q .
{3...}

xnvpufifl

:1.

z. 3 .
3.964“...th
i150 1.7!.Ju :

:4 : fad!

 

 

» .4 again
a .m.‘ 2&3.

 

DE

 

/\.;,\N..
d&/_ .

.23

(U
5603-?
E s 5
[13.9.2
_.£ C

E

 

 

 

This is to certify that the
dissertation entitled

SOIL PROCESSES AND PLANT SPECIES: DOES THE RE-
INTRODUCTION OF NATIVE GRASSES ALTER SOIL
CARBON AND NITROGEN CYCLING?

presented by

Wendy Mae Mahaney

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Department of Plant Biology
and Program in Ecology,
Evolutionary Biology and
Behavior

 

 

Wiifla

Major Professor’s Signature

50W01007~

Uofie

MSU is an affirmative-action, equal-opportunity employer

 

l

i
‘ PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDueindd-p1

 

 

SOIL PROCESSES AND PLANT SPECIES: DOES THE RE-INTRODUCTION OF
NATIVE GRASSES ALTER SOIL CARBON AND NITROGEN CYCLING?

By

Wendy Mae Mahaney

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Plant Biology
and
Program in Ecology, Evolutionary Biology and Behavior

2007

ABSTRACT

SOIL PROCESSES AND PLANT SPECIES: DOES THE RE-INTRODUCTION OF
NATIVE GRASSES ALTER SOIL CARBON AND NITROGEN CYCLING?

By

Wendy Mae Mahaney

Human activities have altered biodiversity on a global scale, but the ecological
implications of shifts in plant species distributions and abundances are poorly understood.
While a number of studies have shown that exotic species can dramatically and rapidly
alter ecosystem properties (Vitousek and Walker 1989, Evans et al. 2001, Mack and
D'Antonio 2003a), little is known about how reintroductions of extirpated species may
impact ecosystem properties in restored systems. My dissertation research focuses on
how the reintroduction of native prairie C4 grasses into abandoned agricultural fields
(old-fields) influences soil carbon (C) and nitrogen (N) cycling compared to non-native
C3 grasses typical of successional communities in southwestern Michigan, USA.

In this dissertation, I explore three main aspects of plant species controls on soil
processes: 1) What are the decadal scale impacts of a shift from a C3- to a C4-dominated
system on soil properties and processes (Chapter Two), 2) How quickly do differences in
species traits and soil conditions arise (Chapter Three), and 3) Which plant traits are
responsible for differences in decomposition rates (Chapter Four)? I addressed these
questions in several old-fields at Michigan State University’s W. K. Kellogg Biological
Station, using previously established experimental plots of C4 grasses in Chapters Two
and Four, and setting up new experimental studies in Chapter Three.

In Chapters Two and Three, I found that C4 species had significantly greater

shoot biomass and more recalcitrant tissue compared to the dominant C3 species, and

these differences became apparent within two years after the species were established. In
contrast, species differences in surface litter and root biomass took longer than two years
to develop but were apparent after 1 1 years. While there was some evidence to suggest
that the C4 species had reduced soil inorganic N levels relative to the C3 species afterjust
two years, many of the changes in soil properties took longer than two growing seasons
to develop. After 1] years, soils under C4 species had significantly lower inorganic N
levels, and slightly lower in situ net N mineralization and nitrification rates when
compared to soils under C3 species. I also found limited evidence for increasing soil C
pools under C4 species 11 years after reintroduction. Nevertheless, the 613C signal of
the C4 species became measurable in the soil within two years.

I examined how litter quality and microclimate affected litter decomposition rates,
and found that while Andropogon gerardii (a C4 prairie grass) differed from C3 species
in its effect on soil moisture and temperature, these differences did not correspond to
differences in decomposition rates. Instead, species litter quality was more important
than microclimate in determining decomposition rates of both C3 and C4 species.

Overall, my results demonstrated that reintroduction of C4 species into old-fields
can alter soil processes related to C and N cycling on relatively short timescales. Process
rates changed first, with changes in pool sizes of C and N taking longer to become
measurable. Improving our understanding of how plant species impact ecosystem
properties and what species traits are driving these changes is imperative if we hope to
predict the ecosystem-level consequences of changes in species distribution or
composition that could occur, and are occurring, as a consequence of changes in

agricultural and land use practices, global change. and species introductions.

ACKNOWLEDGMENTS

I have to first thank my husband, Kurt Smemo, who has been a pillar of support
for me throughout this process. He provided emotional support and encouragement
whenever I needed it, and helped give me the confidence to develop my intellectual
independence. He has shown remarkable patience these past four years by reading
countless proposals and versions of these chapters, as well as listening to my ideas,
questions, and complaints along the way. He helped me collect soil samples on many
nights and weekends without complaint, even though he was tired from doing his own
research. He has been both an advisor and my best friend, with an uncanny ability to
know which one I needed him to be. I could not ask for a more supportive husband, and I
am so grateful to him for all that he has done for me.

I thank my family for all of their never—ending support for anything I set out to
achieve. My parents instilled in me the work ethic and values that made this
accomplishment possible. My sister, Julie, has always been supportive of my
achievements, and I would not be getting this degree without our sibling rivalry pushing
me to excel at school. I thank all of them for years of support and guidance in my life,
and for their patience when I was stressed (and grumpy!) from working too hard. They
have supported me in every aspect of my life, and I cannot begin to convey how much
this has meant to me. They even helped me water my experiment on the day of my
wedding to help my plants survive a drought! I also thank Kurt’s family—especially
Sandy, Tom, Monty, and Erin—~for being so supportive these past four years. I am

grateful to Tom, Monty, and Terry for helping me with field work before my wedding.

I thank my advisor and committee for helping me gain confidence in my abilities
as both a researcher and an independent thinker. Thanks to Katherine Gross for agreeing
to mentor me and supporting my decision to spend these past months in Ohio. Merritt
Turetsky was a wonderful mentor and selflessly gave me the attention and advice that I
really needed in the early stages of my research. Thanks to Peter Murphy for providing
me an academic home while living on main campus, and for his participation on my
committee for almost four years. His guidance throughout this process is greatly
appreciated. I am grateful to G. Philip Robertson for agreeing to serve on my committee
at the last minute, and I thank him and Stephen Hamilton for their thoughtful insights and
comments on my research.

I want to thank Amy Burgin and Terry Loecke for being such great friends. They
made the fun times so much more enjoyable and the tough times so much more tolerable.
I appreciate their friendship and support through these past four years, and they have
enriched my life more than I can convey. I also appreciate Terry’s willingness to teach
me new techniques, and his patience to repeat it several times!

The people who have helped me with field and laboratory work are too numerous
to mention, but I need to thank the following people for their contributions. I thank Carol
Baker for her friendship, help, and patience while teaching me how the Gross lab works.
I thank Pam Moseley and the summer employees in the Gross lab for laboratory and field
assistance. I appreciate that Emily Grman and Todd Robinson took time away from their
work to help me with mine. Thanks to Stu Bassett for digging holes for my Plant Trait

Experiment, and to Greg Parker and Gale Kellam for applying herbicide for the

Monoculture Experiment. I thank Merritt Turetsky for letting me use her lab, and Neville
Millar and Claire Treat for helping me set up and run the Ankom fiber analyzer.

This dissertation was greatly improved by comments from and conversations with
Kurt Smemo, Kay Gross, Peter Murphy, Stephen Hamilton, Merritt Turetsky, Terry
Loecke, G. Philip Robertson, and Bryan Foster. My research benefited from comments
by the Gross lab graduate students and post—docs, particularly Emily Grman, Todd
Robinson, Sarah Emery, Rich Smith, Chad Brassil, and Greg Houseman.

I appreciate the many sources of financial assistance that I received, as this
research could not have been done without it. Financial support for research and travel to
scientific meetings was provided to me in the form of several George H. Lauff Research
Awards, two T. Wayne and Kathryn Porter Research Scholarships, two LTER graduate
student research grants, three Paul Taylor Travel Grants, a Marvin Hensley Endowed
Fellowship, and a Graduate School Travel Grant. Financial assistance was also provided
by the NSF Long-Tenn Ecological Research Program DEB 0423627 to Katherine Gross.
I also want to thank the College of Natural Sciences and the Plant Sciences program for
awarding me recruiting fellowships, the Ecology, Evolutionary Biology and Behavior
Program for awarding me a Summer Fellowship, and the Graduate School for awarding
me a Dissertation Completion Fellowship. These financial awards were instrumental in

my completing this degree and I am extremely grateful to have received them.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. ix
LIST OF FIGURES ................................................................................ xi
CHAPTER ONE: INTRODUCTION ............................................................ 1

CHAPTER TWO: DECADAL SCALE IMPACTS OF C4 GRASS REINTRODUCTION
ON SOIL CARBON AND NITROGEN CYCLING IN SUCCESSIONAL
ECOSYSTEMS

Abstract ...................................................................................... 8
Introduction .................................................................................. 9
Methods ................................................................................... 12
Study Sites ........................................................................ 12

Field Sampling ................................................................... 14
Laboratory Analyses ............................................................. 15
Statistical Analyses .............................................................. 17
Results ..................................................................................... 18
Species Traits: Biomass and Tissue Characteristics ......................... 18

Soil Properties .................................................................... 22
Discussion ................................................................................. 28

CHAPTER THREE: SHORT-TERM IMPACTS OF C4 PRAIRIE GRASS RE-
ESTABLISHMENT ON SOIL CARBON AND NITROGEN IN OLD-FIELDS

Abstract ..................................................................................... 35
Introduction ............................................................................... 36
Methods .................................................................................... 38
Monoculture Experiment ....................................................... 38
Plant Trait Experiment .......................................................... 43
Data Analysis ...................................................................... 44
Results ...................................................................................... 45
Monoculture Experiment ....................................................... 45
Plant Trait Experiment .......................................................... 53
Discussion .................................................................................. 53

CHAPTER FOUR: EFFECTS OF LITTER QUALITY AND SOIL MICROCLIMATE
ON DECOMPOSITION

Abstract .................................................................................... 61
Introduction ................................................................................ 62
Methods ..................................................................................... 64
Study Sites ......................................................................... 64
Reciprocal Transplant Experiment ............................................. 65
Decomposition and Soil Microclimate Experiment .......................... 67

vii

Results ...................................................................................... 69

Reciprocal Transplant Experiment ............................................ 69
Decomposition and Soil Microclimate Experiment .......................... 73

Discussion .................................................................................. 73
CHAPTER FIVE: CONCLUSIONS ............................................................ 82

APPENDIX 1.Location of the three study sites at the WK. Kellogg Biological Station.87

REFERENCES ..................................................................................... 89

viii

LIST OF TABLES

Table 1. Plant community characteristics of the plots at harvest in August 2006. Shoot
biomass (mean 2 SE; n=9) and species percentages of the total biomass are based on the
average of nine samples for each dominant vegetation plot. Species richness is given as
the mean of the plots (n=9; 0.5 m x 0.5 m) and as the total across plots (mean, 2). Plot
names are abbreviated as follows: Ag=A na’ropogon, Ss=Schizachyrium,

Sn=Sorghastrum, Er=Elymus, and Bi=Br0mus. A designation as C3 or C4 species and as
monocot (M) or F orb (F) follows the species names .......................................... 13

Table 2. ANOVA results for various measures of plant production, using Site and
dominant species Plot as main effects. Natural log transformations were used to
normalize the data for all variables. Significant p-values are indicated in bold ........... 19

Table 3. ANOVA model for plant tissue chemistry, using Site and the dominant Species
as main effects. Mean (:SE) values are also shown for the dominant species at each site.
When Species effects are present, superscript numbers indicate significant differences
(p<0.05) between species. For significant interactions, superscript letters denote
significant differences (p<0.05). The Species names are abbreviated as follows:
Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Er=Elymus, and Bi=Br0mus.

Sample sizes for nitrogen (N), carbon (C), C :N, Acid detergent fiber (ADF):N and 613C
were 4 for each species, while n=6 for ADP .................................................... 21

Table 4. ANOVA model results for surface soil (0-10cm) carbon (C) and nitrogen (N)
variables, using Site and Plot (dominant vegetation) as main effects. Mean (:SE) values
are shown below the ANOVA. When Species effects are present, superscript numbers
indicate significant species differences (p<0.05). For significant interactions, superscript
letters denote significant differences (p<0.05). The Plot names are abbreviated as
follows: Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Er=Elymus, and
Bi=Br0mus. The data shown are untransformed values, but inorganic N was In-
transformed prior to analysis ...................................................................... 24

Table 5. ANOVA model for subsurface soil (10-20cm) carbon (C) and nitrogen (N)
variables, using Site and Plot (dominant vegetation) as main effects. Mean (:SE) values
are shown below the ANOVA. For significant Species effects, superscript numbers
indicate significant species differences (p<0.05). The Plot names are abbreviated as
follows: Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Er=Elymus, and
Bi=Br0mus. The data shown are untransformed means, but the following variables were

transformed prior to analysis: inorganic N (In), N (square root) and 613C (square root)..25

Table 6. Site characteristics of old-fields used for the Monoculture Experiment. All data
are means (:SE) from data collected in September 2005; biomass data (g m-Z) was from

harvested plots (n=9) and soil carbon (g mo") and nitrogen were from soils cores (0-10cm
deep; n=4). Species percent of the total shoot biomass is given in parentheses after the
species name. Nomenclature for all species follows the USDA Plants Database
(plants.usda. gov). .’ .................................................................................. 40

Table 7. ANOVA results for aboveground variables for all five species in Louden Field
McKay Field, and Turkey Meadow, using Site and Species as main effects, with Block
(Blk) nested within Site. Significant differences are indicated in bold ..................... 46

Table 8. Mean (:tSE) surface litter mass (g m-z) for each species at each site. Superscript
numbers denote significant differences (p<0.05) between species within each Site. In
addition, McKay Field had significantly more surface litter than Turkey Meadow
(p=0.011) ............................................................................................. 49

Table 9. Plant tissue chemistry (mean:SE) for each species (n=4), averaged across sites.
Elymus data is shown for comparison only. For variables with significant Species main
effects, lowercase letters denote significant differences (p<0.05) determined from post-
hoc Tukey comparisons ............................................................................ 51

Table 10. Soil chemistry data (mean:SE) for soils under each species, averaged across
sites. For variables with significant Species effects, superscript letters denote significant
differences (p<0.05) determined from post-hoc Tukey comparisons. Elymus is shown for
comparison only. .L .................................................................................. 52

Table 11. Plant Trait Experiment root biomass at each depth interval (mean-ESE; n=4).

For each depth interval, significant differences (p<0.05) between species are indicated by
superscript lowercase letters, as determined from post-hoe Tukey comparisons. Biomass
values without letters indicate no significant difference from any of the other species....56

Table 12. Chemistry of the senesced shoot tissue for the species used in the Reciprocal
Transplant Experiment. Data shown are means:SE, based on fraction of oven-dried
weight. Acid detergent fiber (ADF) is primarily lignin and hemicellulose). Superscript
lowercase letters denote significant differences (p<0.001) between species ............... 70

LIST OF FIGURES

Figure 1. Conceptual diagram of how changes in plant species composition can influence
ecosystem processes in terrestrial ecosystems. Plant species may directly (via differences
in litter quality and quantity) and indirectly (via effects on soil microenvironment) exert

important controls over the functioning of microbial communities that determine nitrogen
(N) and carbon (C) cycling .......................................................................... 3

Figure 2. Root and shoot biomass (mean :SE in g m-z) in August 2006 at both sites.
Root biomass is shown separately by depth intervals here for additional detail, but
statistical results are reported for total root biomass from 0-20cm. Lowercase letters
denote significant differences (p<0.05) between plots for shoot (n=9) and root (n=6)
biomass. Species names are abbreviated as follows: Ag=Andr0p0g0n,

Ss=Schizachyrium, Sn=S0rghastrum, and C3=Elymus in McKay Field and Bromus in
Turkey Meadow ..................................................................................... 20

Figure 3. Percentage of total soil carbon (mean :SE) contributed by C4 species at both
surface (O-lOcm; n=6) and subsurface (10-20cm; n=5) soil depths at both sites.
Lowercase letters denote significant differences (p<0.05) between plots separately for
each depth. Plot names are abbreviated as follows: Ag=Andr0p0g0n, Ss=Schizachyrium,

Sn=Sorghastrum and C3=Elymus in McKay Field and Bromus in Turkey Meadow ...... 26

Figure 4. In situ net nitrification and N mineralization rates (mean 2 SE; n=5) for

Andropogon plots (solid lines) and C3 plots (dashed lines) at both Turkey Meadow (black
lines) and McKay Field (gray lines). Rates are calculated from 28-day in situ incubations
performed consecutively from June to November 2006 ....................................... 27

Figure 5. Shoot and root production estimates of each species in each site in the
Monoculture Experiment, averaged across the three blocks in each site. Mean (iSE) root
biomass for each depth interval (n=6) and shoot biomass (n=9) in August 2006 is shown
separately for each site. Root biomass was not collected at Louden Field. Species names
are abbreviated: Ag=Androp0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Bi=Br0mus,
Pp=P0a, and Er=Elymus ........................................................................... 47

Figure 6. Species effects on inorganic N (mean 1- SE) concentrations in the Monoculture
Experiment, averaged across blocks, for the three sites. Species names are abbreviated:
Ag=Andr0p0g0n, Ss= Schizachyrium, Sn=Sorghastrum, Bi=Br0mus, Pp=P0a, and

Er=Elymus. Solid bars represent soils under C4 species and patterned bars represent soils

under C3 species. Lowercase letters denote significant differences (p<0.05) between
plots (Elymus plot data is shown for comparison purposes only) ............................ 50

xi

Figure 7. Above- (n=5) and below- (0-80 cm; n=4) ground biomass per plant for each
species (mean :SE) in the Plant Trait Experiment. Species names are abbreviated:

Ag=Andr0p0g0n, Ss= Schizachyrium, Sn=S0rghastrum, Bi=Br0mus, Pp=P0a, and
Er=Elymus. Solid bars represent C4 species and patterned bars represent C3 species.

Lowercase letters denote significant (p<0.05) differences between species for shoot
biomass and total root biomass ................................................................... 54

Figure 8. Tissue chemistry data for species grown in the Plant Trait Experiment; values
are mean (:SE; n=4). Species names are abbreviated: Ag=Andr0p0g0n, Ss=
Schizachyrium, Sn=Sorghastrum, Bi=Br0mus, Pp=Poa, and Er=Elymus. Acid detergent
fiber (ADF) is a measure of recalcitrant compounds, primarily lignin and hemicellulose.

Solid bars represent C4 species and patterned bars represent C3 species. Lowercase
letters denote significant differences (p<0.05) between species .............................. 55

Figure 9. Litter mass remaining (% of initial) across time for the Reciprocal Transplant
Experiment. Black lines (both solid and dotted) represent Andropogon litter, while Gray

lines represent the C3 species litter (Bromus in Turkey Meadow, Elymus in McKay
Field). Solid lines show decomposition when the litter was placed under Andropogon

and dotted lines show decomposition when the litter was placed under C3 species ...... 71

Figure 10. Comparison of the mass remaining (% of initial) after the first 142-days for
litter placed in the field from November 2005-April 2006 (black bars) compared to litter
placed in the field from April-September 2006 (gray bars). Data are the means (:SE) for

each litter type, averaged across placement location (i.e., und'er C3 and under
Andropogon) ......................................................................................... 72

Figure 11. Soil temperature (degrees C) on the soil surface (a) and at 10cm depth (b) for
Sites in 2006, averaged across the +/-Litter treatment. Black lines represent Andropogon

plots, while Gray lines represent C3 plots. Greenhouse-Geisser adjusted p-values for log-
transformed data are reported for time variables, and untransformed data are shown...74—5

Figure 12. Soil moisture (0-10cm depth) for the two sites in 2006, averaged across the +/-
Litter treatment. Black lines represent Andropogon plots, while Gray lines represent the

C3 plots. Greenhouse-Geisser adjusted p-values for ln-transformed data are reported
below, and untransformed data are shown ....................................................... 76

Figure 13. Mass loss (% of initial) for cellulose in both sites (McKay Field and Turkey

Meadow), placed under either Andropogon or the C3 species (Placement). Black solid
bars represent Andropogon plots with litter left intact (+litter), black striped bars
represent A na’ropogon plots where the litter was removed (-litter), gray solid bars

represent C3 plots with litter left intact (+litter), and gray striped bars represent C3 plots
where the litter was removed (-litter) ............................................................. 77

xii

Appendix 1. Location of the three study sites at the WK. Kellogg Biological Station...87

xiii

CHAPTER ONE

INTRODUCTION

Human activities, such as land use change and the introduction of exotic species, have
altered biodiversity on a global scale (Lawton and May 1995, Pimm et al. 1995, Chapin
et al. 2000, Hooper et al. 2005). However, the ecological implications of shifts in plant
species distributions and abundances are poorly understood. While a number of studies
have shown that exotic species can dramatically and rapidly alter ecosystem properties
(Vitousek and Walker 1989, Evans et al. 2001, Mack and D'Antonio 2003a), little is
known about how reintroductions of extirpated species may impact ecosystem properties
in restored systems. My dissertation research focused on how the reintroduction of C4
grasses into abandoned agricultural fields (old-fields) influenced soil carbon (C) and
nitrogen (N) cycling compared to exotic C3 grasses typical of successional communities
in southwestern Michigan, USA on both short (1-2 years) and decadal timescales.

Total soil C and N loss associated with conversion of prairie soils to agriculture is
dramatic (Burke et al. 1995, Camill et al. 2004, DeGryze et al. 2004), and grassland
cultivation in the Midwestern United States has resulted in a 30-60% loss of soil organic
C and N (Burke et al. 1995). There is some evidence that prairie restorations can begin to
restore soil processes (Baer et al. 2002, Camill et al. 2004) and hasten the buildup of soil
C and N pools. However, it is unknown how quickly soil conditions will return to pre-
agricultural levels, and whether these changes can be accelerated by using particular

species.

Plant species influence soil processes in a variety of ways, and changes in species
composition can have dramatic impacts on C and N cycling (Figure 1). One of the most
commonly examined ways in which plants alter soil processes is through differences in
litter quality (i.e., tissue chemistry) and quantity (Hobbie 1992, Wardle et al. 1998,
Ehrenfeld 2003, Lovett et al. 2004, Hooper et al. 2005, Dijkstra et al. 2006). While the
importance of species litter chemistry in controlling nutrient cycling is well accepted,
how plants impact these processes via their effect on soil microclimate has received less
attention. Several recent papers (Mack and D'Antonio 2003b, Eviner 2004, Eviner et al.
2006) suggest that such effects can be an important determinant of soil process rates.

To examine how plant traits are linked to soil properties, I chose to examine C3
and C4 grass species that differ significantly in tissue production and chemistry. The C4
grasses that I examined typically produce more biomass (both root and shoot) and have
more recalcitrant tissue (i.e., higher CzN, lignin:N) than the C3 grass species (Wedin and
Tilman 1990, Baer et al. 2002, Camill et al. 2004). These differences led me to
hypothesize that soil C and N pools and cycling rates would differ under the two groups
of species. However, I did not know how quickly these soil differences would arise, or if
species within a group would differ in their effects. Because these C3 and C4 species
differ in litter quality, quantity and their influences on soil microclimate, it difficult to
determine if soil process changes are due to one or a combination of these factors.

The next three chapters explore three main aspects of plant species controls on
soil processes in old-fields: 1) What are the decadal scale impacts of a shift from a C3-
dominated system to a C4-dominated system on soil properties and processes (Chapter

Two), 2) How quickly do differences in species traits, and the soil conditions observed

 

Plant

 

 

 

NPP Resource Use ‘~\

 

  
   

  
 

\\\ E

.\ m

\\\ a

' ' Resource Use - - \‘t E
Litter Quantity ;;;;;;;;;;;;;;;;;;;;;;2 Litter Quality 39.;
(Root & Shoot Plant Available Nutrients (tissue CzN, lignin, fig)
biomass. surface litter) root 'exudates) / g
l I x" S.

. ‘. ,fSoil $-

SOII \\ . I, pH I” 3-
Microclimate‘. C & N Mineralization ,' 0’

(moisture& ‘~\
temperature) ‘\\

\

‘A
Soil Microorganisms

(activity, composition, abundance)

 

Soil C & N Pools

(Total, labile & recalcitrant pools)

 

 

 

Figure 1. Conceptual diagram of how changes in plant species composition can influence
ecosystem processes in terrestrial ecosystems. Plant species may directly (via differences
in litter quality and quantity) and indirectly (via effects on soil microenvironment) exert
important controls'over the fiinctioning of microbial communities that determine nitrogen
(N) and carbon (C) cycling.

after 11 years, arise (Chapter Three), and 3) Which plant traits are responsible for
differences in decomposition rates (Chapter Four)? This research was performed at
several old-fields located at the W. K. Kellogg Biological Station (Appendix 1). All sites
were abandoned more than 35 years ago following decades of row crop agriculture
(Burbank et al. 1992, Foster and Gross 1997). The old-fields were dominated by non-
native C3 grasses, but had patches of C4 grass monocultures that were established in
1995 to examine plant competition (Foster 1996, 1999). The C4 monoculture plots were
still intact in 2005 at two sites, allowing me to compare how C4 versus C3 grass-
dominated communities impact soil properties on a decadal time scale at two old-fields.

In Chapter Two, I examine the decadal-scale effects of the reintroduction of C4
grasses on soil C and N cycling in the two old-fields with intact patches of C4 grasses. I
utilized the C4 monocultures established in 1995 to compare plant traits and soil
properties in the C4 monocultures to the surrounding C3 grass-dominated matrix
community. I expected that: l) the three C4 species would produce greater biomass and
consequently would have greater surface litter and total soil C than the surrounding C3-
dominated community, and 2) slowed microbial process rates associated with the
recalcitrant tissue of the C4 species would result in larger total C and N pools, and
inorganic N levelsiwould be lower due to greater microbial N immobilization.

In Chapter Three, I examine whether trait differences between C4 and C3 grass
species develop within the first two growing seasons, and if soil processes begin to reflect
these species differences immediately. 1 established experimental monocultures in all
three old-field sites in 2005, and related tissue production and chemistry to differences in

soil properties after two growing seasons to determine the short-term effects of plant

species on soil properties. I predicted that species biomass and tissue chemistry
differences would be detected within l-2 yrs after establishment, and that these
differences would translate rapidly into detectable changes in soil inorganic nitrogen
pools. In contrast, I expected that changes in soil C and N pools would show a lag
between when process rates change and when the total pool would begin to reflect those
changes. I hypothesized that the magnitude of species biomass and tissue chemistry
differences would determine the rate at which the soil changes are detected.

In Chapter Four, I evaluate the ways in which different traits associated with C4
and C3 species can influence litter decomposition to determine the relative importance of
litter quality and soil microclimate on plant decomposition rates. I performed two
experiments to separate the effects of litter quality and soil microclimate on
decomposition. In the first experiment, the decomposition rates of the two litter types
(C3 and C4) in a common location (either under C3 species or under C4 species) were
compared to examine both litter quality and soil microclimate effects on decomposition
rates. I expected C4 litter to decompose more slowly than C3 litter because C4 litter is
more recalcitrant (higher C:N, lignin:N). I also predicted that decomposition rates would
be slower under C3 than C4 species because the smaller litter layer and lower
aboveground biomass of the C3 species would make these environments less
advantageous for microbial activity (i.e., drier and hotter) in the summer than under C4
species. In the second experiment, I measured the soil microclimate (soil moisture and
temperature) under C3- and C4-dominated plots, with and without surface litter, to
determine if differences in surface litter and aboveground biomass impacted microclimate

in a manner that affected decomposition of a standard substrate, cellulose. Summer and

fall soil moisture levels and soil temperature were expected to follow a gradient of
relative ground cover, with highest soil moisture and lowest temperatures under the C4
species with litter intact, followed by C4 species without litter, then C3 species with litter
and finally C3 species without litter. I expected cellulose decomposition to be faster
under C4 species if soil moisture limited decomposition and the decomposition rates to be
fastest in the C3 species plots without litter if temperature was the primary factor
determining decomposition rates.

The next three chapters in this dissertation describe the results of field
experiments that explore how introductions, or re-introductions, of species that differ in
functionally important traits (e. g., litter quantity and quality) can alter soil processes, and
the timescales over which such changes can be expected to occur. I focused on C3 versus
C4 grass comparisons as a model system to examine how different plant traits influence
soil processes, which has specific application to a variety of areas, including prairie
restoration, climate change alterations of C4-dominated grassland distribution, and the
recent focus on using a C4 prairie grass (switchgrass or Panicum virgatum L) as a biofuel
source. For example, the current interest in biofuels could result in large expanses of the
Midwestern United States being converted from agriculture, old-fields, or conservation
properties (i.e., Conservation Reserve Program lands) into C4 grass monocultures. Given
the magnitude of the landscape that may be affected by such changes in species
distributions, it is important that we understand how those changes may alter soil
properties and processes.

Further, the relationships I show between plant traits and soil processes extend

beyond simply native and non-native comparisons or C3 and C4 comparisons, and are

generally applicable to other species and systems where a “new” dominant species differs
from the previous dominant species by functionally important traits. Improving our
understanding of how plant species impact ecosystem properties and what species traits
are driving these changes is imperative if we hope to predict the ecosystem-level
consequences of changes in species distribution or composition that are predicted to
occur, and are occurring, as a consequence of changes in agricultural and land use

practices, global change, and species introductions.

CHAPTER TWO
DECADAL SCALE IMPACTS OF C4 GRASS REINTRODUCTIONS ON SOIL

CARBON AND NITROGEN CYCLING IN SUCCESSIONAL ECOSYSTEMS

ABSTRACT. While much recent research has focused on the effects of exotic plant species
on ecosystem properties, little is known about how reintroductions of native species may
impact these processes in restored systems. I examined how the reintroduction of three
native C4 grasses into old-fields affected soil carbon (C) and nitrogen (N) cycling 11
years after their reintroduction compared to unmanaged successional communities
dominated by non-native C3 grasses in southwestern Michigan, USA. The C4 species
(Andropogon gerardii (Vitman), Sorghastrum nutans (L), Schizachyrium scoparium
(Michx)) and C3 species (Bromus inermis (Leyss), Elymus repens (L)) examined in this
study differ significantly in many traits that are expected to influence soil C and N
cycling, and led me to hypothesize that soil C and N pools and cycling rates would differ
under the two groups of species. As predicted, the three C4 species had significantly
greater root and shoot biomass, and more recalcitrant tissue compared to the dominant C3
species. Soils under the three C4 species had significantly lower inorganic N levels, and
Andropogon had slightly lower in situ net N mineralization rates than soils under C3
species. I also found little evidence for increasing soil C pools under C4 species 1 1 years
after their reintroduction. Overall, these results show that reintroduction of C4 species
into grasslands can result in alterations of soil processes related to C and N cycling on
relatively short timescales. Improving our understanding of how plant species impact

ecosystem properties and what species traits are driving these changes is imperative if we

hope to predict the ecosystem-level consequences of changes in species distribution or
composition that could occur, and are occurring, as a consequence of changes in

agricultural and land use practices, global change, and species introductions.

INTRODUCTION

Human activities, such as land use change and the introduction of exotic species, are
altering biodiversity on a global scale (Lawton and May 1995, Pimm et al. 1995, Chapin
et al. 2000, Hooper et al. 2005). However, the ecological implications of such shifts in
plant species distributions and abundances are poorly understood. A number of studies
have shown that the introduction of an exotic species can dramatically and rapidly alter
ecosystem properties (Vitousek and Walker 1989, Evans et al. 2001 , Mack and
D'Antonio 2003a), but few studies have focused on the reverse: how the reintroduction of
native species affects ecosystem properties (sensu Hooper et al. 2005). This study
examines how the reintroduction of native prairie grasses into abandoned agricultural
fields alters soil carbon (C) and nitrogen (N) cycling after 11 years.

Agriculture has substantially reduced and fragmented prairie systems throughout
the United States (Mlot 1990, Samson and Knopf 1994), altering plant communities and
ecosystem properties (Camill et al. 2004, DeGryze et al. 2004). In southwestern
Michigan, agricultural development has restricted once-common, native C4 grasses
(Gotshall 1972) to prairie remnants, and these species are now rarely found in abandoned
agricultural fields (old-fields). Old-fields throughout the Midwest are typically colonized
by a successional trajectory of C3 species, many of which are non-native (Inouye and

Tilman 1988, 1995, Foster and Gross 1997, Averett et al. 2004, Gross and Emery 2007).

While C3 and C4 functional groups typically differ in many traits that are expected to
influence ecosystem properties such as C and N cycling, little is known about the
ecosystem-level impacts of the loss of C4 grasses, in part because their extirpation is
confounded with agricultural disturbance. Currently, large tracts of former agricultural
land are being reverted to C4—dominated communities (e.g., prairie restoration and the
USDA Conservation Reserve Program) and many climate change models predict shifts in
C4 species distribution (Collatz et al. 1998, Winslow et al. 2003). In addition, the
growing interest in biofuels as an alternative energy source is likely to increase the
acreage planted to native C4 grass monocultures such as Panicum virgatum (Samson et
al. 2005, Sanderson et al. 2006, Tilman et al. 2006). Thus, it is important to understand
how the re-establishment, or introduction, of particular C4 species influences ecosystem
properties compared to a common C3-dominated old-field community.

There is considerable evidence that species’ functional characteristics drive
important ecosystem properties (Hooper and Vitousek 1998, Reich et al. 2004, Wardle et
al. 2004, Hooper et al. 2005). Two common ways in which plant species influence soil
processes is through the quality and quantity of their litter (Hobbie 1992, Wardle et al.
1998, Dijkstra et al. 2006). Changes in litter quantity (via net primary productivity) and
quality (via allocation and partitioning of C and N into various tissues, determining tissue
chemistry) can directly and indirectly influence microbial community activity,
abundance, and composition (Zak et al. 2003, Carney and Matson 2005, Hooper et al.
2005, Zavaleta and Hulvey 2007), thereby altering nutrient cycling rates, and potentially

feeding back to alter the plant community (Figure 1).

10

The C3 and C4 grasses examined in this study provide an opportunity to examine
how plant traits are linked to soil properties because they have similar growth forms yet
differ in a number of functionally important traits. C4 grasses typically produce more
biomass (above- and below- ground) and have more recalcitrant tissue (i.e., C:N,
lignin:N) than their C3 grass counterparts (Wedin and Tilman 1990, Baer et al. 2002,
Camill et al. 2004). These traits can directly influence soil processes via changes in the
amount and form of substrates available for microbial utilization (Zak et al. 2003, Carney
and Matson 2005, Hooper et al. 2005, Zavaleta and Hulvey 2007). C3 and C4 species
also may indirectly impact the soil microbial community via differential effects on the
soil microenvironment and timing of resource uptake and release back to the soil.
Indeed, several studies have found lower net N mineralization and higher C
mineralization rates over relatively short timescales in grasslands as C4 species become
dominant (Baer et al. 2002, Camill et al. 2004).

Here, I examined the decadal-scale effects of the reintroduction of three C4
grasses on soil C and N cycling in two southwestern Michigan old-fields. Both fields
were abandoned from agriculture over 35 yrs ago and were used for an experiment in
1995 to examine colonization and growth of three native C4 species (Foster 1999). The
C4 monoculture plots were still intact in 2006, allowing me to compare plant traits and
soil properties of the C4 monocultures to the surrounding C3 grass-dominated matrix
community after 11 years. Given the expected differences in functional traits of the C3
and C4 grasses, I predicted that: 1) the C4 species would produce greater biomass (root
and shoot), which would cause an increase in surface litter and total soil C compared to

- the C3-dominated communities, 2) the C4 species would have more recalcitrant tissue,

11

which would slow microbial process rates and result in a buildup of total soil C and N
pools, and 3) the more recalcitrant tissue of the C4 species would cause microbial N

immobilization and therefore result in smaller inorganic N pools.

METHODS
Study Sites
I compared experimental monocultures of three C4 species (A ndropogon gerardii
(Vitman) or Big bluestem, Sorghastrum nutans (L) or Little bluestem, and Schizachyrium
scoparium (Michx) or Indian grass) to the surrounding old-field community dominated
by C3 grasses in two old-fields at Michigan State University’s W. K. Kellogg Biological
Station (KBS) in southwestern Michigan, USA. Non-native C3 grasses dominated both
fields, though the dominant species identity differed: the Turkey Meadow site is
dominated by Bromus inermis (Leyss; Smooth brome), while McKay Field is dominated
by Elymus repens (L; Quackgrass) (Table l). Nomenclature for all species follows the
USDA Plants Database (plants.usda. gov). Both fields were abandoned over 35 years ago
following decades of row crop agriculture (Burbank et al. 1992, Foster and Gross 1997).
This area of Michigan had extensive prairies and savannas prior to agricultural
development, and the C4 species examined in this experiment were common components
of those grasslands (Gotshall 1972, Burbank et al. 1992).

The experimental monocultures were established in 1995 for competition
experiments described in Foster (1999). Species were transplanted into clipped plots
with minimal soil disturbance, weeded for one year, and then abandoned in 1996. After

1 1 years, the experimental monoculture plots in both fields were still dominated by the

17

~

Table 1. Plant community characteristics of the plots at harvest in August 2006. Shoot
biomass (mean t SE; n=9) and species percentages of the total biomass are based on the
average of nine samples for each dominant vegetation plot. Species richness is given as
the mean of the plots (n=9; 0.5 m x 0.5 m) and as the total across plots (mean, 2). Plot
names are abbreviated as follows: Ag=A ndropogon, Ss=Schizachyrium,

Sn=Sorghastrum, Er=Elymus, and Bi=Br0mus. A designation as C3 or C4 species and as
monocot (M) or Forb (F) follows the species names.

 

 

- S ecies
Shoot biomass o p
Site Plot -2 Dominant species 1? Shoot richness
(g m ) Iomass (mean, 2)
Ag 1637 2162 Andropogon gerardii (Vitman) C4 M 97.5 2, 4
Elymus repens (L) C3 M 1.6

$5 577 1- 68 Schizachyrium scoparium (Michx) C4 M 85.5 3, 9

 

McKay Elymus repens C3 M 9.8
Sn 442 x 49 Sorghastrum nutans (L) C4 M 76.7 4, 10
Elymus repens C3 M 16.3
Er 307 2 18 Elymus repens C3 M 98.7 2, 5
Achillea millefolium (L) C3 F 1.0
Ag 1570 :129 Andropogon gerardii C4 M 95.8 3, 8
Bromus inermus (Leyss) C3 M 2.8
SS 577 z 56 Schizachyrium scoparium C4 M 68.9 8, 22
Bromus inermus C3 M 8.7
Turkey Solidago canadensis (L) C3 F 5.0
Sn 959 z 108 Sorghastrum nutans C4 M 84.4 7, l3
Poa pratensis (L) C3 M 5.0
Bromus inermus C3 M 4.7
Bi 282 2 27 Bromus inermus C3 M 74.4 4, 13
Poa pratensis C3 M 21.3

 

l3

C4 species, and the surrounding matrix remains dominated by C3 species (Table 1). Both
fields have sandy loam soils (Foster and Gross 1997), although McKay Field has a higher
sand fraction and appears more drought-prone than Turkey Meadow. By using two
overall similar Sites with different dominant C3 species, I hope to increase my ability to

generalize about the results.

Field Sampling
In 2006, I randomly selected nine of the monoculture plots (0.5m x 0.5m) dominated by
each of the C4 species, Andropogon gerardii (Andropogon plots), Sorghastrum nutans
(Sorghastrum plots) and Schizachyrium scoparium (Schizachyrium plots) in each in field.
Nine additional plots were established in the surrounding C3 matrix community (C3
plots), within 10-12m of the C4 plots. I determined aboveground biomass production in
August 2006 by clipping all vegetation at ground level and separating individual species.
I then collected surface litter from the plots. Two soil cores (0-20cm deep, 3.80m
diameter) were collected from each plot for soil chemical analyses. Both cores were split
into two depths (0-10, 10—20cm), and the soil from each depth interval was combined and
refrigerated until processed in the lab. I determined root biomass by taking a single core
(6.35cm diameter) from a subset of plots (n=6), split it into two depths (0-10 and 10-
200m), and refrigerated it until roots could be separated from the soil. All soil cores were
taken immediately after sampling the litter.

Seasonal patterns of species effects on N-availability were determined using
repeated 28-day in situ net N mineralization incubations throughout the 2006 growing

season. In situ incubations were done in the Andropogon and C3 plots (0-10cm cores,

14

n=5) in both Sites. At the onset of each incubation period, two PVC pipes (3.8cm
diameter, sharpened on one end) were pounded into the ground to a 10cm depth. The
first core (to) was removed and taken to the lab for soil inorganic N analyses. The second
core (Ifmal) was removed, capped on the top and bottom, and placed back into its original
hole (modified from Robertson et al. 1999). After 28 days, the If‘ma] core was removed
from the ground, sealed in a plastic bag, and refrigerated until processed in the lab. On
this same day, a new set of to and If‘ma] cores were collected and installed, respectively.

This process was repeated every 28 days from June through November 2006.

Laboratory Analyses
Aboveground biomass (separated by species) and surface litter were dried for at least 72h
at 65C and weighed (10.01 g). Green tissue samples were taken from a subset of the
harvested biomass (n=5 plots) for each dominant species —Andr0p0gon, Schizachyrium,
Sorghastrum, Bromus (in Turkey Meadow), Elymus (in McKay Field). The dried tissue
was coarse ground in a Wiley Mill, then ground to <2mm on a Cyclotech Grinder, and re-
dried for 48 hours at 65C. Oven-dried tissue (2-3mg) was then packed in tin capsules for
C, N and isotope analyses (analyzed at the UC Davis Stable Isotope Facility). Acid
Detergent Fiber (ADF; recalcitrant compounds, primarily lignin and hemicellulose)
analyses of ground tissue (~0.5 g) were performed on an Ankom 2000 Fiber Analyzer
(Macedon, NY) at Michigan State University.

Root cores were washed in tap water and roots were floated in a pan, removed
with tweezers, and rewashed to remove any remaining soil. Root material was then

placed in metal weigh boats, dried for 48h at 65C, and weighed (:0.0001g). Rootzshoot

15

was calculated by dividing root biomass (scaled to g m-2 for 0-20em depth) by shoot
biomass (g mg) in each plot. Total plant N was estimated by multiplying total biomass
(root + shoot) by shoot %N value of the dominant species. I used the dominant species
shoot %N to approximate total plant N because it comprised the majority (69-99%) of the
total shoot biomass and I did not have root %N data. While root and shoot %N data may
differ, Wedin and Tilman (1990) Show that %N for Elymus, Andropogon and
Schizachyrium shoots was generally similar to that for roots.

Soil cores collected for chemical analyses were sieved through a 2mm soil sieve
to homogenize and remove large debris. Inorganic N was extracted from a 20g sub-
sample of soil using 50ml of 1M KCl, within 24h of sample collection. Extracts were
placed in the freezer until analysis on an 0.1. Analytical Flow Solution IV analyzer.
Gravimetric soil moisture was determined on another sub-sample of soil (~25 g fresh
weight) by drying soils at 105C for 48h. The remaining soil was air-dried and stored in
the laboratory. For C, N and isotope analyses, ~50g of air-dried soil was ground to a
flour-like texture using a roller mill and oven-dried at 65C for 48h. 20-50mg sub-
samples were then packed into tin capsules and sent to the UC Davis Stable Isotope
F aeility for determinations of total C, N and 813C (relative to PeeDee Belemnite (PDB)).
Because C3 and C4 species differ markedly in OBC, the percentage of soil C contributed
by C4 species could be calculated for each plot using a simple end-member mixing model
(using the 613C ofthe dominant C4 and C3 species in each plot):

%C4 signal = (C4 Soil 613C — C3 Soil 613C)/(C4 plant 5'3C - c3 Soil 613C)*100
Soil bulk Density was determined at each site under the Andropogon and C3 plots (n=3)

using an Eijkelkamp root corer (8cm diameter *10cm deep). Bulk density was calculated

l6

as oven-dried mass/volume (g cm-3). Bulk density was used to convert surface soil C
and N to a mass basis.

For each in situ net N mineralization incubation, t0 and lfina] samples were taken
back to the lab and processed according to the inorganic N extraction procedure
described above. To calculate final N pool sizes, the sum of inorganic N in the Ifina] core
was divided by the. mass of the soil in the core (calculated using bulk density). Net N-
mineralization and nitrification rates were calculated as the change in total inorganic N
(N H4+ and N03) and NO3-N (in ,ugN per gram of dry soil), respectively, over the

incubation period.

Statistical Analyses

All data were checked for normality and equal variance, and appropriate transformations
were performed prior to analysis. Plant and soil variables were compared between
dominant species (C3, Andropogon, Schizachyrium, and Sorghastrum) using a Two-Way
ANOVA with Site and Species (or Plot) as main effects using SigmaStat 3.5. Pearson
correlations were performed to determine whether plant traits were correlated with soil
variables. In addition, in situ N mineralization and nitrification rates were compared
between Andropogon and C3 plots using Repeated Measures ANOVA on Systat 11. For
any ANOVA indicating a significant interaction, post-hoe contrasts were made using

Tukey comparisons.

l7

RESULTS

Species Traits: Biomass and Tissue Characteristics

There were Significant differences among the dominant species in shoot biomass and
surface litter with both typically greater in plots dominated by a C4 species than by C3
species (Table 2; Figure 2). A na’ropogon plots had significantly greater root biomass (0-
20cm depth) than C3 and Schizachyrium plots in both sites (Table 2, Figure 2). Analyses
of plant tissue chemistry separated C3 from C4 species. Plant tissue C:N was
significantly higher for the C4 species compared to the C3 species, and %N was
significantly lower for the C4 species (Table 3). Isotope analyses showed that C4 species
had a significantly higher 513C (relative to PDB) compared to the C3 species, indicating a
greater discrimination by C3 species against 13C (Table 3). All C4 species had higher
%Acid Detergent Fiber (ADF) and ADF:N than the C3 species, except Sorghastrum in
McKay Field (Table 3). Estimates of total plant N (Shoot + root) on an area basis were
not significantly different between species in McKay Field (p>0.05), but in Turkey
Meadow, Andropogon had more total N than both Schizachyrium and Sorghastrum
(p30.009), and Schizachyrium had more than Bromus (p=0.029). C3 species had
significantly greater rootzshoot compared to all C4 species, with Andropogon having the
lowest ratio ofthe C4 species (p<0.001).

There were also site differences in the growth and tissue chemistry of the species
(Figure 2, Tables 2 and 3). All species had significantly greater root biomass and
rootzshoot in McKay Field than Turkey Meadow, and C :N and 613C was significantly
higher for all Species in Turkey Meadow. Sorghastrum had higher shoot biomass and

ADF:N in Turkey Meadow, and Significantly higher %C and N in McKay Field.

18

Table 2. ANOVA results for various measures of plant production, using Site and
dominant species Plot as main effects. Natural log transformations were used to
normalize the data for all variables. Significant p-values are indicated in bold.

 

Degrees

 

Variable Factor F of p-value
freedom

Site 6.14 1, 64 0.016

Shoot Biomass Plot 113.43 3,64 <0.001

Site*Plot 9.71 3, 64 <0.001

 

Site 10.05 1, 64 0.002
Surface Littfl Plot 39.38 3, 64 <0.001

Site*Plot 3.72 3, 64 0.016

 

 

Site 33.22 1, 40 <0.001
Root Biomass

Plot 5.59 3, 40 0.003
(0-20cm)

Site*Plot 1.64 3, 40 0.195

Site 26.20 1, 40 <0.001
Root : Shoot Plot 16.63 3, 40 <0.001

Site*Plot 2.61 3, 40 0.064

 

l9

McKay Field Turkey Meadow

 

3000
2000 -
1000 -

Dghoot DRootO-10cm lRoot 10-20cm DShoot El Root O—lOcm I Root 10-20cm
a

A

O
1

Biomass (g m‘2
N H
o o
O O
o o
I l

3000 -
4000 a ab b
Ag 85 Sn C3 Ag 55 Sn C3

Plot Plot

           

 

 

 

 

 

b

 

Figure 2. Root and shoot biomass (mean :SE in g m-Z) in August 2006 at both sites.
Root biomass is shown separately by depth intervals here for additional detail, but
statistical results are reported for total root biomass from O-200m. Lowercase letters
denote significant differences (p<0.05) between plots for shoot (n=9) and root (n=6)
biomass. Species names are abbreviated as follows: Ag=Andr0pogon,
Ss=Sch1'zachyrium, Sn=Sorghastrum, and C3=E1ymus in McKay Field and Bromus in
Turkey Meadow.

20

Table 3. ANOVA model for plant tissue chemistry, using Site and the dominant Species
as main effects. Mean (2SE) values are also shown for the dominant species at each site.
When Species effects are present, superscript numbers indicate significant differences
(p<0.05) between species. For significant interactions, superscript letters denote
significant differences (p<0.05). The Species names are abbreviated as follows:
Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=S0rghastrum, Er=Elymus, and Bi=Br0mus. a
Sample sizes for nitrogen (N), carbon (C), C:N, Acid detergent fiber (ADF):N and 013C
were 4 for each species, while n=6 for ADP.

 

l
Site Species %N %C C:N %ADF ADF:N 0 3C

 

Site F1,24=l7.54 F] 34:83.58 F1924=28.99 F1932=32.47 F1,24=14.79 F1,24=4.67
p<0.001 ' p<0.001 p<0.001 p<0.001 p<0.001 p<0.041
Model Species F3’24=29.97 F3924=56.21 F3’24=l440 F3932=31.23 F3,24=21.62 F3,24=l451.36

p<0.001 p<0.001 p<0.001 p<0.001 p<0.001 p<0.001
Site*

F3’24=3.42 F3,24=35.72 F3,24=2.15 F3’32=3.10 F3,24=5.59 F3924=2.61
Species

p=0.033 p<0.001 p=0 1 20 p=0.040 p=0.005 p=0.075

 

ab a 1 ab a 1
Ag 08020.08 45.5202 58.5263 79.0213 10042109 —12.7201

a a _ l b a 2
Ss 09020.08 45.3203 51.0244 81.1204 91327.7 —l3.520.5
McKay
a a 1 ac ab 2
Sn 1.012009 44720.5 45.2247 77.521.l 77327.5 -l4.020.5
c a 2 c b 3
Er 1.262003 45.7204 36.5206 74.6208 59721.9 -28.7201

 

ab b 3 d ac 4
Ag 07420.06 49.4205 67.7251 74.3205 101.9282 -12.520.l

b a 3 b c 5
Ss 06720.04 46.2208 69.5245 79.121.4 118.8275 -13.9204
Turkey
b c 3 bd c 5
Sn 0.642005 43.6203 69.3246 75921.0 121.8288 -12.8204
c d 4 e b 6
Bi 1.212010 51.9202 43.5234 68321.1 57.9249 -28.120.2

 

21

Schizachyrium had significantly higher %N and surface litter in McKay Field, and higher
ADF:N in Turkey Meadow. Both A ndropogon and the C3 species had significantly
higher %ADF in McKay Field and significantly higher %C in Turkey Meadow. Plant
total N was significantly greater in McKay Field (except Andropogon). Plant Shoot
biomass was highly correlated with surface litter (n=48, p<0.001, F070), and plant 013C
values were highly correlated with ADF:N (n=3-, p<0.001, r=0.74), %N (n=32, p<0.001,

=-0.79), C:N (n=32, p<0.001, r=0.65) and ADF (n=32, p<0.001, r=0.62)

Soil Properties

Surface (0-10cm) and subsurface (10-20cm) soils sampled from under the different
species did not significantly differ in total soil C content (Tables 4 and 5). However, both
surface and subsurface soil 613C values were significantly enriched under the C4 than C3
species, and there were also differences between the soil 013C values among the C4
species (Tables 4 and 5). The proportion of both surface and subsurface total C pools
contributed by C4 species was significantly higher in plots that had C4 species planted to
them a decade ago compared to the C3 plots (Figure 3).

Soil inorganic N was significantly lower in both surface and subsurface soils
under C4 species than under the C3 species in both sites (Tables 4 and 5). Both NH4+
and N03 typically were higher in soils under C3 species compared to C4 species (data
not shown). In situ net nitrification and N mineralization rates (performed in C3 and
Andropogon plots) varied over the growing season (p50002 using Greenhouse-Geisser
adjustment), and typically were higher in soils under C3 species than under A ndropogon.

However, these differences were only marginally significant (p=0.09 and 0.06,

22

respectively; Figure 4) in part due to high variability among replicates. Total subsurface
soil N pools did not Significantly differ between species, but surface N pools were
significantly greater under A ndropogon compared to Sorghastrum (Tables 4 and 5).

In additionto species effects on soil properties, the sites also differed in soil
properties (i.e., significant Site effect; Tables 4 and 5). Across plots, surface soil C and N
pools were significantly larger in Turkey Meadow, although subsurface pools did not
differ between sites. In Turkey Meadow, surface soils had significantly higher 513C, and
subsurface soils had significantly lower C:N compared to McKay Field. In situ net
nitrification differed through time but was significantly higher in McKay Field than
Turkey Meadow for all incubations (p=0.04 using Greenhouse-Geisser adjustment). The
calculated proportion of total surface soil C contributed by C4 species also was
significantly higher in Turkey Meadow than McKay Field (F1,40=4.41, p=0.042).

Several plant traits were highly correlated with particular soil variables. Shoot
biomass was correlated with both soil 613C (surface soil: n=23, p<0.001, r=0.61;
subsurface: n=21, p<0.001, r=0.63) and %C4-carbon contribution to the soil C pool
(surface: n=23, p<0.001, r=0.59; subsurface: n=2l, p<0.001, r=0.62), and subsurface soil
N (n=21, p=0.01, F056) and C content (n=21, p=0.03, r=0.48). Shallow (0-10cm) root
biomass was positively correlated with %C4-carbon in surface soils (n=3 5, p=0.04,
r=0.35), and root biomass from 10-20cm was positively correlated with %C4-carbon in
subsurface soils (n=30, p=0.03, r=0.41) and subsurface soil 013C (n=30, p=0.01, r=050).
Tissue C:N was negatively correlated with inorganic N (surface soil: n=2l , p=0.03,
r=0.48; subsurface: n=21, p=0.03, r=0.48), and tissue N content was positively correlated

with inorganic N (surface: n=32, p=0.02, r=0.4l; subsurface: n=32, p=0.01. r=0.46).

23

Table 4. ANOVA model results for surface soil (0-10cm) carbon (C) and nitrogen (N)
variables, using Site and Plot (dominant vegetation) as main effects. Mean (2SE) values
are shown below the ANOVA. When Species effects are present, superscript numbers
indicate significant species differences (p<0.05). For significant interactions, superscript
letters denote significant differences (p<0.05). The Plot names are abbreviated as
follows: Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Er=Elymus, and
Bi=Br0mus. The data shown are untransformed values, but inorganic N was In-
transformed prior to analysis.

 

 

 

 

Site Plot Inorganic N _1 C:N C (kg m'z) N (kg m'z) 013C
(ugN g dry soil )
Site F1964=3.21 F1’40=1.90 F1,40=22.97 F1,40=21.07 F1,40=4.21
p=0.078 p=0176 p<0.001 p<0.001 p=0.047
Model Plot F3,64=9.37 F3,40=7.33 F1,40=3.08 F1,40=3.68 F1,40=20.39
p<0.001 p<0.001 p=0.038 p=0.020 p<0.001
8116* F3’64=l .81 F3,40=3.20 F1’40=3.l l F1340=2.7I F1,40=0.95
Plot p=0.154 p=0.034 p=0.037 p=0.058 p=0.428
1 ab 3 l 1
Ag 3.912043 113:0.2 1.942018 01720.02 -24.62O.6
McKay 83 3.5520471 11.2201ab 14920.213 01320.0212 25.82052
Fleld Sn 4.1120431 11.5201a 14220.1131 01220.012 45.72032
Er 7.0820602 11.2201ab 20820.20“) 01920.0212 07.12013
4 1 a b 3 4
Ag 3.122037 11.420.3 2.7120.34 02420.03 -23.32O.8
1 a b 34 5
Turkey 83 3.952081 10.9203 2.182013 02020.01 25.5203
Meadow Sn 4.172039l 12420.3b 2.39: 0.10b 0.19: 0.014 04.92045
7 ‘M
Bi 4.832055" 11.20281 20420.15ab 018200134 47.12015

 

24

Table 5. ANOVA model for subsurface soil (10-20cm) carbon (C) and nitrogen (N)
variables, using Site and Plot (dominant vegetation) as main effects. Mean (28E) values
are shown below the ANOVA. For significant Species effects, superscript numbers
indicate significant species differences (p<0.05). The Plot names are abbreviated as
follows: Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=Sorghastrum, Er=Elymus, and
Bi=Br0mus. The data shown are untransformed means, but the follow3ing variables were

transformed prior to analysis: inorganic N (In), N (square root) and 61 C (square root).

 

Inorganic N

13

 

 

 

' . -l -l
3‘“ Plot '(ugNgdry C-N C(gkg 1 thkg ) 5 C
soil-)
Site F],64=2.89 F],32=124.66 F1332=011 F1’32=l.52 F],32=3.44
p=0.094 p<0.001 p=0.745 p=0227 p=0.073
Model p10, 1346425133 F3332=076 1:132:097 131,324.15 1532:1731
p=0.001 p=0.527 p=0.421 p=0.344 p<0.001
Site* F3,64=0.79 F3,32=1 81 13132-1120 F],32=l.32 F1,32=1.24
P1“ p=0.505 p=0.165 p=0326 p=0285 p=o.312
Ag 1.1520251 11.5101 8.832129 0.772012 04.5203]
1
McKay ss 0.922010 17.020.2 7.7021.46 0.652013 -25.120.2:2
. . 1
Fleld Sn 08620.08 11.8202 6.322095 01220.53 24.820312
I) 4»
Er 1.822022" 11.5202 8.262096 0.722008 05.7201"
Ag 0.8320191 10.420.1 7.512022 0.722002 03720.51
1
Turkey Ss 10720.27 10.3203 6.9972057 06820.05 -24.9=o.12
Meadow Sn 0.932021 10.5202 7.742048 07420.05 24.5202”
7 5
Bi 1.292019” 10.5202 7.582038 07220.04 25.8202"

 

25

McKay Field

 

30

—I-—Ag - + -Ss —-k- Sn —0—C3

N
O

 

°/oC from C4 species
ES

 

O
1

 

 

 

 

-10
Surface Soil Subsurface Soil
Soil Depth
40 Turkey Meadow

+Ag - + -Ss -—t— Sn —e—C3

%C from C4 species

 

 

 

 

 

-10
Surface Soil Subsurface Soil

Soil Depth

Figure 3. Percentage of total soil carbon (mean 2SE) contributed by C4 species at both
surface (0-10cm; n=6) and subsurface (10-20cm; n=5) soil depths at both sites.
Lowercase letters denote significant differences (p<0.05) between plots separately for
each depth. Plot names are abbreviated as follows: Ag=Andr0p0g0n, Ss=Schizachyrium,

Sn=Sorghastrum and C3=Elymus in McKay Field and Bromus in Turkey Meadow.

26

 

.0
Ox

0

A
;
>——————4

'

 

 
 

I..___.___

 

.0
N
1

 

 

 

 

N Mineralization (ugN g dry soil’1 day'l)

.0
0

June July Aug Sept‘ Oct

0.4

 

0.3 ~

0.2 -

0.1 -

 

 

 

 

 

Nitrification (ugNO3_N g dry soil"1 day'l)

June July Aug Sept, Oct

Figure 4. In situ net nitrification and N mineralization rates (mean 2 SE; n=5) for

Andropogon plots (solid lines) and C3 plots (dashed lines) at both Turkey Meadow (black
lines) and McKay Field (gray lines). Rates are calculated from 28-day in situ incubations
performed consecutively from June to November 2006.

27

DISCUSSION
Eleven years after the establishment of native C4 Species, there were detectable
differences in soil processes in areas planted as C4 monocultures compared to the
surrounding C3-dominated matrix community. These differences correspond to
differences in biomass production and tissue chemistry between these two functional
groups of grasses. While the magnitude of the differences in soil properties and Species
characteristics betWeen the dominant species plots sometimes varied between the two
sites, the qualitative interpretations (i.e., the relative differences between species) of the
results were remarkably Similar. This suggests that the differences I observed between
species can be generalized across sites. In addition, members of each functional group
(C3 and C4 species) were typically similar to one another for both plant characteristics
and impacts on soil properties, suggesting that the species within each functional group
may be functionally equivalent (for the particular species examined in this study).
Consistent with my expectations, the C3 and C4 species differed in both the
quantity and quality (C:N, ADF:N) of biomass and litter produced. Total shoot biomass
was 0.5 to 8-fold higher in plots with C4 compared to C3 species, and C4-dominated plots
tended to have much larger total root biomass and more investment in deeper root
systems than C3-dominated plots (Figure 2). Many studies (e.g., Baer et al. 2002, Camill
et al. 2004) have found higher biomass (root and shoot) and surface litter accumulation in
plots as C4 species abundance increases. Overall, the predicted increase in tissue
quantity with C4 Species reintroduction was supported by my results, as was the
prediction that C4 species have more recalcitrant tissue (higher C:N and ADF:N) than the

C3 species.

28

I found a strong positive correlation between biomass (both root and Shoot) and
the percentage of soil C contributed by C4 Species, su vgesting that perhaps C4 species are
affecting soil C. The increase in tissue quantity associated with the C4 species, and
_ therefore potential contributions to the soil C pool, might be expected to stimulate
microbial activity; however, the greater recalcitrance of this tissue may suppress C
mineralization rates. The interaction of litter quality and quantity differences on soil
processes in this study confounds my ability to independently examine how litter quality
and quantity impact C cycling, and thus how C pools may change. Several studies have
shown that decomposition is highly correlated with tissue C :N and lignin:N (Wardle et al.
1997, Vinton and Goergen 2006), so C4 tissue should decompose slower than C3 tissue
and result in increased soil C. However, recent work in restored prairies has found higher
C mineralization rates in sites that have high levels of C4 species dominance (Baer et a1.
2002, Camill et al. 2004). One explanation for this apparent contradiction is that the
larger rooting system Of C4 species, compared to C3 species, results in larger quantities of
labile material being released to the microbial community (e.g., fine root turnover and
exudation), subsequently stimulating C mineralization in the rooting zone (Baer et al.
2002). At the same time, the greater total production and recalcitrance of C4 species
shoot and coarse root tissue acts to slow its decomposition relative to C3 Species. Thus,
while C mineralization rates may be higher in the rooting zone of C4 species, overall
decomposition is slower, and soil C subsequently accumulates faster in C4-dominated
plots compared to the C 3-dominated plots.

Despite these differences in tissue quality and quantity, I did not find Significant

increases in total soil C in the C4 plots compared to the C3 plots in this study. It may

29

take longer than 11 years for soil pools to respond to plant-driven changes in microbial
process rates. Kindscher and Tieszen (1998) found no evidence of C accumulation after
5 and 25 years following the re-establishment of C4 species dominance in prairie
restorations in Kansas. Camill et al. (2004) saw a similar lack of soil C accumulation
after 6-8 years of C4 Species dominance in restored prairies in Minnesota. In contrast,
McLauchlan et al. (2006) found increased soil C in grasslands on decadal timescales after
agricultural abandonment, regardless of whether they were planted as C3- or C4-
dominated communities. Conventionally tilled agricultural fields in close proximity to
my sites have much lower soil C (~690 g m.2 in 0-50m depth, Grandy et al. 2006) than in
the old-fields in my study (~2000 g m.2 in 0-10em depth), suggesting that soil C pools
likely are increasing under both C3 and C4 species in my sites, potentially making it
difficult to detect Species differences after just 11 years. I did detect slightly higher
surface soil C in C4 plots compared to C3 plots in Turkey Meadow, suggesting that soil C
is increasing under C4 species, but that more time is needed for the higher production and
slower decomposition of C4 species tissue to result in measurably larger soil C pools.

While I did not find evidence of increasing total soil C pools, species may still be
affecting the stability of the soil C present in those soils. Two factors that can have a
large influence on soil aggregate structure and stability are the amount of roots and the
presence of mycorrhizae (Jastrow et al. 1998, Rillig et al. 2002, Rillig and Mummey
2006, Jastrow et al. 2007). The C4 species examined in this study are known to be
mycorrhizal and tended to have more root biomass (0-2OCm), both of which may
contribute to greater soil C physiochemical protection and thereby lOnger-term

sequestration of C relative to the C3 Species. Jastrow (1987) found that prairie graminoid

30

aboveground biomass was the most Significant predictor of percent aggregates >0.2mm
diameter and >2mm, and that a 14 year old pasture had Significantly fewer small
aggregates (>0.2mm) than an 11 year old restored prairie. In addition, McLauchlan et a1.
(2006) found that soils under C4 Species tended to have larger aggregates than soils under
C3 species. Direct investigation of how individual species may differ in their ability for
direct and indirect physiochemical protection of soil organic matter is needed.

At both sites, C4 species are contributing 9-26% of the total Surface soil C pool
and 6-16% of the total subsurface soil C pool after just 11 years. This suggests a
relatively rapid turnover of the soil C pools, and could indicate that greater quantities of
C are entering the Soil C pool in C4 plots. This fast turnover of C suggests that any
increases in total soil C by the C4 species are likely to become apparent relatively
quickly. Based on the C4 species contributions, complete turnover of surface soil C
could happen in 50 to 100 years, and any increase in total soil C under C4 species should
become measurable prior to complete turnover.

As predicted, the re-introduction of C4 grass species into these old-fields
significantly altered N cycling. My measurements of significantly lower total inorganic
N availability in plots dominated by C4 species compared to C3 plots are consistent with
patterns found in several studies from a broad range of temperate grasslands. Wedin and
Tilman (1990) found that after only three years, inorganic N levels sampled under
monocultures of C4 Species were significantly lower than monocultures of C3 species.
Evans et al. (2001) found that establishment of an invasive C3 grass (Bromus tectorum)
with greater biomass and more recalcitrant tissue compared to the native C3 species

(Bromus tectorum is the functional analog to the C4 Species in my study) Significantly

31

altered N cycling within two years in an arid grassland in Utah. Decreased potential net
N mineralization rates in that study were linked to changes in litter quality and quantity,
which resulted in increased microbial N immobilization (Evans et a1. 2001). The reduced
levels of inorganicN in my C4 plots compared to the C3 plots could be a result of slower
N mineralization rates causing more of the N in the system to be immobilized by
microbes, greater N uptake and storage by C4 species, or a combination of both factors.
Differences in N uptake among species inherently confound the effects on
microbial communities, so it is difficult to determine which mechanism might exert a
stronger influence on N cycling. In my study, the total plant N stocks of C4 and C3
species did not differ on an area basis. This suggests that C4 species are not taking up
more N than C3 species, and thus uptake differences may not be a determining factor for
N availability at these Sites. Despite the high variability, in situ net N mineralization and
nitrification rates tended to be higher in C3 plots compared to Andropogon plots (C4),
particularly from lune-September. I also found significant positive correlations between
tissue %N and C:N and inorganic N levels, suggesting that the Species with the highest
%N (the C3 species) may have increased N cycling rates to increase inorganic N
availability. Vinton et al. (2006) suggested that the lower C:N of Bromus inermus may
encourage rapid and efficient N cycling, which could increase inorganic N availability in
the soil. Even though A ndropogon was the only C4 species examined for N cycling rates,
one might expect similar results for the other C4 species because all three C4 species had
Similar tissue chemistry and productivity. Other studies have found similar results when
comparing C3 and C4 dominated plots (e.g., Wedin and Tilman 1990, Baer et a1. 2002,

Dijkstra et al. 2006), supporting the hypothesis that C4 species alter inorganic N

32

availability by Slowing mineralization rates and/or increasing microbial immobilization.
Total soil N was not significantly different in soils under C4 and C3 species, and a lag in
pool responses to rate changes would be expected given the large size of the soil N pool.
Total soil N was slightly higher under C4 species than in C3 plots in Turkey Meadow,
which provides some evidence to support the prediction that total N pools are increasing
in soil under C4 species relative to C3 species.

Net N mineralization and nitrification rates varied seasonally for both
Andropogon and C3 plots and were higher in the early summer. Mineralization rates are
often high in spring and early summer as a consequence of litter inputs from the previous
fall that are available for microbial utilization, combined with warmer soil temperatures
and adequate moisture tc‘)’ stimulate microbial activity (Eviner et al. 2006). Net
nitrification and mineralization rates were low by October for all plots, however, the
Andropogon plots continued to process N later into the fall than the C3 plots. The higher
surface litter in the Andropogon plots may insulate the soil in these plots and thereby
sustain microbial activity later into the fall. My evidence suggests that Andropogon is
altering the timing and rate of N transformations, but more detailed studies are needed to
examine the community- and ecosystem-level importance of such changes, as well as the
generality of these'results to other C4 species.

Overall, this study demonstrates that the introduction of a species with different
functional traits than the surrounding community can alter soil properties after ten years.
However, because the C3 and C4 grasses examined in this study differ in three main
ways—litter quality, litter quantity, and timing of resource use/release—it is difficult to

determine which traits, or combination of traits, influence particular processes in the

33

field. In this study, I provide evidence that differences between C3 and C4 species in
tissue quantity, tissue quality, and plant phenology are likely to influence soil processes.
Further study is needed in more a controlled setting to tease apart the direct and indirect
impacts of these trait differences on soil properties and processes.

My results also have important implications for understanding the effects of
restoring native species on ecosystem processes, and provide insights into the challenges
of re-establishing ecosystem structure and functions in prairie restorations. Years of
agricultural activity create a legacy of low soil C and N pools, small inorganic N pools,
and poor soil aggregate structure (Camill et al. 2004, DeGryze et al. 2004). Mostefforts
to restore old-fields to prairie and savanna have focused primarily on re-establishing
native plant assemblages using seeds or propagules, however, the success of these
restoration efforts may be limited due to the failure to consider how changes associated
with past land use have altered both soil properties and Species interactions (Suding et al.
2004). Reintroduction of specific species or functional groups that can facilitate
restoration of soil processes may be integral to the success of restoration projects (Suding
et al. 2004). The extent of human impacts on plant communities are ever increasing, and
a mechanistic understanding of how plant Species introductions or losses are likely to
impact ecosystem properties is needed to evaluate which species changes are likely to

have the greatest effect on ecosystem function and properties.

34

CHAPTER THREE
SHORT-TERM IMPACTS OF C4 PRAIRIE GRASS RE-ESTABLISHMENT

ON SOIL CARBON AND NITROGEN IN OLD-FIELDS '

ABSTRACT. Prairie restorations often focus on the re-establishment of diverse native
plant communities rather than on the restoration of soil properties and processes that have
been altered by decades of agricultural activities. C3 species typically dominate
abandoned agricultural fields, likely because they are superior competitors in the
disturbed conditions created by agricultural activities compared to the native C4 prairie
grasses that were common prior to agriculture. Differences in aboveground production
and tissue chemistry between the C3 and C4 grasses examined in this study support the
expectation that C4 grasses will accelerate soil carbon (C) and nitrogen (N) buildup
compared to C3 grasses, although the immediacy of this effect is unknown. I compared
newly established monocultures of three native prairie C4 grasses to similar stands of C3
grasses that typically dominate successional old-fields in the Midwestern United States to
determine whether, and how quickly, C4 grasses affect soil properties. I related Species
differences in productivity and tissue chemistry to changes in soil properties at three sites
in southwestern Michigan, USA after two growing seasons. I examined root production
of these species in a separate study to determine how the development and distribution of
roots throughout a soil profile might influence C, N cycling and pools. I found that the
C4 species consistently produced more shoot biomass, which was more recalcitrant than
C3 tissue (i.e., higher C:N, lignin:N). Although there were few species’ differences in

root biomass, C4 species typically produced more biomass deeper in the soil profile

35

(below 200m) compared to C3 species. Although inorganic N was significantly lower in
soils under the C4 functional group, there was little evidence to suggest that soil total C
or N pools differed after two years. However, based on Observed species differences in
productivity and tissue chemistry, I expect that inorganic N levels from soils beneath C3
and C4 Species will continue to diverge over the next several years as the plants mature,
but decades may be needed before soil pools reflect the plant species effects on these soil

pI'OCCSSCS.

INTRODUCTION
While agricultural activities have dramatically changed both plant cOmmunity
composition and soil properties, most restoration efforts focus primarily on restoring
plant communitiesand species composition (Howe 1994, Sluis 2002, Averett et al. 2004,
Blumenthal et a1. 2005, Martin et al. 2005, Williams et al. 2007) and have not considered
restoring pre-agricultural soil conditions (but see Brye et al. 2002, McLauchlan et al.
2006). Total soil carbon (C) and nitrogen (N) loss associated with agriculture is dramatic
(Burke et a1. 1995, Camill et a1. 2004, DeGryze et a1. 2004), and grassland cultivation in
the Midwestern United States has resulted in a 30-60% loss of soil organic C and N
(Burke et al. 1995). There is evidence that prairie restorations begin to restore these
pools and processes (Baer et al. 2002, Camill et al. 2004), however, it is unknown how
quickly soils will return to pre-agricultural levels, and whether changes in these processes
can be accelerated by using particular plant species.

After abandonment, old-fields are often dominated by a successional sequence of

non-native C3 species; native C4 prairie grasses that once were dominant are Slow, or fail,

36

to establish in these sites (Foster and Gross 1997, Averett et al. 2004, Emery and Gross
2006, Gross and Emery 2007). In this study, I examined how individual Species common
in native prairie or old-field communities influence soil properties and the rate at which
change might occur. C3 and C4 grass species differ in several traits expected to affect C
and N cycling, and there is evidence that re-establishment of particular ‘keystone’ prairie
species (C4 grasses) may help to restore soil processes to pre-agrieultural levels (Baer et
al. 2002, Camill et al. 2004, Chapter Two). Native prairie C4 grasses typically have
higher biomass and more recalcitrant tissue (i.e., high C:N and lignin:N) compared to the
non-native C3 grasses that characterize successional old-fields in the Midwestern USA
(Chapter Two). These differences in species traits have been shown to affect soil C and
N cycling (Wedin and Tilman 1990, Evans et al. 2001, Baer et al. 2002, Chapter Two).
However, it is unknown how quickly these Species’ differences will affect soil process
rates, or how long it will take for those process rates to restore total soil C and N pools.
While productivity and tissue chemistry differences between C3 and C4 grasses
support the expectation of long-term effects on soil ecosystem processes, the Short-term
temporal dynamics of this effect is unknown. In Chapter Two, I found that there were
detectable differences in soil prOperties and processes in soils collected from stands of C4
species that were established in 1995 (decadal timescale) compared to soils from the
surrounding community dominated by non-native C3 species. 1 established monocultures
of several C3 grass species that typically dominate old-fields in southwestern Michigan
and three C4 grass species that were common in the region prior to agricultural
development (Gotshall 1972) to determine whether species can affect soil properties and

processes soon after establishment (i.e., after two growing seasons). To test the

37

generality of these results, I examined these differences in three old-field sites, with
similar agricultural histories,at Michigan State University’s W. K. Kellogg Biological
Station (KBS) in southwestern Michigan. I measured differences in productivity and
tissue chemistry and related these differences to changes in soil properties at three sites
after two growing seasons. Based on known differences in C3 and C4 species
productivity and tissue chemistry, I expected that there would be detectable changes in
inorganic N availability within 1-2 years. However, I expected that changes in soil C and
N pools may take longer than a few years to become apparent because there is a lag
between when rates change (i.e., input and output rates) and the pool sizes (i.e., total
stocks) begin to reflect those changes. I also hypothesized that the magnitude of species
differences in biomass and tissue chemistry within and among functional groups would
determine the rate at which the soil changes occur and can be detected. Given the
relative similarity of the Sites, I did not expect many differences in the qualitative results

between sites.

METHODS
Monoculture Experiment

Species and Study Sites

I established experimental monocultures of six species, three C4 (Andropogon
gerardii (Vitman) or Big bluestem, Sorghastrum nutans (L) or Little bluestem,
Schizachyrium scoparium (Michx) or Indian grass) and three C3 grasses (Bromus inermis
(Leyss) or Smooth brome, Poa pratensis (L) or Kentucky bluegrass, Elymus repens (L)

or Quackgrass planted only in one site)) in spring 2005 at three old-fields at KBS in

38

southwestern Michigan. The three C4 species were all common in the area prior to
agricultural development (Gotshall 1972), and the three C3 grasses are all commonly
found in old-fields in the area (Burbank et al. 1992, Foster 1999). All three sites were
abandoned more than 35 years ago following decades of row crop agriculture (Burbank et
a1. 1992, Foster and Gross 1997). Site characteristics are summarized in Table 6.

I established experimental monocultures of the six species in late spring 2005 in
three randomized blocks (1 1 In by 14 m) at each Site. I treated each block with a
glyphosate herbicide (Roundup®) in early spring 2005, and then removed surface litter
and standing dead tissue by hand clipping. To minimize soil disturbance, I left highly
fragmented surface litter in place. I planted five Species in all sites: Andropogon,
Sorghastrum, Schizachyrium, Bromus and Poa. I wanted to include monocultures of
Elymus because it was the dominant species at McKay Field, but its status as a noxious
weed prevented me from purchasing seed and I had only enough seed to plant it at one
site (Turkey Meadow). In each block, I randomly selected six (1 m2) plots for each
species, with plots separated by 0.5 m. In May 2005, I covered the plots with landscape
fabric to minimizeweed growth, and planted 7-week old transplants (16 plants m-z),
started from seed in the greenhouse, into holes cut in the fabric. The landscape fabric
was kept in place for several months following transplanting and the transplants were
watered as needed for several weeks to aid establishment. In total, each site had 18

replicate monocultures of each species, six in each of three blocks.

39

Table 6. Site characteristics of old-fields used for the Monoculture Experiment. All data

are means (28E) from data collected in September 2005; biomass data (g mg) was from

harvested plots (n=9) and soil carbon (g mg) and nitrogen were from soils cores (0—10cm
deep; n=4). Species percent of the total shoot biomass is given in parentheses after the
species name. Nomenclature for all Species follows the USDA Plants Database

(plants.usda. gov).

 

 

Turkey Meadow McKay Field Louden Field
Dominant Vegetation
Bromus inermis (72%) Elymus repens (94%) Bromus inermis (50%)

Poa pratensis (23%)

Achillea millefolium L (5%) Solidago canadensis L (20%)

Dactylis glomerata L (14%)

 

-7
Shoot Biomass (g m 7)

 

 

 

362230 . 241227 386226
Soil type
Sandy loam Sandy loam Loam
Soil Carbon (g m-z)
28152150 16382225 --
Soil Nitrogen (g m-z)
288212 169218 --

 

40

Field Sampling

Throughout 2005 and 2006, I maintained the monocultures by removing (clipping
at ground level) Species that were not planted in that particular plot (hereafter referred to
as weeds). In 2005, weeding was minimal as the landscape fabric effectively prevented
weed growth. In 2006, I clipped and removed weeds every 6-7 weeks; weed biomass
from each plot was bagged, dried, and weighed (20.01 g). In August 2006, three plots of
each species were randomly selected in each block to sample plant and soil properties
(n=9 per species, 3 per block). I established subplots (0.25 m x 0.25 m) within each plot
for vegetation and soil sampling. All vegetation was clipped at ground level and
separated into the monoculture species and weed biomass. I also collected the surface
litter from these subplots. I then estimated root biomass in a subset of plots at McKay
Field and Turkey Meadow (n=6 per species, 2 per block), by collecting one soil core (0-
20 cm deep, 6.35 cm diameter) from the subplot immediately after vegetation sampling.
The core was Split into two depths (0-1 0 and 10—20 cm) and refrigerated until processed
in the laboratory. After sampling for root biomass, I collected two additional cores (0-10
cm deep, 3.8 cm diameter) from all subplots in each site for soil chemical analyses; the

two cores were combined and refrigerated until processed in the lab.

Sample Processing and Laboratory Analyses

Aboveground biomass and surface litter was dried for at least 72h at 65C and
species biomass, weed biomass, and surface litter mass was recorded (20.01 g). For each
species in each site, Shoot tissue was coarse ground in a .Wiley Mill, then ground to

<2mm in a Cyclotech Grinder. and re-dried for 48h at 65C. Tissue (2-3 mg) was then

41

packed in tin capsules for C, N and 013C analyses (relative to Peedee Belemnite; run by
UC Davis Stable Isotope Facility). Roots >06 mm were removed from soil cores by
floatation in water and collected using tweezers, and then rewashed to remove remaining
soil. Roots from each soil depth were dried for 48h at 65C and weighed (20.0001 g). I
estimated plant total N for each plot by multiplying total biomass (root+shoot) by shoot
tissue %N. Using the tissue %N of the monoculture species was appropriate, as weed
biomass was a minor component of the total shoot biomass (typically less than 15%).
While root and shoot %N data may differ, Wedin and Tilman (1990) showed that %N for
Elymus, Andropogon and Schizachyrium shoots was generally similar to that for roots.

Soils collected for soil chemistry were passed through a 2 mm sieve prior to
analysis to homogenize the sample and remove large roots and rocks. I extracted
inorganic N from 20 g subsamples processed within 24h of sample collection using 50ml
of 1M KCl. The extractions were stored at 3C until analysis on an 0.1. Analytical Flow
Solution IV analyzer. I determined gravimetric soil moisture on a subsample of soil
(~25g fresh weight) by drying soils at 105C for 48h. A subsample of air-dried soil (~50
mg) from six replicates was ground to a flour-like texture on a roller mill, oven—dried, and
analyzed for C, N and 013C at UC Davis. C3 and C4 species 013C differs, so I calculated
the percent soil C contributed by C4 Species with the model:

%C4 signal = (C4 Soil BBC — C3 soil 613C)/(C4 plant 613C - C3 soil 613C)* 100,

using the plot mean 55le from the Bromus plots as the C3 soil 613C.

42

Plant Trait Experiment

To Obtain more detailed information about the traits of the species used in the
Monoculture Experiment, I grew individuals of all six species in open-bottom “pots”
constructed from PVC pipe (15.24 cm diameter, 100 cm deep) filled with a sandzsoil
mixture. I buried the pots to a depth of ~97 cm in a fenced area at KBS, on a grid (10.5
m x 9 m) with a 1.5 in buffer between pots. The pots were filled with a 3:1 mixture of
pure sand and sandy loam topsoil collected from an old-field at KBS. Prior to filling the
pots, the soil was sieved (6.35 mm) to remove large rocks and roots and then combined
with the sand in a cement mixer. After filling, the soil was supersaturated with water to
settle the soil and achieve similar bulk densities among pots (Craine et al. 2003), and
refilled with the sandzsoil to within 3cm of the top of the pot. I randomly assigned five
replicate pots per species and transplanted one 7-week old individual to the center of each
pot in June 2006. Pots were watered as necessary throughout the summer to prevent
desiccation, with all pots receiving equal amounts at each application.

In late-September 2006, I clipped plants at ground level, dried the material for
72h at 65C, andgweighed it (20.01 g). Using the methods described in the Monoculture
Experiment, I analyzed plant tissue for total C and N using an Elemental Analyzer
(Costech Analytical, Ventura, CA). In addition, I performed Acid Detergent Fiber
analyses (recalcitrant compounds composed primarily of lignin and hemicellulose) on the
ground plant tissue (~05 g) using an Ankom 2000 Fiber Analyzer (Macedon, NY). To
determine root biomass, I removed the pots from the ground, cut them in half lengthwise
without disturbing the soil core, and separated the core into four depth intervals (0-10,

10-20, 20-40, and 40-80 cm). I placed the soils from each depth interval in plastic bags

43

and kept them at 3C until processed using the methods described in the Monoculture
Experiment. For each species, I determined the absolute biomass and proportion of total
root biomass in each depth interval to examine whether Species exhibited differential

patterns of root biomass production and depth distribution.

Data Analysis

I checked all data for normality and homoseedasity of variance, andmade appropriate
transformations prior to analysis. I used ANOVA to examine species differences and for
any ANOVA indicating a significant interaction, I performed post-hoe contrasts using
Tukey comparisons. For the Monoculture Experiment, I compared plant biomass and soil
inorganic N using a nested ANOVA with Site and Species as main effects, with Block
nested in Site. I was primarily interested in between (and not within) site variation, so I
ran analyses excluding Block from the model to test for Site and Species effects on tissue
and soil C, N and 613C. Weed biomass was a minor component of the total shoot
biomass collected in August 2006, so I used the combined monoculture species shoot
biomass and weed biomass (total shoot biomass) for statistical analysis. I did not include
Elymus in the analyses because it was present at only Turkey Meadow, but I Show the
data for comparison purposes. To test for species differences in the Plant Trait
Experiment, I used a One-Way ANOVA with post-hoe Tukey comparisons. Pearson’s
Product Moment correlations were performed to determine if the plant traits were highly
correlated with soil variables. I used SigmaStat 3.5 for all analyses except the nested

ANOVA, where I used Systat 1 1.

RESULTS
Monoculture Experiment

Plant Productivity and Chemistry

The differences among the three sites in soil fertility (Table 6) were reflected in
plant production and tissue chemistry differences of species across sites. However, I still
found consistent differences in plant and soil variables between C3 and C4 species. I
p<0.001). C4 species had Significantly higher aboveground biomass than C3 species, and
there were also significant differences among Species for surface litter production and
subsurface root biomass (Table 7, Figure 5). Species differed in surface litter production
across Sites, but Poa generally had Significantly lower surface litter than all other species
(Tables 7 and 8). I did not detect differences among Species for shallow root biomass
within a site, but Poa had significantly less subsurface (1 O-20cm) root biomass than all
three C4 species and Bronze had significantly less subsurface root biomass than
Sorghastrum (Figure 5, Table 7). McKay Field had significantly higher surface litter and
root biomass (both surface and subsurface) than Turkey Meadow (Table 7). Louden
Field had significantly more Shoot biomass than McKay Field, which had significantly
more than Turkey Meadow (Table 7). C4 species typically had significantly higher tissue
C:N, and lower %N content than C3 Species (Table 9), and tissue %N and C:N differed
across sites (F 2345:1272, 14.49, respectively; p<0.001 for both). Species had
significantlyI higher tissue %N and C:N at McKay Field (%N=1.1520.11,
C:N=44.2523.27) than at Turkey Meadow (%N=0.8620.07, C:N=57.8624.16) or Louden

Field (%N=0.9420.1 1, C:N=55.6524.22). Individual species comparisons showed no

45

 

 

Table 7. ANOVA results for aboveground variables for all five species in Louden Field
McKay Field, and Turkey Meadow, using Site and Species as main effects, with Block
(Blk) nested within Site. Significant differences are indicated in bold.

 

 

 

 

 

 

Model Site Species Site*Species Blk(Site)
Total Shoot F 8.91 162.15 1.66 4.03
-2
Biomass (8m ) d.f. 2,111 4,111 8,111 6,111
P value <0.001 <0.001 0.177 0.001
Surface Litter F 4.71 23.06 0.99 3.80
-2
(gm ) d.f. 2,111 4,111 8,111 6,111
P value 0.011 <0.001 0.450 0.002
Surface Root F 19.94 1.22 2.73 1.14
Biomass 0'10““ d.f. 1.45 4,45 4.45 4,45
( n2)

g P value <0.001 0.314 0.041 0.352
Subsurface Root F 9.47 9.18 0.46 2.36
Biomass 1040“” d.f. 1,45 4,45 4,45 4,45

-2
(8 m )
P value 0.004 <0.001 0.762 0.068
Soil Inorganic F 37.05 3.10 1.27 6.25
Nitrogen d.f. 2,111 4,111 8,111 6,111
-1
(ugN 8dry soil )
P value <0.001 0.018 0.267 <0.001

 

Turkey Meadow

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2000
[Shoot BRoot(0-10cm) lRoot (10—20cm)
J‘ 1000 -
'E
3
S
(U
E
.9
m 2000 ~
3000 -
Ag 55 Sn Bi Pp Er
C4 species C3 species
McKay Field
2000
EIShoot lRoot (0-10cm) lRoot (10-20cm)
(9 1000 4
'E
3 0 -
3 1
IO. _
E 000
.9
m 2000 ~
-3000 J
Ag 55 Sn Bi Pp
C4 species 03 species
Louden Field
m 2000
in T [:1 Shoot
‘0 l
g 1000]
a": O I
Ag 55 Sn ' Bi Pp
C4 species C3 species

 

Figure 5. Shoot and root production estimates of each species in each site in the
Monoculture Experiment, averaged across the three blocks in each site. Mean (2SE) root
biomass for each depth interval (n=6) and shoot biomass (n=9) in August 2006 is shown
separately for each site. Root biomass was not collected at Louden Field. Species names
are abbreviated: Ag=Andr0p0g0n, Ss=Schizachyrium, Sn=S0rghastrum, Bi=Br0mus,
Pp=P0a, and Er=Elymus.

47

significant differences in total plant (root+shoot) N stocks at either Turkey Meadow or

McKay Field (p>014).

Soil Properties

Soil properties varied among dominant vegetation plots and Sites, with no Site x
Plot interactions (Tables 7 and 10). Soil inorganic N was significantly higher in soil
under Bromus than under Sorghustrum (Figure 6, Table 7). There were a few Significant
differences between Species in total soil C or N, and soil under Andropogon and
Sorghastrum tended to have significantly higher C:N and 513C than C3 species (Table 9).
Soil 013C, N content, and C content were strongly correlated with surface root biomass
(Pearson r=049, -0.44, -0.43, respectively; p<0.02 for all). C4 Species contributed an
average of 2.320.52% (Range: 0 to 9.6%) of the total soil C in C4 monoculture soils after
just two years.

The three Sites differed significantly for many soil prOperties. Soil N and C
content (g kg-]) was higher at Louden Field (means: N=0.1920.01; C=2.02:l:0.07) than at
the McKay Field (means: N=0.1220.01; C=1.3420.08) or Turkey Meadow (means:
N=0.13:l:0.01; C=l .472006) Sites (F2360=38.77, 32.34, respectively; p<0.001 for both).
Soil C:N was significantly higher in McKay F ield. (mean of 11.5201) than in Turkey
Meadow (mean of 11.1201) and Louden Field (mean of 108201; F2,6O=22.82,
p<0.001). Soil 013C also differed between sites (132,453.90, p=0.026) and was
significantly more depleted in Louden Field (mean of -26.920.1) than in Turkey Meadow
(mean of -26.7:l:0.1). Soil inorganic N was also significantly higher in Louden Field than

at McKay Field or Turkey Meadow (Table 7).

48

Table 8. Mean (28E) surface litter mass (g m-2) for each Species at each site.
Superscript numbers denote significant differences (p<0.05) between Species within each
Site. In addition, McKay Field had significantly more surface litter than Turkey Meadow

 

 

(p=0.011).

Plot Turkey Meadow McKay Field Louden Field
Andropogon 76:18a 78210a 91:15a
Schizachyrium 43.9ab 70:10ab 61 :9ab
Sorghastrum 5 3 +133b 5927ab 5228ab
Bromus 37210b 63+ 1 5b 3526b

Poa 21:9C 27:6C 19:6C
Elymus 3526 -- --

 

49

CD

C41EIAg ESS ISn

C3:Z|Bi Ele EEr
b

O)

aba

ab

Inorganic N
(ug N per g dry soil)
A

N

 

0

Turkey Meadow McKay Field Louden Field

Figure 6. Species effects on inorganic N (mean 2 SE) concentrations in the Monoculture
Experiment, averaged across blocks, for the three sites. Species names are abbreviated:
Ag=Andr0p0g0n, Ss= Schizachyrium, Sn=Sorghastrum, Bi=Br0mus, Pp=P0a, and

Er=Elymus. Solid bars represent soils under C4 species and patterned bars represent soils

under C3 species. Lowercase letters denote significant differences (p<0.05) between
plots (Elymus plot data is shown for comparison purposes only).

50

Table 9. Plant tissue chemistry (mean2SE) for each species (n=4), averaged across sites.
Elymus data is shown for comparison only. For variables with significant Species main
effects, lowercase letters denote significant differences (p<0.05) determined from post-
hoc Tukey comparisons.

 

Species Group %N %C C:N 51 3 C

 

Andropogon C4 0.72:0.05a 45.1 : 0.1a1 65.7 : 4.381 -124 : 0.2a

SCH-7069”” C4 0.71:0.043l 45.0 : 0.2a 65.1 : 3.161 -12.6 : 0.1a

Sorghastrum C4 0.80:0.0481b 44.3 : 0231’ 57.2 : 3.0ab -12.7 : 0.181
Bromus C3 0.98:0.09b 44.6 : 0.23 48.4 : 3.3b -28.0 : 0.2b
Poa C3 1.71:0.11c 43.8 : 0.31’ 26.6 : 1.7C -27.8 : 0.2b
Elymus C3 0.93:0.08 45.3 : 0.9 49.4 : 4.3 -28.3 : 0.2

 

51

Table 10. Soil chemistry data (mean2SE) for soils under each Species, averaged across
sites. For variables with significant Species effects, superscript letters denote significant
differences (p<0.05) determined from post-hoe Tukey comparisons. Elymus is shown for
comparison only.

 

Species Group N (g kg-l) C (g ng) C:N 513C

 

Andromgon C4 1.3:0.1a 14.9:1.3"‘b 11.62023 -26.520la

SChizachyrium C4 1.5:0.1 16.2:09ab ll.l:0.1"" -26.82O.labc

 

Sorghastrum C4 1.5:0.1ab 16.7:12ab 11.4:0.lab -26.720lab
Bromus C3 1.6:0.1b 18.1:14b 111:0.le -27.1:0.1°
1’00 C3 1.3:0.1ab 14.7:09a 11.0:0.1C -26.9:0.2bc
Elymus C3 15:04 16.3:26 10.9:0.2 -27.0:0.2

 

 

52

There were also several significant correlations between plant and soil variables.
Plant tissue C:N was significantly correlated with soil C (n=39, p=0.02, r=036) and N
content (n=39, p=0.03, r=035), and total root biomass was positively correlated with soil
813C (n=29, p<0.001, r=055), and negatively correlated with both soil C (n=29, p=0.01,
r=-0.45) and N content (n=29, p=0.01, r=-0.47). Surface litter was positively correlated

with soil 613C (n=75, p<0.001, r=041).

Plant Trait Experiment

Growing these Species as individuals in pots allowed me to more accurately assess plant
traits, particularly root characteristics, than I could in the field monocultures. I again
found that the three C4 species tended to have greater shoot biomass than the C3 species
(Figure 7), and tissue chemistry also separated the C3 and C4 species (Figure 8). There
were significant differences between Species in root biomass at each depth interval, but
few clear trends emerged (Table 11). While all species had over 50% of their root
biomass in the top 20cm of soil, the three C3 species tended to have a larger proportion Of
their total root biomass (62-68%) in the surface soils (0-20cm) than the C4 species (51-
60%). The C4 Species produced a larger percentage of root biomass (18-21%) deeper in

the soil (20-40cm) than did the C3 species ('1 0-13%).

DISCUSSION

As expected, I was able to detect significant differences in plant production among

species within two years of establishment in all three sites. The greater Shoot biomass for

the C4 species seemed to be a consistent pattern for all C4 species, and P00 had the

53

Shoot I Root (0-80cm)

 

,2 abc
‘8
a be
a:
D.
3
B
(O
E
.9
1:0
ab
Ag ' Ss Sn Bi Er Pp
C4 species C3 species

Figure 7. Above- (n=5) and below- (0-80 cm; n=4) ground biomass per plant for each
species (mean 2SE) in the Plant Trait Experiment. Species names are abbreviated:
Ag=Andr0pogon, Ss= Schizachyrium, Sn=Sorghastrum, Bi=Br0mus, Pp=P0a, and
Er=Elymus. Solid bars represent C4 species and patterned bars represent C3 species.

Lowercase letters denote significant (p<0.05) differences between species for shoot
biomass and total root biomass.

54

C4:EIAg SSS ISn
C3:IBi EEr EPp

a

 

C:N Tissue Chemistry ADF:N

Figure 8. Tissue chemistry data for species grown in the Plant Trait Experiment; values
are mean (2SE; n=4). Species names are abbreviated: Ag=Andr0pog0n, Ss=

Schizachyrium, Sn=Sorghastrum, Bi=Br0mus, Pp=P0a, and Er=Elymus. Acid detergent
fiber (ADF) is a measure of recalcitrant compounds, primarily lignin and hemicellulose.

Solid bars represent C4 species and patterned bars represent C3 species. Lowercase
letters denote significant differences (p<0.05) between species.

55

Table 11. Plant Trait Experiment root biomass at each depth interval (mean2SE; n=4).
For each depth interval, significant differences (p<0.05) between Species are indicated by
superscript lowercase letters, as determined from post-hoe Tukey comparisons. Biomass
values without letters indicate no significant difference from any of the other Species.

 

Root Biomass (g per plant)

 

0-10 cm 10-20 cm 20-40 cm 40-80 cm
Andropogon 62120-98 3.47:0.33a 3.58:0.63a 3.60:0.71a
Schizachyrium 2.7910933 1.332035 1.682054 2.162063
Sorghastrum 6.4721.ll 1.382009 2.34:0.41ab 2.782040
Bromus 1008:290b 2.32:0.71ab 2.64:0.81ab 4.26:0.83a
P061 429:1-33 0.80:0.18C 0.71:0.10C f 1.44:0.23b
Elymus 4.64:1.18 0.8130451)C 1.12:0.27‘” 2.15:0.50

 

56

lowest shoot biomass and surface litter layers. The surface litter in this experiment
reflects a single growing season and so Should strongly reflect difference in shoot
production. Andropogon had 8 to 27% higher aboveground biomass than the other C4
species and this difference was clearly reflected in litter levels, where Andropogon had 10
to 80% more litter mass than the other C4 species. All else being equal (e. g. similar
phenology and litter quality), I would expect that the significantly higher shoot biomass
levels for C4 species would result in faster litter accumulation in C4 plots after a
relatively Short period of time. I found in Chapter Two that surface litter mass in plots of

these C4 species was significantly greater than in the surrounding C3-dominated old-

 

fields Of Turkey Meadow and McKay Field after 11 years.

There were few differences in root biomass among Species after two growing
seasons in the Monoculture Experiment. The Plant Trait Experiment, which allowed me
to examine species rooting patterns in greater detail, also did not reveal clear differences
in root biomass among Species after just one year. Craine et al. (2003), in a three-year

study of 11 grassland species in Minnesota, also found it difficult to generalize about

 

functional group differences in rooting dynamics. While I did not see clear trends in root '3‘
production, I did see differences in allocation patterns in the Plant Trait Experiment. 1
While the majority of root biomass for all species was in the top 20cm of soil, C4 species

allocated on average 40-49% of root biomass to deeper soils (20-80cm depth), compared 1 I

to 32-3 8% for C3 species. Craine et al. (2003) also found comparable results for Bromus,
Poa, A ndropogon, and Schizachyrium in a Similar study. Sampling for roots in the top
20cm may have disproportionately underestimated root biomass for C4 species because

of their higher allocation (almost half) in deeper soil. This greater investment in deeper

57

roots suggests that C4 species may be accessing soil nutrients and water that are
unavailable to the shallower rooted C3 species and potentially gaining a competitive
advantage during periods of drought. This deeper root investment also suggests that C4
species may affect’soil processes at greater depth than C3 species, with implications for C
sequestration in deeper soil.

The lack of clear differences between species root biomass may reflect the fact
that these were relatively short-term experiments and the rooting systems of the species
may not have fully developed after 1-2 growing seasons. In a related study of 11-year-

old monocultures of the same C4 Species in McKay Field and Turkey Meadow, I found

 

higher overall root biomass values, and significantly greater root biomass (0-20cm) for
the C4 species than for Elymus and Bromus (Chapter Two). I expect that the differences
between C3 and C4 Species will become more pronounced as the plants mature and
belowground biomass accumulates.

The C4 species produced more aboveground biomass (potential litter) and this
tissue was higher in C:N, ADF:N and had lower %N than C3 species. As a result, litter

produced by C4 species would be expected to decompose slowly and thus increase soil

"‘1. ., n
'xgfl

organic matter buildup. Indeed, I found strong, positive correlations between soil 013C

and both surface litter and root biomass, suggesting that the soil is beginning to reflect the

 

C inputs of the C4 species after just two growing seasons. In contrast, while C4 species 1 .1
generally had lower soil inorganic N compared to C3 Species, these differences were not
statistically significant except for between Sorghastrum and Bromus. Many studies have
shown lower levels of inorganic N and/or reduced N mineralization rates in communities

dominated by C4 species (Wedin and Tilman 1990, Wedin and Pastor 1993, Baer et al.

58

2002, Camill et al. 2004, Chapter Two). Based on the results in Chapter Two I expect
that inorganic N levels in soils associated with C3 and C4 Species will continue to diverge
and produce detectable difference in the next 10 years.

The mechanism behind these Slight reductions in inorganic N is unclear. C4
species could be reducing inorganic N levels by taking up more N than C3 and therefore
have larger tissue N stocks, or through reduced N cycling rates of its recalcitrant litter.
Evidence from the 11-year-old C4 monoculture stands of these species in McKay Field
and Turkey Meadow supports slowed N cycling as the more likely explanation (Chapter
Two), and data from that study and the current study suggest that C4 species do not
contain more total N in their combined root and shoot tissue than C3 species. Craine et
al. (2003) found similar results; there were no significant differences in aboveground
biomass N between Andropogon, Schizachyrium, Poa and Elymus, although Poa had
significantly lower belowground biomass N content than the other three species.

While there was little indication that species had impacted soil total C and N
pools after two years, there was a significant, positive correlation between plant tissue
C:N and both soil C and soil N pools. This suggests that the higher C:N species (i.e., the
C4 species) are increasing total soil C and N pools. In addition, the isotopic signature of
C4 Species was detectable in the soils after just two growing seasons, indicating that C4
species had contributed a significant amount of C to the soil C pool in a short time period.
C4 species contributed an average 2.3% of the total soil C (0-10cm) in C4 monoculture
soils after just two years. Most of this contribution is likely from roots, as little surface
litter was likely incorporated into the soil after such a short time. I Showed in Chapter

Two that C4 species contributed 9-26% of the total soil C pool after 1 1 years, which

59

suggests that C replacement in my study will steadily increase in the next decade.
However, I do not have clear evidence as to whether this replacement will result in faster
rates of soil C accumulation for C4 species. Baer (2002) showed some evidence of total
soil C levels increasing in C4 dominated restored prairies after 12 years, but McLauchlan
et al. (2006) found that total soil C increased with time since abandonment regardless of
' whether the community was dominated by C4 or C3 species. Soil C:N was higher under
several C4 species than C3 species, suggesting that C and N pools may be changing
slowly to reflect the higher C:N ratios of the C4 species.

The results of this study demonstrate that species differences in biomass and
tissue chemistry can be detected rapidly in the field, but the effects of these species
differences on soil processes are slower to emerge and will depend on the magnitude of
the differences betWeen species. Many of the plant and soil variables were similar for
species within a functional group, including Elymus, which typically fell between Poa
and Bromus in terms of plant traits and soil properties at Turkey Meadow. Although
inorganic N pools were just beginning to Show species effects, I did not see significant
changes in soil total C or N pools after two years. I expect that inorganic N differences
between species will continue to diverge over the next several years as the plants mature
and differences in surface litter and root production become more pronounced. However,
a decade or more may be needed before total soil pools reflect species effects on soil

processes.

60

CHAPTER FOUR
EFFECTS OF LITTER QUALITY AND SOIL MICROCLIMATE ON

DECOMPOSITION

ABSTRACT. Plant Species can affect soil processes in a variety Of ways, including
through the quantity of biomass produced and the chemistry, or quality, of biomass. In
this study I examined whether species differing in litter characteristics may alter
decomposition directly via tissue chemistry differences and indirectly via effects on
microclimate. I set up a reciprocal transplant experiment to compare the decomposition
rates of two C3 grasses (Bromus inermis and Elymus repens) and Andropogon gerardii (a
C4 grass) in two old-field communities in southwest Michigan. My findings suggest that
seasonal controls (i.e., temperature and moisture) on decomposition are stronger than the
effects of litter quality, and litter quality differences become important when examining
decomposition within a common environment (i.e., within a site). Examination of the
effects of these species on microclimate (soil moisture and temperature) indicated that
while soils were warmer under C3 than C4 Species, there was no evidence to suggest that
this affected decomposition rates. This suggests that litter chemistry was the controlling
factor determining the observed differences in decomposition rates between C3 and C4
Species. These results indicate that restoring native C4 prairie grasses into C3-dominated
old-fields will slow decomposition rates of aboveground plant tissue and may ultimately

lead to changes in soil C and N cycling and storage.

61

INTRODUCTION '

Plant Species that differ in above- and below-ground biomass and tissue chemistry can
influence soil processes in various ways, and changes in species composition can have
dramatic impacts on carbon and nutrient cycling. One of the most frequently examined
ways by which plants alter soil processes is through differences in litter quality (i.e.,
tissue chemistry) and quantity (Hobbie 1992, Wardle et al. 1998, Ehrenfeld 2003, Lovett
et al. 2004, Hooper et al. 2005, Dijkstra et al. 2006). Vitousek and Walker (1989)
demonstrated that the invasion of a novel nitrogen fixer (Myricafaya) dramatically alters
N cycling in young Hawaiian forests. Other studies (Evans et al. 2001, Mack et al. 2001,
Drenovsky and Batten 2007) have Shown that plant invasions alter N cycling via
production of more recalcitrant tissue, and that species with higher C:N or lignin:N
typically have slower decomposition rates (Ehrenfeld et al. 2001, Xu and Hirata 2005,
Hobbie et al. 2006, Drenovsky and Batten 2007).

While the importance of Species litter chemistry in controlling decomposition
rates is well accepted, whether plants can impact soil processes via their effect on soil
microclimate has received less attention. However, several recent papers (Mack and
D'Antonio 2003b, Eviner 2004, Eviner et al. 2006) suggest that such effects can be as or
more important than litter chemistry as a determinant of soil process rates. Soil
temperature and moisture are two factors that exert strong controls on soil microbial
function (Aerts 2006). Plants can influence these factors in a variety of ways (reviewed
in Eviner and Chapin 2003), including the rate and timing of water uptake, surface litter
effects on evaporation from the soil, and aboveground production, all of which can buffer

soil temperature fluctuations. Bengston et al. (2005) found that soil water is positively

62

related to soil respiration and N mineralization rates, which suggests that a Shift to a
species with higher water use efficiency or greater litter (i.e., reducing evaporation) could
alleviate microbial water limitation and result in higher process rates. Thus, a shift in
species composition to a new dominant species that alters microclimate could also
influence soil processes such as decomposition.

Even with dramatic differences in litter characteristics among species, it is often
difficult to isolate plant traits responsible for observed influences on soil processes. In
Chapter Two, where I compared the relative effects of C3 and C4 grasses on soil C and N
after 11 years, I found that soils under native prairie C4 grasses, with both greater
biomass and more recalcitrant tissue than C3 grasses, tended to have lower rates of in situ
net nitrification and N mineralization and significantly higher surface litter accumulation
than soils under non-native C3 grasses typical of old-field communities in the Midwest.
Results from that study led me to hypothesize that decomposition rates would be slower
in sites dominated by C4 species. However, because these two functional groups differ in
biomass production and litter quality, I expected that they would also affect soil
microclimate. Thus it is difficult to determine whether effects on soil processes are due
to direct effects of plant traits, microclimate or a combination of these factors.

This study evaluates the ways in which traits of C4 and C3 species can influence
litter decomposition and turnover. I performed two experiments to separate the effects of
litter quality and soil microclimate. In the first experiment, the decomposition rates of
the two litter types (C 3 and C4) were compared in two common locations (both litter
types were placed under C3 species and under C4 species) in two fields. I expected C4

litter to decompose slower than C3 litter because C4 litter is more recalcitrant (higher

63

C:N, lignin:N). Comparing decomposition of a particular litter placed in two locations
allowed me to examine how differences in soil microclimate influences decomposition
rates. I also expected decomposition to be slower under C3 species than C4 species
because the smaller litter layer and lower aboveground biomass would make these
environments less advantageous for microbial activity (i.e., drier and hotter) in the
summer.

In a second experiment, I used cellulose filters as a standard substrate to
determine decomposition rates under C3 and C4 species, with and without litter. Soil
moisture and midday temperatures were measured in these plots to determine the effect
of aboveground biomass and surface litter on these factors that are important
determinants of soil processes such as decomposition. Summer and fall soil moisture
levels and soil temperature were expected to vary with ground cover (litter and above-
ground biomass) and so I predicted highest soil moisture and lowest temperatures under
the C4 species with litter intact, followed by C4 species without litter, then C3 species
with litter and finally C3 species without litter. I expected cellulose decomposition to be
faster under C4 Species if soil moisture limited decomposition and to be fastest in the C3
Species plots without litter if temperature was the primary factor determining

decomposition rates.

METHODS
Study Sites
To determine the relative importance of litter quality and microclimate associated with

different dominant species on decomposition rates, I performed a reciprocal transplant

64

litterbag decomposition experiment using litter from two C3 species and A ndropogon
gerardii (Vitman; a C4 Species), in two Old-fields located at the WK. Kellogg Biological
Station (KBS) in southwestern Michigan, USA. Both sites were abandoned more than 35
years ago following decades of row crop agriculture (Burbank et al. 1992, Foster and
Gross 1997). Plant communities in the two Sites were dominated by C3 non-native
species, but the composition of the communities differed: Turkey Meadow is dominated
by Bromus inermis (Leyss) and McKay Field is dominated by Elymus repens (L).
Patches ofAna’ropogon were established in 1995 for competition experiments described
in Foster (1999), and subsequently abandoned in 1996. Nomenclature for all species
follows the USDA Plants Database (available online: plants.usda.gov). All soils are
sandy loams (Foster and Gross 1997), but McKay Field has a higher sand fraction and

seems to be more drought-prone than Turkey Meadow.

Reciprocal Transplant Experiment

I compared the decomposition dynamics of Andropogon gerardii and Elymus repens at
McKay Field, and Andropogon gerardii and Bromus inermis at Turkey Meadow by
collecting recently senesced, standing aboveground tissue from many plants of each
species in late October 2005. Litter was air-dried in the laboratory for two weeks, out
into 6-8cm long pieces, and gently mixed to homogenize the litter for each species.
Approximately 4g of air-dried litter (3.8g oven-dried equivalent) was placed into
10cmx10cm polyester mesh litterbags (0.17cmx0.l7cm mesh), which were sealed with
an impulse heat sealer. For both species at each site, I placed six replicate sets of Six

litterbags in each of two environments (litter under C4 Species and litter under C3

65

species) in November 2005. In total, I had four treatments in each site—C3 litter under
C4 Species, C3 litter under C3 species, C4 litter under C4 species and C4 litter under C3
speciese—with Six replicates of each treatment for collection at each of Six time intervals.
In late November 2005, I gently slid the litterbags under the barely decomposed surface
litter layer (to avoid contamination, highly fragmented litter was not placed on the
litterbags). Some litterbag replicates were lost to mammal activity, so I collected at only
four dates in 2006: April 10, June 19, August 30, and November 28, after 142, 213, 284
and 374 days.

The litterbags were sealed in plastic bags and transported to the laboratory, where
soil, root, and green tissue contamination was removed, and then weighed to determine a
field-moist weight of the litter (field moisture was used as an indicator of potential
microclimate differences between species and sites). The litterbags were then dried for
48h at 65C, and then reweighed. For each bag, I calculated percent moisture using the
field-moist weight measurement and the oven-dried measurement for the litter. I
calculated the percent mass remaining relative to timeo. This data was then used to
calculate decay constants (k) for each replicate in each treatment using both a linear and
single exponential model (Trofymow et al. 2002). The two models had a Similar fit to the
data (average R2 of 0.92 for the linear model and 0.90 for the exponential model), so I
used the more biologically realistic exponential model for analysis.

I compared mass remaining between treatments over the one-year period using
Repeated Measures ANOVA with Site, Litter type (C4 or C3), and Placement (placed
under C4 or under C3 species) as factors. Decay constants for the exponential model

were compared using a Three-way ANOVA with Site, Litter type (C4 or C3), and

66

Placement (placed under C4 or under C3 species) as factors. For any ANOVA indicating
a significant interaction, 1 performed post-hoe contrasts using Tukey comparisons. I used
Systat 11 for Repeated Measures and SigmaStat 3.5 for all other analyses.

To examine seasonal impacts on decomposition rates, I compared the
decomposition of the first 142 days of the experiment above (November 2005-April
2006; winter decomposition) to a second set of litterbags installed in April 2006 (April
2006-September 2006; summer decomposition). Mass remaining and the exponential
decay rates (k) for winter and summer periods were compared separately for each site
using a Three-way ANOVA with Litter type (C4 or C3), Placement (under C4 or under
C3 species), and Season (Winter or Summer) as factors. For significant interactions, 1
performed post-hoe contrasts using Tukey comparisons.

To determine initial litter chemistry for each species, I analyzed a finely ground
(<2mm on a Cyclotech grinder) litter sub-sample for C and N concentrations using an
Elemental Analyzer (Costech Analytical, Ventura, CA), and for Acid Detergent Fiber
(recalcitrant compounds composed primarily of lignin and hemicellulose) on an Ankom
2000 Fiber Analyzer (Macedon, NY). Data for Species and Site were compared using

Two-way ANOVA.

Decomposition and Soil Microclimate Experiment

The species examined above differ in both their surface litter and aboveground biomass,
and these differences may affect soil microclimate, particularly soil moisture and
temperature. This experiment determines whether microclimate differences exist

between the species, and how these differences affect the decomposition of a standard

67

cellulose substrate. To examine how surface litter contributes to microclimate
differences under C3 and C4 grasses, I removed the surface litter by hand from six plots
(0.5mx0.5m) for both C3 and C4 grasses (-litter treatment) and left the surface litter intact
in six plots for both C3 and C4 grasses (+litter treatment). I placed one 10cmx10cm
polyester mesh litterbag (0.17cmx0.l7cm mesh) filled with ~4g of cellulose filter paper
(Whatman No.1) into each plot in late May 2006. Cellulose was used as a decomposition
standard material to eliminate the confounding factor of litter quality and focus on the
microclimate impacts on decomposition. Bags were placed on the soil surface in the -
litter treatment and under the surface litter layer in the +litter treatment.

I measured soil moisture and temperature in each plot in mid-June, mid-July, mid-
October and mid-November 2006. All measurements were made close to midday on
mostly sunny days. To determine gravimetric soil moisture, I combined 3 soil cores (2cm
diameter, 0-10cm depth) from each plot and dried them at 105C for 48h. I measured soil
temperature (Taylor model 9841) at two depths: surface soil temperature and soil
temperature at 10cm below the soil surface. Three temperature readings were averaged
for each plot at each depth.

I collected the cellulose bags in mid-May 2007 (after 354 days), placed them in
plastic bags for transport to the laboratory, removed soil and plant contamination, and
weighed the filters immediately and after oven-drying for 24h at 65C. I calculated the
percent mass remaining from timeo and moisture content of the cellulose. Both variables
were compared using Three-way ANOVA comparing Site, Placement, and +/- litter as
main effects. Repeated Measures ANOVA was used to determine changes in soil

moisture and temperature over the one-year period, using Site, Placement (under C4 or

68

under C3), and +/- Litter (+litter and -litter) as factors. For any ANOVA indicating a

significant interaction, I performed post-hoe contrasts using Tukey comparisons. I used

Systat 1 l for Repeated Measures ANOVA and SigmaStat 3.5 for Three-way ANOVA.

RESULTS
Reciprocal Transplant Experiment
Andropogon (C4) had significantly higher C:N (F 19:61.72, p<0.001) and ADF:N (ADF
is primarily lignin and hemicellulose; F137=49.22, p<0.001) than the C3 species, and there
were no significant differences in tissue quality between the two Sites (Table 12).
In both sites, mass loss ofAndropogon (the C4 species) was significantly less than for C3
tissue (F 1,39=90.49, p<0.001), and mass loss for Andropogon was lower in Turkey
Meadow compared to McKay Field (F1,39=11.76, p=0.002; Figure 9). Using an
exponential model, k values were Si gnificantly higher for C3 litter compared to
Andropogon in Turkey Meadow but not McKay Field (p<0.001 and p=0.054,
respectively). Moisture contents of litter at field collection were significantly higher in
Turkey Meadow than in McKay Field (F1339=9.03, p=0.005) and significantly higher for
litter located under C4 species than for litter located under C3 Species (F1939=28.51,
p<0.001). However, there was no impact of litter type on moisture content (F1,39=2.50,
p=0. 122).

Comparisons of winter and summer decomposition after 142 days showed that
mass loss and decay rates were significantly influenced by season. but C3 litter still
decomposed faster than Andropogon litter (p<0.001 for all; Figure 10). In McKay Field,

litter placement was also important; litter placed under C3 Species decomposed more

69

Table 12. Chemistry of the senesced shoot tissue for the Species used in the Reciprocal
Transplant Experiment. Data shown are meanS2SE, based on fraction of oven-dried
weight. Acid detergent fiber (ADF) is primarily lignin and hemicellulose). Superscript
lowercase letters denote significant differences (p<0.001) between species.

 

 

 

Site Species C:N ADF:N
Turkey Meadow Andropogm 83.17: 6.77a 151.57:13.56a
Bromus b b
47.02: 4.40 8037:1367
McKay F ield Andropogon 101.14:11643 178.50:19.74a
El mus b b
y 52.65: 2.86 84.89: 4.87

 

70

Turkey Meadow

100 2-3

90-

80-

70-

Mass remaining (%)

60‘

 

 

 

SO _ t .
Nov-05 Feb-06 May-O6 Aug-06 Nov-06
Day 0 142 213 284 374

 

McKay Field
100

90‘
801

703

Mass remaining (%)

601

 

 

 

50 . 2 4
Nov-05 Feb-O6 May-06 Aug—O6 Nov-O6
Day 0 142 213 284 374

 

Figure 9. Litter mass remaining (% of initial) across time for the Reciprocal Transplant
Experiment. Black lines (both solid and dotted) represent Andropogon litter, while Gray

lines represent the C3 Species litter (Bromus in Turkey Meadow, Elymus in McKay
Field). Solid lines Show decomposition when the litter was placed under Andropogon

and dotted lines Show decomposition when the litter was placed under C3 species.

71

 

Litter Type p<0.001

Turkey Meadow Season p<0.001

 

    

 

 

 

 

 

 

100

m IWinter ISummer
55 9
L ‘° 75 —
a a
(U -:
0') O
E “5 50 —
es
2 (I)
m g 25 —
In ‘0
(U
E

0 _

Andropogon _ Bromus
Litter Type
_ Litter Type p<0.001
100 McKay Field Season p<0.001

N IWinter Summer
3 9
a 2 75 -
E g
E) O
'7: M5 50 -
'6 \°
E 9.,
2 U)
U, i; 25 —
VI '13
(U
z

0 _

Andropogon E/ymus
Litter Type

Figure 10. Comparison of the mass remaining (% of initial) after the first 142-days for
litter placed in the field from November 2005-April 2006 (black bars) compared to litter
placed in the field from April-September 2006 (gray bars). Data are the means (2SE) for

each litter type, averaged across placement location (i.e., under C3 and under
Andropogon).

72

quickly than litter placed under C4 species, regardless of litter type or measure of

decomposition (p=0.004 for mass loss and p=0.002 for k).

Decomposition and Soil Microclimate Experiment

Soil temperatures at both the surface and 10cm depth differed significantly between both
Sites and species (Figure 1 1). Surface and 10cm temperatures were typically higher
under C3 species than C4, and were higher in Turkey Meadow compared to McKay Field
(Figure 11). However, there was no significant effect of whether litter was intact or
removed on either temperature measure. Soil moisture was generally higher in Turkey
Meadow than McKay Field (Figure 12). In contrast, there was no evidence of Placement
effects on soil moisture except for one sampling point; after day 213 (June 19, 2006), soil
moisture was significantly higher under C3 Species than C4 species (F],40=15 .60,
p<0.001). In. contrast to the microclimate variables, the presence of surface litter (+/-litter
treatment) had a significant influence on mass loss of the cellulose (F1,39=4.74, p=0.036),
while mass loss did not differ under C3 and C4 species (Placement effect) or between

sites (Figure 13).

DISCUSSION

As expected, the C3 species had significantly lower C:N and ADF:N than Andropogon,
and these two ratios are often correlated with slower decomposition rates. In a study of
125 fresh leaf litter types, including woody Species, forbs and grasses, Cornelissen (1996)
found that decomposition was Strongly negatively correlated with 1i gninzN. Indeed, I

found that decomposition rates were faster for the C3 litter than for the Andropogon litter.

73

a) Surface Soil Temperatures

Turkey Meadow

 

     
 

 

 

 

 

 

 

35
E +Andropogon *Bromus
3
to
L
8 25 -
E
cu
'—
3 Site p=0.018
a, 15 - Species p<0.0001
«,3 Time p<0.001
‘t: Time*Site p<0.001
a Time*Species p=0.049

5 T I I I I
Jun-06 Jul-O6 Aug-O6 Sep-O6 Oct-O6 Nov-O6

35 McKay Held
93 +Andropogon -—°-E|ymus
B
E
3 25 -
E
m
l.—
'8
a, 15 -
S
‘1:
:5
U)

5 I

 

Jun-06 Jul-O6 Aug-O6 Sep-O6 Oct-O6 Nov-06

Figure 11. Soil temperature (degrees C) on the soil surface (a) and at 10cm depth (b) for
Sites in 2006, averaged across the +/-Litter treatment. Black lines represent Andropogon

plots, while Gray lines represent C3 plots. Greenhouse-Geisser adjusted p-values for log-
transformed data are reported for time variables, and untransformed data are Shown.

74

 

Figure l 1 (cont’d)

b) Soil Temperatures at 10cm Depth

Turkey Meadow

 

25

*Andropogon --- Bromus

   
 

10cm Depth Temperature
H
U1

 

 

 

Site p<0.001
Species p=0.005
Time p<0.001
Time*Site p<0.001
5 I l I I r
Jun-06 Jul-06 Aug-O6 Sep-O6 Oct-06 Nov-06
25 McKay Field

 

--Andropogon -- Elymus

10cm Depth Temperature
H
U'I

 

 

 

5 T I l I I
Jun-06 Jul-O6 Aug-O6 Sep-O6 Oct-O6 Nov-06

75

Turkey Meadow

 

     
 

 

 

 

25
g --Andropogon *Bromus
o
T," 20 -
o
e5
a) 15 -
L
B .
.9 Site p<0.001
(Z) 10 3 Species p<0.001
_ Time p<0.001
'6 Time*Site p<0.001
U)

S I I I l I
Jun-06 Jul-06 Aug-O6 Sep-O6 Oct-O6 Nov-06
McKay Field
25

 

+Andropogon --- Elymus

N
O
l

H
O
I

/——-

Jun-06‘ Jul-06 Aug-06 Sep-06 Oct-06 Nov-O6

Soil Moisture (°/o, O-lOcm)
a

 

 

U1

 

Figure 12. Soil moisture (0-10cm depth) for the two sites in 2006, averaged across the
+/-Litter treatment. Black lines represent Andropogon plots, while Gray lines represent

the C3 plots. Greenhouse-Geisser adjusted p-values for ln-transformed data are reported
below, and untransformed data are shown.

76

“rm. khan" ‘51-; I‘ll“,

 

Cellulose

 

80
Site p=0.143
j—g Placement p=0.268
1g 50 - +/-Litter p=0.036
“5
9:4
81
3
B 2
(U
E l

 

 

 

 

 

 

 

Turkey Meadow McKay Field

 

Figure 13. Mass loss (% of initial) for cellulose in both sites (McKay Field and Turkey

Meadow), placed under either Andropogon or the C3 species (Placement). Black solid
bars represent Andropogon plots with litter left intact (+litter), black striped bars
represent Andropogon plots where the litter was removed (-litter), gray solid bars

represent C3 plots with litter left intact (+litter), and gray striped bars represent C3 plots
where the litter was removed (-Iitter).

 

77

In a study comparing decomposition of Bromus inermis and Panicum virgatum (a C4
species similar to A ndropogon), Vinton and Goergen (2006) related its faster
decomposition rates for Bromus to its lower tissue C:N. The invasion of A egilops
triuncialis into serpentene grasslands slowed C and N cycling rates via slower
decomposition rates of its recalcitrant shoot tissue (Drenovsky and Batten 2007). My
results follow the patterns seen in these studies; faster decomposition occurs for the
species with the lower tissue C:N and ADF:N.

The significantly faster summer decomposition rates (April-Sept 2006) compared
to the winter rates (Nov 2005-April 2006) provided evidence of strong seasonal effects
on decomposition, regardless of tissue quality. In fact, initial decomposition rates were
often twice as fast in summer compared to winter (Figure 10). Decomposition proceeded
slowly through the winter months, resulting in low initial mass loss and a more linear
trend for mass remaining after one year. This slow initial decomposition in the winter
months was likely a result of cold temperatures limiting microbial activity. My findings
suggest that seasonal controls (i.e., temperature and moisture) on decomposition are
stronger than the effects of litter quality, and litter quality differences become important
when examining decomposition within a common environment (i.e., within a site).

The moisture levels of the litter at the time of collection were significantly higher
under A ndropogon plants (i.e., a Placement effect) compared to the C3 species,
suggesting that the microclimate under the Species may have differed. However, there
were no differences in decomposition rates when the litter was placed in locations with
different microenvironments (i.e., Placement under C3 and under C4 species), suggesting

that litter quality, and not microclimate, was the primary factor controlling the differences

78

 

 

in decomposition between species in this study. However, based on that experiment
alone, I could not determine whether there were no microclimate differences between C3
and C4 species or whether microclimate differences between C3 and C4 species had no
impact on decomposition.

The second experiment allowed me to determine whether microclimate
differences existed and if so, whether those differences affected decomposition rates. I
focused on soil moisture and temperature, as they are two important variables controlling
microbial activity. While there were few differences in soil moisture between C3 and C4

Species, temperatures at the soil surface and at 100m were significantly higher under C3

 

species. These temperature results were expected based on the small amount of surface
litter and aboveground biomass associated with the C3 species (Chapter Two), which
would provide less shade to moderate midday temperature increases.

The soil moisture results were surprising; based on the finding of greater moisture
content of litter placed under Andropogon, I expected to see soil moisture differences
between soils under Andropogon and soils under the C3 species. Another unexpected
result was that the presence or absence of litter (+/—Litter treatment) had no impact on
either soil moisture or temperature. Eviner (2004) found that species had a large
influence on soil temperature fluctuations, with lower daily fluctuations in plots with

higher graminoid Shoot biomass and litter. Summer afternoon temperatures were

 

negatively correlated with litter, and the relationship was reversed in winter (Eviner
2004). It is possible that the high Andropogon biomass (Chapter Two) provided a
maximum “Shading effect” on soil temperatures, and thus surface litter did not provide

any additional shading benefit. In addition, the relatively sparse surface litter in the C3

79

plots (Chapter Two) may not have Shaded the plots enough to alter temperature or
moisture. Eviner (2004) found differences in Species effects on soil moisture only during
the driest months and related those differences to aboveground biomass. Soil moisture
was only weakly related to surface litter (Eviner 2004). Any decrease in evaporation due
to shading by Andropogon litter and biomass could have been offset by higher
transpiration losses compared to the C3 species, and therefore no differences in soil
moisture were found between treatments.

Although I found differences in soil temperatures between species and treatments,

I found no evidence to suggest that microclimate influenced decomposition rates. There

 

were no significant differences in the decomposition rate of either C3 or C4 species
between the two microenvironments (placement under their own species versus placed
under the opposite Species), nor was there a difference in cellulose decomposition
between the two environments (under C3 and C4). This evidence suggests that while the
microbial communities may differ in many aspects, the microbial decomposers were
functionally similar under Andropogon and the C3 species. The only factor that affected

cellulose decomposition was the presence or absence of surface litter (+/- Litter

 

treatment), and soil moisture and temperature did not differ significantly between these

two treatments. Thus, it is unclear whether some aspect of microclimate other than soil

ramvmm _
. _ ~ . . V 1‘

moisture and temperature influenced cellulose decomposition.

Results presented here demonstrate that both litter quality and environmental
conditions affect decomposition rates. Litter quality controls relative decomposition rates
of species in a site, but seasonal differences in environmental conditions also impact

decomposition rates. While this study focused on only a few species, these results should

80

hold for a variety of C3 species common to old-fields and C4 species common in prairies.
The two C3 species examined here did not differ in their tissue quality, nor did they differ
in their decomposition rate. In Chapter Two, I found no Significant differences between
Bromus and Elymus in surface litter, shoot biomass, and tissue C:N or ADF :N between
the two Sites, suggesting that these two species are Similar in their effects on
decomposition. In addition, I also found no Significant differences in tissue C:N and
ADF:N between A ndropogon and two other C4 grasses (Schizachyrium scoparium and

Sorghastrum nutans) in Chapter Two, suggesting that those two C4 grasses would have

 

similar decomposition rates. This research suggests that communities dominated by C4 Ir-
grasses will impact C and N cycling via these slower decomposition rates compared to C3

grass dominated communities. This has particular implications for predicting how prairie

restorations will impacts soil C and N recovery processes over time. These results have

an application for the prediction of how invasive species may alter ecosystem processes,

and therefore which species we should be most concerned about from a belowground

ecosystem process perspective.

 

‘fl“.‘.~.l- '

81

CHAPTER FIVE

CONCLUSIONS

While much recent research has focused on the effects of exotic plant Species on
ecosystem properties, little is known about how reintroductions of extirpated species may
impact these processes in restored systems. I examined how the reintroduction of C4
grasses into old-fields influenced soil C and N cycling, and the timeframe over which I
these changes become apparent. The previous three chapters demonstrated that C3 and

C4 species differ in litter quality, quantity and their influences on soil microclimate. r: i

 

Biomass production and tissue chemistry traits were related to differences in soil
properties and processes, and the magnitude of species differences, as well as the time
since the Species was established, were important factors influencing the relative changes
in soil properties. However, because species differed in several traits, it was difficult to
isolate which plant traits are responsible for observed influences on soil processes. In
Chapter Two, I examined the decadal scale impacts of a shift from a C3-dominated
system to a C4-dominated system. In Chapter Three, I explored how quickly differences
in Species traits, and the soil process changes observed after 11 years, arise. Finally, in

Chapter Four, 1 evaluated which plant traits were responsible for differences in l

 

decomposition rates. I
In Chapters Two and Three, I showed that C4 species had Significantly greater

shoot biomass and more recalcitrant tissue compared to the dominant C3 species, and

these differences became apparent within two years of their establishment. However,

differences between the two groups of species in surface litter and root biomass tended to

82

take longer to develop, but were apparent after 11 years. While there was some evidence
to suggest that the C4 species had reduced soil inorganic N levels relative to the C3
Species after just two years, many of the changes in soil properties took longer than two
growing seasons to develop. After 1 1 years, soils under C4 Species had significantly
lower inorganic N levels, and slightly lower in situ net N mineralization and nitrification
rates when compared to soils under C3 species. I also found limited evidence for
increasing soil C pools under C4 species 11 years after reintroduction. Nevertheless, the
013C signal of the C4 species became measurable in the soil very quickly. Overall, my

results demonstrated that reintroduction of C4 species into grasslands can result in

 

alterations of soil processes related to C and N cycling on relatively Short timescales.
These results also indicated that process rates tended to change first, with changes in pool
sizes of C and N taking longer to become measurable.

However, because the C3 and C4 grasses examined in this study differ in three
main ways—litter quality, litter quantity, and timing of resource use/release—it is
difficult to determine which traits, or combination of traits, influence particular processes
in the field. The decomposition experiments (Chapter Four) took place in a more m
controlled setting to separate the direct and indirect effects of these trait differences on

decomposition rates. I found that while climate controls decomposition on a regional

scale, litter quality was the most important factor determining decomposition rates within

 

a Site. I also found evidence for differences in soil microclimate under Andropogon
compared to C3 species, but these differences did not appear to have a strong influence
on decomposition rates of either tissue type or of cellulose. Although the greater Shoot

biomass of C4 species is contributing to the larger surface litter layers compared to C3

83

species, Slowed decomposition rates of this more recalcitrant C4-derived litter is also
increasing surface litter mass. It is likely that both litter quality and quantity are
important factors contributing to increased soil C and reduced inorganic N levels under
C4 species. Soil microclimate differences may have some effect on soil processes, such
as extending microbial N mineralization longer into the autumn, but not on surface litter
decomposition rates.

In the future, I plan to follow up on these experiments. I hope to sample the 11-
and 2-year-old plots in the coming years to follow the changes in soil properties and
processes as the plots age. This fall, I will begin a laboratory experiment to examine the
soil microbial communities under Andropogon and the dominant C3 species at Turkey
Meadow and McKay Field, to see whether the microbial communities under these Species
have diverged over the last decade. I also will investigate whether the microbial
community under Andropogon is better adapted to decomposing recalcitrant litter
compared to the C3 species. This will provide an additional component to my research; I
did not examine whether there were changes in the microbial community composition,
abundance and activity under the different species. I will begin exploring this question as
part of my research at the Holden Arboretum in Ohio.

My dissertation research, as well as the research that I plan to continue in these
old-field sites, is centered on how plant species can influence soil processes. Improving
our understanding Of how plant species impact ecosystem properties and what Species
traits are driving these changes is imperative if we hope to predict the ecosystem-level

consequences of changes in species distribution or composition that could occur, and are

84

occurring, as a consequence of changes in agricultural and land use practices, global

change, and Species introductions.

85

 

 

APPENDIX

 

86

Turkey
Meadow

Louden
Field 0

Pond Lab

 

McKay Field

Photo credit: KBS LTER

 

Appendix 1. Location of the three study sites at the W. K. Kellogg Biological Station.

I

87

REFERENCES

88

REFERENCES

Aerts, R. 2006. The freezer defrosting: global warming and litter decomposition rates in
cold biomes. Journal of Ecology 94:713-724.

Averett, J. M., R. A. Klips, L. E. Nave, S. D. Frey, and P. S. Curtis. 2004. Effects of soil
carbon amendment on nitrogen availability and plant grth in an experimental
tallgrass prairie restoration. Restoration Ecology 12:568-5 74.

Baer, S. G., D. J. Kitchen, J. M. Blair, and C. W. Rice. 2002. Changes in ecosystem
structure and function along a chronosequence of restored grasslands. Ecological
Applications 12: 1 688-1701 .

Bengtson, P., U. Falkengren-Grerup, and G. Bengtsson. 2005. Relieving substrate
limitation-soil moisture and temperature determine gross N transformation rates.
OIKOS 111281-90.

 

Blumenthal, D. M., N. R. Jordan, and E. L. Svenson. 2005. Effects of prairie restoration
on weed invasions. Agriculture Ecosystems & Environment 107:221-230.

Brye, K. R., J. M. Norman, and S. T. Gower. 2002. Assessing the progress of a tallgrass
prairie restoration in Southern Wisconsin. American Midland Naturalist 1482218
235. -

Burbank, D. _H., K. S. Pregitzer, and K. L. Gross. 1992. Vegetation of the W.K. Kellogg
Biological Station, Kalamazoo County, Michigan. Research Report 510,
Michigan State University Agricultural Experiment Station, East Lansing.

Burke, 1. C., W. K. Lauenroth, and D. P. Coffin. 1995. Soil Organic-Matter Recovery in pa.
Semiarid Grasslands - Implications for the Conservation Reserve Program. 5‘
Ecological Applications 5:793-801.

Camill, P., M. J. McKone, S. T. Sturges, W. J. Severud, E. Ellis, J. Lirnmer, C. B. Martin,
R. T. Navratil, A. J. Purdie, B. S. Sandel, S. Talukder, and A. Trout. 2004.
Community- and ecosystem-level changes in a species-rich tallgrass prairie
restoration. Ecological Applications 14:1680-1694.

 

Carney, K. M., and P. A. Matson. 2005. Plant communities, soil microorganisms, and soil
carbon cycling: Does altering the world belowground matter to ecosystem
functioning? Ecosystems 8:928-940.

Chapin, F. S., E. S. Zavaleta, V. T. Eviner, R. L. Naylor, P. M. Vitousek, H. L. Reynolds,

D. U. Hooper, S. Lavorel, O. E. Sala, S. E. Hobbie, M. C. Mack, and S. Diaz.
2000. Consequences of changing biodiversity. Nature 405:234-242.

89

Collatz, G. J., J. A. Berry, and J. S. Clark. 1998. Effects of climate and atmospheric CO2
partial pressure on the global distribution of C-4 grasses: present, past, and future.
Oecologia 114:441-454.

Cornelissen, J. H. C. 1996. An experimental comparison of leaf decomposition rates in a
wide range'of temperate plant species and types. Journal of Ecology 84:573-582.

Craine, J. M., D. A. Wedin, F. S. Chapin, and P. B. Reich. 2003. Relationship between
the structure of root systems and resource use for 11 North American grassland
plants. Plant Ecology 165:85-100.

DeGryze, S., J. Six, K. Paustian, S. J. Morris, E. A. Paul, and R. Merckx. 2004. Soil
organic carbon pool changes following land-use conversions. Global Change
Biology 10:1120-1132.

Dijkstra, F. A., S. E. Hobbie, and P. B. Reich. 2006. Soil processes affected by sixteen
grassland species grown under different environmental conditions. Soil Science
Society of America Journal 70:770-777.

Drenovsky, R. E., and K. M. Batten. 2007. Invasion by Aegilops triuncialis (barb

goatgrass) slows carbon and nutrient cycling in a serpentine grassland. Biological
Invasions 9:107-116.

Ehrenfeld, J. G. 2003. Effects of exotic plant invasions on soil nutrient cycling processes.
Ecosystems 6:503-523.

Ehrenfeld, J. G., P. Kourtev, and W. Z. Huang. 2001. Changes In soil functions following
invasions of exotic understory plants In deciduous forests. Ecological
Applications 11: 1287-1300.

Emery, S. M., and K. L. Gross. 2006. Dominant species identity regulates invasibility of
old-field plant communities. OIKOS 115:549-558.

Evans, R. D., R. Rimer, L. Sperry, and J. Belnap. 2001. Exotic plant invasion alters
nitrogen dynamics in an arid grassland. Ecological Applications 11:1301-1310.

Eviner, V. T. 2004. Plant traits that influence ecosystem processes vary independently
among Species. Ecology 85:2215-2229.

Eviner, V. T., and F. S. Chapin. 2003. Functional matrix: A conceptual framework for
predicting multiple plant effects on ecosystem processes. Annual Review of

Ecology Evolution and Systematics 34:455-485.

Eviner, V. T., F. S. Chapin, and C. E. Vaughn. 2006. Seasonal variations in plant species
effects on soil N and P dynamics. Ecology 87:974-986.

90

 

 

Foster, B. L. 1996. Plant competition and diversity in relation to productivity in old-field
plant communities. Michigan State University, East Lansing, Michigan.

Foster, B. L. 1999. Establishment. competition and the distribution of native grasses
among Michigan old-fields. Journal of Ecology 87:476-489.

Foster, B. L., and K. L. Gross. 1997. Partitioning the effects of plant biomass and litter on
Andropogon gerardi in old-field vegetation. Ecology 78:2091-2104.

Gotshall, T. B. 1972. The vegetation of Kalamazoo County at the time of settlement.
Pages 1-21 in R. Brewer, editor. The ecology of Kalamazoo County. Western
Michigan University Press, Kalamazoo, Michigan.

Grandy, A. S., T. D. Loecke, S. Parr, and G. P. Robertson. 2006. Long-term trends in
nitrous oxide emissions, soil nitrogen, and crop yields of till and no-till cropping
systems. Journal of Environmental Quality 35:1487-1495.

Gross, K. L., and S. M. Emery. 2007. Succession and Restoration in Michigan Old-field
Communities. Pages 221-243 in V. A. Cramer and R. J. Hobbs, editors. Old
fields: Dynamics and Restoration of Abandoned Farmland. Island Press.

Hobbie, S. E. 1992. Effects of plant species on nutrient cycling. TREE 71336-339.

Hobbie, S. E., P. B. Reich, J. Oleksyn, M. Ogdahl, R. Zytkowiak, C. Hale, and P.
Karolewski. 2006. Tree species effects on decomposition and forest floor
dynamics in a common garden. Ecology 87:2288-2297.

Hooper, D. U., F. S. Chapin, J. J. Ewel, A. Hector, P. Inchausti, S. Lavorel, J. H. Lawton,
D. M. Lodge, M. Loreau, S. Naeem, B. Schmid, H. Setala, A. J. Symstad, J.
Vanderrneer, and D. A. Wardle. 2005. Effects of biodiversity on ecosystem
functioning: A consensus of current knowledge. Ecological Monographs 7513-35.

Hooper, D. U., and P. M. Vitousek. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 12 1-1 49.

Howe, H. F. 1994. Managing Species-Diversity in Tallgrass Prairie - Assumptions and
Implications. Conservation Biology 8:691-704.

Inouye, R. S., and D. Tilman. 1988. Convergence and Divergence of Old-Field Plant-
Communities Along Experimental Nitrogen Gradients. Ecology 69:995-1004.

Inouye, R. S., and D. Tilman. 1995. Convergence and Divergence of Old-Field
Vegetation after 1 1 Yr of Nitrogen Addition. Ecology 76:1872-1887.

Jastrow, J. D. 1987. Changes in Soil Aggregation Associated with Tallgrass Prairie
Restoration. American Journal of Botany 74:1656-1664.

91

Jastrow, J. D., J. E. Amonette, and V. L. Bailey. 2007. Mechanisms controlling soil
carbon turnover and their potential application for enhancing carbon
sequestration. Climatic Change 8025-23.

Jastrow. J. D.. R. M. Miller, and J. Lussenhop. 1998. Contributions of interacting
biological mechanisms to soil aggregate stabilization in restored prairie. Soil
Biology & Biochemistry 30:905-916.

Kindscher, K., and L. L. Tieszen. 1998. Floristic and soil organic matter changes after
five and thirty-five years of native tallgrass prairie restoration. Restoration
Ecology 6: 1 81-196.

Lawton, J. H., and R. M. May, editors. 1995. Extinction Rates. Oxford University Press,
New York.

Lovett, G. M., K. C. Weathers, M. A. Arthur, and J. C. Schultz. 2004. Nitrogen cycling in
a northern hardwood forest: Do species matter? Biogeochemistry 67:289-308.

Mack, M. C., and C. D'Antonio. 2003a. Exotic grasses alter controls over soil nitrogen
dynamics in a Hawaiian woodland. Ecological Applications 13:154-166.

Mack, M. C., and C. M. D'Antonio. 2003b. The effects of exotic grasses on litter
decomposition in a Hawaiian woodland: The importance of indirect effects.
Ecosystems 61723-73 8.

Mack, M. C., C. M. D'Antonio, and R. E. Ley. 2001. Alteration of ecosystem nitrogen
dynamics by exotic plants: A case study of C-4 grasses in Hawaii. Ecological
Applications 11:1323-1335.

Martin, L. M., K. A. Moloney, and B. J. Wilsey. 2005. An assessment of grassland
restoration success using species diversity components. Journal of Applied
Ecology 42:327-336.

McLauchlan, K. K., S. E. Hobbie, and W. M. Post. 2006. Conversion from agriculture to
grassland builds soil organic matter on decadal timescales. Ecological
Applications 16:143-153.

Mlot, C. 1990. Restoring the Prairie. BioScience 40:804-809.

Pimm. S. L., G. J. Russell, J. L. Gittleman, and T. M. Brooks. 1995. The future of
biodiversity. Science 269:347-350.

Reich, P. B., D. Tilman, S. Naeem, D. S. Ellsworth, J. Knops, J. Craine, D. Wedin, and J.
Trost. 2004. Species and functional group diversity independently influence
biomass accumulation and its response to CO2 and N. Proceedings of the
National Academy of Sciences ofthe United States of America 101:10101-10106.

92

Rillig, M. C.. and D. L. Mummey. 2006. Mycorrhizas and soil structure. New Phytologist
171241-53.

Rillig, M. C., S. F. Wright, and V. T. Eviner. 2002. The role of arbuscular mycorrhizal
fungi and glomalin in soil aggregation: comparing effects of five plant species.
Plant and Soil 238:325-333.

Robertson, G. P., D. C. Coleman, C. S. Bledsoe, and P. Sollins, editors. 1999. Standard
soil methods for long-term ecological research. Oxford University Press, New
York.

Samson, F ., and F. Knopf. 1994. Prairie Conservation in North-America. BioScience
44:418-421.

Samson, R., S. Mani, R. Boddey, S. Sokhansanj, D. Quesada, S. Urquiaga, V. Reis, and
C. H. Lem. 2005. The potential of C-4 perennial grasses for developing global
BIOHEAT industry. Critical Reviews in Plant Sciences 24:461-495.

Sanderson, M. A., P. R. Adler, A. A. Boateng, M. D. Casler, and G. Sarath. 2006.
Switchgrass as a biofuels feedstock in the USA. Canadian Journal of Plant
Science 86:1315-1325.

Sluis, W. J. 2002. Patterns of species richness and composition in re-created grassland.
Restoration Ecology 10:677-684.

Suding, K. N., K. L. Gross, and G. R. Houseman. 2004. Alternative states and positive
feedbacks in restoration ecology. Trends in Ecology & Evolution 19:46-53.

Tilman, D., J. Hill, and C. Lehman. 2006. Carbon-negative biofuels from low-input high-
diversity grassland biomass. Science 314:1598-1600.

Trofymow, J. A., T. R. Moore, B. Titus, C. Prescott, I. Morrison, M. Siltanen, S. Smith, J.
Fyles, R. Wein, C. CamirT, L. Duschene, L. Kozak, M. Kranabetter, and S.
Visser. 2002. Rates of litter decomposition over 6 years in Canadian forests:
influence of litter quality and climate. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere 32:789-804.

Vinton, M. A., and E. M. Goergen. 2006. Plant-soil feedbacks contribute to the
persistence of Bromus inermis in tallgrass prairie. Ecosystems 92967-976.

Vitousek, P. M., and L. R. Walker. 1989. Biological Invasion by Myrica-Faya in Hawaii -

Plant Demography, Nitrogen-Fixation, Ecosystem Effects. Ecological
Monographs 59:247-265.

93

Wardle, D. A., R. D. Bardgett, J. N. Klironomos, H. Setala, W. H. van der Putten. and D.
H. Wall. 2004. Ecological linkages between aboveground and belowground biota.

Science 304: 1629-1633.

Wardle, D. A., G. M. Barker, K. I. Bonner, and K. S. Nicholson. 1998. Can comparative
approaches based on plant ecophysiological traits predict the nature of biotic
interactions and individual plant species effects in ecosystems? Journal of
Ecology 86:405-420.

Wardle, D. A., K. I. Bonner, and K. S. Nicholson. 1997. Biodiversity and plant litter:
Experimental evidence which does not support the view that enhanced species
richness improves ecosystem function. OIKOS 79:247-258.

Wedin, D. A., and J. Pastor. 1993. Nitrogen Mineralization Dynamics in Grass
Monocultures. Oecologia 96: 186-192.

Wedin, D. A., and D. Tilman. 1990. Species effects on nitrogen cycling: A test with
perennial grasses. Oecologia 84:433-441.

Williams, D. W., L. L. Jackson, and D. D. Smith. 2007. Effects of frequent mowing on
survival and persistence of forbs seeded into a species-poor grassland. Restoration
Ecology 15:24-33.

Winslow, J. C., E. R. Hunt, and S. C. Piper. 2003. The influence of seasonal water
availability on global C-3 versus C-4 grassland biomass and its implications for
climate change research. Ecological Modelling 163:153-173.

Xu, X. N., and E. J. Hirata. 2005. Decomposition patterns of leaf litter of seven common
canopy species in a subtropical forest: N and P dynamics. Plant and Soil 273:279-

289.

Zak, D. R., W. E. Holmes, D. C. White, A. D. Peacock, and D. Tilman. 2003. Plant
diversity, soil microbial communities, and ecosystem function: are there any
links? Ecology 84:2042-2050.

Zavaleta, E. S., and K. B. Hulvey. 2007. Realistic variation in species composition affects

grassland production, resource use and invasion resistance. Plant Ecology 188:39-
51.

94

 

 

MICHIGAN STATE UHIVEFISTY LIBR 155
l ‘1 1
l

IIIIIIIIIIIIIII lllllllllllllllllllllllllll

  
     

l

l i

ll . i
3 1293 02955 0459