SUBSTRATE EFFECTS ON SEEDLING DYNAMICS IN MANAGED NORTHERN
HARDWOOD FORESTS
By
John Lawrence Willis

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Forestry - Doctor of Philosophy
2014

ABSTRACT
SUBSTRATE EFFECTS ON SEEDLING DYNAMICS IN MANAGED NORTHERN
HARDWOOD FORESTS
By
John Lawrence Willis
In order to sustain themselves within forests any tree species must be able to
reproduce. Accomplishing this task is often difficult as trees must pass through a variety of
life history stages and transitions in order to reach sexual maturity. Given this complexity,
it is often difficult to identify why some certain species regenerate prolifically within
forests while others often fail to regenerate. This dissertation investigated how different
types of forest substrate (bare mineral soil and decaying coarse woody debris) influence
individual species’ ability to establish seedlings on the forest floor in the managed northern
hardwood forests of the Great Lakes region, where only a small percentage of species are
reported to be reproducing successfully. I consider substrate’s influence on seedling
establishment from three separate starting points: seed added to the forest floor, seed
dispersed from local seed sources and seedlings established on bare mineral soil and
different types of decaying coarse woody debris. Taking this approach allowed me to
investigate not only how substrate influences seedling dynamics, but also how substrate
interacted with several factors which could override the influence of substrate including
light availability, competing non-tree vegetation, site quality (soil and nutrient
availability), local seed source density, and mycorrhizal fungi. Beginning with the presence
of a seed (seed additions), bare mineral soil demonstrated a strong influence over smaller
seeded species germination, but not larger seeded white pine. In addition, increasing light

availability improved seedling establishment for all species. However, under natural
regeneration conditions (without seed additions) the importance of bare mineral soil for
smaller seeded species and light for all species’ seedling establishment became less
apparent as, in many situations, local seed source density and site quality exerted a
stronger influence over species’ initial establishment. Differing types of forest substrate,
however, were strongly influential on individual species growth after establishment, as
bare mineral soil and decaying hemlock, northern white cedar and paper birch consistently
supported greater seedling growth compared to decaying yellow birch, balsam fir, and
sugar maple. In addition, seedlings colonized by mycorrhizal fungi generally shown to
positively influence seedling growth across substrates, and in certain situations, were
associated with dramatic rank changes in species growth performance on different
substrates. Collectively, this dissertation demonstrates how different types of forest
substrate can alter seedling layer species composition through its differing effects on
individual species’ seedling establishment. However, it also demonstrates how a host of
other factors can mitigate the influence of substrate, suggesting that managing for substrate
alone may not reverse the current reproduction failures experienced by several northern
hardwood tree species.

For all those dedicated to forest conservation. Lang lebe der Förster!

iv

ACKNOWLEDGMENTS

Without a doubt, I would not be writing this dissertation without the support of
several individuals who have helped me along the way. Among this group, no one is more
responsible for me reaching this point than my parents. I cannot thank you enough for all
the support you have given me over the years and for teaching me the power of
perseverance. “Tough times don’t last, tough people do.” Gregory Peck.
I acknowledge my committee: Dr. David Rothstein, Dr. Lars Brudvig and Dr. Rich
Kobe for their helpful advice throughout this process. Your contributions and insights have
not only improved this dissertation, but have also facilitated my development as a scientist.
Thank you for your guidance.
No field based forestry dissertation could ever be completed without the help of my
labmates: Evan Farinosi, Ben Daly, Emily Standen, Mike Cook, Ben Schram, Kaitlyn
Snarski and Casey Koleski. It wasn’t always easy and it wasn’t always predictable, but
working with you will always be a fond memory of mine. Thank you for taking stake in
this project and treating it as your own.
Several non-lab members have assisted me throughout this process. I thank Dr.
Laura Marx and Dr. Sarah Tannis for their advice on field work and life as a graduate
student, Tim Greco and David Neumann for all their logistical support, Randy Klevickas
and Paul Bloese for all their assistance at the Tree Research Center and Sasha and Odo for
their help with small mammal research and security in the field. Thank you for your time
and your interest.

v

Tina Ciaramitaro, you are undoubtedly the most patient, understanding and
supportive girlfriend anyone could have. You sacrificed much of what is important to you
so that we could keep “The Unit” intact. There is no way I could have done this without
you. Consequently, I consider this dissertation as much yours as it is mine. Thank you.
Finally, I recognize my advisor Dr. Mike Walters. Nothing I say or do can ever
repay you for what you have done for me. Thank you for seeing potential in me and
stepping in when I really needed help. Thank you for allowing me to run my own ship and
for pushing me when I needed to be pushed. You are the best. Thanks for everything.

vi

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................................... ix
LIST OF FIGURES ........................................................................................................... xi
CHAPTER I……………………………………………………………………………….1
Introduction ............................................................................................................. 1
Dissertation Structure .............................................................................................. 1
Understory Composition ......................................................................................... 3
Major Harvesting Events......................................................................................... 4
LITERATURE CITED ........................................................................................... 7
CHAPTER II - Scarification and gap size have interacting effects on northern temperate
seedling establishment……………………………………………………………10
Abstract ................................................................................................................. 10
Introduction ........................................................................................................... 12
Methods…………………………………………………………………………..18
Study site ................................................................................................... 18
Experiment one.......................................................................................... 19
Experiment two.......................................................................................... 21
Statistical analysis..................................................................................... 22
Results……………………………………………………………………………25
Germination .............................................................................................. 25
Seedling establishment and survival ......................................................... 25
Temporal changes in seedling density in harvest gaps ............................. 26
Discussion………………………………………………………………………..26
Germination .............................................................................................. 26
Seedling establishment and survival ......................................................... 29
Conclusion and Management Recommendations ................................................. 31
Figures ................................................................................................................... 32
Tables .................................................................................................................... 37
LITERATURE CITED ......................................................................................... 43
CHAPTER III - Does substrate availability alter natural regeneration patterns within
harvest gaps?..........................................................................................................51
Abstract ................................................................................................................. 51
Introduction ........................................................................................................... 53
Methods…………………………………………………………………………..57
Study area ................................................................................................. 57
Field methods ............................................................................................ 58
Field measurements .................................................................................. 59
Statistical methods .................................................................................... 60
Results……………………………………………………………………………62

vii

Seedling establishment .............................................................................. 62
Discussion………………………………………………………………………..64
Seedling Establishment ............................................................................. 64
Conclusion and Management Recommendations ................................................. 67
Figures ................................................................................................................... 68
Tables .................................................................................................................... 71
LITERATURE CITED ......................................................................................... 76
CHAPTER IV - The effect of substrate type and mycorrhizal fungi on seedling
development for northern temperate seedlings…………………………………..84
Abstract ................................................................................................................. 84
Introduction ........................................................................................................... 86
Methods…………………………………………………………………………..90
Field methods ............................................................................................ 90
Shade house methods ................................................................................ 90
Laboratory methods .................................................................................. 93
Statistical methods .................................................................................... 94
Results……………………………………………………………………………96
Height growth............................................................................................ 96
Seedling survival ....................................................................................... 97
Mycorrhizal colonization patterns and effects .......................................... 97
Sterilization’s effect on performance rankings ......................................... 98
Discussion………………………………………………………………………..98
Height growth............................................................................................ 98
Seedling survival ..................................................................................... 101
Mycorrhizal colonization ........................................................................ 102
Conclusions and Management Recommendations ............................................. 102
Figures ................................................................................................................. 104
Tables .................................................................................................................. 111
LITERATURE CITED ....................................................................................... 117
CHAPTER V - Conclusions ........................................................................................... 122

viii

LIST OF TABLES

Table 2.1. Results of a generalized linear mixed model for the effects of Scarification,
Light Availability, Species, and their interactions on germination for eastern hemlock,
yellow birch, paper birch, and white pine. Each combination of factors was considered up
to three way interactions in the original model. All interactions Prob ChiSq > 0.25 were
pooled with the error term and the models rerun (Bancroft 1964)………………………...37
Table 2.2. Observed germination response of paper birch, yellow birch, eastern hemlock,
and white pine in Scarified and Unscarified subplots. SD=standard deviation…………...38
Table 2.3. Summary of environmental conditions across different gap sizes classes……..39
Table 2.4. Results of a generalized logistic mixed model for the effects of Clipping,
Light Availability, Species, and their interactions on germinant establishment
(1st year survival) and second year survival for eastern hemlock, yellow birch, paper birch,
and white pine. Each combination of factors was considered up to three way interactions in
the original model. All interactions Prob ChiSq > 0.25 were pooled with the error term
and the models rerun (Bancroft 1964)…………………………………………………….40
Table 2.5. Results of a generalized logistic mixed model for the effects of Clipping,
Light Availability, Species, Fencing, and their interactions on third year eastern hemlock
and paper birch survival. Each combination of factors was considered up to three way
interactions in the original model. All interactions Prob ChiSq > 0.25 were pooled with the
error term and the models rerun (Bancroft 1964)………………………………………....41
Table 2.6. The effect of Clipping on paper birch and eastern hemlock third year survival.
SD=standard deviation…………………………………………………………………….42
Table 3.1. Results of a generalized logistic mixed model for the effects of Light
Availability, Non-Tree Vegetation Density, Habitat Type, Species, and their two way
interactions on seedling survival. All interactions Prob ChiSq > 0.25 were pooled with the
error term and the models rerun (Bancroft 1964)………………………………………....71
Table 3.2. Results of a generalized linear mixed model for the effects of Scarification,
Light Availability, Habitat Type, Species, Seed Production Potential (SPP) and their two
way interactions on seedling establishment. All interactions Prob ChiSq > 0.25 were
pooled with the error term and the models rerun (Bancroft 1964)………………………...72
Table 3.3. Species’ average seedling density (with standard deviation in parentheses) in
228 plots across 19 northern hardwood stands in the Northern Lower and Eastern Upper
Peninsulas of Michigan……………………………………………………………………73

ix

Table 3.4. Species’ average seedling density (with standard deviation in parentheses) in
114 scarified and unscarified plots across 19 northern hardwood stands in the Northern
Lower and Eastern Upper Peninsulas of Michigan. Asterisks represent statistically
significant effects (Prob ChiSq < 05)……………………………………………………...74
Table 3.5. Mean average light availability and competing non-tree vegetation
characteristics (with standard deviation in parentheses) across high (AFOCa and AFOA),
medium (AFO and AFPo) and low (ParVVb and ATFD) quality habitat types (Burger and
Kotar 2004)………………………………………………………………………………..75
Table 4.1. Results of a standard least squares model for the effects of Sterilization, Species,
Substrate, and their interactions on height growth……………………………………….111
Table 4.2. F ratio’s and P > F from standard ANOVA least squares models for the effects
of Sterilization, Substrate, and their interactions on species’ height growth………….....112
Table 4.3. Results of a nominal logistic regression model for the effects of Sterilization,
Species, Substrate, and their interactions on seedling survival…………………………..113
Table 4.4. Overall survival rate of individual species at the end of the experiment……..114
Table 4.5. Results of a standard least squares model for the effects of Species, Substrate
and their interaction on mycorrhizal colonization of roots………………………………115
Table 4.6. Sterilization’s effect on species height growth rankings within individual
substrate types. Values of ρ are Spearman’s rank correlations between sterilized and
unsterilized treatments. Smaller values indicate greater changes in ranks………………116

x

LIST OF FIGURES

Figure 2.1. The effects of scarification (SC) and light on eastern hemlock (A), paper birch
(B), yellow birch (C), and white pine (D) germination. R2 applies to the whole model
including the main effects of scarification and light availability. P values are displayed for
scarification and light in that order………………………………………………………..33
Figure 2.2. The effect of light availability on eastern hemlock (A), paper birch (B), yellow
birch (C), and white pine (D) germinant establishment (1st year survival). R2 applies to the
whole model including the main effects of light availability. P values represent the effect
of light on seedling establishment…………………………………………………………34
Figure 2.3. Average germination (A) and number of second year surviving seedlings (B)
for eastern hemlock (EH), paper birch (PB), yellow birch (YB), and white pine (WP) in
unharvested areas (UND) (0m diameter), single tree gaps (ST) (<23m diameter), large
group gaps (LG) (24-50m diameter), and patch cut gaps (PC) (>50m diameter)…………35
Figure 2.4. Average number of third year surviving seedlings for eastern hemlock (EH)
and paper birch (PB) in unharvested areas (UND) (0m diameter), single tree gaps (ST)
(<23m diameter), large group gaps(LG) (24-50m diameter), and patch cut gaps (PC)
(>50m diameter)…………………………………………………………………………...36
Figure 3.1. Predicted establishment response of yellow birch (YB), white ash (WA),
black cherry (BC), and eastern hemlock (EH) to local (50m) seed production potential
(∑diameter2/distance2). R2 and P values represent the fit of the whole model and
significance of seed production potential, respectively…………………………………...68
Figure 3.2. Seedling establishment as predicted by local seed production potential in high
(AFOCa and AFOA), medium (AFO and AFPo) and low (ParVVb and ATFD) quality
habitat types for red maple (A), black cherry (B), sugar maple (C), and white ash (D)
(Burger and Kotar 2004)……………………………………………………………….….69
Figure 3.3. The effect of varying light availability in 114 harvest gaps on red maple (A)
and sugar maple (B) seedling establishment………………………………………………70

xi

Figure 4.1. Height growth across balsam fir (BF), eastern hemlock (EH),
northern white cedar (NWC), paper birch (PB), sugar maple (SM), mineral soil (MS), and
yellow birch (YB) for balsam fir (A), hemlock (B), paper birch (C), and
white pine (D) seedlings. Error bars represent one standard error, while substrates not
sharing a common letter supported significantly different seedling growth (P <0.05, TukeyKramer HSD)…………………………………………………………………………….104
Figure 4.2. Height growth across sterilized (Maroon) and unsterilized (Gold) balsam fir
(BF), eastern hemlock (EH), northern white cedar (NWC), paper birch (PB), sugar maple
(SM), mineral soil (MS), and yellow birch (YB) for red maple (A), northern white cedar
(B), white spruce (C), and yellow birch (D) seedlings. Error bars represent one standard
error, while asterisks represent statistically significant effects of sterilization on growth
(prob |t|< 0.05)……………………………………………………………………………105
Figure 4.3. Percent colonization of arbuscular mycorrhizal fungi (AMF) on red maple (A)
and northern white cedar (B) seedlings and ectomycorrhizal fungi colonization on yellow
birch (C) and white pine seedlings (D) across unsterilized balsam fir (BF), eastern hemlock
(EH), northern white cedar (NWC), paper birch (PB), sugar maple (SM), mineral soil
(MS), and yellow birch (YB) substrates. Error bars represent one stand error, while
substrates not sharing a common letter supported colonization at significantly different
rates (P <0.05, Tukey-Kramer HSD)……………………………………………………106
Figure 4.4. The relationship between arbuscular mycorrhizal fungi (AMF) colonization and
height growth for red maple (A) and northern white cedar (B) and ectomycorrhizal fungi
(EMF) colonization and height growth for yellow birch (C) and white pine (D)……….107
Figure 4.5. Effect of sterilization (maroon) on species’ height growth performance
on balsam fir CWD. Species were ranked from one to eight with one representing the best
height growth performance. The Spearman’s rank correlation statistic is listed in the top
left corner………………………………………………………………………………...108
Figure 4.6. Effect of sterilization (maroon) on species’ height growth performance
on northern white cedar CWD. Species were ranked from one to eight with one
representing the best height growth performance. The Spearman’s rank correlation statistic
is listed in the top left corner……………………………………………………………..109
Figure 4.7. Effect of sterilization (maroon) on species’ height growth performance
on eastern hemlock CWD. Species were ranked from one to eight with one representing
the best height growth performance. The Spearman’s rank correlation statistic is listed in
the top left corner………………………………………………………………………...110

xii

CHAPTER I

Introduction

Forestry has been defined as the science and art of cultivating, maintaining and
developing forests. Forest regeneration is a critical component of this definition, as forests
with characteristics commensurate with human goals can neither be maintained nor
developed unless cultivation occurs. In species diverse forests, one of the greatest
challenges foresters face is maintaining species diversity, as they must create
environmental conditions that will allow species with varying resource requirements the
opportunity to reproduce. This process is further complicated by the fact that seedlings,
due to their size and lack of resource acquiring structures, are highly vulnerable to a variety
of factors which could affect their survival. Forest substrate is one such factor which may
influence seedling establishment and survival. In addition, forest substrate is a factor that
can realistically be manipulated by forest management. My dissertation investigates how
forest substrate interacts with local seed source density, light availability, substrate
moisture and nutrient availability, non-tree vegetation, and mycorrhizal fungi to affect
seedling establishment and survival in the managed northern hardwood forests of the
Northern Great Lakes region.

Dissertation Structure
I begin, in this chapter, by describing the current state of seedling demographics in
managed northern hardwood forests and identifying some interesting unexplained patterns.
1

I’ll then briefly review the history of northern hardwood forest exploitation and
management in the Northern Great Lakes region and some of the major harvesting events
that have contributed to the forest composition and structures we see today. In doing so, I
will explain how the legacy of these events may be influencing some of the current
regeneration patterns addressed in this dissertation
The second chapter focuses on identifying critical processes which influence
seedling establishment and survival within different sized harvest gaps independent of seed
supply. In this chapter, I demonstrate how substrate type and light availability influence
species germination, and how light availability and exposure to non-tree vegetation (grass,
forbs, shrubs, and ferns) affect seedling survival. In addition to identifying important
factors and processes, this chapter also demonstrates how, via different patterns for
seedling germination and survival, species shift in seedling density across a range of
harvest gap sizes over time.
The third chapter investigates seedling establishment within harvest gaps under
natural regeneration conditions across a collection of forest stands representing a broad
range of northern hardwood forest site qualities and overstory composition. This chapter
highlights the effects of local seed sources, site quality, substrate availability, and light on
species establishment. It also represents a realistic case study of some of the issues land
managers are facing when attempting to regenerate individual species, as it considers how
variation in seed availability affects the regeneration process.
The fourth chapter explores species’ growth and survival across different types of
forest substrate. It also separates the effects of mycorrhizal fungi and substrate mineral
nitrogen content as driving mechanisms for this pattern. In addition, this chapter provides

2

important information for restoring old-growth characteristics in northern hardwood
forests.

Understory Composition
Consistent with species demographic trends observed throughout eastern North
America, the understories of northern hardwood forests in the Great Lakes region are
currently dominated by shade tolerant species (Neuendorff et al. 2007; Fei and Steiner
2009; Matonis et al. 2011.) Among the most common species are ironwood (Ostrya
virginiana (Mill.) K. Koch), red maple (Acer rubrum L.) and sugar maple (A. saccharum
Marsh.). On the other end of this spectrum are several formerly prominent, shade intolerant
to mid-tolerant species including white pine (Pinus strobus L.), yellow birch (Betula
alleghaniensis Britton.) and red oak (Quercus rubra L.) which now comprise only a small
percentage of the seedling layer. While this pattern suggests that light availability may be
exerting a strong influence over seedling layer species composition, extremely shadetolerant eastern hemlock is also declining in the regeneration layer of northern hardwood
forests, (Matonis et al. 2011) suggesting that more is limiting than just light. Consistent
with this notion, in trials where light availability has been increased through cutting larger
harvest gaps, only modest gains of less tolerant species’ seedling layer density have been
observed (Shields et al. 2007; Bolton and D’Amato 2011; Kern et al. 2013). Consequently,
the processes and mechanisms by which some shade tolerant species have come to
dominance and most shade intolerant to mid tolerant species have been replaced remain
undefined.

3

Major Harvesting Events
Containing up to as many as 20 different tree species, northern hardwood forests
are one of the most tree species diverse forest types in the eastern deciduous forest. Prior
to European settlement, northern hardwood forests could generally be characterized as a
mix of early to late successional conifer and deciduous species with white pine, sugar
maple, and eastern hemlock occurring at the greatest density (Whitney 1987; Zhang et al.
2000; Schulte et al. 2007). These dominant species, however, became the target of
exploitive logging operations and were extracted in great quantities throughout the region
in the late nineteenth and early twentieth centuries (Whitney 1987). The effect of this
disturbance was, in many locations, compounded by large high intensity fires which arose
from the combination of large residual fuel loads and open stand conditions created by the
exploitative harvests, which dried the now abundant forest fuels. These fires killed many of
the smaller trees ignored by the loggers and eliminated much of the existing advanced
regeneration, setting the stage for regional species composition shifts. Among the most
dramatic shifts associated with these disturbances was the loss of many formerly prevalent
conifer species, which lacked the ability to sprout or colonize disturbed sites at long
distances via seed dispersal (Whitney et al. 1989; Schulte et al. 2007). In their absence,
many deciduous species possessing these traits such as sugar maple and aspen (Populus
spp.) became more prominent. Consequently, most conifer species are now less common in
the overstory when compared to the late twentieth century and, as a result, may now face
strong seed source limitations to their regeneration (Zhang et al. 2000; Friedman and Reich
2005; Schwartz et al. 2005).

4

Beginning in the 1950’s, single-tree selection harvesting became the dominant
management paradigm for northern hardwood forests (Arbogast 1957). This system was
adopted originally to increase the dominance of economically important shade tolerant
species such as sugar maple. It does so by creating small harvest gaps which only
marginally and ephemerally increase light at the forest floor (Eyre, and Zillgitt 1953;
Klingsporn et al. 2012). As a result, the forests floor has been subjected to decades of low
light conditions, which is incompatible with the comparatively higher light demands of
intolerant and mid-tolerant species (Kobe et al. 1995; Walters and Reich 2000). Therefore,
the scarcity of shade intolerant to mid-tolerant species in the seedling layer may be
explained by an inadequate supply of light to support their survival (Webster and Lorimer
2005; Webster and Jensen 2007; Matonis et al. 2011)
Single tree selection may also be constraining seedling layer diversity by reducing
important seedling establishment substrates. Decaying coarse woody debris (hereafter
referred to as CWD) has been shown to be an important seedling establishment site in
several forest types including northern hardwoods (Harmon and Franklin 1989; Mori et al.
2004; Marx and Walters 2008). CWD may be particularly important for smaller seeded
species, as thick litter layers have been shown to inhibit their establishment on the forest
floor. Unfortunately, CWD is becoming increasingly rare in managed forests, as selection
harvests remove older or diseased trees that are likely to become the next generation of
CWD (Hura and Crow 2004). Mineral soil represents an alternative safe site for small
seeded species to establish; however, its availability has also been diminished by an
increased emphasis on gentle harvesting practices which limits surface disturbance by
harvesting over snow (Shields et al. 2007). As a result, the rarity of smaller seeded species

5

in the seedling layer may be explained by a lack of favorable seedling establishment
substrate (Caspersen and Saprunoff 2005; Bolton and D’Amato 2011).
Broadly, this dissertation asks whether the availability of mineral soil and/or CWD
is a major factor influencing seedling layer species composition. To answer this question, I
conduct a series of manipulative experiments at differing points in the seedling
establishment process and at differing spatial scales. This approach allows me to not only
document mineral soil and CWD’s initial influence on seedling layer species composition,
but also test a variety of other factors which may override its potential influence. My
design also provides land managers with a blueprint for overcoming potential substrate
limitations in northern hardwood forests. In Chapter Two, I begin my investigation by
following the fate of different species’ seeds, dispersed in different types of forest
substrate, spread across a range of light environments.

6

LITERATURE CITED

7

LITERATURE CITED

Arbogast, C., 1957. Marking guides for northern hardwoods under the selection system.
USDA For. Serv. Stat. Paper LS-56, 1-20.
Bolton, N.W., D’Amato, A.W., 2011. Regeneration responses to gap size and coarse
woody debris within natural disturbance-based silvicultural systems in northeastern
Minnesota, USA. For. Ecol. Manage. 262, 1215-1222.
Caspersen, J.P., Saprunoff, M., 2005. Seedling recruitment in a northern temperate forest:
the relative importance of supply and establishment limitation. Can. J. For. Res. 35,
978-989.
Eyre, F.H., Zillgitt, W.M., 1953. Partial cuttings in northern hardwoods of the Lake States:
Twenty-year Experimental Results. USDA Forest Service Technical Bulletin LSTB-1076, St. Paul, MN, USA.
Fei, S., Steiner, K.C., 2009. Rapid capture of growing space by red maple. Can. J. For.
Res. 39, 1444-1452.
Friedman, S.K., Reich, P.B., 2005. Regional legacies of logging: departure from
presettlement forest conditions in northern Minnesota. Ecol. Appl. 15, 726-744.
Harmon, M.E., Franklin, J.F., 1989. Tree seedlings on logs in Picea-Tsuga forests of
Oregon and Washington. Ecology. 70, 48-59.
Hura, C.A., Crow, T.R., 2004. Woody debris as a component of ecological diversity in
thinned and unthinned northern hardwood forests. Nat. Ar. J. 24, 57-64.
Kern, C.C., D’Amato, A.W., Strong, T.F., 2013. Diversifying the composition and
structure of managed, late-succesional forests with harvest gaps: what is the
optimal gap size? For. Ecol. Manage. 304, 110-120.
Klingsporn, S., Webster, C.R., Bump, J.K., 2012. Influence of legacy-tree retention on
group-selection opening persistence. For. Ecol. Manage. 286, 121-128.
Kobe, R.K., Pacala, S.W., Silander, J.A., Canham, C.D., 1995. Juvenile tree survivorship
as a component of shade tolerance. Ecological Applications. 5, 517-532.
Marx, L., Walters, M.B., 2008. Survival of tree seedlings on different species of decaying
wood maintains tree distributions in Michigan hemlock-hardwood forests. Ecology
96, 505-513.
Matonis, M.S., Walters, M.B., Millington, J.D.A, 2011. Gap-, stand,-and landscape-scale
factors contribute to poor sugar maple regeneration after timber harvest. For. Ecol.
Manage. 262, 286-298.

8

Mori, A., Mizumachi, E., Osono, T., Doi, Y., 2004. Substrate-associated seedling
recruitment and establishment of major conifer species in an old-growth subalpine
forest in central Japan. For. Ecol. Manage. 196, 287-297.
Neuendorff, J.K., Nagel, L.M., Webster, C.R., Janowiak, M.K., 2007. Stand structure and
composition in a northern hardwood forest after 40 years of single-tree selection.
N.J. Appl. For. 24, 197-202.
Schulte, L.A., Mladenoff, D.J., Crow, T.R., Merrick, L.C., Cleland, D.T., 2007
Homogenization of northern U.S. Great Lakes forests due to land use. Landscape
Ecol. 22, 1089-1103.
Schwartz, J.W., Nagel, L.M., Webster, C.R., 2005. Effects of uneven-aged management on
diameter distribution and species composition of northern hardwoods in Upper
Michigan. For. Ecol. Manage. 211, 356-370.
Shields, J.M., Webster, C.R., Nagel, L.M., 2007. Factors influencing tree species diversity
and Betula alleghaniensis establishment in silvicultural openings. Forestry 80,
293-307.
Walters, M.B., Reich, P.B., 2000. Trade-offs in low-light CO2 exchange: a component of
variation in shade tolerance among cold temperate tree seedlings. Funct. Ecol. 14,
155-165.
Webster, C.R., Lorimer, C.G., 2005. Minimum opening sizes for canopy recruitment of
midtolerant tree species: a retrospective approach. Ecol. Appl. 15, 1245-1262.
Webster, C.R., Jensen, N.R., 2007. A shift in the gap dynamics of Betula alleghaniensis in
response to single-tree selection. Can. J. For. Res. 37, 682-689.
Whitney, C.G., 1987. An ecological history of the Great Lakes forest of Michigan.
Ecology. 75, 667-684.
Zhang, Q.F., Pregitzer, K.S., Reed, D.D., 2000. Historical changes in the forests of the
Luce District of the Upper Peninsula of Michigan. Am. Midl. Nat. 143, 94-110.

9

CHAPTER II
Scarification and gap size have interacting effects on northern temperate seedling
establishment

Abstract
After decades of promoting only the most economically valuable tree species,
silvicultural practices in northern temperate forests of North America have increasingly
become focused on maintaining tree species diversity. Unfortunately, many formerly
prominent tree species including paper birch, yellow birch, eastern hemlock and white pine
are now rarely found in the seedling layer, raising concerns about their future
sustainability. This study investigates the mechanisms of seedling recruitment failure for
these four species in two related seed addition experiments conducted in 45 variably sized
harvest gaps (220 to 6500m2) and four unharvested locations in two adjacent mesic
northern hardwood stands, located in Emmet County Michigan, USA. In addition to
elucidating mechanisms, these experiments provide evidence for what size harvest gap
provides the best environment for regenerating this particular group of species. The first
experiment uses mixed models to investigate the influence of competing non-tree
vegetation, light availability and browsing on seedling survival, while the second
experiment uses the same statistical approach to explore the influence of scarification, light
availability, and non-tree vegetation on seed germination and seedling establishment.
Eastern hemlock, paper birch and yellow birch, all smaller seeded species, were 12, 17 and
95 times more abundant in scarified plots compared to unscarified plots. In contrast, white

10

pine, the largest seeded species, was unaffected by scarification. Highly shade tolerant
hemlock and highly shade intolerant paper birch both germinated at greater densities in
lower light environments, while both mid-tolerant species, white pine and yellow birch,
were unaffected by light. Each species’ initial establishment significantly increased with
increasing light availability, and with the exception of yellow birch, each species also
survived at a significantly higher rate with increasing light availability. Paper birch and
hemlock third year survival also increased with increasing light. However, both species
survival was also significantly negatively influenced by competition from non-tree
vegetation. At the conclusion of the study, large group selection gaps (24-50m diameter)
contained the highest density of each species’ except white pine, suggesting that large
group selection gaps may provide the best opportunity for reestablishing this particular
group of species in the seedling layer.

11

Introduction
With the human population expected to surpass eight billion in the next decade
(United Nations 2012), demand for goods and services from already heavily exploited
forest ecosystems will continue to increase. Unfortunately, due to declining tree species
diversity and evenness, the variety of goods and services some forest ecosystems can
provide may soon decline (Schuler and Gillespie 2000; Lawrence 2004; Amatangelo et al.
2011), as lower diversity forests are oftentimes limited in the range of materials they offer
for forest products, the types of habitat they provide wildlife and their resilience to
disturbance (Chapin et al. 2000; Folke et al. 2004; Drever et al. 2006; Fischer et al. 2006).
Complicating matters further, many forest ecosystems may soon become less productive as
several factors, including climate change, are expected to expose species to growing
conditions and disturbances to which they are less suitably adapted, and now potentially
less resilient (Sturrock et al. 2011; Anderson et al. 2013;Duveneck et al. 2014).
The northern temperate forests of Eastern North America represent one such forest
type where species composition shifts may soon threaten the sustainability of certain
timber products and elevate the risk of severe disturbance. Some of the most noticeable
compositional changes include a decrease in overstory evenness (Zhang et al. 2000), an
expansion of shade tolerant species (Schulte et al. 2007; Nowacki and Abrams 2008;
Amatangelo et al. 2011; Hanberry 2013), and a near exclusion of conifers (including
hemlock and white pine) and shade intolerant species from the understory (including paper
birch and yellow birch) (Neuendorff et al. 2007; Matonis et al. 2011). Recognizing the

12

problems associated with declining tree diversity, forest managers have begun
investigating silvicultural means to restore tree species diversity to this forest type.
Imitating natural disturbance is one often suggested approach for promoting tree
species diversity in northern temperate forests (Scheller and Mladenoff 2002; Seymour et
al. 2002; Bolton and D’Amato 2011). Harvest disturbance using single-tree selection, the
dominant silvicultural system in northern hardwood forests for over 60 years, creates a
series of environmentally similar small harvest gaps (< 23 meters diameter). Unlike single
tree selection, a natural disturbance based harvesting regime creates a range of
microenvironments within a stand by varying harvest gap sizes. It is generally believed
that the range of environmental conditions produced by a natural disturbance harvesting
regime will promote species diversity, as the lower light environments produced in small
harvest gaps are thought to favor the establishment of shade tolerant species, while the
higher light environments found in progressively larger harvest gaps are thought to
promote the establishment of mid-intolerant to shade intolerant species (Ricklefs 1977;
Denslow 1980). Nevertheless, evidence from recent gap regeneration studies suggests that
gap dynamics are far less predictable, and may be influenced by several factors other than
gap size (Shields et al. 2007; Falk et al. 2010; Gasser et al. 2010; Bolton and D’Amato
2011; Matonis et al. 2011; Kern et al. 2012; Fahey and Lorimer 2013; Kern et al. 2013;
Klingsporn-Poznanovic et al. 2013). Consequently, there are several potential explanations
for why restoration efforts that manipulate gap size alone have failed to produce a more
diverse species mix.
Seedling establishment failure is one potential explanation for the variable patterns
of tree recruitment within harvest gaps (Wright et al. 1998; Caspersen and Saprunoff

13

2005). While harvest gap size can be manipulated to give individual species or groups of
species a competitive advantage based on their shade tolerance, seed must be present and
be able to germinate and establish before any competitive advantage can be realized. As
such, manipulating factors that influence declining tree species seed supply, and/or
germination and establishment may be more important to their restoration than
manipulating factors that affect their growth.
Similar to how tree species vary in their competitive abilities among light
environments, tree species vary in their ability to germinate and establish on a variety of
forest floor substrates (Perala and Alm 1990; Shields et al. 2007a; Marx and Walters 2006;
2008). For instance, smaller-seeded species are more successful on easily-penetrable,
moisture-holding substrates such as decaying coarse woody debris (hereafter referred to as
CWD) compared to the undisturbed litter layer (McGee and Birmingham 1997; Cornett et
al. 2001; Caspersen and Saprunoff 2005; Marx and Walters 2008). In contrast, larger
seeded species, some of which are now dominating northern temperate forest understories
(sugar maple, beech), establish successfully on undisturbed leaf litter substrates (Caspersen
and Saprunoff 2005; Neuendorff et al. 2007; Matonis et al. 2011). Unfortunately, in
managed northern hardwood forests, undisturbed leaf litter substrates dominate and CWD
is becoming increasingly rare, as mortality of large trees has been greatly diminished by
harvest removals (Hura and Crow 2004). As such, smaller-seeded species recruitment
failures within harvest gaps in managed northern temperate forests may actually be the
product of substrate limitation rather than low light availability.
One obvious solution to this potential recruitment bottleneck is to increase the
availability of favorable establishment sites. While CWD may take decades to develop

14

naturally, and is prohibitively labor intensive to restore artificially, bare mineral soil/humus
establishment sites are relatively easy to create and may provide similar establishment
opportunities. For example, the forest floor could be scarified, a silvicultural technique
which disturbs the litter layer and understory vegetation by dragging chains or disks across
the forest floor, to increase bare mineral soil/humus availability. Scarification has even
been shown to be effective at increasing light-seeded species establishment in boreal and
sub-boreal forest biomes with more recalcitrant litter layers (Raymond et al. 2003;
Lorenzetti et al. 2008; Prèvost et al. 2010), but little is known about whether scarification
will increase small-seeded species seedling establishment in more productive northern
temperate forests. Furthermore, little is known about how gap size might interact with
scarification in affecting the establishment of tree seedlings. Given the move toward using
variable harvest gap sizes in northern temperate forest management, these interactions are
important to understand if scarification is to be employed operationally. Another potential
explanation for inconsistent patterns of harvest gap recruitment is the interaction between
harvest gap size and non-tree vegetation (forbs, shrubs, ferns, and graminoids). By
absorbing and/or acquiring light, moisture, and nutrients, non-tree vegetation can modify
growing conditions at the forest floor and potentially minimize environmental differences
that exist between different sized harvest gaps (Royo and Carson 2006). Large gaps
created by the removal of groups of trees (group selection or patch selection harvesting)
may experience the greatest degree of modification as highly competitive early
successional species like raspberry (Rubus idaeus), black berry (Rubus occidentalis), and
sedge (Carex sp.) have been shown to increase in larger sized harvest gaps (Shields and
Webster 2007; Matonis et al. 2011, Kern 2012; Walters et al. in preparation).

15

Consequently, intense competition for resources from non-tree vegetation may shift the
competitive balance in large harvest gaps from resource demanding shade intolerant
species to more resource conserving shade tolerant species that can withstand being
overtopped and shaded by competing vegetation (Montgomery et al. 2010). Smaller group
harvest gaps, on the other hand, may provide a more favorable environment for shade
intolerant species to establish. While lower light availability may constrain shade
intolerant species’ growth potential, early successional non-tree vegetation expansion may
also be muted in the lower light environment of smaller group harvest gaps. As such, some
tree species, perhaps including those that are shade intolerant to mid-tolerant, may actually
have greater access to light in smaller group harvest gaps as a result of reduced
competition if their growth is less sensitive to lower light availability than is competing
non-tree vegetation. This, in turn, may lead to shade intolerant to mid-tolerant species
being most abundant in smaller group harvest gaps.
In addition to substrate limitations and competition from non-tree vegetation,
browsing from white-tailed deer (Odocoileus virginianus (Zimmermann)) could also limit
the efficacy of varying gap size to increase tree diversity. In order to reach the canopy,
seedlings must first pass through the browsing range of deer (to ~ 1.5m tall) (Walters et al.
in preparation). Within this range, seedlings are subject to repeated browsing, which has
been shown to constrain seedling development and survival. Nevertheless, escaping the
browsing range may challenge some species more than others, as deer have been shown to
browse certain species preferentially, and certain species have demonstrated a greater
tolerance to browsing (Horsley et al. 2003; Rooney and Waller 2003; Long et al. 2007;
Witt and Webster 2010; Kain et al. 2011; Randall and Walters 2011; Nuttle et al. 2014;

16

Walters et al. in preparation). On the other hand, newly germinated seedlings may not be
apparent or desirable to foraging deer, as the effects of browsing have been shown to
increase with seedling size (Randall and Walters 2011, Nuttle et al. 2013). Therefore, even
if silvicultural treatments are successful at overcoming establishment substrate and
light/non-tree competition barriers to regeneration. recruitment in harvest gaps may still
ultimately depend on local deer density (Millington et al. 2010.)
In summary, manipulating harvest gap size is an increasingly common silvicultural
strategy aimed at increasing tree diversity in northern temperate forests. However,
unexpected patterns of tree recruitment observed in several gap regeneration experiments
suggest that gap size may not be the only factor affecting tree species recruitment. Given
the combined effects of declining CWD availability and a recent shift to gentle harvesting
practices which minimize surface disturbance (Shields et al. 2007), bare mineral soil
availability may be an important component to the regeneration failures of several small
seeded species. Here, we report how substrate availability, harvest gap size, and exposure
to non-tree vegetation and deer interact to affect the germination, establishment (1 year),
and short-term survival (up to 3 years) of four species that are declining in northern
temperate forests (eastern hemlock, white pine, paper birch, yellow birch) across a gradient
of light availability in a mesic northern hardwood stand. Specifically we predict: 1) small
seeded species germination will be restricted by a lack of bare mineral substrate across
harvest gaps, 2) competition from non-tree vegetation will negatively affect seedling
survival, especially for shade-intolerant species in large group selection gaps, 3) browsing
will not affect seedlings due to their small stature 4) sufficient resource availability and

17

modest competition from non-tree vegetation will allow all species’ seedling survivorship
to be maximized in medium sized harvest gaps.
To test these predictions we established two experiments that collectively tracked
the germination and survival of eastern hemlock (Tsuga Canadensis L.), yellow birch
(Betula. alleghaniensis Britton.), paper birch (B. papyrifera Marsh), and white pine (Pinus
strobus L.) seedlings for up to three years in subplots that were unfenced or fenced to
exclude deer, unscarified or scarified to create bare mineral soil/humus substrate, and
unclipped or clipped to control non-tree vegetation across 45 harvest gaps of varying size
and 4 unharvested areas in two adjacent recently harvested northern hardwood stands. In
the first experiment, annual seedling censusing over three years allowed us to develop a
longer term pattern of how light availability and exposure to deer and non-tree vegetation
affect paper birch and hemlock seedling survivorship. In the second experiment, bi-weekly
censusing coupled with annual censusing allowed us higher resolution data of germination
and seedling survivorship over two years for each of the four species. Collectively, the
information presented in these experiments allow us to accurately depict patterns of
seedling germination and survival within different sized harvest gaps and gain mechanistic
insights into why harvest gap recruitment failures occur for four species declining in the
northern temperate forest.

Methods
Study site
Our experimental sites were located in two adjacent northern hardwood stands in
Emmet County, Michigan, USA (N45.574624 -W85.074373). The stands feature post-

18

glacial moraine topography and a mesic, rich to very rich habitat moisture and habitat type
(AFOca) (Burger and Kotar 2003). Both stands are dominated by sugar maple (Acer
saccharum (Marsh.) with white ash (Fraxinus americana L.), basswood (Tilia americana
L.), paper birch (Betula papyrifera (Marsh.)), black cherry (Prunus serotina (Ehrh.)), and
American beech (Fagus grandifolia (Ehrh.)) representing minor components of the
overstories. In total, 45 harvest gaps ranging in size from 220 to 6500m2 and four
unharvested areas were selected from the two stands (stand 1: 40 gaps + 4 unharvested
areas, stand 2: 5 gaps). To ensure gap independence, each gap was located at least 50m
from another harvest gap. Logging slash was removed from the interior of each gap and
piled along the borders.

Experiment one
The goal of this experiment was to identify the effects of light availability, non-tree
vegetation and deer herbivory on seedling survival. Beginning in the spring of 2011, we
located main plots (13x13m) in the center of each harvest gap and four unharvested areas.
All existing advanced regeneration (existing seedlings and saplings) were then removed
with a brush saw from each main plot. Main plots were then either fenced to exclude deer
(30) or left unfenced (14). Plots were chosen randomly for non-fencing from within bins of
gap sizes (0m diameter), single-tree gaps (<23m diameter), large group gaps (24-50m
diameter), and patch cut gaps (PC) (>50m diameter) to ensure our fencing/non-fencing
treatment occurred across the full range of gap sizes. Four subplots (4m2) were nested
within each main plot. Each subplot was assigned one of four treatments: control, surface
scarification, competing non-tree vegetation (forb, graminoids, ferns, shrubs) removal, and

19

surface scarification with vegetation removal. Scarification was conducted in the spring
and was accomplished by raking away the litter layer to create a bare mineral soil/humus
seedbed. Vegetation removal was accomplished by clipping all non-tree vegetation
monthly throughout the growing season. After scarification, 500 cold-stratified seeds of
eastern hemlock (Michigan source, Michigan Department of Natural Resources, Wyman
Nursery, Manistique Michigan) yellow birch (Pennsylvania source, Sheffield’s seeds,
Locke, New York) and paper birch (Michigan source, Sheffield’s seeds Locke, New York),
(of each species) were sown in late May in each subplot. Prior to sowing, we conducted
germination tests on each species to ensure seed viability. Despite yellow birch
demonstrating adequate germination in the test, poor germination was observed in the field
prompting us to remove yellow birch results from this report.
For each subplot, we assessed non-tree vegetation cover and height, light
availability above non-tree vegetation, and light availability at the forest floor (beneath
non-tree vegetation). Non-tree vegetation density was determined by visually estimating
subplot coverage to the nearest five percent (looking down) and by measuring its height
(cm) at nine pre-determined locations. Canopy openness, a proxy for canopy gap light
availability (Kobe and Hogarth 2007), was determined from analyses of hemispherical
photographs (taken at a height above the non-tree vegetation layer 1.5m) using Gap Light
Analyzer v 2.0 (Frazier et al. 1999) with an automatic threshold value determined for each
photo from Sidelook v. 1.1.01 (Nobis and Hunziker 2005). The light environment at the
forest floor (1cm) was estimated by taking readings with a LAI 2000 plant canopy analyzer
(LI-COR Incorporated, Lincoln, Nebraska) to determine diffuse non-interceptance (DIFN),
which is the fraction of the sky visible to the sensor. This was done to quantify the impact

20

of competing non-tree vegetation on light availability to developing seedlings. To obtain
this measurement, we first took a reading above the competing vegetation layer (1.5m) to
get a maximum light value for the subplot, then three readings from random locations
beneath the non-tree vegetation layer (1cm) to characterize its impact on light
diminishment. The percentage of light available beneath the non-tree vegetation layer was
then multiplied by the canopy openness value to determine light availability to seedlings
established on the forest floor. Each seedling was given a colored tag denoting its cohort
year at the time of its germination. Seedling survivorship for sown species was then
monitored within subplots at the end of September for three years.

Experiment two
The goals of this experiment were to a) acquire higher resolution information than
in Experiment One on how light, substrate availability, and competing non-tree vegetation
affect the germination and survival via more frequent censusing, and b) assess the effects
of these factors on a greater number of species than those that successfully germinated in
Experiment One (paper birch, yellow birch, eastern hemlock, and white pine). To
accomplish this, in spring 2012, we established four additional subplots adjacent to the
main plot in 13 of the 40 harvest gaps and two understory locations. Gaps were selected
from gap size strata in order to assure we captured the entire range of gap sizes. Subplots
were also established in five additional single-tree gaps, located in an adjacent recently
harvested stand (summer 2011), as a lack of available gap space prevented us from
locating subplots outside of the main plot in the single-tree gaps used in Experiment One.
Each of the four new subplots was randomly assigned one of the treatment options

21

described above in Experiment One. After scarification, 500 cold-stratified seeds of
eastern hemlock (Michigan source, Michigan Department of Natural Resources, Wyman
Nursery, Manistique Michigan), yellow birch (Pennsylvania source, Sheffield’s seeds,
Locke, New York), paper birch (Michigan source, Sheffield’s seeds Locke, New York),
and white pine (Michigan source, Michigan Department of Natural Resources, Wyman
Nursery, Manistique Michigan) were spread evenly throughout each subplot. Due to a
shortage in seed, white pine was only added to subplots in 12 of the 20 gaps.
In each subplot in each location, we assessed non-tree vegetation cover and height,
light availability above non-tree vegetation, and light availability below non-tree
vegetation using the same protocol described in Experiment One. In addition, we also
quantified substrate (bare mineral soil/humus mix or hardwood litter) availability by
visually estimating its cover (looking down upon subplots) to the nearest 5%. Germination
was surveyed within each subplot on a bi-weekly basis from early-May through
September. We increased the sampling intensity in this experiment to get a more accurate
measurement of germination and early survival as high seedling mortality was observed in
Experiment One prior to our lone September census. A seed was considered successfully
germinated and established once its first true leaves expanded. Each established seedling
was then tagged and monitored for survival over the next two years.

Statistical analysis
For statistical analysis, we kept the results of each experiment separate. This was
done to exclude potential confounding factors including differences in cohort age,
sampling intensity and variation in climate between years. As such, we used the cohort

22

from Experiment Two (greater sampling intensity and larger species population) to
compare species’ germination, establishment (1st year survival) and second year survival,
while the cohort established in Experiment One (less intense sampling, smaller species
population but greater cohort age) was used to compare third year survival.
We examined the effects of scarification/unscarified, light availability, species, and
their interactions on seed germination for each of our four species (Experiment Two) with
Generalized Linear Mixed Models (GLMM) assuming a Poisson distribution with a
logarithmic link function. This method was adopted because our data were heavily right
skewed due to the large number of plots with few or no germinants. Clipping was not
considered as a factor for germination as the competing vegetation was already altered in
subplots receiving scarification treatments.
To examine seedling establishment (1st year survival) and second and third year
survival we used Generalized Logistic Mixed Models assuming a binomial distribution.
For establishment and second year survival of each of our four species, we considered the
effects of clipping/unclipped, light availability, species, and their interactions. Fencing was
not included as a factor at these intervals because seedling establishment and second year
survival were only evaluated in unfenced subplots (Experiment Two). Fencing was,
however, considered along with each previously mentioned factor in the third year survival
of paper birch and hemlock, as seedling survivorship was evaluated in subplots which had
been either fenced or left unfenced (Experiment One). However, we do not report on the
effect of fencing in this report, as we did not observe browsing effects on seedlings.
Model selection was accomplished through backwards elimination until only
statistically significant (p< .05) factors and/or interactions remained in the model.

23

Significant species interactions were investigated by examining the response of each
individual species to the significant interacting factor. Once we arrived at a final model, we
checked the model for dispersion of the distribution, goodness of fit, and the existence of
potential outliers. Model dispersion was estimated by calculating an over-dispersion factor
(ĉ = chi-sqaure/ df). Models whose dispersion factors exceeded one were considered overdispersed. In such cases, parameter estimates were obtained using a quasi-likelihood
approach which inflated the standard errors of each factor estimate by √ ĉ . The model was
then re-run and checked for significance. Model fit and potential outlier detection was
accomplished by examining plots of studentized deviance residuals. Model fit was
evaluated by visually inspecting how evenly the residuals were distributed around zero.
Points on the plot exceeding two standard deviations of the mean were identified as
potential outliers. To evaluate the influence of such points, we used Cook’s distance
statistic. Distance statistics exceeding one were identified as potentially having strong
leverage effects on the model. To investigate these effects, we removed the potential
outlier from the data set and re-ran the model. If all parameters in the model remained
significant, the potential outlier was kept as part of the data set. However, if changes in
parameter significance occurred without the outlier, the point was then removed (Ramsey
and Schafer 1997). All analyses were performed in (JMP 9.0, SAS Institute, Cary, NC).

Results
Germination
Scarification had strong overall effects on germination, but significant species and
Scarification interactions indicate that species responded differently to Scarification (Table

24

2.1). Eastern hemlock, paper birch and yellow birch, all smaller seeded species, were 12,
17 and 95 times more abundant in scarified plots when compared to unscarified plots
(Table 2.2). In contrast, white pine, the largest seeded species, was unaffected by
scarification (Table 2.2). Light availability, which increased along with harvest gap size,
also had strong effects on germination, but a significant interaction with species also
indicated that species responded differently to Light (Figure 2.1) (Tables 2.1 & 2.3).
Highly shade tolerant hemlock and highly shade intolerant paper birch both germinated at
greater densities in lower light environments, while both mid-tolerant species, white pine
and yellow birch, were unaffected by Light (Figure 2.1).

Seedling establishment and survival
Seedling establishment (1st growing season survival) and second year survival were
strongly affected by the main effects of species and Light availability and modestly
affected by their interaction (Table 2.4). Each species’ initial establishment increased with
increasing Light availability, however, the strength of response varied among species, as
hemlock demonstrated the most dramatic increase in survival in response to increasing
light (Figure 2.2). Similar increases in second year survival were observed at higher light
availabilities for hemlock, paper birch and white pine. Yellow birch survival, however, did
not respond significantly to light by the end of the second year (Data not shown). In the
third year, seedling survival was most strongly influenced by the main effects of Clipping,
Light and species (only hemlock and paper birch tested) (Table 2.5). Clipping (Table 2.6)
and Light (data not shown) both had positive effects on each species, while paper birch
exhibited overall higher third year survivorship (65%) compared to hemlock (21%).

25

Temporal changes in seedling density in harvest gaps
No particular gap size proved to be an optimal germination environment. Among
species, white pine and hemlock germination was maximized in unharvested closed
canopy areas, while paper birch and yellow birch favored single tree and large group
selection gaps, respectively (Figure 2.3a). Similar to germination, no particular gap size
was shown to maximize seedling density after two growing seasons. However, in contrast
to germination, only white pine could be found surviving in subplots beneath unharvested
areas. Among species, white pine and hemlock shifted their gap size of maximum density
from unharvested areas to single tree gaps; while paper birch and yellow birch remained
most dense in single tree and large group selection gaps respectively (Figure 2.3b). In the
third and final year, large group selection gaps supported the highest seedling density of
hemlock and paper birch (Figure 2.4).

Discussion
Germination
Since the seed to seedling transition is one of the earliest steps in the natural
regeneration process, factors affecting germination have the potential to strongly influence
forest stand dynamics. Smaller seeded species (yellow birch, paper birch, and hemlock)
germinants were overwhelmingly more abundant in scarified subplots compared to
unscarified subplots, demonstrating the strong influence of bare mineral soil/humus on
seedling layer composition (Table 2.2). These results support our initial hypothesis of

26

substrate limitation for smaller seeded species, and are consistent with germination
patterns found in other forest types (Valkonen and Maguire 2005), seedling substrate
associations found in northern hardwood forests (Caspersen and Saprunoff 2005;Bolton
and D’Amato 2011; Matonis et al. 2011; Bèland and Chicoine 2013), and post scarification
seedling layer composition from other forest types (Raymond et al. 2003; Lorenzetti et al.
2008; Beguin et al. 2009; and Prèvost et al. 2010). Collectively, these findings suggest that
substrate related germination failures are likely contributing strongly to the scarcity of
smaller seeded species seedlings reported by Neuendorff et al. (2007) and Matonis et al.
(2011). However, it should also be noted that a reduction in local seed sources may also be
contributing strongly to natural regeneration patterns, and were not considered in this seed
addition study (Investigated in Willis et al. Ch2).
In contrast to the germination response of smaller seeded species, white pine was
unaffected by the scarification treatment (Table 2.2). This finding supports our nonsubstrate limitation hypothesis for larger seeded species and the findings of Cornett et al.
(1998), who also found white pine emergence to be insensitive to forest floor disturbance,
but differs with the findings of Raymond (2003) who found white pine germination
improved following scarification. One potential explanation for this discrepancy may
come from a difference in litter depth at both sites. First, with the exception of American
beech (a minor overstory component), the overstory at our site is composed almost
exclusively of species (sugar maple, white ash, basswood) which produce relatively
nutrient rich quickly decomposing litter (Melillo and Aber 1982). In contrast, the dominant
overstory species in Raymond et al. (2003) was white pine (38% volume) which, due to its
lower litter quality, likely produced a more recalcitrant litter layer than the one existing at

27

our site (Melillo and Aber 1982; Rustad and Cronan 1988). This suggests that the litter
layer at our site may have presented a less significant physical barrier to white pine
germination than the one reported by Raymond et al. (2003). Consequently, scarification
may have had a larger effect on seed bed conditions at the site used by Raymond et al.
(2003). Nevertheless, Smith (1951) found that white pine was capable of penetrating up to
5cm of needle litter within two weeks of germination suggesting that differences in litter
depth may not entirely account for observed differences in white pine establishment.
Another factor which may have minimized scarification’s influence on white pine
germination is seed predation. White pine seeds are a preferred food for small mammals
(Abbott 1961; Martell 1979), and have demonstrated greater germination when excluded
from seed predators Cornett et al. 1998; Raymond et al. 2003). In this study, no attempts
were made to exclude seed predators and white pine germinant density was the lowest
among our experimental species (Table 2.2) (Figure 2.3a). Thus, we cannot eliminate the
possibility that seed predation may have overridden any positive effects that scarification
may have had on white pine germination.
In addition to scarification, species responded differently to light availability (Table
2.1) (Figure 2.1). While hemlock’s germination decline in higher light environments was
unsurprising considering its extreme shade tolerance and late successional status, paper
birch’s preference for lower light environments is surprising given its extreme intolerance
of shade and reputation as an early successional pioneer species (Figure 2.1) (Burns and
Honkala 1990). Nevertheless, this finding is consistent with earlier studies which reported
greater hemlock and paper birch germination in shade (Hough 1960; Goerlich and Nyland
2000).

28

In contrast to hemlock and paper birch, light availability had no effect on white
pine and yellow birch germination (Figure 2.1). As previously mentioned, seed predation
may account for the overall low germination observed for white pine; however, yellow
birch’s lack of response is surprising considering that light has been shown to lessen the
effect of a water soluble germination inhibitor present on its seed-coat (Redmond and
Robinson 1954). Nevertheless, previous work has also shown light to have little effect on
yellow birch germination (Houle 1992).
While the limited range of species investigated in this study restricts our ability to
comment on general trends among species, these results do suggest that lower light
availability (associated with single-tree gaps (Table 2.3)) is not a strong constraint on
germination for this particular group of declining species (Figure 2.3a). Thus, moving
away from single-tree selection towards a natural disturbance based silvicultural system is
unlikely to increase these species’ initial representation in harvest gaps.

Seedling establishment and survival
Within harvest gaps, increasing light availability (larger harvest gaps) had a
positive influence on each species’ initial establishment (Figure 2.2) and, with the
exception of yellow birch, each species’ second year survival (Table 2.4). This finding is
unsurprising considering that other work has shown light to be highly limiting in richmesic growing environments like the one which exists at our site (Putz and Canham 1992;
Burger and Kotar 2004). It also supports the notion that canopy gaps provide important
regeneration opportunities in forest systems where large scale disturbances are uncommon
(Runkle 1982; McCarthy 2001). The combination of yellow birch’s mid-tolerance of

29

shade and the short time period for which we tracked its survival (2 years) may account for
the lack of response (Burns and Honkala 1990; Kobe et al. 1995). Nevertheless, white
pine, a species of similar shade tolerance, responded positively to increasing light
availability, suggesting that another unmeasured factor(s) may be obscuring the influence
of light availability on yellow birch survival.
Consistent with our expectations, seedlings established in unclipped subplots were
quickly overtopped by non-tree vegetation, and experienced dramatically reduced light
environments compared to seedlings established in clipped subplots in all but the smallest
harvest gaps and unharvested areas (Table 2.3). In partial agreement with our hypothesis,
clipping non-tree vegetation had a positive influence on third year seedling survivorship
(Tables 2.5 & 2.6). However, the lack of a significant interaction with light indicates that
clipping was equally effective across all gap sizes. Eastern hemlock’s positive response to
clipping was also unexpected given its extremely high tolerance of shade (Burns and
Honkala 1990; Kobe et al. 1995). While limited to just two species (hemlock and paper
birch), this finding suggests that competition from non-tree vegetation can influence
seeding layer composition even at the earliest stages of seedling development. It also
supports the notion that competition for light is more important than competition for
moisture and nutrients on fertile sites (Putz and Canham 1992; Montgomery et al. 2010),
given our site conditions (rich-mesic habitat type) and the fact that our clipping treatment
only controlled aboveground competition. It is our opinion, however, that yellow birch and
white pine’s non-response to clipping is likely more related to the short duration of time in
which they were exposed to non-tree vegetation (2 years) than an indicator of understory
competitiveness, as other investigators have reported negative effects of non-tree

30

vegetation on these species in experiments conducted over greater time intervals (Gasser et
al. 2010; Kern et al. 2012).

Conclusion and Management Recommendations
The collective findings from the two manipulative experiments conducted in this
investigation provide several important insights for natural disturbance based silviculture
in managed northern temperate forests. Central among these findings is the importance of
bare mineral soil/humus for smaller seeded species germination. While we cannot predict
whether seedlings established in scarified plots will ultimately gain dominance in the
understory and capture canopy growing space, scarification should increase the odds of
small seeded species successfully recruiting in harvest gaps simply by increasing their
initial density. Thus, our findings support the notion put forth by Bolton and D’Amato
(2011) suggesting that germination substrate may override the influence of harvest gap size
for smaller seeded species, and demonstrate the importance of incorporating surface
disturbance into silvicultural prescriptions. However, it should be emphasized again that
seed supply issues may override the influence of substrate under natural regeneration
conditions.
Variation in harvest gap size also affected seedling demographics. Although no
harvest gap size provided the optimal environment for seed germination (Figure 2.3a),
large group harvest gaps supported the highest densities of seedlings for all species except
white pine, which germinated at low initial densities (Figure(s) 2.3B & 2.4). Taken
together, our results suggest that large group selection harvest gaps may represent the best
management option for restoring this particular group of species, as they provide generally

31

favorable environments for seed germination (Figure 2.3A) and encourage greater seedling
establishment and survival (Figure 2.2). This pattern also provides initial support for
adopting an expanding gap harvesting system where harvest gaps are expanded once
advanced regeneration has been achieved (Raymond et al. 2009). Nevertheless, our results
also indicate that non-tree vegetation had a negative effect on seedling survival. As such,
restoration efforts may need to consider vegetation control options if they are attempting to
regenerate this collection of species in harvest gaps on high quality sites.

32

Figure 2.1- The effects of scarification (SC) and light on eastern hemlock (A), paper birch
(B), yellow birch (C), and white pine (D) germination. R2 applies to the whole model
including the main effects of scarification and light availability. P values are displayed for
scarification and light in that order.
A

2

50

Seedlings/4m

60

2

R =0.56
p<.0001
p<.0001

Seedlings/4m2

60

SC

40
30
20
10

R2=0.41
p<.0001
p=.0069

B

50
40 SC
30
20
10

UN

UN

0

0
0

10

20

30

40

50

60

70

80

0

10

60

R2=0.31
p<.0001
p=.4284

50
40

Seedlings/4m2

Seedlings /4m2

C

30
20

SC

10

30

40

50

60

70

80

% Light Availability

% Light Availability

60

20

D

R2<.01
p=.9512
p=.6557

50
40
30
20
10

UN

0

0
0

10

20

30

40

50

60

70

80

0

% Light Availability

10

20

30

40

50

60

% Light Availability

33

70

80

Figure 2.2- The effect of light availability on eastern hemlock (A), paper birch (B), yellow
birch (C), and white pine (D) germinant establishment (1st year survival). R2 applies to the
whole model including the main effects of light availability. P values represent the effect
of light on seedling establishment.
100 B

A

Proportion Surviving

Proportion Surviving

100
80
60
40
20

R2=.28
p<.0001

0
0

10 20

30 40 50

60 70

80
60
40
20

2

R =.06
p=.0289

0

0

80

10 20 30 40 50 60 70 80
% Light Availability

% Light Availability

100
C

80
60
40
20

R2=.17
p=.0012

Proportion Surviving

Proportion Surviving

100

10

20

30

40

50

60

70

80
60
40
R2=.11
p=.009

20
0

0
0

D

0

80

% Light Availability

34

10 20 30 40 50 60 70 80
% Light Availability

Figure 2.3- Average germination (A) and number of second year surviving seedlings (B)
for eastern hemlock (EH), paper birch (PB), yellow birch (YB), and white pine (WP) in
unharvested areas (UND) (0m diameter), single tree gaps (ST) (<23m diameter), large
group gaps (LG) (24-50m diameter), and patch cut gaps (PC) (>50m diameter).

Germinants/4m

2

80

A

EH
PB
YB
WP

60
40
20
0
UND

Seedlilngs/4m2

80

ST

LG

PC

B

60
40
20
0
UND ST
LG
PC
Gap Size Classes

35

Figure 2.4- Average number of third year surviving seedlings for eastern hemlock (EH)
and paper birch (PB) in unharvested areas (UND) (0m diameter), single tree gaps (ST)
(<23m diameter), large group gaps(LG) (24-50m diameter), and patch cut gaps (PC)
(>50m diameter).

5
Seedlings/4m2

EH
PB

4
3
2
1
0
UND ST
LG
PC
Gap Size Classes

36

Table 2.1- Results of a generalized linear mixed model for the effects of Scarification,
Light Availability, Species, and their interactions on germination for eastern hemlock,
yellow birch, paper birch, and white pine. Each combination of factors was considered up
to three way interactions in the original model. All interactions Prob ChiSq > 0.25 were
pooled with the error term and the models rerun (Bancroft 1964).
Factor

Germination

Scarification
Species
Scarification x Species
Light
Species x Light
R2=.44

37

L-R
ChiSquare
38.31
3.96
11.17
14.14
12.26

Prob>ChiSq
<.0001
0.2659
0.0108
0.0002
0.0066

Table 2.2- Observed germination response of paper birch, yellow birch, eastern hemlock,
and white pine in Scarified and Unscarified subplots. SD=standard deviation.
Species

Treatment

Subplot
s

Paper birch
Yellow birch
Hemlock
White pine

Scarified
Scarified
Scarified
Scarified

40
40
40
24

Species

Treatment

Subplot
s

Paper birch
Yellow birch
Hemlock
White pine

Unscarified
Unscarified
Unscarified
Unscarified

40
40
40
24

Mean
seedlings/4m
2
17.15
9.53
10.98
2.04

Media
n

Rang
e

SD

7.5
1
2
1

0-128
0-113
0-59
0-13

24.97
20.38
16.02
2.84

Mean
seedlings/4m
2
0.98
0.1
0.88
2

Media
n

Rang
e

SD

0
0
0
1

0-16
0-1
0-18
0-8

2.65
0.3
2.93
2.19

Note1- White pine was seeded in fewer subplots due to seed shortages.

38

Table 2.3- Summary of environmental conditions across different gap sizes classes.
Large
Understory Single Tree
Group
Patch Cut
Variable
(0m dia)
(<23m dia)
(24-50m
(>50m dia)
dia)
Gaps/Understory (n=49)
4
13
24
8
Canopy Openness (%)
7.18
17.21
35
74.12
Non-Tree Vegetation Cover
10
12
66
76
(%)
Non-Tree Vegetation Height
3
4
27
44
(cm)
Rubus spp. Cover (%)
0
1
15
50
Forest Floor Light (%)
6.98
16.15
21.06
36.16

39

Table 2.4- Results of a generalized logistic mixed model for the effects of Clipping,
Light Availability, Species, and their interactions on germinant establishment
(1st year survival) and second year survival for eastern hemlock, yellow birch, paper birch,
and white pine. Each combination of factors was considered up to three way interactions in
the original model. All interactions Prob ChiSq > 0.25 were pooled with the error term
and the models rerun (Bancroft 1964).
Factor

Establishment (1st year)

Survival (2nd year)

Species
Light
Species x Light
R2=.43
Species
Light
Species x Light
R2=.36

40

L-R
ChiSquare
100.2
34.92
8.85

Prob>ChiSq

82.69
19.28
9.37

<.0001
<.0001
0.0247

<.0001
<.0001
0.0314

Table 2.5- Results of a generalized logistic mixed model for the effects of Clipping,
Light Availability, Species, Fencing, and their interactions on third year eastern hemlock
and paper birch survival. Each combination of factors was considered up to three way
interactions in the original model. All interactions Prob ChiSq > 0.25 were pooled with the
error term and the models rerun (Bancroft 1964).
L-R
Factor
Prob>ChiSq
ChiSquare
Clipping
16.55
<.0001
Species
15.98
<.0001
Light
Survival (3rd year)
12.85
0.0003
Clipping x Species
3.1
0.0784
2
R =.39

41

Table 2.6- The effect of Clipping on paper birch and eastern hemlock third year survival.
SD=standard deviation.
Species
Paper birch
Hemlock
Species
Paper birch
Hemlock

Cohort
Age

Treatment

Subplots

Mean
(%)

Median

Range

SD

3
3

Clipped
Clipped

26
32

68
27

82
6

0-100
0-100

39
37

Cohort
Age

Treatment

Subplots

Median

Range

SD

3
3

Unclipped
Unclipped

19
28

Mean
(%)
33
15

0
0

0-100
0-100

46
32

42

LITERATURE CITED

43

LITERATURE CITED

Abbott, H.G., 1961. White pine consumption by small mammals. Forestry. 59, 197-201.
Amatangelo, K.L., Fulton, M.R., Rogers, D.A., Waller D.M., 2011. Converging forest
community composition along an edaphic gradient threatens landscape-level
diversity. Diversity Distrib. 17, 201-213.
Anderson, R.C., Loucks, O.L., 1979. White-tail deer (Odcoileus virginianus) influence on
structure and composition of Tsuga canadensis forests. J. Appl. Ecol. 16, 855-861.
Anderson-Teixeira, K.J., Miller, A.D., Mohan, J.E., Hudiburg, T.W., Duval, B.D.,
DeLucia, E.H., 2013. Altered dynamics of forest recovery under a changing
climate. Global Change Biology. 19, 2001-2021.
Bancroft, T.A., 1964. Analyses and inferences for incompletely specified models involving
the use of preliminary tests of significance. Biometrics. 20, 427-442.
Beguin, J., Pr÷vost, M., Pothier, D., C te, S.D., 2009. Establishment of natural
regeneration under severe browsing pressure from white-tailed deer after group
seed-tree cutting with scarification on Anticosti Island. Can. J. For. Res. 39, 596605.
Bèland, M., Chicoine, B., 2013. Tolerant hardwood natural regeneration 15 years after
various silvicultural treatments on an industrial freehold of northwestern New
Brunswick. For. Chron. 89, 512-524.
Bolton, N.W., D’Amato, A.W., 2011. Regeneration responses to gap size and coarse
woody debris within natural disturbance-based silvicultural systems in northeastern
Minnesota, USA. For. Ecol. Manage. 262, 1215-1222.
Burger, T.L., Kotar, J., 2003. A guide to forest communities and habitat types of Michigan.
The Department of Forest Ecology and Management. University of WisconsinMadison, WI, USA.
Burns, R.M., Honkala, B.H., 1990. Silvics of North America, vol. 1. Agriculture
Handbook, vol. 654. USDA Forest Service, Washington, DC, USA.
Caspersen, J.P., Saprunoff, M., 2005. Seedling recruitment in a northern temperate forest:
the relative importance of supply and establishment limitation. Can. J. For. Res. 35,
978-989.
Chapin, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., 2000.
Consequences of changing biodiversity. Nature. 405, 234-242.

44

Cornett, M.W., Puettmann, K.J., Reich, P.B., 1998. Canopy type, leaf litter, predation, and
competition influence conifer seedling emergence and early survival in two coniferdeciduous forests. Can. J. For. Res. 28, 196-205.
Cornett, M.W., Puettmann, K.J., Frelich, L.E., Reich P.B., 2001. Comparing the
importance of seedbed and canopy type in the restoration of upland Thuja
Occidentalis forests of northeastern Minnesota. Restoraction Ecology. 9, 386-396.
Côte, S.D., Rooney, T.P., Tremblay, J.P., Dussault, C., Waller, D.M., 2004. Ecological
impacts of deer overabundance. Ann. Rev. Ecol. Evol. Syst. 35, 113-147.
Denslow, J.S., 1980. Gap partitioning among tropical rainforest trees. Biotropica. 12, 4755.
Drever, C.R., Peterson, G., Messier, C., Bergeron, Y., Flannigan, M., 2006. Can forest
management based on natural disturbances maintain ecological resilience? Can. J.
For. Res. 36, 2285-2299.
Duveneck, M.J., Scheller, R.M., White, M.A., Handler, S.D., Ravenscroft, C., 2014.
Climate change effects on northern Great Lake (USA) forests: A case for
preserving diversity. Ecosphere. 5, 1-26.
Fahey, R.T., Lorimer, C.G., 2013. Restoring a midtolerant pine species as a component of
late-successional forests: results of gap-based planting trials. For. Ecol. Manage.
292, 139-149.
Falk, K.J., Elliott, K.A., Burke, D.M., Nol, E., 2010. Early seedling response to group
selection harvesting in a northern hardwood forest. Forestry Chron. 86, 100-109.
Fischer, J., Lindenmayer, D.B., Manning, A.D., 2006. Biodiversity, ecosystem function,
and resilience: ten guiding principles for commodity production landscapes.
Frontiers in Ecology and the Environment. 4, 80-86.
Folke, C., Carpenter, S., Walker, B., Scheffer, M., Elmqvist, T., Gunderson, L., Holling,
C.S., 2004. Regime shifts, resilience, and biodiversity in ecosystem management.
Annu. Rev. Ecol. Evol. Syst. 35, 557-581.
Frazer, G.W., Canham, C.D., Lertzman, K.P., 1999. Gap Light Analyzer (GLA) Version
2.0 Users Manual and Program Documentation. Simon Fraser University, Burnaby,
British Columbia, Canada and the Institute of Ecosystem Studies, NY, USA.
Frelich, L.E., Lorimer, C.G., 1991. Natural disturbance regimes in hemlock-hardwood
forests of the upper Great Lakes region. Ecol. Mon. 61, 145-164.
Gasser, D., Messier, C., Beaudet, M., Lechowicz, M.J., 2010. Sugar maple and yellow
birch regeneration in response to canopy opening, liming, and vegetation control in
a temperate deciduous forest of Quebec. For. Ecol. Manage. 259, 2006-2014.

45

Goerlich, D.L., Nyland, R.D., 2000. Natural regeneration of eastern hemlock: a review.
Pages 14-23 in K.A. McManuc. K.S. Shield, Souto, editors. Proceedings:
symposium on sustainable management of hemlock ecosystems in eastern North
America, Durham, New Hampshire, USA, 22-24 June 1999. General Technical
Report NE-267. USDA Forest Service, Newton Square, Pennsylvania, USA.
Hanberry, B.B., 2013. Changing eastern broadleaf, southern mixed, and northern mixed
forest ecosystems of the eastern United States. For. Ecol. Manage. 306, 171-178.
Heinemann, K., Kirtzberger, T., 2006. Effects of position, understory vegetation, and
coarse woody debris on tree regeneration in two environmentally contrasting
forests of north-western Patagonia: a manipulative approach. J. Biogeogr. 33, 13571367.
Horsley, S.B., Stout, S.L., DeCalesta, D.S., 2003. White-tailed deer impact on the
vegetation dynamics of a northern hardwood forest. Eco. Appl. 13, 98-118.
Hough, A.F., 1960. Silvical characteristics of eastern hemlock (Tsuga canadensis). Res.
Pap. NE-132. U.S. Department of Agriculture, Forest Service, Northeastern Forest
Experiment Station. 23 p.
Houle, G., 1992. Spatial relationship between seed and seedling abundance and mortality
in a deciduous forest of north-eastern North America. Ecology. 80, 99-108.
Hura, C.A., Crow, T.R., 2004. Woody debris as a component of ecological diversity in
thinned and unthinned northern hardwood forests. Nat. Ar. J. 24, 57-64.
Kain, M., Battaglia, L., Royo, A., Carson, W.P., 2011. Over-browsing in Pennsylvania
creates a depauperate forest dominated by an understory tree: Results from a 60year-old deer exclosure. J. Torr. Bot. Soc. 138, 322-326.
Kern, C.C., Reich, P.B., Montgomery, R.A., Strong, T.F., 2012. Do deer and shrubs
override canopy gap size effects on growth and survival of yellow birch, northern
red oak, eastern white pine, and eastern hemlock seedlings? For. Ecol. Manage.
267, 134-143.
Kern, C.C., D’Amato, A.W., Strong, T.F., 2013. Diversifying the composition and
structure of managed, late-succesional forests with harvest gaps: what is the
optimal gap size? For. Ecol. Manage. 304, 110-120.
Klingsporn Poznanovic, S., Webster, C.R., Bump, J.K., 2013. Maintaining mid-tolerant
tree species with uneven-aged forest management: 9 year results from a novel
group-selection experiment. Forestry 86, 555-567.
Kobe, R.K., Pacala, S.W., Silander, J.A., Canham, C.D., 1995. Juvenile tree survivorship
as a component of shade tolerance. Ecological Applications. 5, 517-532.
Kobe, R.K., Hogarth, L., 2007. Evaluation of irradiance metrics with respect to predicting
sapling growth. Can. J. For. Res. 37, 1203-1213.
46

Lawrence, D., 2004. Erosion of tree diversity during 200 years of shifting cultivation in
Bornean rain forest. Ecological Applications. 14, 1855-1869.
Long, Z.T., Pendergast IV, T.H., Carson, W.P., 2007. The impact of deer on relationships
between tree growth and mortality in an old-growth beech-maple forest. For. Ecol.
Manage. 252, 230-238.
Lorenzetti, F., Delagrange, S., Bouffard, D., Nolet, P., 2008. Establishment, survivorship
and growth of yellow birch seedlings after site preparation treatments in large gaps.
For. Ecol. Manage. 254, 350-361.
Martell, A.M., 1979. Selection of conifer seeds by deer mice and red-backed voles. Can. J.
For. Res. 9, 201-204.
Marx., L., Walters, M.B., 2006. Effects of nitrogen supply and wood species on Tsuga
canadensis and Betula alleghaniensis seedling growth on decaying wood. Can. J.
For. Res. 36, 2873-2884.
Marx, L., Walters, M.B., 2008. Survival of tree seedlings on different species of decaying
wood maintains tree distributions in Michigan hemlock-hardwood forests. Ecology
96, 505-513.
Matonis, M.S., Walters, M.B., Millington, J.D.A, 2011. Gap-,stand,-and landscape-scale
factors contribute to poor sugar maple regeneration after timber harvest. For. Ecol.
Manage. 262, 286-298.
McCarthy, J., 2001. Gap dynamics of forest trees: A review with particular attention to
boreal forests. Environmental Reviews. 9, 1-59.
McGee, G.G., Birmingham, J.P., 1997. Decaying logs as germination sites in northern
hardwood forests. Nor. J. of App. For. 14, 178-182.
Melillo, J.M., Aber, J.D., Muratore, J.F., 1982. Nitrogen and lignin control of hardwood
leaf litter decomposition dynamics. Ecology. 63, 621-626.
Millington, J.D.A., Walters, M.B., Matonis, M.S., Liu, J., 2010. Effects of local and
regional landscape characteristics on wildlife distribution across managed forests.
259, 1102-1110.
Montgomery, R.A., Reich, P.B., Palik, B.J., 2010. Untangling positive and negative biotic
interactions: views from above and below ground in a forest ecosystem. Ecology.
91, 3641-3655.
Neuendorff, J.K., Nagel, L.M., Webster, C.R., Janowiak, M.K., 2007. Stand structure and
composition in a northern hardwood forest after 40 years of single-tree selection.
N.J. Appl. For. 24, 197-202.

47

Nobis, M., Hunziker, U., 2005. Automatic thresholding for hemisphericald canopyphotographs based on edge detection. Agricultural and Forest Meteorology. 128,
243-250.
Nowacki, G.J., Abrams, M.D., 2008. The demise of fire and “mesophication” of forests in
the Eastern United States. Bioscience. 58, 123-138.
Nuttle, T., Ristau, T.E., Royo, A.A., 2014. Long-term biological legacies of herbivore
density in a landscape-scale experiment: forest understoreys reflect past deer
density treatments for at least 20 years. Ecology. 102, 221-228.
Perala, D.A., Alm, A.A., 1990. Reproductive ecology of birch: A review. For. Ecol.
Manage. 32, 1-38.
Pitt, D.G., Morneault, A., Parker, W.C., Stinson, A., Lanteigne, L., 2009. The effects of
herbaceous and woody competition on planted white pine in a clearcut site.
Prèvost, M., Raymond, P., Lussier, J.M., 2010. Regeneration dynamics after patch cutting
and scarification in yellow birch –conifer stands. Can. J. For. Res. 40, 357-369.
Putz, F.E., Canham, C.D., 1992. Mechanisms of arrested succession in shrublands: root
and shoot competition between shrubs and seedlings. For. Ecol. Manage. 49, 267275.
Ramsey, F., Schafer, D. The Statistical Sleuth: A course in Methods of Data Analysis.
Belmont, CA: Duxbury Press, 1997.
Randall, J.A., Walters, M.B., 2011. Deer density effects on vegetation in aspen forest
understories over site productivity and stand age gradients. For. Ecol. Manage. 261,
408-415.
Raymond, P., Munson, A.D., Ruel, J.C., Nolet, P., 2003. Group and single-tree selection
cutting in mixed tolerant hardwood-white pine stands: Early establishment
dynamics of white pine and associated species. For. Chron. 79, 1093-1106.
Raymond, P., Bèdard, S., Roy, V., Larouce, C., Tremblay, S., 2009. The irregular
shelterwood system: review, classification, and potential application to forests
affected by partial disturbances. Forestry. 107, 405-413.
Redmond, D.R., Robinson, R.C., 1954. Viability and germination in yellow birch. For.
Chron. 30, 79-87.
Ricklefs, R.E., 1977. Environmental heterogeneity and plant species diversity: a
hypothesis. Am. Nat. 111, 376-381.
Rooney, T.P., Waller, D.M., 2003. Direct and indirect effects of white-tailed deer in forest
ecosystems. For. Ecol. Manage. 181, 165-176.

48

Royo, A.A., Carson, W.P., 2006. On the formation of dense understory layers in forests
worldwide: consequences and implications for forest dynamics, biodiversity, and
succession. Can. J. For. Res. 36, 1345-1362.
Runkle, J.R., Patterns of disturbance in some old-growth mesic forests in eastern North
America. Ecology. 63, 1533-1546.
Rustad, L.E., Cronan, C.S., 1988. Element loss and retention during litter decay in a red
spruce stand in Maine. Can. J. For. Res. 18, 947-953.
Schuler, T.M., Gillespie, A.R., 2000. Temporal patterns of woody species diversity in a
Centrral Appalachian forest from 1856 from 1856 to 1997. J. Torr. Bot. Soc. 127,
149-161.
Schulte, L.A., Mladenoff, D.J., 2005. Severe wind and fire regimes in northern forests:
historical variability at the regional scale. Ecology. 86, 431-435.
Schulte, L.A., Mladenoff, D.J., Crow, T.R., Merrick, L.C., Cleland, D.T., 2007.
Homogenization of northern U.S. Great Lakes forests due to land use. Landscape
Ecol. 22, 1089-1103.
Scheller, R.M., Mladenoff, D.J., 2002. Understory species patterns and diversity in oldgrowth and managed northern hardwood forests. Ecological Applications. 12,
1329-1343.
Seymour, R.S., White, A.S., deMaynadier, P.G., 2002. Natural disturbance regimes in
northeastern North America-evaluating silvicultural systems using natural scales
and frequencies. For. Ecol. Manage. 155, 357-367.
Shields, J.M., Webster, C.R., Nagel, L.M., 2007. Factors influencing tree species diversity
and Betula alleghaniensis establishment in silvicultural openings. Forestry 80,
293-307.
Shields, J.M., Webster, C.R., 2007. Ground-layer response to group selection with legacytree retention in a managed northern hardwood forest. Can. J. For. Res. 37, 17971807.
Smith, D.M., 1951. The influence of seedbed conditions on the regeneration of E. white
pine. Conn. Agric. Exp. Stn. Bull. 545.
Sturrock, R.N., Frankel, S.J., Brown, A.V., Hennon, P.E., Kliejunas, J.T., Lewis, K.J.,
Worrall, J.J., Woods, A.J., 2011. Climate change and forest diseases. Pl.Path. 60,
133-149.
Tremblay, J.P., Thibault, I., Dussault, C., Hout, J., Côte, S.D., 2005. Long-term decline in
white-tailed deer browse supply: can lichens and litterfall act as alternative food
sources that preclude density-dependent feedbacks. Can. J. Zool. 83, 1087-1096.

49

United Nations, Department of Economic and Social Affairs, Population Division (2013).
World Population Prospects: The 2012 Revision, Highlights and Advance Tables.
Working Paper No. ESA/P/WP.228.
Valkonen, S., Maguire, D.A., 2005. Relationship between seedbed properties and the
emergence of spruce germinants in recently cut Norway spruce selection stands in
Southern Finland. For. Ecol. Manage. 210, 255-266.
Witt, J.C., Webster, C.R., 2010. Regeneration dynamics in remnant Tsuga canadensis
stands in the northern Lake States: Potential direct and indirect effects of herbivory.
For. Ecol. Manage. 260, 519-525.
Wright, E.F., Coates, K.D., Bartemucci, P., 1998. Regeneration from seed of six tree
species in the interior cedar-hemlock forests of British Columbia as affected by
substrate and canopy.
Zhang, Q.F., Pregitzer, K.S., Reed, D.D., 2000. Historical changes in the forests of the
Luce District of the Upper Peninsula of Michigan. Am. Midl. Nat. 143, 94-110.

50

Chapter III
Does substrate availability alter natural regeneration patterns within harvest gaps?

Abstract
Balancing timber production and non-timber attributes like biodiversity has
increasingly become a goal of forest management. By nature, this is a more challenging
task than timber centric management, as silvicultural prescriptions must account for the
regeneration requirements of several different tree species rather than just those species
which are most economically valued. In the managed northern hardwood forests of the
Great Lakes region, low light availability, resulting from decades of single-tree selection
management, has been implicated as a potential cause of regeneration failures of several
shade intolerant to mid-tolerant species. However, evidence from several harvest
regeneration studies have shown that increasing light availability through the creation of
larger harvest gaps has been only modestly successful in altering seedling layer
composition. This study uses mixed models approach to investigate the influence of local
seed source density, substrate availability, site quality (moisture and nutrient availability),
and harvest gap size (75-650m2) on sugar maple, red maple, white ash, white pine, yellow
birch, paper birch, eastern hemlock, and black cherry seedling establishment over three
years in 19 upland hardwood stands located throughout the Northern Lower and Eastern
Upper Peninsula’s of Michigan. In general, larger seeded shade tolerant to mid-tolerant
species dominated the seedling layer, as sugar maple, red maple, and white ash were at

51

least twice as abundant as any other species. The five other species’ comparatively lower
density was not related to harvest gap size as only sugar maple (lower establishment with
increasing light availability) and red maple (increasing establishment with increasing light
availability) establishment significantly responded to differences in harvest gap size. In
contrast, local seed source density strongly influenced yellow birch, eastern hemlock, black
cherry, and white ash seedling layer presence, as each species significantly and uniquely
increased in density with increasing local seed source production potential. In a similar
fashion, white pine and paper birch also experienced a significant increase in seedling
establishment in plots that had been scarified (bare mineral soil/humus dominated) when
compared to unscarified plots (hardwood litter dominated). Finally, site quality was an
important contributing factor to seedling establishment, as at a given level of seed
production potential, white ash and sugar maple occurred at higher densities in gaps with
higher moisture and nutrient availability, while red maple and black cherry were more
abundant in gaps with lower moisture and nutrient availability. Collectively, these results
suggest that local seed production, substrate availability and site quality may be exerting a
stronger influence on shade tolerant species seedling layer dominance than is light
availability.

52

Introduction
Balancing timber production and non-timber attributes like biodiversity has
increasingly become a goal of forest management (Lindenmayer et al. 2000; Lindenmayer
et al. 2006; Heller and Zavaleta 2009). By nature, this is a more challenging task than
timber centric management, as silvicultural prescriptions must account for the regeneration
requirements of several different tree species rather than just those species which are most
economically valued. Further complicating matters, silvicultural techniques aimed at
increasing species diversity remain untested in several forest ecosystems.
One such community where concern over tree species diversity is growing and
experimentation with alternative silvicultural tools is needed is the managed northern
hardwoods forests of the Great Lakes region. Here, for nearly six decades, silvicultural
practices have concentrated on regenerating economically valuable shade tolerant species
by using the single-tree selection system. The single-tree selection system relies on natural
regeneration to fill the spaces created by the harvest of individually marked trees at 10-15
year harvest intervals (15-20% reduction in basal area per harvest) (Arbogast 1957;
O’Hara 2002; Neuendorff et al. 2007). In the process, harvesting typically creates small
harvest gaps, which only marginally and ephemerally increase light availability to
regenerating tree seedlings (Eyre, and Zillgitt 1953; Klingsporn et al. 2012). Thus, it is
widely believed that the low light environments created in single-tree selection harvests
have likely contributed to the expansion of shade tolerant species and the decline of less
shade tolerant species in this forest type (Zhang et al. 2000; Schwartz et al. 2005;
Neuendorff et al. 2007; Amatangelo et al. 2011; Matonis et al. 2011).

53

Recognizing that declining diversity could lead to decreased forest resilience and
sustainability, operational and experimental efforts have been made to increase the
representation of less shade tolerant species by creating larger, higher-light harvest gap
sizes by selecting groups of trees for harvest. However, these trials have been generally
unsuccessful, as the same shade tolerant species dominating small gaps generally dominate
large gaps as well (Shields et al. 2007; Bolton and D’Amato 2011; Kern et al. 2012; Kern
et al. 2013; Klingsporn-Poznanovic et al. 2013). This suggests that factors in addition to
light availability are constraining the recruitment of some species within harvest gaps.
Among these factors, seedling substrate availability, local seed sources, advanced
regeneration, and competition from non-tree vegetation may be some of the most
important.
Due to a decline in coarse woody debris (Hura and Crow 2004), and an increase in
gentle harvesting practices which seek to minimize surface disturbance (e.g. winter
harvesting), bare mineral soil/humus often covers only a small fraction of the forest floor
in harvest gaps (Shields et al. 2007). Among substrates, bare mineral soil/humus may be
particularly important for some species regeneration, as its combination of adequate
nutrient availability, moisture holding capacity and receptiveness to root penetration
provide a favorable environment for seedling establishment (Collis-George and Sands
1959; Gray et al. 1997; Prescott et al. 2000; Oleskog and Sahlèn 2000). Smaller seeded
species in particular may be dependent on bare mineral soil, as their stored carbohydrate
reserves provide little energy for the development of roots to find reliable moisture and
nutrient supplies whilst establishing through leaf litter (Burton et al. 1969; Kidson and
Westoby 2000; Moles and Westoby 2004). Therefore, the regeneration of smaller seeded

54

species, several of which are less shade tolerant (e.g. yellow birch, paper birch) but also
eastern hemlock, which is extremely shade tolerant (Kobe et al. 1995), may be limited
more by seedling establishment substrates than low light environments (Casperson and
Saprunoff 2005; Willis et al. ch1).
Lack of local seed production may also limit the regeneration of several species in
northern hardwood forest harvest gaps. While overstory species richness has remained
largely unchanged over the past century, several factors including exploitative logging,
invasive pests and pathogens, and decades of poor recruitment by several species have
combined to reduce canopy layer evenness at the landscape scale (Zhang et al. 2000;
Schulte et al. 2007; Amatangelo et al. 2011). As such, several previously abundant species
are now likely contributing less seed to the annual seed rain within stands. Three such
species which have experienced sharp long-term declines are eastern hemlock (Tsuga
canadensis L.), white pine (Pinus strobus L.) and yellow birch (Betula alleghaniensis
Britton.) (Whitney 1987; Frelich 1995; Woods 2000; Zhang et al. 2000; Schulte et al.
2002; Friedman and Reich 2005; Schulte et al. 2007; Rhemtulla et al. 2009). While their
reduction in the overstory likely results in local reductions of seed availability, the effects
likely vary by species because dispersal characteristics vary among species. For instance,
compared to birch, hemlock and white pine seeds are less broadly dispersed by wind and
thus, unless vectored by wildlife, are more likely to be locally dispersal limited (Ribbens et
al. 1994). Therefore, the lack of conifer representation within harvest gaps may be
facilitated to a greater extent by a lack of local seed sources rather than a lack of favorable
establishment substrate or light availability.

55

Competition from non-tree vegetation (shrubs, forbs, graminoids and ferns) may
represent an additional explanation for failed species recruitment within harvest gaps. By
occupying establishment sites and competing for light, moisture, and nutrients, non-tree
vegetation has been shown to influence seedling survival (George and Bazzaz 1999;
Beckage and Clark 2003; Royo and Carson 2006; Royo and Carson 2008; Montgomery et
al. 2010). Shade intolerant to mid-tolerant species may be the functional group most
threatened by dense layers of non-tree vegetation, as the low light levels created by
competing vegetation may prevent them from energetically offsetting their comparatively
higher respiration costs (Bazzaz 1979; Walters and Reich 2000; Crane and Reich 2005;
Valladares and Niinemets 2008). However, competition may not be a significant factor in
all harvest gaps, as non-tree vegetation cover and height have been shown to positively
correlate with increasing gap size (Shields et al. 2007; Matonis et al. 2011; Kern et al.
2012). Likewise, competition intensity may vary geographically, as local soil moisture and
nutrient regimes (site quality) may modify non-tree vegetation’s response to increasing
light availability. Therefore, larger harvest gaps on lower quality sites may provide better
recruitment opportunities for shade intolerant to mid-tolerant species than larger harvest
gaps on higher quality sites if competition intensity varies with site quality (Davis et al.
1999).
In conclusion, low light availability created by small harvest gaps has been
considered as the primary factor driving the expansion of shade tolerant species in the
Great Lakes region. However, attempts to restore seedling diversity by varying only
harvest gap size to affect light availability have been largely unsuccessful at promoting less
shade tolerant species and other declining species, suggesting that other factors within

56

harvest gaps may be preventing their regeneration. In this report, we examine the effects of
local seed tree density, mineral soil substrate availability, light availability, and non-tree
vegetation on seedling establishment and first-year survival of eight tree species in 114
harvest gaps located in 19 different stands of varying site quality in the Northern Lower
and Eastern Upper Peninsulas of Michigan. Specifically we predict: 1) local seed source
proximity will limit seedling establishment for conifers and larger seeded species 2) bare
mineral soil/humus availability will limit the establishment of lighter seeded species 3)
competition from non-tree vegetation will limit shade intolerant species establishment in
large harvest gaps on high quality sites but not on lower quality sites.

Methods
Study area
In May, 2011, 11 upland hardwood or mixed-upland forest stands entering their
first growing season following a winter harvest were located throughout the Northern
Lower and Eastern Upper Peninsula’s of Michigan with the aid of a harvest database
maintained by the Michigan Department of Natural Resources (MDNR). Eight additional
stands fitting the same criteria were located in May of 2012. Overstory composition in
most stands was dominated by sugar maple (Acer saccharum Marsh.) with lesser and
varying amounts of red maple (A. rubrum L.), white ash (Fraxinus americana L.),
American basswood (Tilia americana L.), white pine (Pinus strobus L.), red oak (Quercus
rubra L.), eastern hemlock (Tsuga canadensis L.), yellow birch (Betula alleghaniensis
Britton.), paper birch (Betula papyrifera Marsh), and black cherry (Prunus serotina Ehrh.).
Habitat type, a classification system that uses assemblages of understory vegetation as a

57

proxy for nutrient and moisture availability, was used to classify the site quality of each
harvest gap (Burger and Kotar 2003). This method was chosen due to the logistical
challenges of quantifying moisture and nutrient availability across 19 sites, and due to the
demonstrated correspondence of this system to in situ measurements of nitrogen and
moisture availability (Walters and Reich 1997). It was also chosen for its relevance to
forest management, as the habitat type classification system is commonly used as a
management tool throughout the Great Lakes region. The three habitat classes encountered
in the Northern Lower Peninsula in this study are AFOCa (mesic to very-mesic soil
moisture availability; rich-very rich nutrient availability), AFO (dry-mesic to mesic;
medium to rich), and ParVVb (dry to dry-mesic; poor to medium); while the three
encountered in the Eastern Upper Peninsula include AFOA (mesic to wet-mesic; medium
to rich), AFPo (dry-mesic to mesic; medium), and ATFD (dry-mesic to mesic; poor to
medium).

Field methods
Within each stand, six harvest gaps were selected to track seedling establishment
and survival. Gaps were selected non-randomly with the goal of varying gap size and
maximizing local seed source (within 50m radius) diversity. Gap size was determined by
measuring the tree bole to tree bole distance for edge trees using the methods of Runkle
(1981). Two additional requirements for gap selection were that they were of harvest origin
(i.e. at least one tree harvested) and at least 50m from a forest edge or another gap included
in the study. Once gaps were chosen, all advanced regeneration was removed from the gap
with a brush saw. Two plots (3x3m) were then established within each gap, with one

58

randomly selected to receive a soil disturbance treatment. In late May, treated plots were
scarified with rakes until the forest floor was dominated by a combination of bare mineral
soil and humus.

Field measurements
Within each plot, we visually estimated coverage of bare mineral soil/humus
(combined), litter, forbs, shrubs, Rubus spp., grass/sedge, and non-receptive substrate (e.g.
rocks) to the nearest 5%. In addition, we measured non-tree vegetation height (cm) at nine
pre-determined locations to calculate an average non-tree vegetation height value for each
plot. In order to estimate seed dispersal within harvest gaps, local seed production potential
was determined for each species included in this study in each gap by summing the
quotient of the squared diameter-at-breast-height (1.4m) and the squared distance (m) to
plot center of each tree greater than 25.4 (cm) in diameter and within 25m of the gap
center, but outside of the gap itself (∑diameter2/distance2) (Ribbons et al. 1994; Matonis et
al. 2011). Species determined to be declining from previous demographic studies
conducted in the region (Zhang et al. 2000; Friedman and Reich 2005; Schwartz et al.
2005; Schulte et al. 2007; Amatangelo et al. 2011; Matonis et al. 2011) (yellow birch,
white pine and hemlock) or shade intolerant species (paper birch and black cherry) were
measured within 50m of gap center to consider a greater range of local seed production
potential and also to account for Betula’s greater seed dissemination distance potential
(Ribbens et al. 1994). Canopy openness, a proxy for light availability (Kobe and Hogarth
2007), was determined for each plot by taking a hemispherical photograph at 1.5 m (which
was always above the understory vegetation layer) above plot center during overcast

59

conditions or twilight. (Canham et al. 1990). Each photo was analyzed with Gap Light
Analyzer v. 2.0 (Frazer et al. 1999) using an automatic threshold value provided by SideLook v. 1.1.01 (Nobis and Hunziker 2005). Seedling establishment, forest floor coverage,
and non-tree vegetation height were then assessed in each plot at the end of the growing
season (September). All established seedlings were tagged and checked in September for
survival. Each plot was then revisited annually to check for the establishment of new
seedlings, changes in substrate coverage, non-tree vegetation coverage and height, and the
survival of previously established seedlings for three (2011 established sites) or two (2012
established sites) growing seasons.

Statistical methods
We examined the effects of species, substrate availability, canopy openness, local
seed production potential, site quality and their two way interactions on seedling
establishment with Generalized Linear Mixed Models (GLMM) assuming a Poisson
distribution with a logarithmic link function. We selected this approach to account for the
high number of biologically significant zeroes which prevented us from transforming the
data to meet the assumption of normality for general linear modeling. Stand was not
incorporated as a factor despite gaps being nested within stands. This was done for two
reasons. First, considering stands as a factor would have severely limited degrees of
freedom for testing the factors of interest. Second, gap plots can reasonably be considered
experimental units as all measurements we took (habitat class, seed sources) were gap
specific. Prior to modeling, site quality was standardized by categorizing individual
habitat types into high (AFOCa and AFOA), medium (AFO and AFPo), and low site

60

quality (ParVVb and ATFD) groups, and analyzed as a continuous variable. While we
recognize that habitat type could also be categorized as an ordinal variable, which would
avoid the potentially invalid assumption of a continuous relationship of moisture/nutrient
availability among habitat types, we chose to treat habitat type as a continuous variable in
order to preserve degrees of freedom, allowing us to examine a broader collection of
species. The influence of each factor and their two-way interactions were then investigated
in a full factorial modeling design. Three and four-way interactions were not considered
due to a loss of degrees of freedom. Model selection was accomplished through backwards
elimination. Insignificant interactions ProbChiSq > 0.25 were pooled with the error term
and rerun as a new model until only statistically significant factors (ProbChiSq <.05) and
interactions (ProbChiSq <0.25) could be reached (Bancroft 1964). Once we arrived at a
final model, we checked the model for dispersion of the distribution, goodness of fit and
potential outliers. Model dispersion was estimated by calculating an over-dispersion factor
(ĉ = chi-square/ df). Models with dispersion factor values > 1 were considered overdispersed. In such cases, parameter estimates were obtained using a quasi-likelihood
approach which inflated the standard errors of each factor estimate by √ ĉ . The model was
then re-run and checked for significance. Model fit and potential outlier detection was
accomplished by examining plots of studentized deviance residuals. Model fit was
evaluated by visually inspecting how evenly the residuals were distributed around zero.
Points on the plot exceeding two standard deviations of the mean were identified as
potential outliers. We evaluated the influence of such points with Cook’s distance statistic.
Values > 1 indicate points potentially having significant effects on the model. To
investigate these effects, we removed the potential outlier from the data set and re-ran the

61

model. If all parameters in the model remained significant, the potential outlier was kept as
part of the data set. If changes in parameter significance occurred while the outlier was
excluded, we removed the observation from the model (Ramsey and Schafer 1997). This
issue, however, was not encountered.
The effects of species, canopy openness, non-tree vegetation density, site quality,
and their two-way interactions on seedling survivorship were investigated through
Generalized Logistic Mixed Models assuming a binomial distribution with a logit link
function. The same model fitting and model diagnostic procedures that were used to model
seedling establishment were repeated for seedling survival. While statistically significant
relationships were found for seedling survival, our final model predictive strength was low
overall (Table 3.1). Furthermore, no compelling main effects or meaningful interactions
were discovered during the analysis. As such, we will not discuss seedling survival in this
report. All analyses were performed in JMP 9.0 (SAS Institute, Cary, NC).

Results
Seedling establishment
The eight species examined in this study varied strongly in their rates of seedling
establishment (Table 3.2). In general, larger seeded shade tolerant to mid-tolerant species
dominated the seedling layer as sugar maple, red maple, and white ash were at least twice
as abundant as any other species (Table 3.3). This pattern, however, was not related to the
interaction between Local Seed Production Potential (SPP) and Habitat Type (site quality)
as, at a given level of SPP, all species were found at higher densities on higher quality

62

Habitat Types (Table 3.2). Species also responded similarly to the interaction of Light and
Habitat Type, as at a given level of Light Availability, seedling establishment was
consistently greater on higher quality Habitat Types (Table 3.2). Similar responses were
also observed among species to the interaction of Light and SPP, as seedling density was
consistently higher for low light environments at a given level of SPP for all species (Table
3.2). However, Species did differ strongly in their interactions with SPP and Habitat Type,
respectively (Table 3.2). Local seed source density was an important constraint on eastern
hemlock, white ash, black cherry, and yellow birch establishment, as each species
increased significantly but uniquely with increasing SPP (Figure 3.1). Independent of
Local Seed Production Potential, black cherry and red maple established at greater
densities on lower quality sites, while sugar maple and white ash were found in greater
abundance on higher quality sites (Figure 3.2). This relationship for black cherry and white
ash, however, was largely driven by variation in SPP (Figure 3.2). In addition to Local
Seed Production Potential and Habitat Type, Species also differed significantly in their
response to Scarification (bare mineral substrate) and Light Availability (Table 3.2).
Scarification was a significant factor for paper birch and white pine, as seedling density
increased for both species in scarified compared to unscarified plots (Table 3.4). Light, on
the other hand, strongly influenced red maple and sugar maple, as red maple seedling
density increased along with increasing Light Availability, while sugar maple seedling
density decreased with increasing Light Availability (Figure 3.3).

63

Discussion
Seedling establishment
One of the basic assumptions of gap based silviculture is that targeted tree species
will have the opportunity to establish within harvest gaps. While logistical constraints
prevented us from measuring seed rain in harvest gaps directly, local seed production
potential exerted a strong influence on seedling establishment for several species including
eastern hemlock, white ash, yellow birch and black cherry (Figure 3.1). These findings are
in agreement with our initial hypothesis of seed limitation for larger seeded or conifer
species and, with the exception of white ash, are also consistent with seed dispersal trends
reported in the literature (Burns and Honkala 1990; Ribbons et al. 1994; Casperson and
Saprunoff 2005; Pairon et al. 2006). In contrast to our hypothesis, however, local seed
production potential was also a significant seedling establishment constraint for yellow
birch. This finding was surprising considering Betula’s reported ability to disperse seeds
long distances by wind over air or snow (Matlack 1989; Ribbons et al. 1994).
Nevertheless, in forests, low wind velocity has been shown to constrain yellow birch’s
seed dispersal over snow (Greene and Johnson 1997), and seed density has also been
shown to be vastly reduced at a distance beyond 125(m) (Hughes and Fahey 1988).
Consequently, our finding of local seed source limitation for yellow birch is consistent
with the seed dispersal findings of Greene and Johnson (1997) and Hughes and Fahey
(1988), and may help explain the low rates of seedling recruitment reported by other
investigators at the gap and landscape scales (McEuen and Curran 2004; Shields et al.
2007; Bolton and D’Amato 2011; Kern et al. 2013).

64

Independent of local seed source production potential, site quality (habitat type)
also factored strongly into individual species seedling establishment (Table 3.2). White ash
and sugar maple established at higher densities in gaps that were of higher site quality
while red maple and black cherry established at greater densities in gaps of lower site
quality (ParVVb and ATFD) (Figure 3.2). Collectively, this pattern corresponds with each
species’ reported regeneration requirements (Burns and Honkala 1990) and is consistent
with other studies which have reported enhanced performance of sugar maple and white
ash on high quality sites and/or moister topographic positions (Host et al. 1987; Walters
and Reich 1997; Shreeg et al. 2005; Frey et al. 2007), and red maple and black cherry’s
reported tolerance of more xeric sites and landscape positions (Abrams 1998; Burger and
Kotar 2004; Frey et al. 2007).
Bare mineral soil/humus availability was yet another factor which modestly
influenced seedling establishment (Table 3.2). Consistent with our expectations, and the
results of other forest regeneration studies, scarification significantly increased paper birch
seedling establishment in harvest gaps (Raymond et al. 2003; Beguin et al. 2009; Prèvost
et al. 2010; Willis et al. Ch1) (Table 3.4). However, contrary to our expectations, and our
previous report (Willis et al.) Ch1, white pine seedling establishment also significantly
increased in scarified plots (Table 3.4). One potential explanation for these seemingly
contradictory patterns may be related to a difference in organic layer depth in the high
quality (AFOca) gaps examined in Willis et al. (ch1) and the mostly medium (AFO and
AFPo) to lower ParVVb and ATFD quality gaps where white pine established in this study
(Data not shown), as the potential combination of a greater proportion of slowly
decomposing nutrient deficient litter and drier soil moisture conditions may have slowed

65

decomposition in harvest gaps on medium and lower quality sites, resulting in deeper,
more recalcitrant organic layers. In support of this theory, Raymond et al. (2003) reported
a similar increase in white pine establishment following scarification in lower quality
harvest gaps than the ones reported in Willis et al. (ch1). As such, scarification may have a
more significant impact on larger seeded species establishment in harvest gaps on lower
quality sites. Nevertheless, no other larger seeded species demonstrated a significant
increase in scarified plots in lower quality gaps following scarification, and white pine
establishment was low overall compared to most other species (Table 3.3).
Light availability was the final factor which had modest significance on seedling
establishment (Table 3.2). Independent of local seed source productivity, red maple
established at greater abundance in high light environments, while sugar maple was more
abundant in low light environments (Figure 3.3). This pattern coincides with each species’
reported shade tolerance (Burns and Honkala 1990; Kobe et al. 1995) and provides limited
support for the notion that gap size can be manipulated to control species composition
based on their differing functional abilities (Messier et al. 1999). Nevertheless, several
species of lesser or equal shade tolerance including paper birch, black cherry, yellow birch,
white ash and white pine did not establish at greater density in higher light environments.
This suggests that creating larger harvest gaps may not increase the representation of
declining shade intolerant to mid-tolerant species unless advanced regeneration is already
present. It also adds to a growing body of evidence in the Great Lakes region which
indicates that the effect of gap size can be marginalized by several different potentially
interacting factors (Shields et al. 2007; Bolton and D’Amato 2011; Matonis et al. 2011;
Kern et al. 2013; Walters et al. in preperation). However, it is also possible that the range

66

of harvest gaps (75-650m2) (Table 5) encountered in this report was too small to incite
significant seedling establishment response from shade intolerant to mid tolerant species,
as Willis et al. (ch1) reported increasing rates of seedling establishment with increasing
light availability across a larger range of gap sizes (220 to 6500m2).

Conclusions and Management Recommendations
Gap regeneration is a complex process involving several steps from seed creation
all the way up through sapling and pole growth. Due to seedling establishment being one
of the earliest steps in this process, the factors which affect its outcome have the potential
to alter regeneration outcomes. Managing for biodiversity complicates this process further,
as land managers must account for each species’ particular regeneration requirements.
Above all, the results of this study emphasize the importance of local seed sources
and site quality for individual species’ seedling establishment. In managed northern
hardwood forests, locating harvest gaps in proximity to seed sources may prove
challenging, as several species are currently less abundant in the overstory compared to
previous decades (Zhang et al. 2000; Schwartz et al. 2005; Schulte et al. 2007; Amatangelo
et al. 2011). In such situations, planting seedlings (Walters et al. in press) or artificial
seeding (Willis et al. Ch1) may help overcome potential seed source limitations. These
actions may also be aided greatly by prior knowledge of site quality and the stress
tolerance of the targeted species. In conclusion, management efforts which ignore these
critical factors, and simply alter harvest gap sizes to manipulate species composition, may
be setting the stage for future regeneration failures unless advanced regeneration is already
established.

67

Figure 3.1- Predicted establishment response of yellow birch (YB), white ash (WA),
black cherry (BC), and eastern hemlock (EH) to local (50m) seed production potential
(∑diameter2/distance2). R2 and P values represent the fit of the whole model and
significance of seed production potential, respectively.

Seedlings/9m2

160
WA
2
R =0.49
p<.0001

120
80

BC
2
R =0.3
p<.0001

YB
2
R =0.13
p=.0003

40

EH
2
R =0.28
p<.0001

0
0

50

100

150

200

Seed Production Potential

68

250

Figure 3.2- Seedling establishment as predicted by local seed production potential in high
(AFOCa and AFOA), medium (AFO and AFPo) and low (ParVVb and ATFD) quality
habitat types for red maple (A), black cherry (B), sugar maple (C), and white ash (D)
(Burger and Kotar 2004).
30

50
R2=0.01
P=.0043

30
20

25

Seedlings/9m2

Seedlings/9m 2

40

15
10
5

0

0

0

100

200

300

400

500

0

20

140

80

100

120

140

160

60
40

Seedlings/9m2

2
R =0.4
P<.0001

D

Higher Quality
Medium Quality
Lower Quaulity

140

80

60

160

C

Higher Quality
Medium Quality
Lower Quality

100

40

Seed Production Potential

Seed Production Potential

Seedlings/9m2

2
R =0.43
P<.0001

20

10

120

B

Higher Quality
Medium Quality
Lower Quality

A

Medium Quality
Lower Quality

120

2

R =0.49
P<.0001

100
80
60
40

20
20

0
0

100

200

300

400

0

500

0

20

40

60

80

100

120

140

160

Seed Production Potential

Seed Production Potential

Note-1 Higher quality sites were not included in this analysis for red maple, as it only had a
viable seed production potential in one harvest gap.

69

Figure 3.3- The effect of varying light availability in 114 harvest gaps on red maple (A)
and sugar maple (B) seedling establishment.

160
140
2
Seedlings /9m

2

R =0.04
p=.0134

A

120
100
80
60
40
20
0
10

20

30

40

50

Light Availability (%)

160

B

R2= 0.13
p<.0001

Seedlings/9m

2

140
120
100
80
60
40
20
0
0

10

20

30

Light Availability (%)

70

40

50

Table 3.1- Results of a generalized logistic mixed model for the effects of Light
Availability, Non-Tree Vegetation Density, Habitat Type, Species, and their two way
interactions on seedling survival. All interactions Prob ChiSq > 0.25 were pooled with the
error term and the models rerun (Bancroft 1964).
L-R
Factors
Chi
Prob>Chisq
Square

Survival

Light
Non-Tree Vegetation
Habitat Type
Species
Light x Non-Tree Vegetation
Light x Species
Non-Tree Vegetation x Habitat
Rank
Habitat Type x Species
R2 = .09

71

1.86
12.71
3.12
8.98
13.14
13.87

0.1726
0.0004
0.0771
0.2539
0.0003
0.0536

10.34
11.93

0.0013
0.1029

Table 3.2- Results of a generalized linear mixed model for the effects of Scarification,
Light Availability, Habitat Type, Species, Seed Production Potential (SPP) and their two
way interactions on seedling establishment. All interactions Prob ChiSq > 0.25 were
pooled with the error term and the models rerun (Bancroft 1964).
L-R
Factors
Chi
Prob>Chisq
Square

Seedling Establishment

Scarification
Light
Habitat Type
SPP
Species
Scarification x Light
Scarification x SPP
Scarification x Species
Light x Habitat Type
Light x SPP
Light x Species
Habitat Type x SPP
Habitat Type x Species
SPP x Species
R2 = .50

72

2.45
0.02
1.8
2.14
192.29
3.46
3.18
25.25
9.88
4.64
17.67
4.88
237.7
223.56

0.1177
0.9025
0.1799
0.1436
<.0001
0.0628
0.0747
0.0007
0.0017
0.0312
0.0135
0.0272
<.0001
<.0001

Table 3.3- Species’ average seedling density (with standard deviation in parentheses) in
228 plots across 19 northern hardwood stands in the Northern Lower and Eastern Upper
Peninsulas of Michigan.
Species
Eastern Hemlock
Yellow Birch
Paper Birch
Red Maple
Sugar Maple
White Pine
Black Cherry
White Ash

Mean± SD
0.12 (0.7)
1.78 (7.78)
0.88 (3.48)
5.07 (10.43)
6.08 (19.46)
0.2 (.69)
2.16 (6.03)
5.26 (19.98)

73

Range
0-8
0-71
0-32
0-82
0-143
0-4
0-73
0-118

Table 3.4- Species’ average seedling density (with standard deviation in parentheses) in
114 scarified and unscarified plots across 19 northern hardwood stands in the Northern
Lower and Eastern Upper Peninsulas of Michigan. Asterisks represent statistically
significant effects (Prob ChiSq < 05).
Scarified

Mean ± SD

Range

Eastern Hemlock
Yellow Birch
Paper Birch *
Red Maple
Sugar Maple
White Pine*
Black Cherry
White Ash

0.12 ± 0.6
2.4 ± 8.32
1.5 ± 4.73
5.44 ± 12.2
7.03 ± 23.51
0.28 ± 0.88
1.6 ±3.17
4.27 ±16.43

(0-5)
(0-56)
0-32)
(0-82)
(0-143)
(0-4)
(0-21)
(0-103)

Mean

Range

0.11 ± 0.79
1.15 ± 7.17
0.25 ± 1.05
4.68 ± 8.33
5.12 ± 14.38
0.11 ± .39
2.7 ± 7.89
6.25 ± 2.16

(0-8)
(0-71)
(0-9)
(0-50)
(0-75)
(0-2)
(0-73)
(0-118)

Unscarified
Eastern Hemlock
Yellow Birch
Paper Birch
Red Maple
Sugar Maple
White Pine
Black Cherry
White Ash

74

Table 3.5- Mean average light availability and competing non-tree vegetation
characteristics (with standard deviation in parentheses) across high (AFOCa and AFOA),
medium (AFO and AFPo) and low (ParVVb and ATFD) quality habitat types (Burger and
Kotar 2004).

Stands
Gaps
Canopy Openness Range (%)
Canopy Openness (%)
Vegetation Cover (%)
Vegetation Height (cm)

High Quality

Medium Quality

Low Quality

5
30
5.73-23.42
14.37 (4.03)
51.16 (37.62)
28.63 (36.49)

7
42
7.91-26.66
16.7 (3.89)
37.14 (36.09)
8.84 (14.47)

7
42
6.45-50.32
21.58 (10.43)
28.39 (31.89)
8.98 (16.24)

75

LITERATURE CITED

76

LITERATURE CITED

Abrams, M. D., 1998. The red maple paradox. BioScience. 48, 355-364.
Amatangelo, K.L., Fulton, M.R., Rogers, D.A., Waller D.M., 2011. Converging forest
community composition along an edaphic gradient threatens landscape-level
diversity. Diversity Distrib. 17, 201-213.
Arbogast, C., 1957. Marking guides for northern hardwoods under the selection system.
USDA For. Serv. Stat. Paper LS-56, 1-20.
Bancroft, T.A., 1964. Analyses and inferences for incompletely specified models involving
the use of preliminary tests of significance. Biometrics. 20, 427-442.
Bazzaz, F.A., 1979. The physiological ecology of plant succession. Ann. Rev. Ecol. Syst.
10, 351-371.
Beckage, B., Clark, J.S., Clinton, B.D., Haines, B.L., 2000. A long-term study of tree
seedling recruitmenet in southern Appalachian forests: the effects of canopy gaps
and shrub understories. Can. J. For. Res. 30, 1617-1631.
Beckage, B., Clark, J.S., 2003. Seedling survival and growth of three forest tree species:
the role of spatial heterogeneity. Ecology. 84, 1849-1861.
Beguin, J., Pr÷vost, M., Pothier, D., C te, S.D., 2009. Establishment of natural
regeneration under severe browsing pressure from white-tailed deer after group
seed-tree cutting with scarification on Anticosti Island. Can. J. For. Res. 39, 596605.
Bolton, N.W., D’Amato, A.W., 2011. Regeneration responses to gap size and coarse
woody debris within natural disturbance-based silvicultural systems in northeastern
Minnesota, USA. For. Ecol. Manage. 262, 1215-1222.
Burger, T.L., Kotar, J., 2003. A guide to forest communities and habitat types of Michigan.
The Department of Forest Ecology and Management. University of WisconsinMadison, WI, USA.
Burns, R.M., Honkala, B.H., 1990. Silvics of North America, vol 1. Agriculture
Handbook, vol 654. USDA Forest Service Washington DC, USA.
Burton, D.H., Anderson, H.W., Riley, L.F., 1969. Natural regeneration of yellow birch in
Canada. In Proceedings, Birch Symposium. USDA Forest Service, Northeast Forest
Experiment Station, Upper Darby, PA. 55-73.

77

Burton, J.L., Zenner, E.K., Frelich, L.E., Cornett, M.W., 2009. Patterns of plant
community structure within and among primary and second-growth northern
hardwood forest stands. For. Ecol. Manage. 258, 2556-2568.
Canham, C.D., Denslow, J.S., Platt, W.J., Runkle, J.R., Spies, T.A., White, P.S., 1990.
Light regimes beneath closed canopies and tree-fall gaps in temperate and tropical
forests. Can. J. For. Res. 20, 620-631.
Caspersen, J.P., Saprunoff, M., 2005. Seedling recruitment in a northern temperate forest:
the relative importance of supply and establishment limitation. Can. J. For. Res. 35,
978-989.
Craine, J.M., Reich, P.B., 2005. Leaf-level light compensation points in shade-tolerant
woody seedlings. New Phytol. 166, 710-713.
Collis-George, N., Sands, J.E., 1959. The control of seed germination by moisture as a soil
physical property. Austr. J. Agricult. Res. 10, 628-636.
Crow, T.R., Buckley, D.S., Nauertz, E.A., Zasada, J.C., 2002. Effects of management on
the composition and structure of northern hardwood forests in Upper Michigan.
Forestry Sci. 48, 129-145.
Davis, M.A., Wrage, K.J., Reich, P.B., Tjoelker, M.G., Schaeffer, T., Muermann, C., 1999.
Survival, growth, and photosynthesis of tree seedlings competing with herbaceous
vegetation along a water-light-nitrogen gradient. Plant Ecology. 145, 341-350.
Eyre, F.H., Zillgitt, W.M., 1953. Partial cuttings in northern hardwoods of the Lake States:
Twenty-year Experimental Results. USDA Forest Service Technical Bulletin LSTB-1076, St. Paul, MN, USA.
Falk, K.J., Elliott, K.A., Burke, D.M., Nol, E., 2010. Early seedling response to group
selection harvesting in a northern hardwood forest. Forestry Chron. 86, 100-109.
Frazer, G.W., Canham, C.D., Lertzman, K.P., 1999. Gap Light Analyzer (GLA) Version
2.0 Users Manual and Program Documentation. Simon Fraser University, Burnaby,
British Columbia, Canada and the Institute of Ecosystem Studies, NY, USA.
Frelich, L.E., 1995. Old forest in the Lake States today and before European settlement.
Nat. Areas. J. 15, 157-167.
Frey, B.R., Ashton, M.S., McKenna, J.J., Ellum, D., Finkral, A., 2007. Topographic and
temporal patterns in tree seedling establishment, growth, and survival among
masting species of southern New England mixed-deciduous forest. For. Ecol.
Manage. 245, 54-63.
Friedman, S.K., Reich, P.B., 2005. Regional legacies of logging: departure from
presettlement forest conditions in northern Minnesota. Ecol. Appl. 15, 726-744.

78

Gasser, D., Messier, C., Beaudet, M., Lechowicz, M.J., 2010. Sugar maple and yellow
birch regeneration in response to canopy opening, liming, and vegetation control in
a temperate deciduous forest of Quebec. For. Ecol. Manage. 259, 2006-2014.
George, L.O., Bazzaz, F.A., 1999. The fern understory as an ecological filter: emergence
and establishment of canopy-tree seedlings. Ecology. 80, 833-845.
Gray, A.N., Spies, T.A., Easter, M.J., 2002. Microclimatic and soil moisture responses to
gap formation in coastal Douglas-fir forests. Can. J. For. Res. 32, 332-343.
Goodman, R.M., Krefting, L.W., 1960. Factors important to yellow birch establishment in
Upper Michigan. Ecology. 41, 18-28.
Greene, D.F., Johnson, E.A., Dispersal of tree seeds on snow. Ecology. 85, 329-340.
Heller, N.E., Zavaleta, E.S., 2009. Biodiversity management in the face of climate change:
A review of 22 years of recommendations. Biological Conservation. 142, 14-32.
Host, G.E., Pregitzer, K.S., Ramm, C.W., Hart, J.B., Cleland, D.T., 1987. Landformmediated differences in successional pathways among upland forest ecosystems in
Northwestern Lower Michigan. For. Sci. 13, 445-457.
Hughes, J.W., Fahey, T.J., 1988. Seed dispersal and colonization in a disturbed northern
hardwood forest. Bulletin of the Torrey Botanical Club. 115, 89-99.
Hura, C.A., Crow, T.R., 2004. Woody debris as a component of ecological diversity in
thinned and unthinned northern hardwood forests. Nat. Ar. J. 24, 57-64.
Kern, C.C., Reich, P.B., Montgomery, R.A., Strong, T.F., 2012. Do deer and shrubs
override canopy gap size effects on growth and survival of yellow birch, northern
red oak, eastern white pine, and eastern hemlock seedlings? For. Ecol. Manage.
267, 134-143.
Kern, C.C., D’Amato, A.W., Strong, T.F., 2013. Diversifying the composition and
structure of managed, late-succesional forests with harvest gaps: what is the
optimal gap size? For. Ecol. Manage. 304, 110-120.
Kidson, R., Westoby, M., 2000. Seed mass and seedling dimensions in relation to seedling
establishment. Oecologia. 125, 11-17.
Klingsporn, S., Webster, C.R., Bump, J.K., 2012. Influence of legacy-tree retention on
group-selection opening persistence. For. Ecol. Manage. 286, 121-128.
Klingsporn-Poznanovic, S., Webster, C.R., Bump, J.K., 2013. Maintaining mid-tolerant
tree species with uneven-aged forest management: 9 year results from a novel
group-selection experiment. Forestry 86, 555-567.
Kobe, R.K., Pacala, S.W., Silander, J.A., Canham, C.D., 1995. Juvenile tree survivorship
as a component of shade tolerance. Ecological Applications. 5, 517-532.

79

Kobe, R.K., Hogarth, L., 2007. Evaluation of irradiance metrics with respect to predicting
sapling growth. Can. J. For. Res. 37, 1203-1213.
Lindenmayer, D.B., Margules, C.R., Botkin, D.B., 2000. Indicators of Biodiversity for
Ecologically sustainable forest management. Conservation Biology. 14, 941-950.
Lindenmayer, D.B., Franklin, J.F., Fischer, J., 2006. General management principles and a
checklist of strategies to guide forest biodiversity conservation. Biological
Conservation. 131, 433-445.
Lorenzetti, F., Delagrange, S., Bouffard, D., Nolet, P., 2008. Establishment, survivorship
and growth of yellow birch seedlings after site preparation treatments in large gaps.
For. Ecol. Manage. 254, 350-361.
Marks, P.L., 1974. The role of pin cherry (Prunus pensylvanica L.) in the maintenance of
stability in northern hardwood ecosystems. Ecological Mongraphs. 44, 73-88.
Marquis, D.A., 1975. Seed storage and germination under northern hardwood forests. Can.
J. For. Res. 5, 478-484.
Matlack, G.R., 1989. Secondary dispersal of seed across snow in Betula Lenta, a gapcolonizing tree species. Ecology. 77, 853-869.
Marx, L., Walters, M.B., 2008. Survival of tree seedlings on different species of decaying
wood maintains tree distributions in Michigan hemlock-hardwood forests. Ecology
96, 505-513.
Matonis, M.S., Walters, M.B., Millington, J.D.A, 2011. Gap-,stand,-and landscape-scale
factors contribute to poor sugar maple regeneration after timber harvest. For. Ecol.
Manage. 262, 286-298.
McEuen, A.B., Curran, L.M., 2004. Seed dispersal and recruitment limitation across
spatial scales in tempereate forest fragments. Ecology. 85, 507-518.
Messier, C., Doucet, R., Ruel, J.C., Claveau, Y., Kelly, C., Lechowicz, M.J., 1999.
Functional ecology of advance regeneration in relation to light in boreal forests.
Can. J. For. Res. 29, 812-823.
Moles, A.T., Westoby, M., 2004. Seedling survival and seed size: a synthesis of the
literature. Ecology. 92, 372-383.
Montgomery, R.A., Reich, P.B., Palik, B.J., 2010. Untangling positive and negative biotic
interactions: views from above and below ground in a forest ecosystem. Ecology.
91, 3641-3655.
Neuendorff, J.K., Nagel, L.M., Webster, C.R., Janowiak, M.K., 2007. Stand structure and
composition in a northern hardwood forest after 40 years of single-tree selection.
N.J. Appl. For. 24, 197-202.

80

Nilsen, E.T., Clinton, B.D., Lei, T.T., Miller, O.K., Semones, S.W., Walker, J.F., 2001.
Does Rhododendron maximum L. (Ericaceae) reduce the availability of resources
above and belowground for canopy tree seedlings? AM. Midl. Nat. 145, 325-343.
Nobis, M., Hunziker, U., 2005. Automatic thresholding for hemisphericald canopyphotographs based on edge detection. Agricultural and Forest Meteorology. 128,
243-250.
Ni, B.R., Pallardy, S.G., 1992. Stomatal and nonstomatal limitations to net photosynthesis
in seedlings of woody angiosperms. Plant Physiol. 99, 1502-1508.
O’Hara, K., 2002. The historical development of uneven-aged silviculture in North
America. Forestry. 75, 339-345.
Oleskog, G., Sahlèn, K., 2000. Effects of seedbed substrate on moisture conditions and
germination of Scotts pine (Pinus sylvestris) seeds in a mixed conifer stand. New
Forests. 20, 119-133.
Pairon, M., Jonard, M., Jacquemart, A.L., 2006. Model seed dispersal of black cherry, an
invasive forest tree: how microsatellites may help? Can. J. For. Res. 36, 1385-1394.
Phillips, D.L., Shure, D.J., 1990. Patch-size effects on early succession in southern
Appalachian forests. Ecology. 71, 204-212.
Prescott, C.E., Maynard, D.G., Laiho, R., 2000. Hums in northern forests: friend or foe?
For. Ecol. Manage. 133, 23-36.
Prèvost, M., Raymond, P., Lussier, J.M., 2010. Regeneration dynamics after patch cutting
and scarification in yellow birch –conifer stands. Can. J. For. Res. 40, 357-369.
Ramsey, F., Schafer, D. The Statistical Sleuth: A course in Methods of Data Analysis.
Belmont, CA: Duxbury Press, 1997.
Raymond, P., Munson, A.D., Ruel, J.C., Nolet, P., 2003. Group and single-tree selection
cutting in mixed tolerant hardwood-white pine stands: Early establishment
dynamics of white pine and associated species. Forestry. Chron. 79, 1093-1106.
Rhemtulla, J.M., Mladenoff, D.J., Clayton, M.K., 2009. Legacies of historical land use on
regional forest composition and structure in Wisconsin, USA (mid-1800s-1930’s2000s) Ecological Applications. 19, 1061-1078.
Ribbons, E., Silander, J.A., Pacala, S.W., 1994. Seedling recruitment in forests: Calibrating
models to predict patterns of tree seedling dispersion. Ecology. 75, 1794-1806.
Royo, A.A., Carson, W.P., 2006. On the formation of dense understory layers in forests
worldwide: consequences and implications for forest dynamics, biodiversity, and
succession. Can. J. For. Res. 36, 1345-1362.

81

Royo, A.A., Carson, W.P., 2008. Direct and indirect effects of a dense understory on
seedling recruitment in temperate forests: habitat-mediated predation versus
competition. Can. J. For. Res. 38, 1634-1645.
Runkle, J.R., 1981. Gap regeneration in some old-growth forests of the Eastern United
States. Ecology. 62, 1041-1051.
Schreeg, L.A., Kobe, R.K., Walters, M.B., 2005. Tree seedling growth, survival, and
morphology in response to landscape-level variation in soil resource availability in
northern Michigan. Can. J. For. Res. 35, 263-273.
Schulte, L.A., Mladenoff, D.J., Nordheim, E.V., 2002. Quantitative classification of a
historic northern Wisconsin (U.S.A.) landscape: mapping forests at regional scales.
Can. J. For. Res. 32, 1618-1638.
Schulte, L.A., Mladenoff, D.J., Crow, T.R., Merrick, L.C., Cleland, D.T., 2007
Homogenization of northern U.S. Great Lakes forests due to land use. Landscape
Ecol. 22, 1089-1103.
Schwartz, J.W., Nagel, L.M., Webster, C.R., 2005. Effects of uneven-aged management on
diameter distribution and species composition of northern hardwoods in Upper
Michigan. For. Ecol. Manage. 211, 356-370.
Shields, J.M., Webster, C.R., Nagel, L.M., 2007. Factors influencing tree species diversity
and Betula alleghaniensis establishment in silvicultural openings. Forestry 80,
293-307.
St. Clair, S.B., Sharpe, W.E., Lynch, J.P., 2008. Key interactions between nutrient
limitation and climatic factors in temperate forests: a synthesis of the sugar maple
literature. Can. J. For. Res. 38, 401-414.
Walters, M.B., Reich, P.B., 1997. Growth of Acer saccharum seedlings in deeply shaded
understories of northern Wisconsin: Effects of nitrogen and water availability. Can.
J. For. Res. 27, 237-247.
Walters, M.B., Reich, P.B., 2000. Trade-offs in low-light CO2 exchange: a component of
variation in shade tolerance among cold temperate tree seedlings. Funct. Ecol. 14,
155-165.
Walters, M.B., Willis, J.L., Gottschalk, K.W., 2014. Shade tolerance, nutrient affinity, and
phylogeny predict tree seedling growth responses to light and mineral N form in
mesic northern hardwood forest harvest gaps. Can. J. For. Res. (In press).
Webster, C.R., Lorimer, C.G., 2005. Minimum opening sizes for canopy recruitment of
midtolerant tree species: a retrospective approach. Ecol. Appl. 15, 1245-1262.
Webster, C.R., Jensen, N.R., 2007. A shift in the gap dynamics of Betula alleghaniensis in
response to single-tree selection. Can. J. For. Res. 37, 682-689.

82

Whitney, C.G., 1987. An ecological history of the Great Lakes forest of Michigan.
Ecology. 75, 667-684.
Woods, K.D., 2000. Dynamics in late-succsional hemlock-hardwood forests over three
decades. Ecology. 81, 110-126.
Valladares, F., Niinemets, Ü., 2008. Shade tolerance, a key plant feature of complex nature
and consequences. Annu. Rev. Ecol. Evol. Syst. 39, 237-257.
Zhang, Q.F., Pregitzer, K.S., Reed, D.D., 2000. Historical changes in the forests of the
Luce District of the Upper Peninsula of Michigan. Am. Midl. Nat. 143, 94-110.

83

Chapter IV
The effect of substrate type and mycorrhizal fungi on seedling development for northern
temperate seedlings

Abstract
Gaining access to light and soil resources is a challenging task for seedlings on the
forest floor. Due to their high moisture holding capacity and elevated position off the forest
floor, decaying coarse woody debris (CWD) is thought to provide seedlings a favorable
environment to establish and develop. Nevertheless, species of CWD have been shown to
differ in their ability to support seedling development, and seedlings have been shown to
differ in their ability to develop on individual species of CWD. This study investigates
mycorrhizal fungi as a potential explanation for this pattern in a potted pot experiment
where 11 tree species (yellow birch, paper birch, red maple, white ash, eastern hemlock,
balsam fir, white spruce, northern white cedar, white pine, and red oak) were established
on seven different types of substrate (bare mineral soil, yellow birch, sugar maple, paper
birch, eastern hemlock, northern white cedar, and balsam fir CWD) that had either been
sterilized or left unsterilized. In general, species grew better on bare mineral soil, paper
birch, eastern hemlock, and northern white cedar CWD compared to sugar maple, yellow
birch, and balsam fir CWD. Seedling survivorship was high for all conifer species (>90%),
but varied widely among deciduous species with yellow birch (89%) surviving at the
highest rate, and white ash (28%) surviving at the lowest rate. Mycorrhizal colonization
had a positive effect on red maple, northern white cedar, white pine, and yellow birch’s
height growth, and was positively associated with large changes in species’ performance

84

rankings on individual species of CWD. Collectively, this research demonstrates that
individual species of CWD differ strongly in their ability to support seedling development,
and thus should be considered at the species level, rather than a generic substrate on the
forest floor. In addition, it identifies mycorrhizal colonization as an important factor
contributing to seedling development on CWD.

85

Introduction
The transition between seedling and sapling is one of the most uncertain stages in
the natural regeneration process. In order to pass through this vulnerable period, seedlings
must gain access to adequate amounts of light and soil resources needed for constructing
additional structural and resource acquiring biomass. For young seedlings, acquiring
adequate resources may be particularly challenging as they often lack the necessary root
system and or canopy area needed to exploit resources beyond their immediate proximity.
As such, the resource environment of a seedlings initial establishment site may play an
important role in determining whether a seedling successfully develops into a sapling.
Heavily decayed coarse woody debris (hereafter referred to as CWD) is one type of
establishment site which may be particularly conducive to seedling development given its
easily penetrated surface and high moisture holding capacity. (Harmon 1986; Cornett et al.
2001; Caspersen and Saprunoff 2005; Shields et al. 2007; Marx and Walters 2008; Bolton
and D’Amato 2011). In addition to its role as a moisture source, CWD may increase
seedling survival through elevating seedlings off the forest floor and thus providing refuge
from many of the factors which limit seedling development there; i.e. competition from
established vegetation, leaf litter smothering, high water tables, and potentially any species
specific pests or pathogens which may be more abundant on the forest floor/mineral soil
(Harmon and Franklin 1989;Packer and Clay 2000 &2003; Simard et al. 2003; O’HanlonManners and Kotanen 2004; Wang and Kembell 2005). Indeed, CWD has been
documented as an important resource for seedling establishment/development in forests
throughout the world (Harmon and Franklin 1989; McGee and Birmingham 1997; Simard

86

et al. 2003; Mori et al. 2004; Marx and Walters 2008; Sanchez et al. 2009). Nevertheless,
compelling evidence also exists indicating that, despite all its beneficial properties, species
differ in the capability for CWD to support seedling development, and that not all tree
species are capable of long term survival on CWD (Marx and Walters 2006; Marx and
Walters 2008). Consequently, management efforts seeking to promote old-growth
structural characteristics by creating/maintaining CWD on the forest floor may need to
focus more intently on conserving or restoring individual species CWD rather than
considering it as a generic entity if maintaining/restoring ecosystem function is the primary
objective.
Variation in CWD’s ability to supply seedlings with nutrients is one potential
explanation for the conflicting patterns of seedling development, as species have been
shown to differ in their decay rates, nutrient ratios, nutrient concentrations, and rate at
which nutrients become mineralized for seedling consumption (Arthur et al. 1993;
Takahashi et al. 2000; Marx and Walters 2006; Strukelj 2013; Shorohova and Kapitsa
2014; Klockow et al. 2014). As such, certain species of CWD may simply be more capable
of supporting development for all seedlings, no seedlings or only certain seedlings based
on their nutrient environments and the nutrient demands/growth strategies of individual
seedlings (Grime 1977; Chapin 1980). If true, seedling growth/survival on CWD may
proceed in a predictable manner where fast growing nutrient demanding strategies
demonstrate the greatest growth on nutrient rich CWD, but also experience the lowest
survival and growth on nutrient poor CWD, reflecting a potential trade-off between
maximizing growth potential versus stress tolerance (Grime; 1977; Chapin 1980; Aerts
1999). In partial support of this hypothesis, Marx and Walters (2006 & 2008) found that

87

fast growing yellow birch survived and grew best on eastern hemlock (Tsuga canadensis
L.) CWD, which also had the highest nitrogen mineralization rates compared to sugar
maple (Acer saccharum Marsh.) and yellow birch (Betula alleghaniensis Britt.) CWD.
Nevertheless, the trends developed by Marx and Walters (2006 & 2008) are based on the
performance of only a small range of species’ seedlings growing on a small range of CWD
species, and thus provide little generality and only limited evidence for the existence of a
nutrient induced trade-off response.
Differences among seedlings in their ability to access nutrients via a symbiotic
relationship with mycorrhizal fungi may also account for observed differences in seedling
growth on CWD independent of differences in measured mineral N dynamics. In
exchange for carbon from the seedling, mycorrhizal fungi improve seedling access to
nutrients by increasing the substrate volume in which a seedling can forage for nutrients
and by breaking down previously insoluble compounds (Smith and Read 1996; PerezMoreno and Read 2000). Furthermore, mycorrhizal fungi are frequently found within
CWD, suggesting that species’ colonization is not limited by mycorrhizal presence
(Harvey et al. 1979; Amaranthus and Perry 1994; Tedersoo et al. 2008). Provided this
nutrient acquisition advantage, colonized seedlings may grow faster and survive at higher
rates than uncolonized seedlings (Marx and Walters 2006). This, in turn, could cause
species to grow and survive at rates that differ from those which would be predicted by a
nutrient trade-off alone. Indeed, Marx and Walters (2006) also reported seedlings growing
on CWD achieved greater mass when colonized with mycorrhizal fungi compared with
uncolonized seedlings, but did not observe rank changes in performance.

88

One factor which may determine the importance of seedlings becoming colonized
is the type of mycorrhizal fungus individual tree species associate with. Due to their
limited saprophytic capabilities, arbuscular mycorrhizae (AMF) may be less abundant in
CWD than ectomycorrhizal fungi (EMF), which are saprophytic (Treseder et al. 2005;
Smith and Smith 2011). Species which associate with EMF may also receive a greater
increase in nutrient availability. Via their production of oxidative extracellular enzymes,
EMF associated species can access nitrogen and phosphorous from a variety of pools
which are inaccessible to AMF fungi (Turner 2008; Courty et al. 2010; Phillips et al.
2013). Consequently, seedlings associated with EMF may have a higher probability of
becoming colonized and may also gain a greater functional advantage as a result.
In this shade house potted plant experiment, we investigate the effect of nutrient
availability and mycorrhizal colonization on 11 tree species established across seven
substrates (six species of CWD and soil) that have either have or have not been sterilized to
kill naturally-occurring mycorrhizae. This approach allows us to not only develop a
broader pattern of seedling development across different seedling establishment sites, but
also allow us to experimentally separate the effects of nutrient availability and mycorrhizal
colonization on seedling development. Specifically we predict: 1) seedling height growth
will reflect mineral nitrogen availability on sterilized substrates with fast growing tree
species demonstrating the greatest growth on nitrogen rich substrates 2) height growth will
depend on mineral nitrogen availability and mycorrhizal association type on unsterilized
substrates, with EMF colonized seedlings demonstrating greater growth than noncolonized seedlings at a given level of nitrogen availability, leading to possible rank
changes in species growth between sterilized and unsterilized substrates.

89

Methods

Field methods
In summer 2012, CWD of sugar maple, yellow birch, paper birch (B. papyrifera
Marsh), eastern hemlock, northern white cedar (Thuja occidentalis L.), and balsam fir
(Abies balsamifera (L.) Mill.) were located in four locations in the northern Lower and
eastern Upper Peninsula’s of Michigan. The four locations included an old growth
hemlock hardwood dominated forest preserve (The Huron Mountain Club), two managed
lowland conifer swamps and a sugar maple dominated managed upland hardwood stand
(Michigan Department of Natural Resources. All CWD was in decay stage 3 or 4 (Graham
and Cromack 1982) and was identified to species by bark and branching pattern. Samples
of CWD were bagged and transported to the Tree Research Center at Michigan State
University for temporary storage (4ºC). With the exception of cedar, for which only 11
logs could be located, at least 16 different logs of each species were sampled. In addition
to CWD, soil was also obtained from each site using soil cores (20cm).

Shade house methods
In late May 2013, CWD and soil samples were divided into two equally sized
populations (by species for CWD) with one being randomly selected for a gamma
irradiation sterilization treatment (30-60kGy) (Sterigenics Inc., Schaumburg, IL,USA).
Sterilized and unsterilized substrate was then embedded in pots (Stuewe and Sons Inc.,
Corvallis Oregon) with sterilized coarse silica sand (Best Sand Inc., Chardon Ohio) (2/3

90

substrate). Sand sterilization was accomplished by drying the sand at 110 ºC for 72 hours.
Each pot was then seeded with sterilized pre-stratified seeds of sugar maple, red maple (A.

rubrum L.), northern white cedar, eastern hemlock, yellow birch, paper birch, white pine
(Pinus strobus L.), white ash (Fraxinus americana L.), red oak (Quercus rubra L.), white
spruce (Picea glauca (Moench)), and American elm (Ulmus americana L. ). Seed
sterilization was accomplished by submerging seeds in a water bleach mixture (10%
bleach) for one minute. All seeds were obtained from the Michigan Department of Natural
Resources (MDNR) or Sheffield’s Seed Company Inc, (Locke, New York, USA).
Unfortunately, germination was poor for sugar maple, American elm, white ash, and red
maple. To overcome this issue we transplanted newly germinated wildings of red maple
and white ash seedlings into pots and removed sugar maple from the experiment. Prior to
transplant the root systems of each seedling was sterilized in 10% bleach. In addition to the
planted/seeded plots, three pots of each unique substrate treatment combination were left
empty for inorganic nitrogen extractions (methods described in forthcoming laboratory
methods).
The experiment itself was located in three hoop houses built within a lath house
(~50% shade). Each hoop house was covered with clear plastic to exclude rain water. To
quantify the light environment we measured photosynthetically active radiation (PAR)
within each house with a quantum sensor (Apogee Instruments Inc., Logan Utah). PAR
ranged from 254-344 µmol m-2s-1 across houses under cloudless mid-morning conditions.
To mitigate any potential issues associated with differences in PAR, we spread all unique
treatment combinations equally across houses. Seedlings were also rotated across houses to
further mitigate environmental differences. In addition, all pots were kept at least five

91

meters from the edge of each house to minimize edge effects. Throughout the experiment,
seedlings were watered as needed with deionized water (DI). DI was chosen to minimize
the addition of external nutrients to the pots.
Initial heights (cm) of germinated/planted seedlings were taken after planting in
early July 2013. Seedling growth was then reassessed and survival checked at the end of
September. At this time, four replicates of yellow birch, paper birch and red oak and three
replicates of each species except sugar maple and red pine (due to low sample size) were
randomly selected from each unique treatment substrate combination and destructively
harvested to examine mycorrhizal colonization and biomass development (methods
described in forthcoming laboratory methods). This was done to preserve our ability to
examine mycorrhizal colonization as a potential mechanism, and to accurately depict
biomass development in case high mortality occurred over winter. Harvest intensity was
greatest for birch and oak because these species had already shown large differentiation in
height growth, thus providing us with the most data to investigate the effects of nutrients
and mycorrhizae on 1st year growth for these species. Fewer replicates of the other species
were harvested in order to preserve our ability to detect future (2nd year) changes in growth
and survival, as seedlings had yet to differentiate by the end of the first growing season. By
this same logic, we decided against investigating 1st year mycorrhizal colonization for this
latter group of species.
Harvested seedlings were placed in plastic bags and kept frozen until final
processing for mycorrhizae. The remaining seedlings were overwintered in the hoop
houses and allowed to grow until the end of June 2014. At that time, all seedlings were
assessed for survival, height, and total mass. Seedlings dead at the time of final harvest

92

were not collected for further analysis. At that time we also selected four species for
mycorrhizal analysis (yellow birch, white pine, red maple, and northern white cedar).
Selection criteria was based on observed patterns of height development, type of
mycorrhizal association (AMF vs EMF) and growth strategies (ranging from rapidly
growing nutrient demanding to slow growing nutrient conserving). Four replicates of each
unique unsterilized substrate combination and one replicate of each unique sterilized
substrate combination were randomly selected from this group of species for colonization
analysis.

Laboratory methods
To analyze substrate inorganic nitrogen (NH4+-N, and NO3—N) concentration we
obtained two 10g samples from each of the three unplanted pots from each unique
sterilization substrate combination. On each sample, we determined gravimetric moisture
content by drying the first sample at 105º C in a container of a known volume. The second
sample was extracted with 50mL of 2 M KCL and analyzed colorimetrically for NH4+N,
and NO3-N content. Following the logic of Marx and Walters (2006) we chose to express
the initial [N] content on a volume basis based on the fact that soil is denser than wood and
because roots forage a given volume of substrate.
For each seedling selected for mycorrhizal colonization analysis, we first carefully
removed all residual CWD and/or soil from the root system by floating them in water.
Fine roots (<2mm) were then randomly harvested from the entire root system for
colonization analysis. Before analysis, all fine roots were first cleared with KOH and
stained with black ink (Sheaffer Incorporated, Ft. Madison, Iowa) to identify AMF

93

colonization. Roots were then mounted on slides and examined for colonization beneath a
compound microscope. AMF colonization was investigated at intersections between an
inserted hairline graticule and the root at 200x magnification, while EMF colonization was
examined at individual root tips at 100x magnification. At least 50 root intersections/tips
were examined for each seedling. Both EMF and AMF colonization was examined for
each seedling as several species included in this study have been shown to support both
AMF and EMF colonization (Personnel Communication Sarah Newman).
To quantify biomass development, each seedling was separated into above (leaves
and stems) and below ground biomass (roots) and dried for 48 hours at 65 ºC. Seedlings
were then weighed to determine the dry weight biomass of each section. This work is
currently ongoing and will be reported in a future publication of this study. In addition,
foliar nutrient analysis will also be performed and reported in a future publication.

Statistical methods
We examined the effects of Species, Sterilization and Substrate on height growth
using a one-way ANOVA. In order to standardize our analysis, all seedlings that died or
were harvested prior to the final harvest in June 2014 were removed from consideration.
We also removed red pine, white ash, and red oak from the analysis due to poor
representation across all substrates/treatments. For the remaining species, final height
growth measurements were log-transformed to meet the assumption of normality prior to
the analysis. Model selection was accomplished through backwards elimination.
Interaction terms exceeding the suggested threshold for pooling variances (F > .25) were
removed from the model and pooled in the error term (Bancroft, 1964). The model was

94

then re-run until a final model containing only significant factors (F<.05) and interactions
(F<0.25) could be reached. Once we arrived at a final model, significant differences
among means were explored using Tukey-Kramer Honestly Significant Difference Test
(HSD). This same statistical approach was also used to investigate the effects of Substrate,
Species and their interaction on mycorrhizal colonization for the four species in which we
investigated root colonization in 2014 (red maple, yellow birch, northern white cedar,
white pine). To gain further insight into mycorrhizae’s influence (EMF and AMF
separately) on seedling development, we examined species height growth vs. colonization
rates using linear regression. Finally, to gauge Sterilization’s effect on species performance
ranks, we compared the mean height growth of individual species growing on Sterilized vs.
Unsterilized substrate on each type of Substrate. Spearman’s Rho correlations were used
to determine height growth rank changes between sterilized and unsterilized substrate,
where lower correlation values corresponded to larger changes in ranks. Finally, we
examined the influence of Species, Sterilization and Substrate on seedling survivorship
using nominal logistic regression models. Model selection was accomplished using the
same procedures described for seedling height. Wildlings that died within the first 6 weeks
following transplant were removed from the analysis. Red pine was not considered in this
analysis due to low sample size. All statistical analysis was performed in JMP 9.0 (SAS
Institute, Cary, NC). All graphics were produced in Sigma Plot version 11 (Systat Software
Incorporated, San Jose, California).

95

Results

Height growth
Height growth varied significantly among Species (Table 4.1). In general, fast
growing shade intolerant to mid-tolerant species (paper birch, yellow birch, white pine and
red maple) outgrew the more shade tolerant conifer species (Figure 4.1). In addition to
differing in height growth, Species also responded differently to the combination of
Sterilization and Substrate treatments (Tables 4.1 & 4.2). Independent of Sterilization
effects, Substrate type was the strongest factor affecting balsam fir, eastern hemlock, paper
birch and white pine, as each species grew best on bare mineral soil (Figure 4.1) (Table
4.2). Paper birch and northern white cedar CWD also consistently supported greater height
growth for this group of species compared to sugar maple, balsam fir and yellow birch
CWD (Figure 4.1). For red maple, northern white cedar, yellow birch and white spruce,
height growth depended significantly on the interaction of Substrate Type and Sterilization
(Table 4.2). Sterilization had strong negative effects for red maple growth on all substrates
except northern white cedar and balsam fir (Figure 4.2). In a similar fashion, Sterilization
had strong negative effects on northern white cedar’s growth across all substrates except
northern white cedar CWD (Figure 4.2). Sterilization also produced significantly less
growth on mineral soil for white spruce (Figure 4.2). Yellow birch demonstrated a mixed
response to Sterilization as its growth was negatively affected on eastern hemlock CWD
but strongly positively affected on balsam fir CWD (Figure 4.2).

96

Seedling survival
Species differed significantly in their survival (Table 4.3). In general, shade
tolerant to mid-tolerant conifer species experienced the greatest survival, as balsam fir,
hemlock, white pine, northern white cedar and white spruce survival equaled or exceeded
90% (Table 4.4). In contrast, among the deciduous species, only yellow birch and paper
birch survived at greater than 75% (Table 4.4). The lowest survival occurred in red oak and
white ash which each survived at 30 and 18%, respectively (Table 4.4). Nevertheless, no
factor or interaction significantly predicted these patterns (Table 4.3).

Mycorrhizal colonization pattern and effects
Root colonization was strongly influenced by Substrate and Species by Substrate
interactions (Table 4.5). With the exception of white pine, which was colonized most
frequently on cedar CWD, root colonization was highest on mineral soil (Figure 4.3). For
primarily AMF associated species (red maple and northern white cedar), mycorrhizal
colonization was also consistently higher on balsam fir and hemlock CWD than on any
other species of CWD (Figure 4.3). For primarily EMF associated species (yellow birch
and white pine), colonization was consistently higher on mineral soil and northern white
cedar CWD than any other substrate (Figure 4.3). Across species, increasing root
colonization was positively associated with increasing height growth, with northern white
cedar showing the strongest correlation between colonization and height growth (Figure
4.4).

97

Sterilization’s effect on performance rankings
Sterilization’s influence on species’ height growth performance rankings varied
across substrates (Table 6). Strong rank changes were observed on balsam fir CWD, as
shade mid-tolerant EMF associated deciduous species dominated the Sterilized treatment,
while primarily shade mid-tolerant to tolerant AMF associated conifer species gained
dominance in the Unsterilized treatment (Figure 4.5). In contrast, Sterilization had very
little influence on species’ height growth on northern white cedar CWD, as shade
intolerant to mid-tolerant EMF associated species dominated both treatments (Figure 4.6).
For each of the five remaining substrates, Sterilization led to only modest changes in
height growth rankings (Table 4.6).

Discussion

Height growth
While we are currently limited in our abilities to identify mechanisms, our results
demonstrate that different types of forest substrate have a strong effect on species height
growth. Among species, height growth was generally better on mineral soil, and
significantly so for balsam fir, eastern hemlock, paper birch and white pine (Figure 4.1).
This finding is not surprising, considering that soil has been shown to support higher rates
of nitrogen mineralization (Nmin) and inorganic nitrogen concentrations [N] compared to
yellow birch, eastern hemlock and sugar maple CWD (Marx and Walters 2006). In
addition, by removing seedlings from several of the factors which influence seedling

98

access to resources on the forest floor our experimental design likely enhanced a nutrient
influence effect.
Among species of CWD, seedling height growth was also generally higher on
northern white cedar, paper birch and hemlock compared to balsam fir, yellow birch and
sugar maple (Figure 4.1). This finding again concurs with the growth trends reported by
Marx and Walters (2006), and strengthens the notion that individual species of CWD vary
in their availability to support seedlings, and thus should be considered individually rather
than a generic forest floor resource. In addition, these findings also align with the
observations of Bolton and D’Amato (2011) and Cornett et al. (2001) who reported
northern white cedar and paper birch as important substrates for seedling establishment.
Contrary to our initial hypothesis, mycorrhizal colonization appears to be an
important factor for red maple and northern white cedar (AMF associated species).
Evidence supporting this conclusion includes both species’ generally negative response to
sterilization (Figure 4.2) and significantly positive response to increasing AMF %
colonization (Figure 4.4). In addition, cedar also grew best on the substrate where it was
colonized most often (mineral soil) (Figure 4.3) and experienced a dramatic height growth
decline (2nd best to worst) in response to sterilization on balsam fir (Figure 4.5), where it
was also frequently colonized (Figure 4.3). The one exception to this pattern for northern
white cedar came when it was growing on its own CWD, where growth increased in
response to sterilization (Figure 4.2). While, at this point, we cannot dismiss the possibility
that this pattern was driven by a treatment related increase in available nutrients, it is also
possible that cedar growth is limited on its own CWD by specialized host specific
pathogens. Nevertheless, it should also be pointed out that northern white cedar’s

99

mycorrhizal colonization rate was lowest on cedar CWD, thus a lack of mycorrhizal fungi
in cedar CWD may also be driving this pattern (Figure 4.3).
In addition to red maple and northern white cedar, mycorrhizal colonization was
also important for yellow birch. Much like northern white cedar, yellow birch grew best on
the two substrates where it was most likely to be colonized (mineral soil and cedar CWD)
(Figures 4.2 & 4.3), and responded positively to increasing EMF% colonization.
Moreover, supporting the findings of Marx and Walters (2006 & 2008) and our initial
hypothesis, yellow birch grew significantly better on unsterilized hemlock CWD, where it
was third most likely to be colonized (Figures 4.2 & 4.3). Mycorrhizal colonization was
potentially so important for yellow birch development that sterilization caused a shift in
ranks from best growing species on unsterilized hemlock CWD to second best on sterilized
hemlock CWD (Figure 4.7). Despite the fact that it generally forms a different type of
association (AMF vs. EMF), red maple experienced a similar dramatic shift, where it went
from the second best growing species on sterilized hemlock CWD to fifth best on the
sterilized treatment (Figure 4.7). Yellow birch’s response to sterilization, however, was not
consistent across substrates, as height growth was unaffected by sterilization on cedar
CWD, where it remained dominant (Figure 4.6), and was strongly positively affected by
sterilization on balsam fir CWD, where it went from the species that grew least in the
unsterilized treatment to species that grew best in sterilized treatment (Figure 4.5). While
we are still unsure as to the exact mechanism for this dramatic shift on balsam fir, a lack of
mycorrhizal fungi is less likely to be the cause, as colonization on balsam fir were neither
high nor low for yellow birch (Figure 4.3). Collectively, these findings demonstrate that

100

mycorrhizal fungi may not be as important on all substrates, and may be strongly differing
effects on different species of CWD.

Seedling survival
Large differences occurred in individual species survival (Table 4.3).
Corresponding with their slow growing nutrient conserving growth strategy, each conifer
species experienced high survival (Grime 1977; Chapin 1980). Consistent with our
observations of seedling demographics on CWD in the field, red maple survival was low
compared to the majority of other species (Table 4.3). While, not surprising to us, this
result does contradict red maple’s reputation in the literature as a “super generalist” which
can tolerate a variety of environmental conditions (Abrams 1998). Nevertheless, it is hard
to say whether red maple’s low survival was related to it being a largely transplanted
species, as the majority of seedlings survived through the first growing season, but failed
to emerge the following spring. White ash, our other transplanted species, followed a
similar pattern of survival. Again, however, this result is consistent with our personal
observations of white ash rarely surviving on CWD in the field. Red oak followed a
similar pattern of high initial survival followed by low spring emergence. This result was
surprising, however, considering that red oak was not transplanted, has a moderately
conservative growth strategy (Crow 1988; Kolb et al. 1990; Kaelke et al. 2001) and was
not shaded at any point during this experiment. Interestingly, paper birch and yellow birch,
our two smallest-seeded, fastest-growing species (Walters et al. 1993; Beaudet and Messier
1998), survived at relatively high rates. While we cannot make any definitive statements
about substrate nutrient content at this point, this finding demonstrates that even our most

101

nutrient depauperate substrate was able to support our most nutrient demanding species for
more than a single growing season.

Mycorrhizal colonization
For the four species examined in 2014, mycorrhizal colonization was consistently
higher on mineral soil than CWD (Figure 4.3). Although no attempt was made to quantify
root density or surrounding vegetation proximity at our substrate collection sites, this
pattern may reflect mineral soil containing a higher density of mycorrhizal infected roots
than CWD. Root density, in turn, may correlate with substrate nutrient concentrations, as
roots have been shown to respond to differences in resource availability (Gersani and
Sachs 1992; Robinson 1994; Hutchings and John 2004). This theory may also contribute to
the differing rates of mycorrhizal colonization observed across different species of CWD
(Figure 4.3).

Conclusions and Management Recommendations
Oftentimes one of the greatest challenges in regenerating or restoring tree species is
promoting their transition from the seedling to the sapling stage. Although our results
indicate that bare mineral soil typically supported the greatest development of species,
species face several challenges to their development on the forest floor (Shields et al. 2007;
Matonis et al. 2011; Kern et al. 2013;Willis et al. ch1 & 2;Walters et al. in preparation).
CWD offers refuge from several of these factors; however, our results indicate that
individual species of CWD can have strongly differing effects on seedling development.
This both echoes the results of Marx and Walters (2006 & 2008) and calls for a change in

102

perception of CWD being a generic resource. One of the challenges in making this
transition is the legacy of reduced mortality in managed stands. In such situations CWD
may already be in decline (Goodburn and Lorimer 1998; Hura and Crow 2004).
Furthermore, restarting this cycle naturally could take decades naturally and, in most
situations, would be prohibitively labor intensive to attempt artificially. As a result, greater
efforts need to be put forth to identify important species of CWD and restart their natural
formation if shortages are to be avoided in the future.

103

Figure 4.1- Height growth across balsam fir (BF), eastern hemlock (EH),
northern white cedar (NWC), paper birch (PB), sugar maple (SM), mineral soil (MS), and
yellow birch (YB) for balsam fir (A), eastern hemlock (B), paper birch (C), and
white pine (D) seedlings. Error bars represent one standard error, while substrates not
sharing a common letter supported significantly different seedling growth
(P<0.05,TukeyKramer HSD).

50

50

B

A
40

Height (cm)

Height (cm)

40

30

20

10

C

BC ABC AB BC

A

10

C

AB
BC ABC
BC

A
BC

0

BF EH NWC PB SM MS YB

BF EH NWC PB SM MS YB

Substrate

Substrate

50

A

C

D

40

Height (cm)

40

Height (cm)

20

BC

0

50

30

30

ABC

20

AB

BC
10

C

BC

30

20

A
10

BC

0

C

BC

ABC AB

BC

BC

0

BF EH NWC PB SM MS YB

BF EH NWC PB SM MS YB

Substrate

Substrate

104

Figure 4.2- Height growth across sterilized (Maroon) and unsterilized (Gold) balsam fir
(BF), eastern hemlock (EH), northern white cedar (NWC), paper birch (PB), sugar maple
(SM), mineral soil (MS), and yellow birch (YB) for red maple (A), northern white cedar
(B), white spruce (C), and yellow birch (D) seedlings. Error bars represent one standard
error, while asterisks represent statistically significant effects of sterilization on growth
(prob |t|< 0.05)
35

25
20
15
10

*

*

5

25
20
15
10

*

5

0

*

0

BF EH NWC PB SM MS YB

BF EH NWC PB SM MS YB

Substrate

Substrate

35

35

C

30

30

25

25

20
15

*

10

Height (cm)

Height (cm)

B

30

Height (cm)

Height (cm)

30

35

A

Sterilized
Unsterilized

5

D

*

20
15

*

10
5

0

0

BF EH NWC PB SM MS YB

BF EH NWC PB SM MS YB

Substrate

Substrate

105

80

A

70
60
50

A

40

A

30

A

20

A

A
A

10

AMF Colonization (%)

AMF Colonization (%)

Figure 4.3- Percent colonization of arbuscular mycorrhizal fungi (AMF) on red maple (A)
and northern white cedar (B) seedlings and ectomycorrhizal fungi colonization on yellow
birch (C) and white pine seedlings (D) across unsterilized balsam fir (BF), eastern hemlock
(EH), northern white cedar (NWC), paper birch (PB), sugar maple (SM), mineral soil
(MS), and yellow birch (YB) substrates. Error bars represent one stand error, while
substrates not sharing a common letter supported colonization at significantly different
rates (P <0.05, Tukey-Kramer HSD).

0

80

B

70
60

A

50

AB
40

ABC
30

BC BC
C

10
0

BF

EH NWC PB

SM

MS

BF

EH NWC PB

A

C

70

AB
60
50

BC

40
30

BC

BC
BC

20

SM MS YB

Substrate

C

10

EMF Colonization (%)

EMF Colonization (%)

Substrate

80

BC

20

0

80

D

70

A

60

AB

50
40

AB
30

AB
AB

20

B

B

10
0

BF

EH NWC PB SM MS YB

BF

Substrate

EH NWC PB

SM MS YB

Substrate

Note-2- Yellow birch was not analyzed a substrate for red maple due to our randomly
sample population failing to meet the minimum number of intersections to be considered
for analysis.

106

Figure 4.4- The relationship between arbuscular mycorrhizal fungi (AMF) colonization
and height growth for red maple (A) and northern white cedar (B) and ectomycorrhizal
fungi (EMF) colonization and height growth for yellow birch (C) and white pine (D).

30

35

A

30

25

Height (cm)

Height Growth (cm)

35

20
15
10
2
R =.2
P=.042

5

B

25
20
15
10
5

0

2
R =.51 P= <.001

0
0

20

40

60

80

0

% AMF Colonization
35

C

30

30

25

25

20
15
10
5
0

20

40

60

60

80

D

20
15
10
5

2
R =.24
P=.016

0

40

% AMF Colonization

Height (cm)

Height (cm)

35

20

2
R =.1
P=.103

0

80

% EMF Colonization

0

20

40

60

% EMF Colonization

107

80

Figure 4.5- Effect of sterilization (maroon) on species’ height growth performance
on balsam fir CWD. Species were ranked from one to eight with one representing the best
height growth performance. The Spearman’s rank correlation statistic is listed in the top
left corner.

14

Height (cm)

12
10

Unsterilized
Sterilized

1

0.119

8
2

6

3

2

3

4
2

1

8
8

0
BF EH NWC PB RM WP WS YB

Seedlings

108

Figure 4.6- Effect of sterilization (maroon) on species’ height growth performance
on northern white cedar CWD. Species were ranked from one to eight with one
representing the best height growth performance. The Spearman’s rank correlation statistic
is listed in the top left corner.
40
Height Growth (cm)

Unsterlized
Sterilized

30

0.9048
1
1
2

20

10

4

3 3

2 4

0
BF EH NWC PB RM WP WS YB

Seedlings

109

Figure 4.7- Effect of sterilization (maroon) on species’ height growth performance
on eastern hemlock CWD. Species were ranked from one to eight with one representing
the best height growth performance. The Spearman’s rank correlation statistic is listed in
the top left corner.
20
Unsterilized
Sterilized

Height (cm)

15

1

.4551
2
1

10
3

4
3

5

2

5

0
BF

EH NWC PB

RM WP WS

Seedlings

110

YB

Table 4.1- Results of a standard least squares model for the effects of Sterilization,
Species, Substrate, and their interactions on height growth.
Sum of
F
Prob >
Factor
Squares Ratio
F
1.03
5.42
0.0203
Sterilization
35.89
26.91 <.0001
Species
7.65
5.73 <.0001
Sterilization x Species
Height Growth
80.68
70.57 <.0001
(cm)
Substrate
10.25
8.97 <.0001
Sterilization x Substrate
71.26
8.9 <.0001
Species x Substrate
Sterilization x Species x
21.06
2.63 <.0001
Substrate
R Square Adj = .75

111

Table 4.2- F ratio’s and P > F from standard ANOVA least squares models for the effects
of Sterilization and Substrate, and their interactions on species’ height growth.
Species
Sterilization
Substrate
Sterilization x Substrate
0.41
5.59
0.66
0.5256
<.0001
0.6856
BF
0.02
4.09
2.03
0.8824
0.002
0.0789
EH
21.63
13.41
2.68
<.0001
<.0001
.0218
NWC
1.05
12.96
2.31
0.3113
<.0001
0.0546
PB
22.92
7.22
2.88
<.0001
<.0001
0.0229
RM
2.71
2.82
0.93
0.1049
0.0175
.4744
WP
0.24
11.22
2.88
0.623
<.0001
0.0148
WS
0.5466
28.83
4.21
0.4633
<.0001
.0018
YB

112

Table 4.3- Results of a nominal logistic regression model for the effects of Sterilization,
Species and Substrate, and their interactions on seedling survival.

Survival

Factor
Sterilization
Species
Sterilization x Species
Substrate
Sterilization x Species
Species x Substrate
Sterilization x Species x Substrate
R Square Adj = .54

113

L-R
Chi Square
0
244.29
>.01
>.01
2.95
67.88
52.68

Prob>ChiSq
1
<.0001
1
1
0.81
0.13
0.52

Table 4.4- Overall survival rate of individual species at the end of the experiment.
Survival
Species
Total
Living
Dead
(%)
Balsam Fir
83
81
2
98
Eastern Hemlock
77
69
8
90
Northern White Cedar
82
80
2
98
Red Oak
69
21
48
30
Paper Birch
66
50
16
76
Red Maple
80
47
33
59
White Ash
67
19
48
28
White Pine
84
78
6
93
White Spruce
86
82
4
95
Yellow Birch
70
62
8
89

114

Table 4.5- Results of a standard least squares model for the effects of Species, Substrate
and their interaction on mycorrhizal colonization of roots.
Sum of
Squares
1359
14269
13260

Factor

F
Ratio
1.84
9.65
2.99

Species
Root Colonization (%)
Substrate
Species x Substrate
R Square Adj = .51
Note-1 Mycorrhizal colonization was not observed on Sterilized substrate.

115

Prob > F
0.1476
<.0001
0.0005

Table 4.6- Sterilization’s effect on species height growth rankings within individual
substrate types. Values of ρ are Spearman’s rank correlations between sterilized and
unsterilized treatments. Smaller values indicate greater changes in ranks.
Substrate
Balsam Fir
Hemlock
Northern White Cedar
Paper Birch
Sugar Maple
Bare Mineral Soil
Yellow Birch

Spearman's ρ

Prob > |ρ|

0.119
0.4551
0.9048
0.6905
0.6429
0.5952
0.6946

0.7789
0.2572
0.002
0.058
0.0856
0.1195
0.0559

116

LITERATURE CITED

117

LITERATURE CITED

Abrams, M. D., 1998. The red maple paradox. BioScience. 48, 355-364.
Aerts, R., 1999. Interspecefic competition in natural plant communities: mechanisms,
trade-offs, and plant-soil feedbacks. Journal of Experimental Botany. 50, 29-37.
Amaranthus, M.P., Perry, D.A., 1994. The functioning of ectomycorrhizal fungi in the
field: linkages in space and time. Plant and Soil. 159, 133-140.
Arthur, M.A., Tritton, L.M., Fahey, T.J., 1993. Dead bole mass and nutrients remaining 23
years after clear-felling of a northern hardwood forest. Can. J. For. Res. 23, 12981305.
Bancroft, T.A., 1964. Analyses and inferences for incompletely specified models involving
the use of preliminary tests of significance. Biometrics. 20, 427-442.
Beaudet. M., Messier, C., 1998. Growth and morphological responses of yellow birch,
sugar maple, and beech seedlings growing under a natural light gradient. Can. J.
For. Res. 28, 1007-1015.
Bolton, N.W., D’Amato, A.W., 2011. Regeneration responses to gap size and coarse
woody debris within natural disturbance-based silvicultural systems in northeastern
Minnesota, USA. For. Ecol. Manage. 262, 1215-1222.
Caspersen, J.P., Saprunoff, M., 2005. Seedling recruitment in a northern temperate forest:
the relative importance of supply and establishment limitation. Can. J. For. Res. 35,
978-989.
Chapin, S., The mineral nutrition of wild plants. 1980. Annual review of ecology and
systematic. 11, 233-260.
Cornett, M.W., Puettmann, K.J., Frelich, L.E., Reich P.B., 2001. Comparing the
importance of seedbed and canopy type in the restoration of upland Thuja
Occidentalis forests of northeastern Minnesota. Restoration Ecology. 9, 386-396.
Courty, P.E., Buee, M., Diedhiou, A.G., Frey-Klett, P., Le Tacon, F., Rineau, F., Turpault,
M.P., Uroz, S., Garbaye, J., 2010. The role of ectomycorrhizal communities in
forest ecosystem processes: new perspectives and emerging concepts. Soil Biology
and Biochemsitry. 42, 679-698.
Crow, T.R., 1988. Reproductive mode and mechanisms for self-replacement of northern
red oak (Quercus ruba)- A review. Forest Science. 34, 19-40.
Gersani, M., Sachs, T., 1992. Developmental correlations between roots in heterogeneous
environments. Plant Cell and Environment. 15, 463-469.

118

Goodburn, J.M., Lorimer, C.G., 1998. Cavity trees and coarse woody debris in old-growth
and managed northern hardwood forests in Wisconsin and Michigan. Can. J. For.
Res. 28, 427-438.
Graham, R.L., Cromack, K., Jr., 1982. Mass, nutrient content, and decay rate of dead boles
in rain forests of Olympic National Park. Can. J. For. Res. 12, 511-521.
Grime, J.P., 1977. Evidence for the existence of three primary strategies in plants and its
relevance to ecological and evolutionary theory. American Naturalist. 111, 11691194.
Harmon, M.E., Franklin, J.F., Swanson, F.J., Sollins, P., Gregory, S.V., Lattin, J.D.,
Anderson, N.H., Cline, S.P., Aumen, N.G., Sedell, J.R., Lienkaemper, G.W.,
Cromack, K., Jr., Cummins, K.W., 1986. Ecology of coarse woody debris in
temperate ecosystems. Advances in Ecological Research. 15, 133-302.
Harmon, M.E., Franklin, J.F., 1989. Tree seedlings on logs in Picea-Tsuga forests of
Oregon and Washington. Ecology. 70, 48-59.
Harvey, A.E., Larsen, M.J., Jurgensen, M.F., 1979. Comparative distribution of
ectomycorrhizae in soils of three western Montana forest habitat types. For. Sc. 25,
350-360.
Hura, C.A., Crow, T.R., 2004. Woody debris as a component of ecological diversity in
thinned and unthinned northern hardwood forests. Nat. Ar. J. 24, 57-64.
Hutchings, M.J., John, E.A., 2004. The effects of environmental heterogeneity on root
growth and root/shoot partitioning. Annals of Botany. 94, 1-8.
Kaelke, C.M., Kruger, E.L., Reich, P.B., 2001. Trade-offs in seedling survival, growth, and
physiology among hardwood species of contrasting succesional status along a light
availability gradient. Can. J. For. Res. 31, 1602-1616.
Kern, C.C., D’Amato, A.W., Strong, T.F., 2013. Diversifying the composition and
structure of managed, late-succesional forests with harvest gaps: what is the
optimal gap size? For. col. Manage. 304, 110-120.
Klockow, P.A., D’Amato, A.W., Bradford, J.B., Fraver, S., 2014. Nutrient concentrations
in coarse and fine woody debris of Populus tremuloides Michx.-dominated forests,
northern Minnesota, USA. Silva Fennica. 48, 1-24.
Kolb, T.E., Steiner, K.C., McCormick, L.H., Bowersox, T.W., Growth response of
northern red-oak and yellow-poplar seedlings to light, soil moisture and nutrients in
relation to ecological strategy. 38, 65-78.
Marx., L., Walters, M.B., 2006. Effects of nitrogen supply and wood species on Tsuga
canadensis and Betula alleghaniensis seedling growth on decaying wood. Can. J.
For. Res. 36, 2873-2884.

119

Marx, L., Walters, M.B., 2008. Survival of tree seedlings on different species of decaying
wood maintains tree distributions in Michigan hemlock-hardwood forests. Ecology
96, 505-513.
Matonis, M.S., Walters, M.B., Millington, J.D.A, 2011. Gap-,stand,-and landscape-scale
factors contribute to poor sugar maple regeneration after timber harvest. For. Ecol.
Manage. 262, 286-298.
McGee, G.G., Birmingham, J.P., 1997. Decaying logs as germination sites in northern
hardwood forests. Nor. J. of App. For. 14, 178-182.
Mori, A., Mizumachi, E., Osono, T., Doi, Y., 2004. Substrate-associated seedling
recruitment and establishment of major conifer species in an old-growth subalpine
forest in central Japan. For. Ecol. Manage. 196, 287-297.
O’Hanlon-Manners, D.L., Kotanen, P.M., 2004. Logs as refugees from fungal pathogens
for seeds of eastern hemlock (Tsuga canadensis). Ecology, 85, 284-289.
Packer, A., Clay, K., 2000. Soil pathogens and spatial patterns of seedling mortality in a
temperate tree. Nature. 404, 278-281.
Packer, A., Clay, K., 2003. Soil pathogens and Prunus serotina seedling and sapling
growth near conspecific trees. Ecology. 84, 108-119.
Perez-Moreno, J., Read, D.J., 2000. Mobilization and transfer of nutrients from litter to
tree seedlings via the vegetative mycelium of ectomycorrhizal plants. New Phytol.
145, 301-309.
Phillips, R.P., Brzostek, E., Midgley, M.G., 2013. The mycorrhizal-associated nutrient
economy: a new framework for predicting carbon-nutrient couplings in temperate
forests. New Phytologist. 199, 41-51.
Robinson, D., 1994. The response of plants to non-uniform supplies of nutrients. New
Phytologist. 127, 635-674.
Sanchez, E., Gallery, R., Dalling, J.W., 2009. Importance of nurse logs as a substrate for
the regeneration of pioneer tree species on Barro Colorado Island, Panama. Journal
of Tropical Ecology. 25, 429-437.
Shields, J.M., Webster, C.R., Nagel, L.M., 2007. Factors influencing tree species diversity
and Betula alleghaniensis establishment in silvicultural openings. Forestry 80,
293-307.
Shorohova, E., Kapista, E., 2014. Mineralization and fragmentation rates of bark attached
to logs in a northern boreal forest. For. Ecol. Manage. 315, 185-190.
Simard, M.J., Bergeron, Y., Sirois, L., 2003. Substrate and litterfall effects on conifer
seedling survivorship in southern boreal stands of Canada. Can. J. For. Res. 33,
672-681.

120

Smith, S.E., Read, D.J., 1997. Mycorrhizal Symbiosis. Academic Press, London.
Smith, S.E., Smith, A.F., 2011. Roles of arbuscular mycorrhizas in plant nutrition and
growth: New paradigms from cellular to ecosystem scales. An. Rev. of Pl. Bio. 62,
227-250.
Strukelj, M., Brais, S., Quideau, S.A., Angers, V.A., Kebli, H., Drapeau, P., Oh, S.W.,
2013. Chemical transformations in downed logs and snags of mixed boreal species
during decomposition. Can. J. For. Res. 43, 785-798.
Takahashi, M., Sakai, Y., Ootomo, R., Shiozaki, M., 2000. Establishment of tree seedlings
and water-soluble nutrients in coarse woody debris in an old-growth Picea-Abies
forest in Hokkaido, northern Japan. Can. J. For. Res. 30, 1148-1155.
Tedersoo, L., Subi, T., Jairus, T., Kõljalg, U., 2008. Forest microsite effects on community
composition of ectomycorrhizal fungi on seedlings of Picea abies and Betula
pendula. Environmental Microbiology. 10, 1189-1201.
Treseder, K.K., Allen, M.F., Ruess, R.W., Pregitzer, K.S., Hendrick, R.L., 2005. Lifespans
of fungal rhizomorphs under nitrogen fertilization in a Pinyon-Juniper woodland.
Plant and Soil. 270, 249-255.
Turner, B.L., 2008. Resource partitioning for soil phosphorous: a hypothesis. J. Ecol. 96,
698-702.
Walters, M.B., Kruger, E.L., Reich, P.B., 1993. Growth, biomass distribution and CO2
exchange of northern hardwood seedlings in high and low light: relationships with
succesional status and shade tolerance. Oecologia. 94, 7-16.
Wang, G.G., Kembell, K.J., 2005. Balsam fir and white spruce seedling recruitment in
response to understory release, seedbed type, and litter exclusion in trembling
aspen stands. Can. J. for. Res. 29, 187-193.

121

CHAPTER V

Conclusions
Elucidating the factors involved with seedling establishment is important for
advancing our understanding of forest stand dynamics and is essential for maintaining
species diversity in managed forest systems. Forest substrate is a particularly important
factor to understand, as it has the potential to influence species composition by affecting
which species become established in the seedling layer. Moreover, substrate availability
can easily be manipulated by existing silvicultural techniques, making understanding
substrate’s effect on seedling dynamics highly relevant to forest management.
In this study, bare mineral soil/humus was critical for the germination of smaller
seeded species provided the availability of seed on the forest floor. This is an important
finding, as forest management can simply scarify the forest floor or harvest in the absence
of snow to increase bare mineral soil availability. Light availability is another such factor
that can be easily manipulated by management and was influential to seedling
establishment; species generally germinated best in lower light environments, but
established and survived at higher rates in higher light environments. Nevertheless,
providing bare mineral soil/humus and creating a favorable light environment by cutting a
particular harvest gap size may not guarantee seedling recruitment, as several species’
initial establishment was also influenced by local seed source density (50m), competition
from non-tree vegetation, and site quality (soil moisture and nutrient availability). In such
scenarios, forest management could turn to artificial seeding or planting, herbicides, and
local site knowledge to overcome these regeneration obstacles.

122

Decaying coarse woody debris (CWD) is another type of substrate on which
seedlings often develop. In this study, conifer species demonstrated the highest survival
across substrates (bare mineral soil and six types of CWD), while deciduous species
survival was more variable. This suggests that CWD may be more important seedling
establishment substrate for conifers in northern hardwood forests. Nonetheless, individual
species’ of CWD varied strongly in their ability to support seedling development, as paper
birch (Betula. papyrifera Marsh), northern white cedar (Thuja occidentalis L.) and eastern
hemlock (Tsuga canadensis L) CWD generally supported greater height growth (cm) than
yellow birch (B. alleghaniensis Britt.), sugar maple (Acer saccharum Marsh.) and balsam
fir (Abies balsamifera (L.) Mill.). This finding suggests that scientists and forest managers
should take into consideration the specific type of CWD that is available on the forest
floor, much like we consider the particular type of soil that exists at a site, rather than
consider CWD as a generic seedling establishment substrate. Mycorrhizal colonization was
also shown to strongly influence seedling development across substrates. In addition,
seedling colonization was also shown to vary across substrate, providing at least
correlative evidence that mycorrhizal fungi may be strongly contributing to the differing
patterns of seedling growth across different substrates.
Seedling establishment is a complex process in which species must pass through a
number of developmental stages before they can develop into saplings. This complexity is
magnified in forests where several factors can inhibit development at each stage. The
evidence presented in this dissertation demonstrates that substrate can have strong effects
over seedling layer composition at the germination stage. However, it was also shown that
several factors, acting on stages before and after germination, may override substrate’s

123

influence. Collectively, these findings suggest that a lack of bare mineral soil/humus or
eastern hemlock, paper birch, and northern white cedar CWD is contributing to the
regeneration failures reported for several tree species in the northern hardwood forests of
the Great Lakes region. However, these findings also suggest that manipulating substrate
availability alone may not be enough to reestablish many of these species to the seedling
layer.

124