.
Sloan, K.R., and D.E. Zastrow. 1985. Red oak regeneration trials on the
Navarino Wildlife Area. In Proc. Conf. Columbia, MO. Challenges in oak
management and utilization. Univ. Wisconsin-Madison Coop. Ext. Serv.
pp. 52-161.
Sutherland, E.K. 1997. History of fire in a southern Ohio second-growth mixedoak forest. In: Pallardy, S.G.; Cecich, R.A.; G.H, Gene; Johnson, P.S.,
(eds.) Proceedings, 11th central hardwood forest conference. 1997 March
23-26; Columbia, MO. Gen. Tech. Rep. NC-188. St. Paul, MN: U.S.
Department of Agriculture, Forest Service, North Central Forest
Experiment Station: pp. 172-183.
Sutherland, E.K, B.J. Hale, D.M. Hix. 2000. Defining species guilds in the Central
Hardwood Forest, USA. Plant Ecology. pp. 1-19.
Sutherland, E.K. T.F. Hutchinson, D.A. Yaussy. 2003. Introduction, study area
description, and experimental design. In: Sutherland, E.K. and
Hutchinson, T.F. (eds.), Characteristics of Mixed-oak Forest Ecosystems
in Southern Ohio Prior to the Reintroduction of Fire. U.S.D.A. Forest
Service, Northeastern Research Station, General Technical Report NE299. Newtown Square, PA. pp. 1-16
Swan, F. R., Jr. 1970. Post-fire response of four plant communities in southcentral New York State. Ecology. pp. 1074-1082.
Van Lear, D.H. 2004. Upland oak ecology and management. In Spetich, Martin
A., ed. 2004. Upland oak ecology symposium: history, current conditions,
and sustainability. Gen. Tech. Rep. SRS-73. Asheville, NC: U.S.
Department of Agriculture, Forest Service, Southern Research Station. pp.
65-71.
Van Lear, D.H., and T.A. Waldrop. 1988. Effects of fire on natural regeneration in
the Appalachian Mountains. In Proc. Guidelines for regenerating
Appalachian hardwood stands. Society of American Foresters Publication
pp. 56-68.
Williams, R. A. 2003. Use of stand density index as an alternative to stocking
percent in upland hardwoods. Northern Journal of Applied Forestry pp.
137-142.
77
Wolf, W.W., Jr. 1988. Shelterwood cutting to regenerate oaks - the Glatfelter
experience. In Proc. Guidelines for regenerating Appalachian hardwood
stands. Society of American Foresters Publication pp. 210-218.
78