SOIL NUTRIENT DYNAMICS AND SHIFTING MICROBIAL PROCESSES WITH
INCREASING ROTATIONAL COMPLEXITY IN ROW CROPS
By
Brendan Edward Neill

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Crop and Soil Sciences – Doctor of Philosophy
Ecological Food and Farming Systems – Dual Major
2017

ABSTRACT
SOIL NUTRIENT DYNAMICS AND SHIFTING MICROBIAL PROCESSES WITH
INCREASING ROTATIONAL COMPLEXITY IN ROW CROPS
By
Brendan Edward Neill
Agriculture faces pressure to supply increased food, fiber, and fuel, while provisioning
ecosystem services. Intensive agriculture has contributed to altered global carbon (C) and
nitrogen (N) cycles, and is the leading source of nitrous oxide (N2O), a powerful greenhouse gas.
New approaches are needed to improve soil nutrient management while sustaining agricultural
productivity. In natural systems a strong relationship exists between aboveground diversity and
belowground ecosystem controls on nutrients. While increasing the number of crops in rotation
in agroecosystems can have multiple benefits, the linkages between rotational diversity, nutrient
cycling pathways, and the soil ecosystem remain poorly understood. My dissertation addresses
the effects of increasing crop rotational complexity on soil C and N cycles and the soil ecosystem
process that regulate them. I extend this work by testing ecosystem-based measures of soil
health on Michigan farms.
Long term cropping system experiments are ideal sites to examine how rotational legacy
shapes belowground processes. Chapters 1, 2, and 3, focus on a cropping biodiversity
experiment in place for 10 years at the initiation of my study. The gradient ranges from
continuous summer annuals – corn (Zea mays L.) and soybean (Glycine max.) – to rotations that
also include a winter annual (wheat; Triticum aestivum L.), to complex rotations with
overwintering cover crops. This experiment has not received any external inputs such as
fertilizer or pest control agents, so that changes across the gradient are narrowed to the effect of
crop rotational complexity. In Chapter 1 I examine how labile C and N pools, soil enzyme

activities, and soil respiration respond to increasing rotational complexity. I found that pools of
potentially mineralizable C were nearly twice as high on fields with a history of cover crops,
compared to those without. Rotations with a legacy of cover crops sustained higher enzyme
activities, significantly higher soil respiration and accumulated significantly higher total soil
organic matter.
In chapter 2, I tested the relationship between increasing the number of crops in rotation
and the species diversity of soil bacteria. I found no significant difference in species diversity of
bacteria, but a shift in the community between rotations with and without cover crops. Taxa
responsible for this shift were mainly from the Acidobacteria and the Proteobacteria which are
characterized by contrasting growth and energy use strategies. I focused on denitrification in
Chapter 3, a process carried out by soil microbes that produces N2O. Rotations with cover crops
had significantly higher mean N2O flux over two growing seasons. Enzyme assays showed that
denitrification was more efficient on these rotations, and rotations with cover crops also had a
significantly higher proportion of genes in the N2O-production pathway that derived from
ammonia oxidizing bacteria.
Finally, in Chapter 4 I tested soil health on fields that Michigan farmers had designated as
having either good or poor soil quality. Testing captured soil variability on farmer fields, but
interviews with farmers revealed caveats to implementing soil health testing. My work on
increasing rotational complexity revealed novel microbial controls on soil C cycling and N2O
flux, but ultimately implementing practices that enhance soil ecosystem function depends on
human decisions about land use, crop production, and environmental outcomes.

Copyright by
Brendan Edward Neill
2017

Dedicated to my parents

v

ACKNOWLEDGEMENTS

The completion of this dissertation would not have been possible without support and
guidance from an extraordinary set of mentors, colleagues, friends, and family. First and
foremost as an advisor Dr. Thomas Schmidt always guided me to think critically and strive for
the crux of every unique scientific problem. I am grateful for his superior insights as a
microbiologist, his great appetite for learning, and for his endless patience and positive feedback.
Dr. G. Philip Robertson, with generosity and wisdom, provided guidance on the smallest details
of my work while shaping my development as a biogeochemist and acting as a model of
consummate scientific scholarship.
I would also like to thank the other members of my committee. Dr. Stuart Grandy led the
work that drew me to Michigan State University and offered incisive feedback and critical
support that guided the core of my research. Dr. John Kerr constantly challenged me to think
about how my research was relevant to the challenges of sustaining natural resources. Dr. Sasha
Kravchenko sharpened my critical analysis, especially in the complex endeavor that is soil
science, and also generously provided laboratory space for my work.
My dissertation journey spanned multiple laboratories and institutions which served to
enrich my experience with many excellent lab mates and colleagues. From the Schmidt lab I
thank Kwi Kim, Tracy Teal, Clegg Waldron, Heli Juottonen, Kevin Theis, John Dover, Bjorn
Ostman, Keara Towery, Steve Stoddard, Ben Roller, Chris Wright, Arvind Venkataraman,
Jessica Sieber, Rachel Morris, Alex Schmidt, Dishari Mukherjee, Robert Hein, and Byron Smith.
From the Robertson lab and affiliates at the Kellogg Biological Station I am grateful to Bonnie
McGill, Steve Hamilton, Suzanne Sippel, Sven Bohm, Dustin Kinkaid, Kate Glanville, Di Liang,

vi

Adam Reimer, Sarah Roley, Ilya Gelfand, Mike Abraha, Leilei Ruan, and critical support from
Joe Simmons, Kevin Kahmark, Neville Millar, Stacey VanderWulp, Cathy McMinn, and Jane
Schuette. I was also lucky to have excellent technicians and REUs, Marcia Croft, Alessandra
Zuniga, and Andrew Galimberti.
I received invaluable support from Dr. James Kells for my work in the department of
Plant, Soil and Microbial Sciences. In addition, I thank Dean Bass, MSU Extension staff Paul
Gross and James DeDecker, and Colleen Forestieri of the Van Buren Conservation District as
well as the 13 farmers that participated in research for this dissertation. For my involvement in
the KBS GK-12 program I thank Dr. Tom Getty and Dr. Sarah Bodbyl Roels, as well as Phil
Robertson. I feel incredibly fortunate to have been a part of KBS, and to have interacted with all
who work there, especially Dr. Julie Doll and Dr. Katherine Gross. I also owe special gratitude to
several collaborators and friends: Christine Sprunger, Krista Isaacs, Steve Culman, Ariane
Peralta, Marshall McDaniel, Erin Haramoto, and Carolyn Lowry.
Finally I received ceaseless support and encouragement from my parents, Mary Ellen and
Kerby, my sister Elizabeth, brothers Kerby, Matthew, Sean and Julio, and from Jennifer Blesh.
I love you all.
Lastly, I would like to thank my funding sources: The MSU Center for Regional Food
the USDA, the LTER, the KBS GK-12 Bioenergy Sustainability Project, the USDA-SARE
graduate student grant, and KBS summer fellowship.

vii

TABLE OF CONTENTS

LIST OF TABLES .........................................................................................................................xi
LIST OF FIGURES........................................................................................................................xv
CHAPTER 1: DISSERTATION INTRODUCTION.......................................................................1
OVERVIEW..............................................................................................................................1
DISSERTATION OBJECTIVES .............................................................................................3
REFERENCES .........................................................................................................................4
CHAPTER 2: CROP ROTATIONAL COMPLEXITY SHIFTS MICROBIAL
PROCESSING AND STORAGE OF SOIL ORGANIC CARBON ...............................................7
ABSTRACT ...............................................................................................................................7
INTRODUCTION .....................................................................................................................8
METHODS ..............................................................................................................................12
Site descriptions and experimental layout .........................................................................12
Sample collection ...............................................................................................................13
Soil carbon and nitrogen ....................................................................................................14
Soil enzymes ......................................................................................................................15
In situ soil respiration.........................................................................................................16
Statistical approach ............................................................................................................17
RESULTS ................................................................................................................................17
Soil carbon and nitrogen ....................................................................................................17
Soil enzyme activity...........................................................................................................18
In situ soil respiration.........................................................................................................20
Labile nitrogen ...................................................................................................................21
DISCUSSION ..........................................................................................................................22
Synthesis ............................................................................................................................27
Implications........................................................................................................................28
ACKNOWLEDGEMENTS .....................................................................................................29
APPENDIX ..............................................................................................................................30
REFERENCES ........................................................................................................................48
CHAPTER 3: COVER CROPS CHANGE SOIL BACTERIAL COMPOSITION BUT
NOT DIVERSITY .........................................................................................................................55
ABSTRACT .............................................................................................................................55
INTRODUCTION ...................................................................................................................55
METHODS ..............................................................................................................................58
Site description...................................................................................................................58
Soil sampling .....................................................................................................................59
Soil DNA extraction and PCR amplification .....................................................................59
Analysis of 16S bacterial communities..............................................................................60
RESULTS ................................................................................................................................61

viii

DISCUSSION ..........................................................................................................................63
ACKNOWLEDGEMENTS .....................................................................................................67
APPENDIX ..............................................................................................................................68
REFERENCES ........................................................................................................................81
CHAPTER 4: LEGUME COVER CROPS ALTER MICROBIAL CONTROLS ON SOIL
DENITRIFICATION RATES AND EFFICIENCY .....................................................................88
ABSTRACT .............................................................................................................................88
INTRODUCTION ...................................................................................................................89
METHODS ..............................................................................................................................92
Site description and experimental layout ...........................................................................92
Soil sampling .....................................................................................................................93
Soil carbon and nitrogen ....................................................................................................93
Soil enzymes ......................................................................................................................95
In situ nitrous oxide ...........................................................................................................95
Denitrification enzyme assay (DEA) ................................................................................ 96
Nitrite reductase (nirK) PCR and quality filtering .............................................................97
Statistical approach ............................................................................................................98
RESULTS ................................................................................................................................98
Soil C and N .......................................................................................................................98
Nitrous oxide......................................................................................................................99
DEA .................................................................................................................................100
nirK PCR amplicons ........................................................................................................101
DISCUSSION ........................................................................................................................101
ACKNOWLEDGEMENTS ...................................................................................................107
APPENDIX ............................................................................................................................108
REFERENCES ......................................................................................................................126
CHAPTER 5: ALINGING SOIL HEALTH TESTING AND FARMER KNOWLEDGE
FOR IMPROVED ON-FARM SOIL MANAGEMENT .............................................................134
ABSTRACT ...........................................................................................................................134
INTRODUCTION .................................................................................................................135
METHODS ............................................................................................................................138
Experimental approach ....................................................................................................138
Field sampling ..................................................................................................................139
Soil testing .......................................................................................................................140
Statistical approach ..........................................................................................................143
Farmer interviews ............................................................................................................143
RESULTS ..............................................................................................................................144
Comparing soil test parameters across field types ...........................................................144
Farmer interviews on soil health test results ....................................................................145
DISCUSSION ........................................................................................................................147
Comparing soil testing approaches ..................................................................................147
How farmers make soil management decisions ...............................................................148
Aligning soil quality with soil testing ..............................................................................149
Emerging themes .............................................................................................................150

ix

I. Soil health methods need to be solid and demonstrable .........................................150
II. Testers need to know how to use the test, and what its limits are.........................150
III. Soil health tests need to adapt to regional differences in soil and farm
practices .....................................................................................................................151
Synthesis ..........................................................................................................................151
Conclusions ......................................................................................................................153
ACKNOWLEDGEMENTS ...................................................................................................154
APPENDIX ............................................................................................................................155
REFERENCES ......................................................................................................................175

x

LIST OF TABLES

Table 2.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment .......................31
Table 2.2 Soil organic C (SOC) and N (SON) and their ratio (C:N), in the biodiversity gradient
treatments (see table 2.1 for treatment acronyms). Values are means (standard error),
n=4 replicate blocks. Significance of treatment effect: * P ≤ 0.05 ....................................32
Table 2.3 Beta-glucosidase (BG), N-acetyl-glucosaminidase (NAG), tyrosine amino peptidase
(TAP), acid phosphatase (PHOS), phenol oxidase (PHENOX), peroxidase (PEROX)
in the biodiversity gradient treatments (see table 2.1 for treatment acronyms) for the
two study years. Values are means (standard error), n=4 replicate blocks. Significance
of treatment effect: * P ≤ 0.05, ** P = 0.01 to 0.001. Each year was analyzed
separately ...........................................................................................................................33
Table 2.4 Potentially mineralizable C (PMC), ratio of phenol oxidase and peroxidase
(OX) to PMC ( OX:PMC), to beta-glucosidase BG:OX, to N-acetylglucosaminidase and tyrosine amino peptidase (NAG+TAP:OX), to acid
phosphotase (PHOS:OX), in the biodiversity gradient treatments (see table 2.1 for
treatment acronyms) for the two study years. Values are means (standard error), n=4
replicate blocks. Significance of treatment effect: * P ≤ 0.05, ** P = 0.01 to 0.001.
Each year was analyzed separately ....................................................................................34
Table 2.5 In situ carbon dioxide fluxes in the biodiversity gradient treatments (see table 2.1 for
treatment acronyms) for the two study years. Median, maximum (Max) and mean
(standard error) for n=4 replicate blocks. Significance of treatment effect: * P ≤ 0.05,
** P = 0.001 to 0.01, *** P ≤ 0.001. Each year was analyzed separately ........................35
Supplemental Table 2.1 Inorganic N (NH4+-N and NO3--N), potentially mineralizable N (PMN),
in the biodiversity gradient treatments (see table 2.1 for treatment acronyms) for the two
study years. Values are means (standard error), n=4 replicate blocks. Significance of
treatment effect: * P <0.05, ** P<0.01. Each year was analyzed separately...................42
Supplemental Table 2.2 In situ carbon dioxide fluxes in the biodiversity gradient
treatments for individual time points over two years with mean (standard error)
for n=4 replicate blocks, followed by letters denoting significant differences
between treatments. See table 2.1 for treatment acronyms. ...............................................43
Table 3.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment .......................69
Table 3.2 Estimated population parameters 16S rDNA from rotations across the biodiversity
gradient experiment (see table 3.1 for treatment acronyms), including mean ± se for
coverage, estimated OTU richness (Chao1) and diversity (Inverse Simpsons).................70

xi

Table 3.3 Multivariate analysis of 16S rDNA from the biodiversity gradient experiment
using perMANOVA with two factors different factors, rotation and presence of a
cover crop, showing the sources of variation, degrees of freedom (DF), sums of
squares (SS) mean square (MS), F-statistic (F) and related coefficient of
determination (R2) and probability (Prob.). Analysis of similarities (ANOSIM)
between grouped factors showing the correlation coefficient (R) and probability
(Prob) .................................................................................................................................71
Table 3.4 Analysis of molecular variance (Amova) of 16S OTUs, with treatment-level
(Overall) F-statistic (P) and table of pair-wise comparisons of individual treatments
across the biodiversity gradient experiment. See table 3.1 for treatment acronyms..........72
Table 3.5 Vector analysis of correlation between matrix of edaphic factors (Euclidean)
and 16S rDNA (Bray-Curtis) for two years on the biodiversity gradient experiment.
For each year the goodness of fit (r2) and the significance, P, was calculated for soil
paramaters: inorganic N (NH4+-N and NO3--N), potentially mineralizable N (PMN)
and C (PMC), permanganate oxidizable C (POXC), enzyme activities for betaglucosidase (BG), peroxidase (PEROX) + phenol oxidase (PHENOX), N-acetylglucosaminidase (NAG) + tyrosine amino peptidase (TAP), acid phosphatase
(PHOS)...............................................................................................................................73
Table 3.6 The 20 the most influential OTUs ranked by their abundance (after normalizing

and sub-setting across all treatments and replicates) and the significance (P≤0.05)
of their influence on distinguishing cover crop and non-cover crop treatments................74
Table 4.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment .....................109
Table 4.2 Potentially mineralizable N (PMN), N-acetyl-glucosaminidase + tyrosine
amino peptidase (NAG+TAP), inorganic N (NH4+-N and NO3--N), potentially
mineralizable C (PMC) and permanganate oxidizable (POX) C on the biodiversity
gradient treatments (see table 4.1 for treatment acronyms) for the two study years.
Values are means (standard error), n=4 replicate blocks. Significance of treatment
effect: * P ≤ 0.05, **P<0.01 ............................................................................................110
Table 4.3 In situ nitrous oxide fluxes in the biodiversity gradient treatments for the two
study years. Median, maximum (Max) and mean (standard error) for n=4 replicate
blocks, followed by letters denoting significant differences between treatments. See
table 4.1 for treatment acronyms .....................................................................................111
Table 4.4 Model reduction to best fit with N2O as response variable and predictor variables
presence of cover (0,1,2), moisture content, nitrate (NO3-), ammonium (NH4+),
potentially mineralizable N (PMN) and N-acetyl-glucosaminidase + tyrosine amino
peptidase (NAG+TAP). Random variables in the equation are treatment block and
time point of flux measurement (block:time point) .........................................................112

xii

Supplemental Table 4.1 Soil organic C (SOC) and N (SON), and their ratio (C:N), in the
biodiversity gradient treatments. Values are means (standard error), n=4 replicate
blocks. Significance of treatment effect: * P ≤ 0.05. See table 4.1 for treatment
acronym............................................................................................................................117
Supplemental Table 4.2 Denitrification enzyme potential assay (DEA) showing N2O
flux, total denitrification (N2O + N2) and relative rates of N2O production
(rN2O) on rotations in the biodiversity gradient experiment. Values are means
(standard error), n=4 replicate blocks, with overall treatment significance
above values and letters denoting significant difference between rotational
treatments at P ≤ 0.05. See table 4.1 for treatment acronym ..........................................118
Supplemental Table 4.3 Total number of nirK sequence reads, and within that pool the
number of sequences that clustered with known sequences of nirK from ammonia
oxidizing bacteria (AOB). Treatments are from the biodiversity cropping
experiment followed by a simultaneous comparison with sequence from a nearby
deciduous forest (DF) and conventional row crop system (T1), for the years 2008
and 2009. See table 4.1 for treatment acronym ...............................................................119
Supplemental Table 4.4 In situ nitrous oxide fluxes in the biodiversity gradient treatments
for individual time points over two years with mean (standard error) for n=4
replicate blocks, followed by letters denoting significant differences between
treatments. See table 4.1 for treatment acronym..............................................................120
Table 5.1 Percent sand and soil texture class from 13 farms in three Michigan counties for
four field types: “Best Field” and “Worst Field” as characterized by farmers, a
non-row crop field (“NRC”) and a “Choice” field also selected by each farmer ............156
Table 5.2 Percent aggregate stability (AS), available water capacity (AWC), soil
compaction measures as resistance at two soil depths shown as means ± se for
across Michigan farms (n=13). Field types are “Best Field” and “Worst Field” as
characterized by farmers, a non-row crop field (“NRC”) and a “Choice” field also
selected by each farmer. Field comparisons are t-tests comparing paired field
types for each farm for each soil parameter. Significance of treatment effect: * P ≤
0.05, ** P = 0.01 to 0.001, ***P<0.001 ..........................................................................158
Table 5.3 Percent organic matter (OM), permanganate oxidizable carbon (POXC),
potentially mineralizable carbon (PMC) and nitrogen (PMN) and nitrification
potential (NIT) shown as means ± se for across Michigan farms (n=13). Field
types are “Best Field” and “Worst Field” as characterized by farmers, a non-row
crop field (“NRC”) and a “Choice” field also selected by each farmer. Field
comparisons are t-tests comparing paired field types for each farm for each soil
parameter. Significance of treatment effect: * P ≤ 0.05, ** P = 0.01 to 0.001,
***P<0.001 ......................................................................................................................159
Table 5.4 Soil pH, soil concentration Bray1 Phosphorus (P), potassium (K), calcium (Ca),
magnesium (Mg), ammonium (NH4+) and nitrate (NO3-) shown as means ± se

xiii

for across Michigan farms (n=13). Field types are “Best Field” and “Worst Field”
as characterized by farmers, a non-row crop field (“NRC”) and a “Choice” field also
selected by each farmer. Field comparisons are t-tests comparing paired field types
for each farm for each soil parameter. Significance of treatment effect: * P ≤ 0.05 ......160
Table 5.5 Field observations from sampling on 4 field types on 13 farms across Michigan ......161
Table 5.6 Responses from 13 farmer interviews regarding soil testing and information
used for soil management ................................................................................................168
Table 5.7 Responses from 13 farmer interviews regarding their selection of field types for
soil sampling on their farms .............................................................................................167
Table 5.8 Responses from 13 farmer interviews regarding their management of fields
selected for soil sampling on their farms ........................................................................171
Table 5.9 Condensed responses from 13 farmer interviews regarding soil health test results
from their farms ...............................................................................................................172
Table 5.10 Condensed responses from 13 farmer interviews regarding approach to soil
management on their farms ..............................................................................................173

xiv

LIST OF FIGURES

Figure 2.1 Potentially mineralizable C (PMC) in the biodiversity gradient treatments (see
table 2.1 for treatment acronyms) for a corn year (2010) and a soybean year (2011).
Values are means with standard error for each month, n=4 replicate blocks....................36
Figure 2.2 Enzyme activities for beta-glucosidase (BG), peroxidase (PEROX) +
phenol oxidase (PHENOX), N-acetyl-glucosaminidase (NAG) + tyrosine
amino peptidase (TAP), acid phosphatase (PHOS). All points are means
across blocks (n=4) with standard errors for a corn year (left) and following
soybean year (right). See table 2.1 for treatment acronyms...............................................37
Figure 2.3 Ratio of soil phenol oxidase (PHENOX) + peroxidase (PEROX) enzyme
activities to potentially mineralizable carbon (PMC). All points are means
across blocks (n=4) with standard errors for a corn year (left) and following
soybean year (right). Letters indicate statistical difference (P≤0.05) based on
post hoc Tukey’s test. See table 2.1 for treatment acronyms.............................................38
Figure 2.4 In situ carbon dioxide flux measurements for KBS biodiversity gradient (see table
2.1 for treatment acronyms) for two years. Time points are means with standard error,
n=4. Asterisks (*) show significant differences (P≤ 0.05) by treatments at a specific
time point. In 2010, cS was left out of treatment comparisons, as was cC in 2011..........39
Figure 2. 5 Canonical correspondence analysis plot of KBS biodiversity gradient for
combined two years of study with differences in rotations constrained to
measured soil parameters and the two axes representing the greatest explained
variation (on a percentage basis). Blue ellipse encircles treatments with cover
crops. Figures are from cumulative measures over the season for all parameters
except pH (which represents a mean value). Other parameters are Moisture
(soil moisture), inorganic N (NH4+ and NO3-), potentially mineralizable N
(PMN) and C (PMC), permanganate oxidizable C (POXC), enzyme activities for
beta-glucosidase (BG), peroxidase (PEROX) + phenol oxidase (PHENOX), Nacetyl-glucosaminidase (NAG) + tyrosine amino peptidase (TAP), acid
phosphatase (PHOS). All ordinations were created with a correlational matrix
to give equal weights to all soil parameters. See table 2.1 for treatment acronyms. .........40
Figure 2.6 Model for C transformation in contrasting continuous ‘Monoculture’ (cC) and
‘Complex Rotations’ (CSW2cov) on the KBS biodiversity gradient experiment.
Boxes and solid arrows for specific parameters are in relative proportion to
measured values from this study. Dashed arrows represent posited, but
unknown C flows. Green arrows at the top are plant inputs entering a labile
pool of potentially mineralizable C (PMC). Yellow arrows represent the
magnitude of hydrolytic enzyme activity (H E) and gray arrows are oxidative
enzyme activities (Ox E), which draw soil C substrates into microbes (gray box).
Passing out of microbial cells is respired CO2 (black arrow) as well as some

xv

unknown quantity of C that passes back into a labile C fraction via a trophic
cascade (red dashed arrow) and microbial-processed C which complexes with soil
as ‘Mineral bound C’ (dark green dashed arrow, also unknown). Stable C which
forms the bulk of SOC (light brown box) derives both from microbially-processed
C and from ‘Recalcitrant plant C’ (dashed black arrow) which is not readily
accessed by microbial activity and passes into longer term SOC pools ............................41
Supplemental Figure 2.1 Ratio of Potentially mineralizable C (PMC) to soil enzyme
activities, beta-glucosidase (BG:PMC), to N-acetyl-glucosaminidase and tyrosine
amino peptidase (NAG+TAP:PMC), to acid phosphotase (PHOS:PMC), in the
biodiversity gradient treatments. See table 2.1 for treatment acronyms. All points are
means across blocks (n=4) with standard errors for a corn year (left) and following
soybean year (right)...........................................................................................................44
Supplemental Figure 2.2 Ratio of soil phenol oxidase (PHENOX) + peroxidase (PEROX)
enzyme activities to beta-glucosidase (BG:OX), to N-acetyl-glucosaminidase
and tyrosine amino peptidase (NAG+TAP:OX), to acid phosphatase (PHOS:OX), in
the biodiversity gradient treatments. All points are means across blocks (n=4) with
standard errors for a corn year (left) and following soybean year (right). Letters
indicate statistical difference (P≤0.05) based on post hoc Tukey’s test. See table 2.1 for
treatment acronyms. ...........................................................................................................45
Supplemental Figure 2.3 Inorganic N (NH4+-N and NO3--N) in the biodiversity gradient
Treatments (see table 2.1 for treatment acronyms) for the two study years. All points are
means across blocks (n=4) with standard errors for a corn year (left) and following
soybean year (right)...........................................................................................................46
Supplemental figure 2.4 Net primary productivity (NPP) with standard error from the KBS
Biodiversity Gradient Experiment. Estimated NPP calculated as mean of three years
(longest period for a full rotation) for 2010 is the mean NPP of 2008, 2009 and 2010
only for the plots examined in this study. A 6 year mean is estimated NPP for all
phases of each rotation from years 2008 to 2014 (each appearing every year for all
treatments on the experimental site). Both these calculations are determined from the
harvest index following Bolinder et al. (2007) and excluding biomass removed as
harvested grain. Non-cash crop aboveground NPP (including cover crop and weeds)
was also calculated from fall biomass collections. See table 2.1 for treatment
acronyms............................................................................................................................47
Figure 3.1 Non-metric multidimensional scaling (NMDS) plot of 16S rDNA of KBS
biodiversity gradient for combined two years of study with rotations in different
colors, and years in different shapes (Stress = 0.1681) Blue ellipse encircles
treatments with cover crops. See table 3.1 for treatment acronyms ..................................76
Figure 3.2 Non-metric multidimensional scaling (NMDS) plot of 16S rDNA of KBS
biodiversity gradient with all rotations shown by sampling year. (Stress = 0.1681) .........77
Figure 3.3 Canonical correspondence analysis plot of KBS biodiversity gradient for

xvi

combined two years of study with differences in rotations constrained to
measured soil parameters and the two axes representing the greatest explained
variation (on a percentage basis) . Figures are from cumulative measures over the
season for all parameters except pH (which represents a mean value). Other
parameters are Moisture (soil moisture), inorganic N (NH4+-N and NO3--N),
potentially mineralizable N (PMN) and C (PMC), permanganate oxidizable C
(POXC), enzyme activities for beta-glucosidase (BG), peroxidase (PEROX) +
phenol oxidase (PHENOX), N-acetyl-glucosaminidase (NAG) + tyrosine amino
peptidase (TAP), acid phosphatase (PHOS). All ordinations were created with a
correlational matrix to give equal weights to all soil parameters. See table 3.1 for
treatment acronyms ............................................................................................................78
Figure 3.4 Phyla of 20 most abundant OTUs, significantly (P≤0.05) related to treatments
with cover crops (A) or with no cover crops (B). Values are mean sequences
across treatments of cover vs. no cover from normalized, subsampled 16S rDNA
libraries ..............................................................................................................................79
Figure 3.5 Relative abundance, after subsampling OTUs across rotations in the
biodiversity experiment gradient of bacteria phyla (Acidobacteria,
Verrucomicrobia) and class (Alphaproteobacteria) for two years. Box and whisker
plots are shown with points from individual treatments (n=4). See table 3.1 for
treatment acronyms ............................................................................................................80
Figure 4.1 In situ N2O flux and gravimetric moisture (θg) for KBS biodiversity gradient
for two years. Time points are means with standard error for N2O, n=4, and
means across all treatments for moisture. Asterisks (*) show significant
differences (P ≤ 0.05) by treatments at a specific time point. See table 4.1 for
treatment acronyms ..........................................................................................................113
Figure 4.2 Denitrification enzyme assay (DEA) on the KBS biodiversity gradient. Bars
represent the mean rN2O or the relative rate of nitrous oxide flux per total
denitrification (including estimated N2) with standard error (n=4). See table 4.1 for
treatment acronyms ..........................................................................................................114
Figure 4.3 Non-metric multidimensional scaling (NMDS) plot of nitrite reductase (nirK)
gene amplicons from crop rotations, (n=3) on the biodiversity gradient experiment (see
table 4.1 for treatment acronym). Blue ellipse encircles treatments with cover crops.
Amplicon count data were transformed using Bray-Curtis dissimilarity matrix. Stress
was 0.12349 .....................................................................................................................115
Figure 4.4 Proportion of nitrite reductase (nirK) gene amplicons AOBs for crop rotations
on the biodiversity gradient experiment (n=3). See table 4.1 for treatment acronyms ....116
Supplemental Figure 4.1 Potentially mineralizable N (PMN) on the biodiversity gradient
treatments for the two study years. All points are means across blocks (n=4) with
standard errors for a corn year (left) and following soybean year (right). See table 4.1

xvii

for treatment acronym ......................................................................................................123
Supplemental Figure 4.2 Permanganate oxidizable (POX) C on the biodiversity gradient
treatments for the two study years. All points are means across blocks (n=4) with
standard errors for a corn year (left) and following soybean year (right). See table 4.1
for treatment acronym ......................................................................................................124
Supplemental Figure 4.3 Proportion of nitrite reductase (nirK) gene amplicons from AOBs
in crop rotations on the biodiversity gradient experiment (n=3), (see table 4.1 for
treatment acronym) shown with treatments from the KBS Main Cropping
Experiments (T1, conventional management for 2008 and 2009) and Deciduous
Forest (DF).......................................................................................................................125

xviii

CHAPTER 1: DISSERTATION INTRODUCTION
OVERVIEW
Agriculture faces the combined challenges of supplying food, fuel, and fiber to an
increasing world population (Pachauri, 2007) while preserving environmental quality. The last
half century of extraordinary growth in global agricultural production stems from improved
cultivars and increased inputs, primarily fertilizer, water, and pest control agents. Nitrogen (N)
fertilizer inputs alone have increased the carrying capacity of the earth by several billion people
(Smil, 1997). Yet these intensive inputs to agriculture have come with widespread
environmental costs. Synthetic fertilizer, frequently applied in excess, is the largest source of
non-point source pollution to waterways, leading to the proliferation of ‘dead zones’ in coastal
ecosystems (Diaz and Rosenberg, 2008). Nitrogen inputs to agriculture are also the largest
source of N2O emissions, a potent greenhouse gas and the primary degrader of stratospheric
ozone (Pachauri, 2007). The widespread “leakiness” of the N cycle indicates poor nutrient use
efficiency in agriculture and the breakdown of soil ecological processes known to function in
systems with greater species diversity. As a result, practices that increase species diversity in
cropping systems are a proposed approach to reduce chemical inputs to agriculture, enhance soil
nutrient cycling processes, and reduce losses to the environment (Hooper et al., 2005).
Soil is perhaps the most complex known biological medium and contains by far the
greatest species diversity on Earth (Whitman et al., 1998). Within soil, essential nutrient cycling
processes occur which ultimately sustain agriculture. Soil nutrient turnover, and N cycling in
particular, occur primarily via a series of microbially-mediated transformations. Understanding
underlying soil ecology and key steps in nutrient transformations is critical for improving N use

1

in agriculture, and for promoting approaches to sustainable crop production and environmental
quality.
Many of these soil processes have been altered by the displacement of biologically-based
nutrient cycling with industrially-produced inputs. Before incorporation of chemical inputs to
agriculture, soil fertility was sustained via plant rotation by alternating grains, legumes, and
fallows. Increasing plant diversity has been shown to increase plant productivity over
monocultures (Tilman et al., 2001), however this relationship may be less certain if the diversity
is divided temporally and spatially with crop rotations. Plant litter is the ultimate driver of
belowground nutrient and energy pathways (Hättenschwiler et al., 2005), and changes to plant
litter diversity can dictate belowground ecological functions (Meier and Bowman, 2008). Still,
most annual crops such as corn and soybean, occupy a field for only 3-4 months of the year.
Fields in bare fallow, have no litter inputs and constantly lose organic matter via decomposition,
and mineralize N, which can be leached to groundwater or denitrified, releasing N2O in the
process.
Ecological theory and evidence from natural systems indicate that greater plant diversity
increases ecosystem function and soil fertility (Dybzinski et al., 2008; Tilman and Snell-Rood,
2014; Zak et al., 2003). Whether increasing spatial and temporal diversity in cropping systems
through rotational complexity similarly alters soil ecosystem function is less clear.
Understanding these effects could offer alternatives to agricultural systems which have
fundamentally altered global ecosystems (Galloway et al., 2004; Vitousek et al., 2009) and yet
are under pressure to increase production (Balmford et al., 2012).

2

DISSERTATION OBJECTIVES
The overarching objective of this work is to understand how increased cropping system
diversity in row crops restores soil ecosystem processes, leading to sustained crop growth and
improved ecosystem health. Specifically, I address the following questions in this work: How
does increasing rotational complexity and plant diversity in row-crop agriculture alter soil C and
N accumulation, transformation, and gaseous losses? How does rotational diversity affect the
belowground microbial diversity that carries out soil nutrient cycling processes? What are the
implications of increased rotational complexity for nutrient management and environmental
benefits in row-crop agriculture, and what are the challenges to farmer adoption for best
management practices? To address these questions my dissertation integrates methods from
ecosystem ecology, soil science, molecular biology and the social sciences.

3

REFERENCES

4

REFERENCES

Balmford, A., Green, R., Phalan, B., 2012. What conservationists need to know about farming.
Proc. R. Soc. B Biol. Sci. 279, 2714 –2724. doi:10.1098/rspb.2012.0515
Diaz, R.J., Rosenberg, R., 2008. Spreading Dead Zones and Consequences for Marine
Ecosystems. Science 321, 926 –929.
Dybzinski, R., Fargione, J.E., Zak, D.R., Fornara, D., Tilman, D., 2008. Soil fertility increases
with plant species diversity in a long-term biodiversity experiment. Oecologia 158, 85–
93. doi:10.1007/s00442-008-1123-x
Galloway, J.N., Dentener, F.J., Capone, D.G., Boyer, E.W., Howarth, R.W., Seitzinger, S.P.,
Asner, G.P., Cleveland, C.C., Green, P.A., Holland, E.A., Karl, D.M., Michaels, A.F.,
Porter, J.H., Townsend, A.R., Vöosmarty, C.J., 2004. Nitrogen Cycles: Past, Present, and
Future. Biogeochemistry 70, 153–226–226.
Hättenschwiler, S., Tiunov, A.V., Scheu, S., 2005. Biodiversity and Litter Decomposition in
Terrestrial Ecosystems. Annu. Rev. Ecol. Evol. Syst. 36, 191–218.
doi:10.1146/annurev.ecolsys.36.112904.151932
Meier, C.L., Bowman, W.D., 2008. Links between plant litter chemistry, species diversity, and
below-ground ecosystem function. Proc. Natl. Acad. Sci. 105, 19780–19785.
doi:10.1073/pnas.0805600105
Pachauri, R.K., 2007. Climate Change 2007: Synthesis Report. Contribution of Working Groups
I, II and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. IPCC.
Smil, V., 1997. Global population and the nitrogen cycle. Sci. Am. 277, 76–81.
Tilman, D., Reich, P.B., Knops, J., Wedin, D., Mielke, T., Lehman, C., 2001. Diversity and
Productivity in a Long-Term Grassland Experiment. Science 294, 843 –845.
doi:10.1126/science.1060391
Tilman, D., Snell-Rood, E.C., 2014. Ecology: Diversity breeds complementarity. Nature 515,
44–45. doi:10.1038/nature13929

5

Vitousek, P.M., Naylor, R., Crews, T., David, M.B., Drinkwater, L.E., Holland, E., Johnes, P.J.,
Katzenberger, J., Martinelli, L.A., Matson, P.A., Nziguheba, G., Ojima, D., Palm, C.A.,
Robertson, G.P., Sanchez, P.A., Townsend, A.R., Zhang, F.S., 2009. Nutrient Imbalances
in Agricultural Development. Science 324, 1519 –1520.
Whitman, W.B., Coleman, D.C., Wiebe, W.J., 1998. Prokaryotes: The unseen majority. Proc.
Natl. Acad. Sci. 95, 6578 –6583.
Zak, D.R., Holmes, W.E., White, D.C., Peacock, A.D., Tilman, D., 2003. Plant Diversity, Soil
Microbial Communities, and Ecosystem Function: Are There Any Links? Ecology 84,
2042–2050. doi:10.1890/02-0433

6

CHAPTER 2: CROP ROTATIONAL COMPLEXITY SHIFTS MICROBIAL
PROCESSING AND STORAGE OF SOIL ORGANIC CARBON

ABSTRACT
Increasing rotational complexity in agroecosystems can have multiple benefits, from
reducing disease pressure to increasing crop yield. Increased rotational complexity can also alter
soil nutrient cycling, but linkages between rotational diversity, altered nutrient cycling pathways,
and the microbial communities that carry out these processes remain poorly understood. I
tracked multiple pools and fluxes of carbon (C) and nitrogen (N) along a gradient of increasing
cropping system complexity from continuous corn or soybean monoculture to up to five plant
species in 3-year rotations – including winter cover crops – in treatments that have not received
external inputs. Under fields with different rotational histories, but under the same crop, I
measured labile C pools, soil enzyme activities, mineralization potentials, and soil respiration
over two growing seasons. I found that the history of crop rotation influenced soil organic
carbon (SOC), which was higher under cover cropped treatments. In addition, the manner in
which soil C was processed also differed by rotation. Over two seasons, rotations with cover
crops respired at least 20% more CO2 but total SOC did not reflect estimated C returns to soil
from crop growth. Over the growing season, treatments with cover crops supported greater
labile C pools, potential mineralization rates and patterns in microbial enzyme potentials. Results
from this research show the strong effect of increasing rotational complexity on C and N pools
and fluxes, as well as an unexpectedly clear and consistent reshaping of the microbial processes
that underlie different crop rotations.

7

INTRODUCTION
To meet ever-increasing pressure to supply food, fiber and fuel, row crop agriculture has
achieved higher yields through a combination of improved crop genetics and agronomic
practices. In the U.S. Midwest, this has resulted in intensified grain production primarily in
monocultures, with plant diversity largely limited to annuals grown in short rotations (e.g.
continuous corn or soybeans) (Broussard and Turner, 2009; Klopatek et al., 1979). The
management inputs needed to maintain these cropping systems have substantial environmental
costs, including soil erosion, pollution of ground and surface waters, and release of greenhouse
gases including nitrous oxide (IPCC, 2013). Nitrogen (N) pollution is a particularly important
concern in the U.S. Midwest (Robertson and Vitousek, 2009), where N is the nutrient that most
frequently limits crop yields. The Corn Belt region of the upper Midwest contributes the majority
of the nitrate pollution causing a hypoxic zone to form in the Gulf of Mexico (David et al., 2010;
Rabalais et al., 2002), and the large N fertilizer inputs also contribute to greenhouse gas
emissions (Shcherbak et al., 2014).
A proposed method to sustain productivity while reducing N fertilizer and other
agronomic inputs and environmental costs is to increase cropping system diversity to enhance
ecosystem functions (McDaniel et al., 2013; Shennan, 2008). A strong body of evidence shows
a relationship between biological diversity and multiple ecosystem functions such as energy
flow, nutrient cycling, pest and disease suppression, and net primary productivity, due to a range
of biotic interactions (Hooper et al., 2005). Soil nutrient cycling, for example, encompasses the
breakdown and release of matter and energy, which is largely a function of microbial activity
(Wardle, 2002), as well as the uptake and recycling of nutrients. These nutrient cycling
processes are more efficient in ecosystems with diverse plant communities such as grasslands

8

and forests due to the broad array of ecological strategies by which species compete for and
capture scarce resources (Smith et al., 2013; Tilman et al., 2001).
By contrast, agricultural systems with low plant diversity have reduced capacity to either
capture or retain nutrients while also undergoing frequent disturbances that further disrupt
nutrient cycles. Indeed, an increased dependence on externally supplied nutrients in
agroecosystems with reduced capacity to cycle and retain nutrients results in inefficient, ‘leaky’
systems (Liebman et al., 2013). Soil organic carbon (SOC) is the pivotal resource for
maintaining and supplying crop nutrients (Robertson and Grandy, 2006). Even in temperate
cropping systems with high inputs of industrial fertilizer, grain crops rely on decomposition of
SOC for approximately half of their aboveground N (Gardner and Drinkwater, 2009). Since
SOC ultimately derives from plants, cropping system diversity strongly influences the quantity
and quality of this input. Low-diversity agricultural systems can have highly degraded SOC
stocks, and, as shown at cropping system research sites, net C inputs can vary by an order of
magnitude due to management practices such as plant species selected for crop rotations, harvest
and tillage practices, or fertilization (Robertson and Grandy, 2006; Sanford et al., 2012).
The composition of SOC ranges from labile pools of fresh and partially decomposed
plant litter to active microbial biomass and dissolved organic matter, to more stable pools of
physically protected, mineral-bound, and chemically recalcitrant SOC which can remain in soil
from years to decades and longer (Schmidt et al., 2011). The energy in SOC substrates is the key
source of most microbial growth and metabolism. Soil microbes access SOC resources by
releasing extracellular enzymes which reflect a diverse spectrum of biochemical metabolism of
labile, plant polymers (e.g. carbohydrates and proteins) via hydrolytic enzymes, and complex,
humic and colloidal substances via oxidative enzymes (Sinsabaugh et al., 2009). As microbes

9

access energy from SOC, they release nutrients (e.g. N and phosphorus (P)) in plant available
forms. The sustained supply of mineralized nutrients from SOC directly influences plant
productivity (Andrén et al., 1995; Borer et al., 2005) and over time, the activity of the soil
microbial community drives the fate of SOC pools: from respiration rates to eventual depletion
or accrual.
Mineralization processes also inform contemporary understanding of how soils
accumulate and store soil C over longer time spans (Schmidt et al., 2011), with the microbial
community acting as a conduit for flows of C and N from plant litter into stabilized SOC
(Grandy and Neff, 2008). Long-term stabilization of C in soil occurs by its complexation with
the surfaces of clay and silt particles. Rather than this stable C deriving from recalcitrant plant
C, an emergent hypothesis suggests that higher quality plant litter, which soil microbes can
process more efficiently (i.e., less is respired as CO2), is a dominant source of microbial
byproducts which complex with mineral soil (Cotrufo et al., 2013).
Cropping practices which can increase the timeframe of primary productivity and supply
more diverse root and plant residue inputs to soil may increase microbial activity and associated
processes that stabilize C in soil (Gregorich et al., 2001; Kallenbach et al., 2015; 2016; Liang and
Balser, 2011). Adding rotational complexity to grain-based agroecosystems by cultivating a
greater number of plant functional types such as legumes, perennial forages, or overwintering
cover crops may maintain the SOC reservoirs that mineralize and supply nutrients to the ensuing
crop. Further, comparisons in long-term cropping systems have demonstrated that ecological
processes can replace synthetic inputs in grain cropping systems to sustain productivity while
reducing environmental impacts (Clark et al., 1998; Drinkwater et al., 1998; Gregorich et al.,
2001; Robertson et al., 2014) For example, Smith et al. (2008) showed a pattern of higher crop

10

yields with increasing crop diversity in the absence of synthetic inputs, concluding that increased
rotational complexity can maintain productivity. The results suggested that enhanced soil
nutrient cycling under more diverse crop rotations increased crop yield, however the mechanisms
leading to this pattern are unclear. Since row crop systems have less aboveground biodiversity
than natural ecosystems, changes to crop species composing a rotation (separated temporally), as
well as the degree of spatial or temporal overlap (e.g., inclusion of overwintering crops), may
lead to large changes in ecosystem functions – such as soil C storage, crop productivity, and
nutrient retention (Shennan, 2008; Tiemann et al., 2015; Wood et al., 2016).
Here we investigate how increasing crop rotational complexity impacts microbial enzyme
activity and SOC oxidation. In a long term cropping system study we measured belowground
nutrient cycling processes for two full growing seasons along a rotational complexity gradient.
The gradient ranges from simple rotations of continuous summer annuals – corn (Zea mays L.)
and soybean (Glycine max.) – to rotations that also include a winter annual (wheat; Triticum
aestivum L.), to more complex rotations that include overwintering cover crops (i.e., nonharvested legumes and grasses). In this experiment, the only factor that differs among the
treatments is crop rotation—tillage practices are the same, and there are no synthetic inputs to the
systems. I sampled along a range of simple to complex rotations that had been in place for ten
years at initial sampling. In order to test hypotheses about how crop rotational complexity alters
SOC dynamics over time, I measured labile organic matter pools; the activity of microbiallyproduced enzymes that break down organic matter; and respired CO2.
Specifically, we hypothesize that SOC will increase in crop rotations with greater plant C
inputs to soil, larger labile soil C pools, and altered processing of SOC by soil microbes. We
expect that the activities for hydrolytic soil enzymes will increase in proportion to the size of

11

labile C pools. By contrast, in cropping systems that result in smaller labile C pools, we expect
higher activities of oxidative enzymes, which break down more recalcitrant soil C substrates.
Further, we predict that larger pools of labile C will occur under complex crop rotations leading
to higher soil respiration rates and but also gains in SOC, compared to reduced soil respiration
and reduced SOC in simpler rotations. In this experiment, total C inputs are greater in the more
complex rotations and they are also more continuous (i.e., the cover cropped treatments have
year-round plant cover). We predict that the more frequent plant C inputs under more complex
rotations will sustain larger labile C pools, and result in distinct microbial C processing patterns
and greater mineralization potentials throughout peak crop growth periods.

METHODS
Site descriptions and experimental layout
We conducted our study at Michigan State University’s Kellogg Biological Station
(KBS) Long Term Ecological Research experiment located in Hickory Corners, Michigan (42°
24′ N, 85° 24′ W, elevation 288 m). KBS lies in SW Michigan, in the eastern part of the U.S.
Corn Belt. Prior to 1988, the site had been conventionally managed for row crop agriculture
(Robertson et al., 1997). Soils developed on glacial outwash are composed of Kalamazoo (Fineloamy) and Oshtemo (Coarse-loamy) mixed, mesic, Typic Hapludalfs (FAO soil order:
Luvisols). The climate is temperate, with approximately 1005 mm rainfall and an average
snowfall of ~1.3 m and mean annual temperature is 10.1° C (Robertson and Hamilton, 2015).
This work was undertaken on the KBS Biodiversity Gradient Experiment
(http://lter.kbs.msu.edu/research/long-termexperiments/biodiversity-gradient/) which consists of
a gradient of rotational complexity of harvested annual grains including corn, soybean, and

12

wheat ranging from continuous monoculture (e.g., continuous corn, or continuous soybean) to
various rotations that also include a non-harvested overwintering cover crop of red clover
(Trifolium prantense L.), crimson clover (Trifolium incarnatum), or annual ryegrass (Lolium
multiflorum Lam.) either alone or in combinations. The study was established in 2000 and is
composed of plant species treatments from bare ground, to 1, 2, 3,4, 6 and 10 species, in a
randomized complete block experiment with four replicates within 9 x 30 m treatment plots and
with every phase of each crop rotation represented every year. Since establishment, none of the
treatments have received synthetic chemical inputs (i.e. no N fertilizer or pest control agents) and
weeds are controlled mechanically.

Sample collection
For this study we selected seven treatments (Table. 2.1): continuous corn (cC) and
continuous soy (cS), corn and soybean in rotation (CS), CS and wheat in rotation (CSW), CSW
with inclusion of clover varieties as a winter cover crop (CSW1cov), CSW with clovers and
annual ryegrass (CSW2cov), and a fallow plot (Fspring) tilled once each spring to stimulate a
diverse, early successional vegetation. We followed these same treatments through two growing
seasons, first in 2010 with all treatments in the corn phase of the rotation (except for cS and
Fspring), and again in 2011, with all treatments planted to soybeans (except for cC and Fspring).
Each year sampling began at planting, in June, and continued over four months and included all
treatment blocks. Within a central row of each plot soil cores were taken to a 10 cm depth from
five locations beginning from 6m to 18m from the start of a row, alternating cores from within
and between crop rows. Cores were composited by plot and placed on ice for transport to the
lab. Within 24 hours of collection soils were sieved (to < 4mm), mixed, and weighed for soil

13

moisture, enzymes and mineral N analyses. A portion of the soil was air-dried for approximately
30 days.

Soil carbon and nitrogen
A portion of air-dried soils was finely ground to determine total SOC and total soil
organic nitrogen (SON) by dry combustion on a Costech ECS 4010 CHNSO Analyzer (Costech
Analytical Technologies, Valencia, CA). Potentially mineralizable carbon (PMC) was adapted
from Franzluebbers et al. (2000) and Haney (2008) using air-dried soil brought to 50% waterfilled pore space (WFPS). Deionized water was added to 15 g of air-dried soil to 50% waterfilled pore space in canning jars with lids fitted with airtight, rubber septa. Lids were placed
loosely on the jars and soils were incubated in the dark at 25° C. At 24-hours, jars were removed
and soil respiration determined using a LI-COR LI-820 infrared gas analyzer (LI-COR
Biosciences, Lincoln, NE): jars were sealed and 0.5 mL of headspace gas was immediately
removed and injected into the gas analyzer, followed by three additional measurements over
approximately 90 min. PMC was calculated as CO2 flux per gram of incubated soil.
Soil mineral N was extracted from fresh soil within 24 hours of sampling. Triplicate
10.0 g subsamples of sieved soil were extracted with 1M KCl for one hour on a rotary shaker
(120 rpm). Samples were filtered and stored at -20 °C until analysis for ammonium (NH4+) and
nitrate (NO3-). Potentially mineralizable nitrogen (PMN) was assessed in a short term anaerobic
incubation according to Drinkwater et al. (1996). Briefly, 10 mL deionized water was added to
triplicate 8 g fresh, sieved soil in conical tubes. Dinitrogen gas was used to replace tube
headspace air and bubbled into the slurry for one minute prior to sealing with butyl rubber
stoppers. Sealed tubes were incubated at 25 °C for seven days. After seven days the stoppers

14

were removed, buffer was added to bring the slurry to 1M KCl, and samples were shaken,
filtered, and stored in the same manner as fresh samples. Concentrations of NH4+ and NO3- were
determined colorimetrically on a 96-well plate-reader. Soil NO3- was determined according to
Doane and Horwáth (2003) and read at 540 nm, and soil NH4+ concentration determined at 630
nm (Sinsabaugh et al., 2000). After incubation and extraction, PMN was determined based on
the concentration of NH4+ incubated soil minus NH4+ from initial soil extraction of the same soil.

Soil Enzymes
We measured soil enzyme potentials from slurries of fresh soils following established
methods (German et al., 2011; Sinsabaugh et al., 2002). First, soil pH was determined using 10
g fresh soil mixed with 20 mL of deionized water, which was allowed to equilibrate for one hour
with occasional stirring. Following this determination, the pH of a 50 mM sodium acetate buffer
was adjusted to the pH of treatment soils. Approximately half of 100 mL of sodium acetate
buffer was added to 1 g of treatment soil, and homogenized vigorously in a blender for exactly
one minute. The rest of the buffer was used to rinse remaining soil contents from the blender
into the slurry and then stored at 4° C until analysis (within 1-2 hours). We evaluated the
breakdown of organic compounds in soil slurries by measuring enzymatic potentials for four
hydrolytic enzymes, β-1-4-glucosidase (BG), β-1-4-N-acetyl-glucosaminidase (NAG), acid
phosphatase (PHOS), and tyrosine amino peptidase (TAP), and for two oxidative enzymes phenol oxidase (PHENOX) and peroxidase (PEROX). At initiation of enzyme assays, soil
slurries were constantly stirred and the suspension was pipetted into 96-well plates. Hydrolytic
enzyme activity was monitored using substrate specific polymers containing fluorescent labels
(either 4-methylumbelliferone or methylcoumarin). Phenol oxidase and peroxidase activity were
assessed by monitoring oxidation of L-DOPA, with 0.3% hydrogen peroxide added to the latter

15

assay. For all enzyme assays, the exact time was noted when substrate was added to the buffered
soil slurry. All enzymes were incubated in the dark at 15° C. At 4-6 hours, hydrolytic enzyme
reactions were stopped by adding 10 µL of 1M sodium hydroxide, and the exact time was noted
and plates were read on a Fisher Scientific Thermo Fluoroskan AscentTM microplate fluorometer
(Waltham, MA). Between 6-8 hours, color change of oxidized L-DOPA was assessed
colorimetrically at 450nm. Background absorbance-emission or fluorescence was accounted for
with controls for buffer, buffer + soil, buffer + substrate.

In situ soil respiration
At approximately 2-week intervals soil CO2 fluxes were measured using static 25 cm
diameter chambers placed in situ in each treatment. Using a method previously described by
Hoben et al. (2011), PVC chambers were inserted in the field mid-way between rows and central
furrows to a depth of 8 cm. The soil surface inside the chamber was maintained free of living
plant material. At each sampling time, chamber height from soil surface was measured to
account for headspace volume and then chambers were sealed with an air-tight, O-ring PVC lid,
fitted with a Vacutainer serum vial septa (Becton-Dickinson, East Rutherford, NJ). Using a
syringe, and mixing the headspace at each sampling, 10 mL gas samples were removed via the
lid septa, at time zero and three additional 20-minute intervals, and stored in 5.9 mL gas
Exetainers (Labco, Ceredigion United Kingdom). Samples were analyzed using gas
chromatography (Hewlett Packard 5890 Series II, Rolling Meadows, IL, USA) with gases
separated on a Poropak Q column (1.8 m, 80/100 mesh) at 80 °C. CO2 was analyzed using an
infrared gas absorption analyzer (LI-820 CO2 analyzer; LI-COR, Lincoln, NE, USA) and N2O
was analyzed with a 63Ni electron capture detector at 350 °C.

16

Statistical approach
Data for all soil analyses were examined for normality and heterogeneity of variance
assumptions following Zuur et al. (2010). Data deviating from normality were log-transformed
prior to analysis, and homogeneity of variance checked with Levine’s test. Analysis of variance
was calculated in R (The R Foundation for Statistical Consulting, Vienna, Austria) using the
lme4 package for linear, mixed-effect (Bates et al., 2016), with sampling time and block as
random effects and crop rotation as a fixed effect. Results are reported as statistically significant
at α = 0.05.

RESULTS
Soil Carbon and Nitrogen
Across the rotation, SOC differed significantly by rotation (P < 0.019), (Table 2.2), with
comparisons showing cC to be the most deficient in SOC when compared to CSW1cov (P <
0.0703) and to CSW2cov (P < 0.0510). The more simplified rotations (cS, CS, CSW) were all
significantly lower than rotations with cover crops, with mean SOC ranging between 7.93 and
7.95 g-C kg -1 soil (Table 2.2). For SON, the effect of rotation was not significant. Among
treatments, comparisons showed a similar trend of cC’s having the smallest SON pool and covercropped rotations the highest values (Table 2.2). Soil C:N was not different across rotations, nor
were there any clear trends by rotation, except for cS tending to have a lower C:N than all other
rotations (Table. 2.2).
Mineralization potentials for C followed similar patterns across treatments for both corn
and soybean years,; they were significantly higher in cover cropped treatments than in no-cover
treatments (Fig. 2.1, Table 2.4). PMC was higher in the corn year for all rotations compared to

17

the soybean year, but the magnitude of the difference between cover and non-cover was larger in
the corn year than in the soybean year. The magnitude of PMC under corn shifted dramatically
over the season, more than doubling between June and July for all treatments except in cC (Fig.
2.1). PMC was higher for CS and CSW than in cC in July and August. In both corn and soybean
years, the different cover crop treatments had nearly identical PMC over the course of the
season, but in the soybean year, the July increase in PMC was not as high as under corn, and the
overall mean PMC was lower in the soybean year (Table 2.4).

Soil Enzyme Activity
In both 2010 and 2011, beta-glucosidase (BG) activity was consistently higher under
cover cropped treatments (Table 2.3), and activities for all treatments spiked and fell off during
the growing season, leveling off between August and September with activities considerably
lower than in 2010 (Fig. 2.2). The pattern for the N-acquiring enzymes, N-acetylglucosaminidase and tyrosine amino peptidase (NAG+TAP), was clearly different in cover vs.
no-cover treatments (Fig. 2.2). In both years, for nearly all crop rotations, these activities spiked
in July and then fell. For acid phosphatase (PHOS), activities differed in both magnitude and
pattern between years (Fig. 2.2). In 2010, PHOS activities increased dramatically between June
and August, doubling for no-cover treatments, and nearly tripling for cover cropped treatments,
before falling in September. In 2011, PHOS activities were generally steadier and fell slightly
over the growing season. PHOS activities under cover cropped treatments were again
significantly higher than for no-cover rotations, except for CSW in 2011 (Table 2.3).
The activities of the oxidative enzymes peroxidase and phenol oxidase (PEROX and
PHENOX), contrasted with the hydrolytic enzymes with increasing rotational complexity. For

18

both corn and soybean years, rotations with winter cover tended to have similar or lower PEROX
+ PHENOX enzyme activities compared with the no-cover treatments (Fig. 2.2). In the corn
year of the study, the oxidative enzymes spiked early and then fell: in July the highest enzyme
activity occurred in cC, whereas CSW1cov had the lowest activity of the oxidative enzymes.
However, for all other months during the corn year, the difference in oxidative enzyme activity
by rotation was nearly indistinguishable. In the soybean year, the highest oxidative enzyme
activity tended to occur in cS, except for in August, when there was higher oxidative activity
under CS and CSW. These two rotations had the greatest fluctuation in oxidative enzyme
activities over both growing seasons. In contrast, the cover crop treatments tended to be less
variable and were consistently lower than continuous cS, and in no month did the oxidative
enzyme activity in either cover crop treatment exceed the highest activities in any of the nocover treatments (Fig. 2.2). The magnitude of oxidative enzyme activities was much greater in
the soybean year than in the corn year, and patterns in activity differed between PHENOX and
PEROX (Table 2.3). The pronounced spikes and month-to-month variability in both years make
rotational-level differences less clear overall. However, the highest oxidative enzyme activities
tended to occur in cC and cS, and peroxidase activities were also generally higher in less-diverse
rotations than they were in the more complex rotations with cover crops.
The mean ratio of hydrolytic enzymes to PMC did not differ for any potential activities
for either year, and no clear trends emerged in this metric based on crop rotation history
(supplemental Fig. 2.1). In contrast, the ratio of oxidative enzymes to PMC (OX:PMC, Table
2.4) was strongly significant by rotation for both 2010 (P < 0.0001) and 2011 (P < 0.0001). In
2010, this ratio was significantly lower for the cover-cropped rotations. A similar pattern was

19

observed in 2011, with the highest OX:PMC ratios in continuous monocultures, followed by CS
and CSW, and the lowest ratios in the cover cropped treatments (Fig. 2.3).
The ratio of hydrolytic enzymes to oxidative enzymes was strongly dependent on crop
rotation history. The ratio of BG, NAG+TAP and PHOS to OX was higher in rotations with
cover crops for both years (Table 2.4, supplemental Fig. 2.2). The relationship was strongest for
N-acquiring enzymes ([NAG+TAP]:OX), with treatments with winter cover crops having a
significantly higher ratio under both corn (P<0.001) and soybean (P<0.0023). The ratio of
BG:OX was strongly significant by rotation in the corn year (P<0.0038), and was less strong in
the soybean year, although there was a similar pattern by rotation. Ratios for PHOS:OX also
tended to be higher under rotations with winter cover, and were marginally significant in both
2010 (P<0.0453) and 2011 (P<0.0847). These trends are even more notable given the
fluctuation in different enzyme activities over the course of both growing seasons.

In situ soil respiration
Clear patterns emerged in soil respiration by rotational history. In 2010, in situ soil CO2
flux was similar for all of the rotations that included corn (Table 2.5). Indeed, through most of
the season CO2 fluxes were not significantly different across treatments at most time points –
even when including cS, which had significantly lower soil respiration rates than did rotations
with corn, (Fig. 2.4). In 2010, only the first and last fluxes differed across all crop rotations (Fig.
2.4), though over the season the highest fluxes were observed in the most diverse rotations
(CSW2cov) and the lowest fluxes in the rotations with lowest diversity (cC and cS; supplemental
Table 2.2). In 2011, soil respiration in the cover cropped treatments was even more
distinguishable from no-cover rotations, with time points earliest in the season leading to the

20

greatest divergence in flux measures between rotations (Fig 2.4, supplemental Table 2.2).
Overall, mean soil respiration rates were much lower in 2011 than in 2010, but the significantly
higher respiration in cover-cropped treatments was more apparent in 2011 (Table 2.5).

Labile nitrogen
In 2010, inorganic N as both NH4+-N and NO3--N was significantly higher in the
CSW1cov rotation, containing only a legume cover (supplemental, Table 2.1). Beginning from a
high concentration, inorganic N in CSW1cov generally remained higher than in all other
treatments over the course of the growing season, until September when it fell below CSW2cov
(supplemental Fig. 2.3). NH4+-N was significantly higher in CSW2cov than in the less diverse
rotations, cC and CS, but not as high as it was in CSW (supplemental, Table 2.1). CSW2cov had
a more pronounced drop and recovery in NH4+-N concentration over the growing season
compared to NO3--N, which retained a steadier concentration over time. Of the rotations without
cover crops, CS generally had the least available N as either NH4+-N or NO3--N, and CSW
generally had higher NH4+-N over the season (supplemental Fig. 2.3). In the rotations without
cover crops, cC had intermediate levels of labile N species over the growing season. In both
years, NH4+-N concentrations varied less throughout the season than the cover cropped
treatments. The CSW2cov treatment barely fluctuated around a mean of 2 µg NH4+-N per gram
of soil over the season, CSW1cov was generally higher in NH4+-N, and fluctuated a bit more,
and the greatest variations in NH4+-N across the season were in the cS and CS rotations
(supplemental Fig. 2.3). For NO3--N, the CSW1cov treatment generally had far higher
concentrations over the season except in August when it fell dramatically, and then recovered in
September. The CSW2cov treatment showed a similar drop followed by an increase, but

21

generally followed a closer pattern to the less complex treatments in NO3--N across the season.
In both years of the study, PMN was not different among no-cover rotations, but significantly
higher by a factor of ~2 in treatments with cover crops (supplemental Table 2.1).

DISCUSSION
Introducing more plant diversity into grain crop rotations can improve crop productivity
and agronomic performance, but the soil ecosystem processes underlying these observations are
poorly understood. I examined how increased crop rotational complexity alters mechanisms that
control microbial metabolism and relative SOC pool size. Results support the hypothesis that
SOC pools increase, and soil C metabolism changes, with increasing rotational complexity, but
almost all differences were due to the presence of a cover crop in rotation rather than crops in
rotation per se. I found that during both corn and soybean growing seasons PMC and hydrolytic
enzyme activities were significantly higher in crop rotations with cover crops than in rotations
without cover. By contrast, no-cover treatments had higher oxidative enzyme activities,
especially relative to PMC. This pattern in enzyme potentials suggests that the microbes in cover
crop treatments are processing more labile C. Even though total plant residue returns,
normalized across all multi-year rotation types, differed little for some of the rotations with and
without cover (supplemental Fig. 2.4), complex rotations with cover crops respired more CO2
over the growing season and have accumulated more SOC over time. The quality and/or
frequency of C inputs from overwintering cover crops appear to be the critical factors altering
microbial metabolism, soil respiration, and net C accrual.

22

In this long term experiment, inputs from an overwintering cover crop shaped patterns in
litter decomposition (flows, rates, turnover), more than the current or previous grain crop in
rotation. In the corn year of this study, CS had been in soybean, CSW, CSW1cov and CSW2cov
had been in wheat, and in the soybean year, all rotations, except cS, had previously been in corn
(Table 2.1). While Smith et al. (2008) presented data from this experiment showing a strong
yield response with increasing rotation complexity, particularly for corn, a comparison of NPP
across a full crop rotation shows a different pattern. I used grain harvest data collected on this
site (http://lter.kbs.msu.edu/datatables) and followed Bolinder et al. (2007) to estimate plant C
returning to soil from aboveground non-harvested biomass, root tissue, and root deposition for
each crop type to determine mean annual NPP across a full crop rotation cycle (e.g. total C
returned to soil from the mean of a CSW rotation vs. three years of cC). Mean NPP calculated
for 2008 through 2010 specifically for plots examined in this study differed little from a
calculation of six years of mean annual NPP across all phases of the rotations occurring on the
experimental site (supplemental Fig. 2.4). For 2010, the rotations with the lowest mean annual
NPP were cS and cC (1799 and 1836 kg C ha-1 yr -1, respectively) while CS was 2751 kg C ha-1
yr -1, which was similar to CSW1cov and CSW2cov (2677 and 2811 kg C ha-1 yr -1,
respectively).
From the same data archive I estimated C inputs from weed and cover crop biomass,
although these time points were only in the fall (e.g. collected as aboveground NPP) and with
limited specificity of inputs by different plant species. The mean ANPP estimate for non-grain
crop (weed+cover crop) biomass ranged from a low of 382 kg C ha-1 yr -1 in CS to a high of 677
kg C ha-1 yr -1 in CSW2cov (supplemental Fig. 2.4) with a mean across all rotations of 552 kg C
ha-1 yr -1, and of the rotations with cover crops, only 13% of the biomass was identified as clover

23

cover crop species. Among no-cover rotations, including C inputs from both grain and auxiliary
plants, the CS rotation had between 22-26% greater residue returns than other treatments (cS, cC,
CSW). Compared to rotations with cover, CS had 4% less C inputs than CSW1cov and 10% less
than CSW2cov. Given this range of residue inputs both among no-cover rotations and between
cover and no-cover treatments, the contrast in belowground C processing, transformation, and
respiration detailed in this study is striking. These results suggest that the input of cover crops
over the winter (e.g., as living, active roots) and into spring prior to plowing, is a primary factor
driving microbial activity and SOC dynamics.
Rotations with cover crops sustained higher baseline labile C pools in soil, in spite of
similar seasonal fluctuations for all treatments over the growing season (Fig. 2.1). This pool of
labile C is indicative of active microbial metabolism (Fierer and Schimel, 2003), reflects changes
in management and environmental conditions (Ladoni et al., 2015), and is strongly correlated
with soil microbial biomass C (Franzluebbers et al., 2000). Root inputs from cover crops have
been shown to contribute larger C inputs to soil than aboveground biomass (Kong and Six, 2010;
Rasse et al., 2005), and likely shaped mineralization patterns in this study. Microbial biomass has
a similar ratio of nutrient content (C:N:P) across a large range of aboveground ecosystems and
inputs (Kallenbach and Grandy, 2011) and when normalized to PMC, hydrolytic enzyme
activities did not differ across treatments (supplemental Fig. 2.1), indicating that these enzyme
activities were tied to the relative size of labile C pools across all treatments. Hydrolytic
enzymes for C, N and P acquisition (BG, NAG+TAP, PHOS, respectively) were all significantly
higher in the cover cropped treatments (Table 2.3), and sustained higher enzyme potentials
throughout the growing season for both study years (Fig. 2.2).

24

The ratios of hydrolytic to oxidative enzyme potentials (PEROX and PHENOX) reflect
that a metabolic shift in microbial C use occurred between treatments with cover crops compared
to those without cover (Table 2.4) toward a disproportionate breakdown of more stable or
recalcitrant C substrates (Sinsabaugh, 2010). Different C substrate acquisition patterns by
microbes in the no-cover treatments appeared to be less related to plant litter chemistry (e.g.
comparing cS and cC) than to a general response to oxidize recalcitrant C to acquire nutrients or
energy. PMC values and patterns in microbial C acquisition over both growing seasons suggest
a means by which total SOC diverged between cover and no-cover treatments. Higher
mineralization potential in the cover crop treatments indicated that a greater proportion of
respired C derived from labile pools, and that a greater flux of microbially processed C into a
stable pool bond to mineral soil (Cotrufo et al., 2013; Kallenbach et al., 2015, 2016). By contrast,
over two years under the no-cover treatments the microbial community allocated resources that
favored recalcitrant SOC oxidation resulting in a greater amount of respired C derived from
degrading SOC pools that might otherwise be stable under cover crops.
Apart from distinct metabolic patterns in soil C use between cover and no-cover
treatments, root inputs from cover crops may facilitate other SOC pathways. On the same site,
Tiemann et al. (2015) found significantly greater aggregate formation under cover cropped
treatments, in spite of regular tillage across all treatments during the growing season. Greater
physical protection of C in soil aggregates preserves plant C from oxidative degradation (Grandy
and Robertson, 2007) more so than chemical recalcitrance (Dungait et al., 2012; Rasse et al.,
2005), which would bolster a more passive pathway for SOC accrual. In addition, root C that
passes into microbial biomass can ascend rapidly through the soil food web (Bradford et al.,

25

2012) resulting in a larger trophic cascade (Borer et al., 2005) under cover crops and an
augmented flux of root C back into labile PMC pools.
In spite of a broad range in the amount of inputs from NPP across rotations, only
treatments under cover crops gained more SOC while respiring more CO2 (Table 2.5). Mean in
situ CO2 fluxes in no-cover treatments were 21% and 20% less compared to treatments with
cover crops during each respective cropping season. A temporal component may influence this
outcome as the largest divergence in respiration rates occurred in the earlier parts of each season
(Fig. 2.5, supplemental Table 2.2). Also, the source of respired C may be distinct for cover and
no-cover treatments. While recalcitrant soil C is disproportionately metabolized in no-cover
treatments, greater soil aggregation under cover crops may spatially structure soil habitat in a
way that determines which soil C pools are more readily respired. Soils with high SOC and
larger labile C pools may have slower decay constants – thus lower respiration per unit SOC, due
to physical protection (Colman and Schimel, 2013). Finally even though the majority of time
points showed no significant difference between fluxes, microbial C use efficiency may also
differ (Sinsabaugh et al., 2013) such that the conversion of plant and soil C into microbial matter
and energy differs between cover and no-cover treatments (Kallenbach et al., 2015). Differences
in C use efficiency between treatments could be due to substrate biochemistry (e.g. legume
inputs) or the physiology of microbial growth (Lee and Schmidt, 2014), or both litter quality and
microbial physiology.
Inclusion of cover crops was the greatest factor in altering belowground ecosystem
processes that underpin differences in agronomic performance across the treatments in this study.
Constraining the measured soil parameters into a matrix-based ordination analysis reveals that
the differences in soil ecosystem processes between continuous monocultures and the most

26

complex rotations are best explained by oxidative enzyme potentials (PHENOX and PEROX) at
one extreme, and parameters relating to potentially mineralizable N (PMN) and NAG (Fig. 2.5)
at the other. In addition to altered microbial C metabolism, cover cropped treatments had nearly
double the magnitude of PMN compared to no-cover rotations (supplemental Table 2.1). Along
with similar patterns in PMC and hydrolytic enzyme activities, the soils in the cover cropped
treatments sustained nutrient mineralization potentials throughout the growing season that
explained greater potential for root uptake and crop growth.

Synthesis
Our conceptual model (Fig. 2.6) attempts to bridge values of SOC – which are a
reflection of long term management history – with measures of C pools and fluxes, and microbial
metabolic activity, that turnover much more quickly. Monoculture plots had significantly less
SOC than plots with cover crops (7.95 and 9.34 g C kg soil-1, respectively). Complex rotations
received more C inputs and respired more CO2, but also led to SOC accumulation over time.
Two other measurements showed consistent differences between cover and no-cover treatments.
First, mineralizable C was significantly higher under cover crops, indicating a larger stock of C
that is poised for microbial processing (i.e., as this C is processed, more is turned into protected
C and more is respired). Second, enzyme activity also reflected different patterns in microbial C
metabolism between the cover and no-cover treatments. In general, we found that hydrolytic
enzymes for C, N and P acquisition were similar across treatments when normalized to PMC –
thus reflecting a similar proportional allocation to sustain relative sizes of the microbial biomass
pool. In contrast, the oxidative enzyme activity was much higher in monoculture plots,
reflecting greater accessing of more stable C, associated with mineral-bound C. This reflects a
different pattern in metabolic C acquisition along the gradient, with the monoculture plots more

27

reliant on recalcitrant forms of SOC, and the cover cropped plots less so, which should be
reflected in both the resulting stores of SOC and fluxes of respired CO2. The presence of cover
crops has been shown to increase SOC over time (Kong and Six, 2010; Syswerda et al., 2011)
even compared to rotations with greater NPP, and this work indicates possible mechanisms for
soil C processing which could explain this finding.

Implications
The crop rotations in this study do not represent the dominant row crop management
systems found in the U.S. Midwest. For example, continuous monoculture corn (cC) would not
be grown in the absence of substantial fertilizer inputs (which would greatly increase NPP), and
the most diverse rotations in this experiment (CSW2cov) are much more rare on farms due to
present socioeconomic challenges for farmers including labor and energy costs (Davis et al.,
2012). However, cropping systems similar to these do fall within the spectrum of agroecosystem
management that both sustain grain yields and lead to improvements in ecosystem services
(Robertson et al., 2014). More importantly, the more complex rotations represent an alternative
to the common winter bare fallow practice used in most Midwestern cropping systems, which
can lead to soil C degradation pathways (Nunan et al., 2015).
Adding cover crops into row crop agriculture has multiple effects on soil ecosystem
function, including increased N supplied to crops (Tonitto et al., 2006) and reduced downstream
impacts from leaching (Plaza-Bonilla et al., 2015). This study also suggests mechanisms by
which cover crops shift the flow of C by expanding a pool of readily metabolizable C and
altering microbial metabolic processing of SOC. This may change the magnitude of trophic
level flows of labile C (DuPont et al., 2009) and microbially processed C that is stabilized on soil

28

minerals (Cotrufo et al., 2015; Kallenbach et al., 2015). Our results suggest that more than crop
composition per se, or the magnitude of C inputs from NPP, the changes in soil ecosystem
function observed in cover cropped treatments may be due to: i) temporal effects of cover crop
inputs over winter leading to a large labile C pool at planting in the spring, ii) spatial effects of
the accessibility of cover crop residues, especially roots, to degradation iii) chemical composition
of residue inputs (or lack thereof), or some combination of these three factors. The resulting soil
ecosystem is poised for mineralization of nutrients from internal cycling processes, which can
sustain higher crop growth and multiple ecological functions similar to more diverse natural
ecosystems (Zavaleta et al., 2010).

ACKNOWLEDGEMENTS

Support for this research was provided by a U.S. Department of Agriculture Soil Processes
Program, grant # 2009-65107-05961, and the National Science Foundation Long-term Ecological
Research Program (DEB 1027253) at the Kellogg Biological Station.

29

APPENDIX

30

Table 2.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment

ID
1
2
3
4
5
6

Treatment
cC
cS
CS
CSW
CSW1cov
CSW2cov

Rotation
corn
soy
corn + soy
corn + soy + wheat
corn + soy + wheat + clover
corn + soy + wheat + clover + rye

Species / yr.
1
1
1
1-2
2-3
3-4

Species / rot.
1
1
2
3
4
5

1
2
3
4
5

6

Month M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D
C

S

W

red clover

31

crimson clover

rye

Table 2.2 Soil organic C (SOC) and N (SON) and their ratio (C:N), in the biodiversity gradient
treatments (see table 2.1 for treatment acronyms). Values are means (standard error), n=4
replicate blocks. Significance of treatment effect: * P ≤ 0.05.

Rotation
cS
cC
CS
CSW
CSW1cov
CSW2cov

SOC
g C kg soil-1
7.93(0.96)
7.56(0.61)
7.95(0.96)
7.93(0.60)
9.30(0.35)*
9.38(0.38)*

32

SON
g N kg soil-1
0.80(0.09)
0.73(0.06)
0.76(0.12)
0.77(0.07)
0.91(0.05)*
0.93(0.06)*

C:N
9.9(0.13)
10.5(0.46)
10.8(0.58)
10.5(0.54)
10.3(0.45)
10.2(0.38)

Table 2.3 Beta-glucosidase (BG), N-acetyl-glucosaminidase (NAG), tyrosine amino peptidase (TAP), acid phosphatase (PHOS),
phenol oxidase (PHENOX),peroxidase (PEROX) in the biodiversity gradient treatments (see table 2.1 for treatment acronyms) for the
two study years. Values are means (standard error), n=4 replicate blocks. Significance of treatment effect: * P ≤ 0.05, ** P = 0.01 to
0.001. Each year was analyzed separately.

Year

Rotation

2010

cS
cC
CS
CSW
CSW1cov
CSW2cov

2011

cS
cC
CS
CSW
CSW1cov
CSW2cov

BG
nmol g-1 h-1
59.9 (5.7)
58.8 (5.8)
68.9 (8.5)
63.0 (6.5)
91.7(11.2)*
98.5 (9.8)**
67.9
75.4
66.9
66.6
92.3
95.0

(6.9)
(9.8)
(7.9)
(7.1)
(9.6)**
(9.0)**

NAG
nmol g-1 h-1
17.0(1.7)
18.9(1.6)
20.9(2.2)
23.3(2.7)
33.1(3.3)**
38.5(4.1)**

TAP
nmol g-1 h-1
10.1(2.3)
8.4(1.5)
9.6(1.8)
7.7(1.3)
11.6(3.1)*
12.0(2.6)*

PHOS
nmol g-1 h-1
160.4(44.2)
123.2(20.8)
128.1(18.1)
118.0(22.2)
182.4(32.8)**
171.3(26.9)**

PHENOX
µmol g-1 h-1
1.1(0.1)
1.0(0.2)
0.9(0.2)
0.9(0.2)
0.8(0.2)
1.2(0.2)

PEROX
µmol g-1 h-1
2.7(0.4)
2.0(0.5)
2.0(0.4)
2.2(0.3)
1.7(0.3)
1.6(0.3)

21.0(2.4)
25.1(2.5)
20.5(2.6)
23.3(2.9)
34.7(4.2)*
40.7(6.6)**

17.6(3.1)
13.4(3.3)
11.8(1.8)
17.2(4.3)
18.9(3.5)
20.5(3.1)

96.5 (8.1)
104.8(12.3)
96.1(11.1)
110.5(14.9)
151.4(14.6)**
138.4(14.6)**

3.2(0.6)
3.4(0.7)
2.3(0.3)
2.4(0.3)
2.8(0.4)
2.5(0.4)

6.0(0.7)
6.4(0.7)
5.3(0.7)
5.2(0.7)
4.6(0.5)
5.4(0.5)

33

Table 2.4. Potentially mineralizable C (PMC), ratio of phenol oxidase and peroxidase (OX) to PMC ( OX:PMC), to beta-glucosidase
BG:OX, to N-acetyl-glucosaminidase and tyrosine amino peptidase (NAG+TAP:OX), to acid phosphotase (PHOS:OX), in the
biodiversity gradient treatments (see table 2.1 for treatment acronyms) for the two study years. Values are means (standard error), n=4
replicate blocks. Significance of treatment effect: * P ≤ 0.05, ** P = 0.01 to 0.001. Each year was analyzed separately.

Year

Rotation

PMC
µg CO2-C g-1
2010 cS
31.3(3.6)
cC
29.3(3.8)
CS
36.0(5.5)
CSW
37.4(5.7)
CSW1cov 55.4(6.8)*
CSW2cov 52.9(6.8)*

0.16(0.03)
0.10(0.01)
0.13(0.03)
0.09(0.01)
0.05(0.01)**
0.06(0.01)**

22.9 (4.5)
36.7(12.1)
32.3 (6.1)
30.0 (6.1)
57.4(14.5)*
56.2(14.9)*

2011 cS
cC
CS
CSW
CSW1cov
CSW2cov

0.31(0.04)
0.33(0.03)
0.24(0.03)*
0.28(0.04)
0.16(0.01)**
0.19(0.02)**

9.5
10.6
12.3
10.2
13.5
14.3

32.6(3.6)
29.4(2.7)
34.5(3.3)
31.9(2.8)
47.6(4.4)*
45.8(3.8)*

OX:PMC

BG:OX

34

(1.7)
(2.2)
(2.7)
(1.7)
(1.5)
(2.3)

NAG+TAP:OX
9.1 (1.7)**
16.4( 5.2)
14.1 (2.6)
14.0 (2.5)
26.8 (6.6)*
27.6 (7.1)*
5.1 (0.8)
5.1 (0.8)
5.4 (0.9)
5.8 (0.9)
7.7 (0.7)*
8.5 (1.0)**

PHOS:OX
62.9(18.9)
76.4(25.2)
62.7(13.0)
60.6(17.0)
113.2(27.2)*
114.8(41.0)*
14.4
13.5
17.2
16.5
22.7
21.1

(3.1)
(2.2)
(3.6)
(2.9)
(2.8)*
(3.8)*

Table 2.5 In situ carbon dioxide fluxes in the biodiversity gradient treatments (see table 2.1 for
treatment acronyms) for the two study years. Median, maximum (Max) and mean (standard
error) for n=4 replicate blocks. Significance of treatment effect: * P ≤ 0.05, ** P = 0.001 to
0.01, *** P ≤ 0.001. Each year was analyzed separately.
Year

Rotation

2010

cS
cC
CS
CSW
CSW1cov
CSW2cov

2011

cS
cC
CS
CSW
CSW1cov
CSW2cov

Median
Max
Mean (SE)
-1
-1
kg CO2-C - ha -day
21.8
43.2
22.2 (2.4) **
26.3
87.1
33.8 (4.7)
30.9
86.2
35.8 (3.9)
29.5
73.8
34.9 (3.8)
31.1
100.3
39.1 (5.4)
31.1
90.1
41.2 (4.8) *
13.4
14.6
13.7
12.6
16.2
20.0

35

22.7
21.3
23.3
21.1
23.5
30.7

14.1 (1.2)
14.0 (1.1)
14.4 (1.1)
13.5 (1.0)
16.2 (1.0) **
18.8 (1.3) ***

Figure 2.1. Potentially mineralizable C (PMC) in the biodiversity gradient treatments (see table
2.1 for treatment acronyms) for a corn year (2010) and a soybean year (2011). Values are means
with standard error for each month, n=4 replicate blocks.

36

Figure 2.2. Enzyme activities for beta-glucosidase (BG), peroxidase (PEROX) + phenol oxidase
(PHENOX), N-acetyl-glucosaminidase (NAG) + tyrosine amino peptidase (TAP), acid
phosphatase (PHOS). All points are means across blocks (n=4) with standard errors for a corn
year (left) and following soybean year (right). See table 2.1 for treatment acronyms.

37

Figure 2.3. Ratio of soil phenol oxidase (PHENOX) + peroxidase (PEROX) enzyme activities to
potentially mineralizable carbon (PMC). All points are means across blocks (n=4) with standard
errors for a corn year (left) and following soybean year (right). Letters indicate statistical
difference (P≤0.05) based on post hoc Tukey’s test. See table 2.1 for treatment acronyms.

38

Figure 2.4. In situ carbon dioxide flux measurements for KBS biodiversity gradient (see table
2.1 for treatment acronyms) for two years. Time points are means with standard error, n=4.
Asterisks (*) show significant differences (P≤ 0.05) by treatments at a specific time point. In
2010, cS was left out of treatment comparisons, as was cC in 2011.

39

CA1 (23%)

CA1 (47%)

Figure 2.5. Canonical correspondence analysis plot of KBS biodiversity gradient for combined
two years of study with differences in rotations constrained to measured soil parameters and the
two axes representing the greatest explained variation (on a percentage basis). Blue ellipse
encircles treatments with cover crops. Figures are from cumulative measures over the season for
all parameters except pH (which represents a mean value). Other parameters are Moisture (soil
moisture), inorganic N (NH4+ and NO3-), potentially mineralizable N (PMN) and C (PMC),
permanganate oxidizable C (POXC), enzyme activities for beta-glucosidase (BG), peroxidase
(PEROX) + phenol oxidase (PHENOX), N-acetyl-glucosaminidase (NAG) + tyrosine amino
peptidase (TAP), acid phosphatase (PHOS). All ordinations were created with a correlational
matrix to give equal weights to all soil parameters. See table 2.1 for treatment acronyms.

40

Complex Rotation

Monoculture
CO2
CO2
PMC

PMC
Trophic cascade

Ox E

Microbes

Ox E

Microbes

Recalcitrant plant C
Mineral bound C

SOC
SOC
Figure 2.6. Model for C transformation in contrasting continuous ‘Monoculture’ (cC) and
‘Complex Rotations’ (CSW2cov) on the KBS biodiversity gradient experiment. Boxes and solid
arrows for specific parameters are in relative proportion to measured values from this study.
Dashed arrows represent posited, but unknown C flows. Green arrows at the top are plant inputs
entering a labile pool of potentially mineralizable C (PMC). Yellow arrows represent the
magnitude of hydrolytic enzyme activity (H E) and gray arrows are oxidative enzyme activities
(Ox E), which draw soil C substrates into microbes (gray box). Passing out of microbial cells is
respired CO2 (black arrow) as well as some unknown quantity of C that passes back into a labile
C fraction via a trophic cascade (red dashed arrow) and microbial-processed C which complexes
with soil as ‘Mineral bound C’ (dark green dashed arrow, also unknown). Stable C which forms
the bulk of SOC (light brown box) derives both from microbially-processed C and from
‘Recalcitrant plant C’ (dashed black arrow) which is not readily accessed by microbial activity
and passes into longer term SOC pools.

41

Supplemental Table 2.1 Inorganic N (NH4+-N and NO3--N) and potentially mineralizable N
(PMN) in the biodiversity gradient treatments (see table 2.1 for treatment acronyms) for the two
study years. Values are means (standard error), n=4 replicate blocks. Significance of treatment
effect: * P ≤ 0.05, ** P = 0.001 to 0.01. Each year was analyzed separately.

cS
cC
CS
CSW
CSW1cov
CSW2cov

NH4+
µg NH4+-N g-1
2.5(0.6)
1.7(0.4)
1.2(0.2)
2.2(0.6)
3.1(0.7)*
1.9(0.2)

NO3µg NO3--N g-1
2.9(0.5)
1.3(0.3)
0.9(0.1)
1.2(0.3)
3.6(0.8)**
2.4(0.4)*

PMN
µg NH4+-N g-1
5.2(0.6)
4.0(0.5)
5.3(0.6)
5.2(0.6)
8.7(0.7)*
9.0(0.8)*

cS
cC
CS
CSW
CSW1cov
CSW2cov

1.7(0.3)
1.7(0.2)
1.9(0.3)
1.7(0.2)
2.3(0.4)
2.0(0.3)

2.4(0.3)
1.6(0.2)
2.5(0.2)
2.1(0.2)
3.6(0.6)
2.5(0.3)

4.1(0.5)
3.6(0.4)
4.0(0.5)
4.9(0.6)
8.5(0.8)*
8.9(0.9)*

Year

Rotation

2010

2011

42

Supplemental Table 2.2. In situ carbon dioxide fluxes in the biodiversity gradient treatments for individual time points over two years
with mean (standard error) for n=4 replicate blocks, followed by letters denoting significant differences between treatments. See table
2.1 for treatment acronyms.

2010

g CO2-C ha-1 day-1

cS
cC
CS
CSW
CSW1cov
CSW2cov

07/13/10
27.1 (1.8)a
38.9 (9.3)ab
41.7 (3.7)ab
47.1 (4.5)ab
45.9 (5.9)ab
65.3 (7.2)b

2011

07/21/10
43.2 (9.4)
87.1 (17.8)
114.5 (28.4)
73.8 (8.8)
100.3 (19.1)
90.1 (10.0)

07/28/10
28.4 (6.8)
38.5 (7.6)
43.5 (6.0)
44.1 (5.4)
45.4 (4.2)
51.8 (10.8)

08/04/10
22.8 (3.6)
28.7 (4.2)
34.3 (9.4)
35.1 (7.1)
33.9 (3.0)
29.5 (6.1)

08/31/10
11.9 (2.2)
19.7 (3.3)
25.0 (5.4)
22.4 (5.4)
22.9 (5.2)
22.1 (2.1)

09/09/10
7.3 (1.1)
10.8 (1.5)
14.8 (3.0)
9.4 (2.9)
13.0 (2.3)
14.5 (3.5)

09/23/10
20.7 (4.7)
22.6 (4.8)
23.4 (4.5)
24.0 (2.6)
18.9 (2.9)
23.4 (4.1)

09/30/10
16.2 (0.6)a
23.9 (4.9)ab
27.5 (2.3)ab
23.7 (4.4)ab
28.2 (3.6)ab
32.7 (3.9)b

g CO2-C ha-1 day-1

06/21/11
cS
13.4 (1.5)a
cC
12.9 (1.8)a
CS
13.7 (0.9)a
CSW
12.6 (1.4)a
CSW1cov 23.2 (2.5)b
CSW2cov 23.0 (4.4)b

06/29/11
6.0 (0.9)a
5.1 (0.4)a
6.8 (0.4)a
6.3 (1.1)a
11.2 (1.5)b
11.6 (1.4)b

07/13/11
17.3 (2.7)
21.3 (0.9)
14.3 (2.1)
15.7 (2.0)
20.3 (0.7)
22.1 (3.1)

07/20/11
13.0 (2.4)a
15.4 (1.9)ab
13.5 (1.6)a
12.2 (0.6)a
16.5 (0.5)ab
20.0 (1.8)b

43

08/04/11
22.7 (2.6)
14.6 (6.1)
21.9 (2.9)
20.6 (2.7)
15.6 (1.4)
23.7 (1.6)

08/12/11
10.0 (1.1)
12.9 (1.0)
12.8 (2.3)
10.3 (1.8)
11.0 (1.2)
12.4 (0.7)

08/17/11
14.5 (2.1)
16.9 (1.9)
15.2 (1.4)
15.2 (2.0)
16.2 (1.1)
17.1 (1.0)

08/31/11
22.2 (5.3)
20.0 (1.0)
23.3 (3.6)
21.1 (2.5)
23.5 (0.9)
30.7 (3.8)

09/23/11
8.0 (1.9)
7.4 (2.4)
7.6 (1.3)
8.0 (1.0)
8.2 (1.8)
8.3 (2.0)

Supplemental Figure 2.1. Ratio of Potentially mineralizable C (PMC) to soil enzyme activities,
beta-glucosidase, BG:PMC, to N-acetyl-glucosaminidase and tyrosine amino peptidase
(NAG+TAP):PMC, to acid phosphatase, PHOS:PMC, in the biodiversity gradient treatments.
See table 2.1 for treatment acronyms. All points are means across blocks (n=4) with standard
errors for a corn year (left) and following soybean year (right).

44

Supplemental Figure 2.2 Ratio of soil phenol oxidase (PHENOX) + peroxidase (PEROX)
enzyme activities to beta-glucosidase (BG:OX), to N-acetyl-glucosaminidase and tyrosine amino
peptidase (NAG+TAP:OX), to acid phosphotase (PHOS:OX), in the biodiversity gradient
treatments. All points are means across blocks (n=4) with standard errors for a corn year (left)
and following soybean year (right). Letters indicate statistical difference (P<0.05) based on post
hoc Tukey’s test. See table 2.1 for treatment acronyms.

45

Supplemental Figure 2.3 Inorganic N (NH4+-N and NO3--N) in the biodiversity gradient (see
table 2.1 for treatment acronyms) treatments for the two study years. All points are means across
blocks (n=4) with standard errors for a corn year (left) and following soybean year (right).

46

Supplemental Figure 2.4 Net primary productivity (NPP) with standard error from the KBS
Biodiversity Gradient Experiment. Estimated NPP calculated as mean of three years (longest
period for a full rotation) for 2010 is the mean NPP of 2008, 2009 and 2010 only for the plots
examined in this study. A 6 year mean is estimated NPP for all phases of each rotation from
years 2008 to 2014 (each appearing every year for all treatments on the experimental site). Both
these calculations are determined from the harvest index following Bolinder et al. (2007) and
excluding biomass removed as harvested grain. Non-cash crop aboveground NPP (including
cover crop and weeds) was also calculated from fall biomass collections. See table 2.1 for
treatment acronyms.

47

REFERENCES

48

REFERENCES

Andrén, O., Bengtsson, J., Clarholm, M., 1995. Biodiversity and species redundancy among litter
decomposers, in: Collins, H.P., Robertson, G.P., Klug, M.J. (Eds.), The Significance and
Regulation of Soil Biodiversity, Developments in Plant and Soil Sciences. Springer Netherlands,
pp. 141–151. doi:10.1007/978-94-011-0479-1_12
Bolinder, M.A., Janzen, H.H., Gregorich, E.G., Angers, D.A., VandenBygaart, A.J., 2007. An approach
for estimating net primary productivity and annual carbon inputs to soil for common agricultural
crops in Canada. Agriculture, Ecosystems & Environment 118, 29–42.
doi:10.1016/j.agee.2006.05.013
Borer, E.T., Seabloom, E.W., Shurin, J.B., Anderson, K.E., Blanchette, C.A., Broitman, B., Cooper, S.D.,
Halpern, B.S., 2005. What Determines the Strength of a Trophic Cascade? Ecology 86, 528–537.
doi:10.1890/03-0816
Bradford, M.A., Strickland, M.S., DeVore, J.L., Maerz, J.C., 2012. Root carbon flow from an invasive
plant to belowground foodwebs. Plant Soil 359, 233–244. doi:10.1007/s11104-012-1210-y
Broussard, W., Turner, R.E., 2009. A century of changing land-use and water-quality relationships in the
continental US. Frontiers in Ecology and the Environment 7, 302–307.
Clark, M.S., Horwath, W.R., Shennan, C., Scow, K.M., 1998. Changes in Soil Chemical Properties
Resulting from Organic and Low-Input Farming Practices. Agronomy Journal 90, 662–671.
doi:10.2134/agronj1998.00021962009000050016x
Colman, B.P., Schimel, J.P., 2013. Drivers of microbial respiration and net N mineralization at the
continental scale. Soil Biology and Biochemistry 60, 65–76. doi:10.1016/j.soilbio.2013.01.003
Cotrufo, M.F., Wallenstein, M.D., Boot, C.M., Denef, K., Paul, E., 2013. The Microbial EfficiencyMatrix Stabilization (MEMS) framework integrates plant litter decomposition with soil organic
matter stabilization: do labile plant inputs form stable soil organic matter? Glob Change Biol 19,
988–995. doi:10.1111/gcb.12113
David, M.B., Drinkwater, L.E., McIsaac, G.F., 2010. Sources of nitrate yields in the Mississippi River
Basin. Journal of Environmental Quality 39, 1657–1667.
Davis, A.S., Hill, J.D., Chase, C.A., Johanns, A.M., Liebman, M., 2012. Increasing Cropping System
Diversity Balances Productivity, Profitability and Environmental Health. PLoS ONE 7, e47149.
doi:10.1371/journal.pone.0047149

49

Doane, T.A., Horwáth, W.R., 2003. Spectrophotometric Determination of Nitrate with a Single Reagent.
Analytical Letters 36, 2713–2722. doi:10.1081/AL-120024647
Drinkwater, L.E., Cambardella, C.A., Reeder, J.D., Rice, C.W., 1996. Potentially mineralizable nitrogen
as an indicator of biologically active soil nitrogen, in: Methods for Assessing Soil Quality. Soil
Science Society of America, Madison, WI, pp. 217–229.
Drinkwater, L.E., Wagoner, P., Sarrantonio, M., 1998. Legume-based cropping systems have reduced
carbon and nitrogen losses. Nature 396, 262–265. doi:10.1038/24376
Dungait, J.A.J., Hopkins, D.W., Gregory, A.S., Whitmore, A.P., 2012. Soil organic matter turnover is
governed by accessibility not recalcitrance. Glob Change Biol 18, 1781–1796.
doi:10.1111/j.1365-2486.2012.02665.x
DuPont, S.T., Ferris, H., Van Horn, M., 2009. Effects of cover crop quality and quantity on nematodebased soil food webs and nutrient cycling. Applied Soil Ecology 41, 157–167.
doi:10.1016/j.apsoil.2008.10.004
Fierer, N., Schimel, J.P., 2003. A Proposed Mechanism for the Pulse in Carbon Dioxide Production
Commonly Observed Following the Rapid Rewetting of a Dry Soil. Soil Sci. Soc. Am. J. 67,
798–805.
Franzluebbers, A.J., Haney, R.L., Honeycutt, C.W., Schomberg, H.H., Hons, F.M., 2000. Flush of carbon
dioxide following rewetting of dried soil relates to active organic pools. Soil Science Society of
America Journal 64, 613–623.
Gardner, J.B., Drinkwater, L.E., 2009. The fate of nitrogen in grain cropping systems: a meta-analysis of
15N field experiments. Ecological Applications 19, 2167–2184.
German, D.P., Weintraub, M.N., Grandy, A.S., Lauber, C.L., Rinkes, Z.L., Allison, S.D., 2011.
Optimization of hydrolytic and oxidative enzyme methods for ecosystem studies. Soil Biology
and Biochemistry 43, 1387–1397. doi:10.1016/j.soilbio.2011.03.017
Grandy, A.S., Robertson, G.P., 2007. Land-Use Intensity Effects on Soil Organic Carbon Accumulation
Rates and Mechanisms. Ecosystems 10, 59–74. doi:10.1007/s10021-006-9010-y
Gregorich, E.G., Drury, C.F., Baldock, J.A., 2001. Changes in soil carbon under long-term maize in
monoculture and legume-based rotation. Canadian journal of soil science 81, 21–31.
Haney, R., Brinton, W., Evans, E., 2008. Soil CO2 respiration: Comparison of chemical titration, CO2
IRGA analysis and the Solvita gel system. Renewable Agriculture and Food Systems 23, 171.

50

Hoben, J.P., Gehl, R.J., Millar, N., Grace, P.R., Robertson, G.P., 2011. Nonlinear nitrous oxide (N2O)
response to nitrogen fertilizer in on-farm corn crops of the US Midwest. Global Change Biology
17, 1140–1152. doi:10.1111/j.1365-2486.2010.02349.x
Hooper, D.U., Chapin, F.S., Ewel, J.J., Hector, A., Inchausti, P., Lavorel, S., Lawton, J.H., Lodge, D.M.,
Loreau, M., Naeem, S., Schmid, B., Setälä, H., Symstad, A.J., Vandermeer, J., Wardle, D.A.,
2005. Effects of Biodiversity on Ecosystem Functioning: A Consensus of Current Knowledge.
Ecological Monographs 75, 3–35.
IPCC, 2013. , The Physical Science Basis. Summary for Policymakers. United Nations.
Kallenbach, C., Grandy, A.S., 2011. Controls over soil microbial biomass responses to carbon
amendments in agricultural systems: A meta-analysis. Agriculture, Ecosystems & Environment
144, 241–252. doi:10.1016/j.agee.2011.08.020
Kallenbach, C.M., Frey, S.D., Grandy, A.S., 2016. Direct evidence for microbial-derived soil organic
matter formation and its ecophysiological controls. Nature Communications 7, 13630.
doi:10.1038/ncomms13630
Kallenbach, C.M., Grandy, A.S., Frey, S.D., Diefendorf, A.F., 2015. Microbial physiology and necromass
regulate agricultural soil carbon accumulation. Soil Biology and Biochemistry 91, 279–290.
doi:10.1016/j.soilbio.2015.09.005
Klopatek, J.M., Olson, R.J., Emerson, C.J., Joness, J.L., 1979. Land-use conflicts with natural vegetation
in the United States. Environmental Conservation 6, 191–199.
Kong, A.Y.Y., Six, J., 2010. Tracing Root vs. Residue Carbon into Soils from Conventional and
Alternative Cropping Systems. Soil Science Society of America Journal; Madison 74, 1201–
1210.
Ladoni, M., Basir, A., Kravchenko, A., 2015. Which Soil Carbon Fraction is the Best for Assessing
Management Differences? A Statistical Power Perspective. Soil Science Society of America
Journal 79, 848–857. doi:10.2136/sssaj2014.10.0426
Lee, Z.M., Schmidt, T.M., 2014. Bacterial growth efficiency varies in soils under different land
management practices. Soil Biology and Biochemistry 69, 282–290.
doi:10.1016/j.soilbio.2013.11.012
Liang, C., Balser, T.C., 2011. Microbial production of recalcitrant organic matter in global soils:
implications for productivity and climate policy. Nat Rev Micro 9, 75–75.
doi:10.1038/nrmicro2386-c1

51

Liebman, M., Helmers, M.J., Schulte, L.A., Chase, C.A., 2013. Using biodiversity to link agricultural
productivity with environmental quality: Results from three field experiments in Iowa.
Renewable Agriculture and Food Systems 28, 115–128. doi:10.1017/S1742170512000300
Matson, P.A., Parton, W.J., Power, A.G., Swift, M.J., 1997. Agricultural Intensification and Ecosystem
Properties. Science 277, 504–509. doi:10.1126/science.277.5325.504
McDaniel, M.D., Tiemann, L.K., Grandy, A.S., 2013. Does agricultural crop diversity enhance soil
microbial biomass and organic matter dynamics? A meta-analysis. Ecological Applications 24,
560–570. doi:10.1890/13-0616.1
Plaza-Bonilla, D., Nolot, J.-M., Raffaillac, D., Justes, E., 2015. Cover crops mitigate nitrate leaching in
cropping systems including grain legumes: Field evidence and model simulations. Agriculture,
Ecosystems & Environment 212, 1–12. doi:10.1016/j.agee.2015.06.014
Rabalais, N.N., Turner, R.E., Scavia, D., 2002. Beyond Science into Policy: Gulf of Mexico Hypoxia and
the Mississippi RiverNutrient policy development for the Mississippi River watershed reflects the
accumulated scientific evidence that the increase in nitrogen loading is the primary factor in the
worsening of hypoxia in the northern Gulf of Mexico. BioScience 52, 129–142.
doi:10.1641/0006-3568(2002)052[0129:BSIPGO]2.0.CO;2
Rasse, D.P., Rumpel, C., Dignac, M.-F., 2005. Is soil carbon mostly root carbon? Mechanisms for a
specific stabilisation. Plant and Soil 269, 341–356.
Robertson, G.P., Grandy, A.S., 2006. Soil System Management in Temperate Regions, in: Biological
Approaches to Sustainable Soil Systems, Books in Soils, Plants, and the Environment. CRC
Press, pp. 27–39.
Robertson, G.P., Gross, K.L., Hamilton, S.K., Landis, D.A., Schmidt, T.M., Snapp, S.S., Swinton, S.M.,
2014. Farming for Ecosystem Services: An Ecological Approach to Production Agriculture.
BioScience 64, 404–415. doi:10.1093/biosci/biu037
Robertson, G.P., Hamilton, S.K., 2015. Long-term ecological research in agricultural landscapes at the
Kellogg Biological Station LTER site: conceptual and experimental framework, in: The Ecology
of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. Oxford
University Press, New York, New York, USA, pp. 1–32.
Robertson, P.G., Vitousek, P.M., 2009. Nitrogen in Agriculture: Balancing the Cost of an Essential
Resource. Annual Review of Environment and Resources 34, 97–125.
doi:10.1146/annurev.environ.032108.105046

52

Sanford, G.R., Posner, J.L., Jackson, R.D., Kucharik, C.J., Hedtcke, J.L., Lin, T.-L., 2012. Soil carbon
lost from Mollisols of the North Central U.S.A. with 20 years of agricultural best management
practices. Agriculture, Ecosystems & Environment 162, 68–76. doi:10.1016/j.agee.2012.08.011
Schmidt, M.W.I., Torn, M.S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I.A., Kleber, M.,
Kögel-Knabner, I., Lehmann, J., Manning, D.A.C., Nannipieri, P., Rasse, D.P., Weiner, S.,
Trumbore, S.E., 2011. Persistence of soil organic matter as an ecosystem property. Nature 478,
49–56. doi:10.1038/nature10386
Shcherbak, I., Millar, N., Robertson, G.P., 2014. Global metaanalysis of the nonlinear response of soil
nitrous oxide (N2O) emissions to fertilizer nitrogen. PNAS 111, 9199–9204.
doi:10.1073/pnas.1322434111
Shennan, C., 2008. Biotic interactions, ecological knowledge and agriculture. Philosophical Transactions
of the Royal Society of London B: Biological Sciences 363, 717–739.
doi:10.1098/rstb.2007.2180
Sinsabaugh, R.., Reynolds, H., Long, T.., 2000. Rapid assay for amidohydrolase (urease) activity in
environmental samples. Soil Biology and Biochemistry 32, 2095–2097. doi:10.1016/S00380717(00)00102-4
Sinsabaugh, R.L., 2010. Phenol oxidase, peroxidase and organic matter dynamics of soil. Soil Biology
and Biochemistry 42, 391–404. doi:10.1016/j.soilbio.2009.10.014
Sinsabaugh, R.L., Carreiro, M.M., Alvarez, S., 2002. Enzyme and microbial dynamics of litter
decomposition. Enzymes in the Environment, Activity, Ecology, and Applications. Marcel
Dekker, New York, Basel 249–265.
Sinsabaugh, R.L., Hill, B.H., Follstad Shah, J.J., 2009. Ecoenzymatic stoichiometry of microbial organic
nutrient acquisition in soil and sediment. Nature 462, 795–798. doi:10.1038/nature08632
Smith, C.M., David, M.B., Mitchell, C.A., Masters, M.D., Anderson-Teixeira, K.J., Bernacchi, C.J.,
DeLucia, E.H., 2013. Reduced nitrogen losses after conversion of row crop agriculture to
perennial biofuel crops. Journal of environmental quality 42, 219–228.
Smith, R., Gross, K., Robertson, G., 2008. Effects of Crop Diversity on Agroecosystem Function: Crop
Yield Response. Ecosystems 11, 355-366–366.
Syswerda, S.P., Corbin, A.T., Mokma, D.L., Kravchenko, A.N., Robertson, G.P., 2011. Agricultural
Management and Soil Carbon Storage in Surface vs. Deep Layers. Soil Science Society of
America Journal 75, 92–101. doi:10.2136/sssaj2009.0414

53

Tiemann, L.K., Grandy, A.S., Atkinson, E.E., Marin-Spiotta, E., McDaniel, M.D., 2015. Crop rotational
diversity enhances belowground communities and functions in an agroecosystem. Ecol Lett 18,
761–771. doi:10.1111/ele.12453
Tilman, D., Reich, P.B., Knops, J., Wedin, D., Mielke, T., Lehman, C., 2001. Diversity and Productivity
in a Long-Term Grassland Experiment. Science 294, 843–845. doi:10.1126/science.1060391
Tonitto, C., David, M.B., Drinkwater, L.E., 2006. Replacing bare fallows with cover crops in fertilizerintensive cropping systems: A meta-analysis of crop yield and N dynamics. Agriculture,
Ecosystems & Environment 112, 58–72. doi:10.1016/j.agee.2005.07.003
Wardle, D.A., 2002. Communities and ecosystems: linking the aboveground and belowground
components. Princeton Univ Pr.
Wood, S.L.R., Rhemtulla, J.M., Coomes, O.T., 2016. Cropping history trumps fallow duration in longterm soil and vegetation dynamics of shifting cultivation systems. Ecol Appl n/a-n/a.
doi:10.1002/eap.1462
Zavaleta, E.S., Pasari, J.R., Hulvey, K.B., Tilman, G.D., 2010. Sustaining multiple ecosystem functions in
grassland communities requires higher biodiversity. PNAS 107, 1443–1446.
doi:10.1073/pnas.0906829107
Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common statistical
problems. Methods in Ecology and Evolution 1, 3–14. doi:10.1111/j.2041-210X.2009.00001.x

54

CHAPTER 3: COVER CROPS CHANGE SOIL BACTERIAL COMPOSITION BUT
NOT DIVERSITY

ABSTRACT
In natural systems a strong relationship exists between aboveground biodiversity and
numerous ecosystem functions. Here I examined how increased crop diversity in row crop
rotations affects the diversity and composition of soil bacteria. I predicted that inclusion of more
crops in rotation, including overwintering cover crops, would increase bacterial diversity in
tandem with increases in soil ecosystem function. Along a gradient of increasing crops in
rotation and inclusion of cover crops, we sampled 16S rDNA from soil bacterial communities.
Sequencing and screening revealed a range of 3275 and 3459 distinct operational taxonomic
units (OTUs) among distinct rotations. I found no effect of crop rotation on bacterial richness or
diversity, but a strong divergence in bacterial community composition in rotations with cover
crops, whereas all other rotations without cover crops were strongly similar. The change in
community between rotations with and without cover fell along clear taxonomic lines. Among
the top ten percent of most abundant OTUs, Proteobacteria dominated in the rotations with
cover crops and while Acidobacteria predominated in rotations without cover. These
communities correlated strongly to measures of carbon and nitrogen mineralization and enzyme
activity, and indicated a physiological basis for the shift in community composition in response
to inclusion of leguminous cover crops in rotation.

INTRODUCTION
Intensification of agriculture is a major factor leading to global environmental change
(Balmford et al., 2012; Foley et al., 2011; West et al., 2014). This transformation has resulted in

55

large landscapes in the U.S. Midwest devoted to production of a few primary annual grain crops.
Productivity in these agroecosystems is often dependent on large nonrenewable inputs and
frequent soil disturbances, which contribute to soil erosion and other downstream impacts to air
and water. With respect to nutrient pollution, these low plant diversity agroecosystems lack
many of the soil ecosystem functions that drive stable and efficient nutrient cycling in natural
ecosystems (Hooper et al., 2005). However, it is not clear if introducing greater plant diversity
into agricultural systems will lead to belowground ecosystem functions similar to those that
occur in diverse, natural ecosystems (Zak et al., 2003).
Soils represent one of the most biologically diverse habitats (Orgiazzi et al., 2016) but
belowground species diversity does not follow a clear relationship with aboveground diversity
(Shange et al., 2012). For instance, nitrogen (N) fertilization reduces aboveground diversity
(Bobbink et al., 2010), but this is not necessarily reflected in belowground microbial
communities (Ramirez et al., 2010). Diverse plant residue inputs alter belowground nutrient
cycles carried out by microbes (Wardle, 2002); however, the links between aboveground
diversity and belowground diversity may depend on the specific microbial process, such as
methane oxidation (Levine et al., 2011) or denitrification (Powell et al., 2015).
Aboveground communities can shape the composition of soil microbial communities as
well as alter the function of microbial processes, such as decomposition, which are common
across these communities (Strickland et al., 2009). In cropping systems, management practices
alter both microbial growth efficiency and mineralization of soil carbon (C) (Lee and Schmidt,
2014), as well as the fate of specific C substrates (Kallenbach et al., 2015). While aboveground
plant diversity can enhance soil fertility (Zak et al., 2003), high fertility and microbial biomass in
agricultural systems might not be directly tied to aboveground plant species per se, but rather to

56

number of crops in rotation and especially to presence of a cover crop (McDaniel et al., 2013).
Over time, specific cropping systems can alter soil microbial diversity, but the reasons for this
observation are not clear; for example, whether these changes are due to specific plant inputs or
to rotational composition (Venter et al., 2016).
To help address this gap, we investigated the effect of increasing crop rotational
complexity on bacterial community composition. We focus on a long term cropping system
experiment that has received no external inputs (e.g. fertilizer or pest control agents), and varies
only in the crop rotation, which includes a gradient from continuous summer annuals, corn (Zea
mays L.) and soybean (Glycine max.), to summer annuals with a winter annual (wheat; Triticum
aestivum L.), to inclusion of non-harvested cover crops (legumes and grasses) in rotation. In this
cropping system experiment, increasing complexity of crops in rotation has resulted in greater
crop productivity (Smith et al., 2008), altered microbial processing of soil substrates (McDaniel
et al., 2016), increased soil C, enzyme activity, and soil aggregation (Tiemann et al., 2015), and
has increased labile C pools and patterns in soil C oxidation (see chapter 2). We expect that
increased crop productivity, diversity, and abundance of plant inputs entering the soil, and
resulting changes in the spatial heterogeneity of the soil habitat, have also restructured soil
bacterial communities. Along a cropping systems gradient from continuous grain monoculture to
annual grains in rotation with cover crops, we expect that rotations with greater plant biomass
inputs will restructure microbial community composition and increase bacterial diversity.
Specifically, we expect that the effect of cover crops in rotation will have the greatest effect on
bacterial community structure due to their strong effect on increasing mineralizable substrate
(see chapter 2) and stimulating soil microbial biomass (McDaniel et al., 2013).

57

METHODS
Site description
We conducted our study at Michigan State University’s Kellogg Biological Station
(KBS) Long Term Ecological Research experiment located in Hickory Corners, Michigan (42°
24′ N, 85° 24′ W, elevation 288 m). KBS lies in SW Michigan, in the eastern part of the U.S.
Corn Belt. Prior to 1988, the site had been conventionally managed for row crop agriculture
(Robertson et al., 1997). Soils developed on glacial outwash are composed of Kalamazoo (Fineloamy) and Oshtemo (Coarse-loamy) mixed, mesic, Typic Hapludalfs (FAO soil order:
Luvisols). The climate is temperate, with approximately 1005 mm rainfall and an average
snowfall of ~1.3 m. Mean annual temperature is 10.1° C (Robertson and Hamilton, 2015).
We used the KBS Biodiversity Gradient Experiment
(http://lter.kbs.msu.edu/research/long-termexperiments/biodiversity-gradient/), which consists of
a gradient of rotational complexity of harvested annual grains including corn, soybean, and
wheat ranging from continuous monoculture (e.g., continuous corn, or continuous soybean) to
various rotations that also include an unharvested overwintering cover crop of red clover
(Trifolium prantense L.), crimson clover (Trifolium incarnatum), or annual ryegrass (Lolium
multiflorum Lam.) either alone or in combinations. The study was established in 2000 and is
composed of plant species treatments from bare ground, to 1, 2, 3, 4, 6 and 10 species in a
randomized complete block experiment with four replicates. Treatment plots are 9 x 30 m and
every phase of each crop rotation is represented every year. Since establishment, none of the
treatments have received synthetic chemical inputs (i.e. no N additions of fertilizer or pest
control agents) and weeds are controlled mechanically.

58

Soil sampling
For this study we selected six treatments (Table 3.1): continuous corn (cC) and
continuous soy (cS), corn and soybean in rotation (CS), CS and wheat in rotation (CSW), CSW
with inclusion of clover varieties as a winter cover crop (CSW1cov), and CSW with clovers and
annual ryegrass in a cover crop mixture (CSW2cov). At initial sampling the rotations had been in
place for ten years. Soil samples were collected after harvest each year: on November 18, 2010
and October 27, 2011. At each sampling, cores were taken along a crop row, offset by one meter
from the central row in the plot, and at equal spacing along a 20 meter length, approximately
four meters in from either edge of the plot. One-inch soil cores were taken to a depth of 10 cm.
Two sub-samples, one at the furrow center and one from the middle of the furrow slope were
taken from five locations, along a crop row. The 10 cores were sieved to 2 mm, homogenized,
and maintained on ice during transport to the lab. In the laboratory, composited soil from each
plot was subsampled for different analysis, and approximately fifty grams were frozen in
Whirlpack ® bags and stored at -80° C.

Soil DNA extraction and PCR amplification
DNA was extracted from 0.25 g of each treatment soil using the MO-BIO PowerMag
Soil DNA Isolation kit (MO BIO, Carlsbad, CA) on an Eppendorf epMotion 5075 TMX robot
(Hamburg, Germany). The V4 region of the 16S rRNA gene was amplified using a dualindexing approach (Kozich et al., 2013). PCR amplification and sequencing was performed
using 1 µl of template DNA with the Illumina MiSeq platform, with the MiSeq Reagent Kit V2
500 cycles (Cat# MS-102-2003), according to the manufacturer’s instructions except that
Accuprime High Fidelity Taq (Life Cat # 12346094 ) was used instead of Accuprime Pfx
supermix. Conditions for the reaction were 2 m at 95° C followed by 30 cycles at 95° C for 20 s,

59

55° C for 15 s and 72° C for 5 m, followed by 10 m at 72° C and then maintained at 4° C.
Amplified products were visualized using an E-Gel 96 with SYBR Safe DNA Gel Stain, 2%
(Life technologies cat# G7208-02) and normalized using Life technologies SequalPrep
Normalization Plate Kit (cat # A10510-01) following the manufacturer’s protocol. Samples
were pooled and concentration determined using Kapa Biosystems Library Quantification kit for
Illumina platforms (KapaBiosystems KK4824). The sizes of the amplicons in the library were
determined using the Agilent Bioanalyzer High Sensitivity DNA analysis kit (cat# 5067-4626) so
that the final library consists of equal molar amounts. Final library preparation and sequencing
was carried out following the manufacture’s protocol, MiSeq (part# 15039740 Rev. D), for 2nM
or 4nM libraries. After DNA extraction, all sample preparation and DNA sequencing were
carried out by the Host Microbiome Initiative at the University of Michigan.

Analysis of 16S bacterial communities
Sequences were sorted by sample and quality filtered using the mothur software (Schloss
et al., 2009). Individual operational taxonomic units (OTUs) were classified using the SILVA
database (Quast et al., 2013), and individual communities were sub-sampled to the number of
sequences in the sample with the fewest amplicons for all further community analysis. Within
mothur, richness was assessed using the Chao 1 estimator and diversity using Shannon and
Inverse Simpson indices. Analysis of molecular variance (Amova) was used as a non-parametric
method of testing for significant differences among populations. Finally, within mothur,
differential abundance of taxa across treatments was assessed via Metastats, Lefse and Indicator
species (McCune et al., 2002; Segata et al., 2011; White et al., 2009) to find consistent patterns
in key taxa occurring in different rotations. Further community analysis was carried out using
the Vegan package (Oksanen et al., 2013) in R (R Core Team, 2012), including non-metric

60

multidimensional scaling (using Bray-Curtis), analysis of similarity (ANOSIM) and
perMANOVA (both with permutations=999), Mantel tests, and fitting of environmental variable
onto bacterial communities.

RESULTS
Following quality filtering of 16S rDNA amplicons, the treatments were subsampled to
17,255 OTUs, the size of the smallest library across all treatments for both years. This resulted
in an estimated coverage of no less than 94.8 ± 0.3 percent across all treatments (Table 3.2).
Chao 1 OTU richness estimates ranged from a low of 3275 ± 120 in CSW2cov to 3459 ± 83 in
CS with no significant difference across treatments. For estimates of diversity using Inverse
Simpson’s index, cS had the lowest value (299.8 ± 12.1), and cC the highest (324.1 ± 21.5),
again with no significant difference or noticeable trend across crop rotation treatments.
Bacterial communities of treatments did differ (P=0.003) when variation in OTUs was
transformed into a distance matrix (Bray-Curtis) with variance partitioned by treatment
(controlling for block and year), and estimated using perMANOVA (Table 3.3), after checking
for homogeneity for variance. The same analysis, when partitioned instead by the presence or
absence of a cover crop, showed even stronger significance (P=0.001), but with higher residual
error (Table 3.3). A similar result was found in ANOSIM, comparing similarities within
grouped treatments (Table 3.3). Another distance-based estimate (Amova), showing pairwise
comparisons, revealed no significant difference among treatments without cover crops, and the
strongest divergence in communities between cC and CSW1cov (P=0.020), cC and CSW2cov
(P=0.034), and CS and CSW1cov (P=0.0029) (Table 3.4). The patterns are clearly visible when

61

scaled and shown in two dimension using NMDS (Figure 3.1). The effect of year (and thus crop
in rotation except for mS and mC) had no effect on this structure (Figure 3.2).
Using a matrix of soil nutrient cycling parameters along the rotational gradient (Figure 3.3,
see also Chapter 2) we used a Mantel test for correlation between two separate years of
environmental variables (transformed using Euclidean distances) and OTUs by rotation. For
2010, the correlation was not significant (r=0.128, P=0.127) and for 2011, this test was
significant (r=0.187, P=0.033). Fitting vectors from soil environmental variables onto OTU
distance matrix of crop rotations for 2010 showed significant goodness of fit (Table 3.5) with
potentially mineralizable N (PMN; P=0.008), and C (PMC; P=0.013) as well as the enzyme Nacetyl-glucosaminidase (NAG) activity (P=0.033). For 2011, the stronger correlation of
environmental and 16S communities matrices is evident in stronger relationships in fitting
environmental factors: PMN (P=0.009), PMC (P=0.001), NAG (P=0.017), POXC (permanganate
oxidizable C; P=0.006), a measure of labile C, as well as soil nitrate, ammonium, acid
phosphatase activity and soil pH (Table 3.5).
Differences in the relative abundance in major bacteria phyla were not significant and
showed no clear pattern among treatments or years (Figure 3.5). We ranked bacteria by the
significance of their influence on distinguishing cover crop and non-cover crop treatments using
a consensus of significance effects (P≤0.05) between techniques in mothur: Lefse, Metastat,
Indicator species. The 20 most influential OTUs representing cover and no-cover treatments
represent approximately 11.0% of mean amplicons across all treatments (with coefficient of
variation between individual treatments of 12%). The most influential OTUs in distinguishing
cover crop rotation came from the Proteobacteria, while the most predominant OTUs
distinguishing non-cover treatments were in the Acidobacteria (Figure 3.4, Table 3.6).

62

DISCUSSION
The legacy of crop rotation in this long term experiment significantly altered composition
of soil bacterial, especially in rotations including cover crops which strongly influenced the
community structure compared to treatments with no cover. In two successive years, the
bacterial community did not shift based on the identity of the grain crop in rotation, and among
treatments, the relative abundance of bacterial phyla was not significantly different. There was
no evidence for differences in bacterial species richness or diversity based on rotation, as I had
hypothesized. However, a clear pattern emerged in the taxa, which differentiated rotations with
and without cover crops. Among unique OTUs, Proteobacteria taxa distinguished bacterial
communities in cover crops while in rotations without cover crops taxa from the Acidobacteria
dominated the bacterial community. The distinct taxa shaping bacterial community structure
strongly correlated with soil ecosystem functions, especially for N and C mineralization between
cover and no cover treatments.
While the linkages between aboveground plant diversity and belowground ecosystem
functions are strong (Wardle, 2002), the effects on the diversity of soil organisms that carry out
many of these functions is less clear. In agricultural systems, increasing crop rotation
complexity has alternately been shown to increase bacterial diversity over the long term
(Figuerola et al., 2015), to have no effect (Navarro-Noya et al., 2013), or to have a negative
effect on diversity and richness (Yin et al., 2010). A recent meta-analysis assessing the effect of
crop rotation on belowground microbial communities found evidence for a small increase in
diversity with rotation, although findings differed in part due to methodological approach
(Venter et al., 2016), with newer sequencing technologies revealing a decrease in diversity with
increased crop rotation diversity. Similarly, controlled studies manipulating plant species

63

richness in grassland systems have demonstrated strong effects of richness on soil ecosystem
function (Steinauer et al., 2015), and associated shifts in microbial community composition, but
have found a negative correlation between plant species richness and soil bacterial diversity
(Schlatter et al., 2015). Thus, the lack of any clear pattern in belowground bacterial diversity
(Table 3.2), in spite of large differences in plant species richness and evenness, suggests that
other factors shape belowground communities.
Agriculture has an strong effect on soil bacterial communities that persists for years after
land is taken out of crop production (Buckley and Schmidt, 2001; Jangid et al., 2011). While
soil bacterial communities may exhibit pronounced heterogeneity in the relative abundance of
different phyla over very small spatial scales, at the ecosystem scale common structural patterns
emerge, particularly linked to management (O’Brien et al., 2016). Seasonal row cropping of
monocultures along with tillage practices may contribute to this ecosystem-scale community
homogeneity. In this study, the relative abundance of major phyla did not differ across crop
rotations (Figure 3.5), but bacterial community structure was significantly different by rotation,
especially for treatments with cover crops (Table 3.3). However, this difference in bacterial
communities by rotation, which remained stable regardless of most recent grain crop in rotation)
without any change in bacterial diversity or richness may overlay a more dynamic and shifting
bacterial community.
While a large segment of species composing estimates of bacterial diversity in soils may
actually derive from dormant cells (Jones and Lennon, 2010), even in soils under decades-long
monoculture cultivation, the introduction of new plant species can rapidly alter the soil bacterial
community structure (Maul and Drinkwater, 2010). And, within ecosystems, the relative
abundance of specific soil bacterial taxa shift with temporal and seasonal patterns (Buckley and

64

Schmidt, 2003). Furthermore, spatial heterogeneity created by soil aggregation is an important
determinant of microbial community structure (Blackwood et al., 2006), and cover cropped
treatments within this rotational gradient experiment have significantly higher soil aggregate
formation than those without cover (Tiemann et al., 2015). Thus an interaction of short term
plant species effects, seasonal effects, and spatial heterogeneity of the soil ecosystem may act to
distinguish microbial community structure among rotations, especially for treatments with cover
crops (Table 3.4).
However microbial communities in rotations with cover crops also appear to process soil
C in a manner distinct from rotations without cover crops (Chapter 2), such that longer-term
changes in substrate quality may be a more important factor shaping the microbial community
structure and physiology (Frey et al., 2013) than the proximal influence of different plant species
inputs (Maul and Drinkwater, 2010). Sampling for bacterial communities for each year of this
study occurred post-harvest - November, 2010 after corn in and October, 2011 after soybean. In
the corn year, cover cropped treatments were planted to overwintering species, with red clover
interseeded into corn and rye planted post-harvest into CSW2cov (Table 3.1). In the soybean
year cover cropped treatments were interseeded with crimson clover, which was terminated at
soybean harvest and followed by wheat planting. Despite these short-term differences in the
plant species present over the two years of sampling, the soil bacterial community structure was
stable among similar treatments. That is, differences in community structure were driven not by
the most recent crop species in rotation, but instead differed based on the long-term history of
presence or absence of winter cover (Figure 3.1, Figure 3.2).
Increased cropping system diversity, and especially the use of cover crops, supports
higher microbial biomass (McDaniel et al., 2013) and larger labile soil C and N pools (Chapter

65

2) readily accessible to microbes. Here, the bacterial taxa that most strongly distinguished
treatments with cover crops from those without winter cover belonged to the Proteobacteria
(Figure 3.4), while taxa that differentiated treatments without cover crops belonged to the
Acidobacteria. Bacteria in these two phyla are among the most common in soil (Janssen, 2006)
and their relative abundance often varies based on ecosystem type. For example forest
ecosystems generally have a greater abundance of Acidobacteria taxa relative to agroecosystems,
while the reverse is true for Proteobacteria (Jangid et al., 2008, 2011). Relative abundance of
Proteobacteria tend to increase with N addition to soil (Ramirez et al., 2010) as many of these
taxa appear to carry genes for numerous N cycling pathways (Nelson et al., 2015). These trends
in the abundance of distinct taxa relative to available soil substrates align with patterns in
nutrient cycling in treatments with and without cover crops (Chapter 2), but also point toward
altered microbial metabolism and resource use across different rotations.
Distinct taxa in rotations with and without cover crops are an indication of the soil
ecosystem function of these communities such as catabolic potential, and rates of C and N
mineralization (McDaniel and Grandy, 2016). Rapid, inefficient growth is a characteristic of
many Proteobacteria taxa (Roller et al., 2016), which respire more CO2 per unit biomass
production than slower-growing organisms (Roller and Schmidt, 2015). By contrast, the bacteria
within the Acidobacteria are characterized by slow doubling times associated with more efficient
growth in oligotrophic environments (Eichorst et al., 2007). In spite of this contrast in growth
efficiency, microbial communities in croplands may have a greater overall bacterial growth
efficiency (BGE) than those in adjacent deciduous forest soils (Lee and Schmidt, 2014). Yet it is
notable that even though forest soils generally have greater soil C than adjacent cropland, within
agricultural and managed grasslands, Acidobacteria abundance tends to increase as total soil C

66

declines (Schmidt and Waldron, 2015). Finally, N inputs to soil are known to lower the relative
abundance of Acidobacteria (Cederlund et al., 2014), alter microbial respiration (Matulich and
Martiny, 2014) and slow soil C decomposition (Ramirez et al., 2010). This may mean that while
microbial communities are fundamentally altered by cropping, they respond differently to short
term inputs of critical resources.
Rotations with cover crops have larger pools of labile C and N that rise and fall over the
growing season in greater magnitude than rotations without cover crops (Chapter 2), and
variation in these substrate fluxes are strongly correlated with the underlying bacterial
community structure (Table 3.5). BGE in annual cropping systems also shifts dramatically over
the growing season (Lee and Schmidt, 2014). Thus as substrate availability changes in annual
cropping systems it may strongly reshape bacterial community structure and at different times
favor taxa which process substrates with different efficiencies. Inclusion of cover crops in row
crop rotations has been found to significantly alter the process of soil C decomposition
(Kallenbach et al., 2015) leading to disproportionately higher soil C accumulation than rotations
without cover (Syswerda and Robertson, 2014). While over the long term, increasing rotational
complexity, with or without cover crops, had no effect on overall bacterial species richness or
diversity, my results suggest that cover crops shift soil microbial communities along bacterial
phylogenies with distinct physiologies which suggests a basis functional changes in C and N
cycling under cover crops.
ACKNOWLEDGEMENTS
Support for this research was also provided by U.S. Department of Agriculture Soil Processes
Program, grant # 2009-65107-05961, and the National Science Foundation Long-term Ecological
Research Program (DEB 1027253) at the Kellogg Biological Station.

67

APPENDIX

68

Table 3.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment

Treatment
cS
cC
CS
CSW
CSW1cov
CSW2cov

Rotation
soy
corn
corn + soy
corn + soy + wheat
corn + soy + wheat + clover
corn + soy + wheat + clover + rye

Species / yr.
1
1
1
1-2
2-3
3-4

Species / rot.
1
1
2
3
4
5

1
2
3
4
5

6

Month M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D
C

S

W

red clover

69

crimson clover

rye

Table 3.2 Estimated population parameters 16S rDNA from rotations across the biodiversity
gradient experiment (see table 3.1 for treatment acronyms), including mean ± se for coverage,
estimated OTU richness (Chao1) and diversity (Inverse Simpsons).

Rotation
cS
cC
CS
CSW
CSW1cov
CSW2cov

Coverage (%)
95.1 ± 0.2
94.8 ± 0.3
94.8 ± 0.1
94.8 ± 0.2
94.8 ± 0.2
95.1 ± 0.2

Chao 1
3279 ± 93
3437 ± 123
3458 ± 83
3408 ± 107
3389 ± 117
3275 ± 119

70

Inverse Simpsons
300 ± 12
315 ± 26
324 ± 21
289 ± 12
301 ± 19
312 ± 16

Table 3.3 Multivariate analysis of 16S rDNA from the biodiversity gradient experiment using
perMANOVA with two different factors, rotation and presence of a cover crop, showing the
sources of variation, degrees of freedom (DF), sums of squares (SS) mean square (MS), Fstatistic (F) and related coefficient of determination (R2) and probability (Prob.). Analysis of
similarities (ANOSIM) between grouped factors showing the correlation coefficient (R) and
probability (Prob).

perMANOVA
Source
Factor
Rotation Rotation
Residuals
Total
Cover

Cover
Residuals
Total

DF
5

SS
0.55

MS
0.1108

42
47

3.58
4.13

0.0852
1

1
46
47

0.27
3.86
4.13

0.27
0.084
1

F
R2
1.3 0.134

Prob.
0.003

ANOSIM
R
Prob.
0.0928
0.003

0.001

0.281

0.866

3.21 0.065
0.935

71

0.001

Table 3.4 Analysis of molecular variance (Amova) of 16S OTUs, with treatment-level (Overall)
F-statistic (P) and table of pair-wise comparisons of individual treatments across the biodiversity
gradient experiment. See table 3.1 for treatment acronyms.

Overall Among
Populations 1.272
(0.008*)
cS
cC
CS
CSW
CSW1cov
CSW2cov

cS

cC

CS

CSW

CSW1cov

CSW2cov

0.870 (0.717)
0.976 (0.474)
0.879 (0.689)
1.710 (0.010)
1.355 (0.065)

0.813 (0.836)
0.768 (0.918)
1.661 (0.021)
1.260 (0.122)

0.884 (0.685)
2.329 (0.001)
1.756 (0.001)

1.681 (0.01)
1.257 (0.092)

0.86 (0.768)

-

72

Table 3.5 Vector analysis of correlation between matrix of edaphic factors (Euclidean) and 16S
rDNA (Bray-Curtis) for two years on the biodiversity gradient experiment. For each year the
goodness of fit (r2) and the significance, P, was calculated for soil parameters: inorganic N
(NH4+-N and NO3--N), potentially mineralizable N (PMN) and C (PMC), permanganate
oxidizable C (POXC), enzyme activities for beta-glucosidase (BG), peroxidase (PEROX) +
phenol oxidase (PHENOX), N-acetyl-glucosaminidase (NAG) + tyrosine amino peptidase
(TAP), acid phosphatase (PHOS).

2010
Soil Factors
PMN
PMC
NAG
POXC
NO3NH4+
PHOS
BG
TAP
PEROX
PHENOX
pH

2

r
0.37
0.36
0.28
0.16
0.21
0.18
0.10
0.20
0.07
0.08
0.02
0.17

P
0.008
0.013
0.033
0.166
0.081
0.113
0.332
0.111
0.449
0.452
0.825
0.121

73

2011
r Pr(>r)
0.38 0.009
0.62 0.001
0.35 0.017
0.41 0.006
0.33 0.024
0.25 0.050
0.24 0.046
0.17 0.155
0.23 0.071
0.13 0.234
0.00 0.961
0.28 0.035
2

**
*
*
.

**
***
*
**
*
*
*

.
*

Table 3.6. The 20 the most influential OTUs ranked by their abundance (after normalizing and

sub-setting across all treatments and replicates) and the significance (P≤0.05) of their influence
on distinguishing cover crop and non-cover crop treatments.
Rotation
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover
Cover

SILVA classification of OUT
Bacteria(100);Proteobacteria(100);Alphaproteobacteria(100);Sphingomonadales(100);
Sphingomonadaceae(100);unclassified(100);
Bacteria(100);Bacteroidetes(100);Sphingobacteriia(100);Sphingobacteriales(100);
Chitinophagaceae(100);Flavisolibacter(91);
Bacteria(100);Actinobacteria(100);Actinobacteria(100);Actinomycetales(100);
Micrococcaceae(100);Arthrobacter(100);
Bacteria(100);Proteobacteria(100);Betaproteobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Verrucomicrobia(100);Spartobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Proteobacteria(100);Gammaproteobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp16(100);Gp16(100);
unclassified(100);unclassified(100);
Bacteria(100);Proteobacteria(100);Deltaproteobacteria(100);Myxococcales(100);
unclassified(100);unclassified(100);
Bacteria(100);Actinobacteria(100);Actinobacteria(100);Actinomycetales(100);
Micromonosporaceae(100);unclassified(84);
Bacteria(100);Proteobacteria(100);Deltaproteobacteria(100);Myxococcales(100);
Myxococcaceae(79);Corallococcus(71);
Bacteria(100);Proteobacteria(100);Alphaproteobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp3(100);Gp3(100);unclassified(100);un
classified(100);
Bacteria(100);Verrucomicrobia(100);Spartobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Proteobacteria(100);Alphaproteobacteria(100);Rhizobiales(100);
Hyphomicrobiaceae(59);Rhodoplanes(59);
Bacteria(100);Proteobacteria(100);Deltaproteobacteria(100);Myxococcales(100);
unclassified(100);unclassified(100);
Bacteria(100);candidate_division_WPS1(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
Bacteria(100);Verrucomicrobia(100);Subdivision3(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);
unclassified(100);unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);

74

Table 3.5 (cont’d)
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover
No cover

Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp16(100);Gp16(100);
unclassified(100);unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);
unclassified(100);unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp6(100);Gp6(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);unclassified(100);
unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp7(100);Gp7(100);unclassified(100);
unclassified(100);
Bacteria(100);Gemmatimonadetes(100);Gemmatimonadetes(100);
Gemmatimonadales(100);Gemmatimonadaceae(100);Gemmatimonas(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Verrucomicrobia(100);Spartobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Proteobacteria(100);Betaproteobacteria(100);unclassified(100);
unclassified(100);unclassified(100);
Bacteria(100);Nitrospirae(100);Nitrospira(100);Nitrospirales(100);Nitrospiraceae(100);
Nitrospira(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp4(100);Gp4(100);unclassified(100);
unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp6(100);Gp6(100);unclassified(100);
unclassified(100);
Bacteria(100);unclassified(100);unclassified(100);unclassified(100);unclassified(100);
unclassified(100);
Bacteria(100);Actinobacteria(100);Actinobacteria(100);Actinomycetales(100);Geoderm
atophilaceae(100);Geodermatophilus(65);
Bacteria(100);Acidobacteria(100);Acidobacteria_Gp17(100);Gp17(100);
unclassified(100);unclassified(100);

75

Figure 3.1 Non-metric multidimensional scaling (NMDS) plot of 16S rDNA of KBS
biodiversity gradient for combined two years of study with rotations in different colors, and
years in different shapes (Stress = 0.1681). Blue ellipse encircles treatments with cover crops.
See table 3.1 for treatment acronyms.

76

Figure 3.2 Non-metric multidimensional scaling (NMDS) plot of 16S rDNA of KBS
biodiversity gradient with all rotations shown by sampling year. (Stress = 0.1681).

77

CA1 (23%)

CA1 (47%)

Figure 3.3 Canonical correspondence analysis plot of KBS biodiversity gradient for combined
two years of study with differences in rotations constrained to measured soil parameters and the
two axes representing the greatest explained variation (on a percentage basis) . Figures are from
cumulative measures over the season for all parameters except pH (which represents a mean
value). Other parameters are Moisture (soil moisture), inorganic N (NH4+-N and NO3--N),
potentially mineralizable N (PMN) and C (PMC), permanganate oxidizable C (POXC), enzyme
activities for beta-glucosidase (BG), peroxidase (PEROX) + phenol oxidase (PHENOX), Nacetyl-glucosaminidase (NAG) + tyrosine amino peptidase (TAP), acid phosphatase (PHOS). All
ordinations were created with a correlational matrix to give equal weights to all soil parameters.
See table 3.1 for treatment acronyms.

78

A

B

Figure 3.4 Phyla of 20 most abundant OTUs, significantly (P≤0.05) related to treatments with
cover crops (A) or with no cover crops (B). Values are mean sequences across treatments of
cover vs. no cover from normalized, subsampled 16S rDNA libraries

79

2011

Relative abundance (%)

Relative abundance (%)

Relative abundance (%)

2010

Figure 3.5 Relative abundance, after subsampling OTUs across rotations in the biodiversity
experiment gradient of bacteria phyla (Acidobacteria, Verrucomicrobia) and class
(Alphaproteobacteria) for two years. Box and whisker plots are shown with points from
individual treatments (n=4). See table 3.1 for treatment acronyms.

80

REFERENCES

81

REFERENCES

Balmford, A., Green, R., Phalan, B., 2012. What conservationists need to know about farming.
Proceedings of the Royal Society B: Biological Sciences 279, 2714–2724.
doi:10.1098/rspb.2012.0515
Blackwood, C.B., Dell, C.J., Smucker, A.J.M., Paul, E.A., 2006. Eubacterial communities in different soil
macroaggregate environments and cropping systems. Soil Biology and Biochemistry 38, 720–
728. doi:10.1016/j.soilbio.2005.07.006
Bobbink, R., Hicks, K., Galloway, J., Spranger, T., Alkemade, R., Ashmore, M., Bustamante, M.,
Cinderby, S., Davidson, E., Dentener, F., 2010. Global assessment of nitrogen deposition effects
on terrestrial plant diversity: a synthesis. Ecological Applications 20, 30–59.
Buckley, D.H., Schmidt, T.M., 2003. Diversity and dynamics of microbial communities in soils from
agro-ecosystems. Environmental Microbiology 5, 441–452. doi:10.1046/j.14622920.2003.00404.x
Buckley, D.H., Schmidt, T.M., 2001. The structure of microbial communities in soil and the lasting
impact of cultivation. Microb Ecol 42, 11–21. doi:10.1007/s002480000108
Cederlund, H., Wessén, E., Enwall, K., Jones, C.M., Juhanson, J., Pell, M., Philippot, L., Hallin, S., 2014.
Soil carbon quality and nitrogen fertilization structure bacterial communities with predictable
responses of major bacterial phyla. Applied Soil Ecology 84, 62–68.
doi:10.1016/j.apsoil.2014.06.003
Eichorst, S.A., Breznak, J.A., Schmidt, T.M., 2007. Isolation and Characterization of Soil Bacteria That
Define Terriglobus gen. nov., in the Phylum Acidobacteria. Appl. Environ. Microbiol. 73, 2708–
2717. doi:10.1128/AEM.02140-06
Figuerola, E.L.M., Guerrero, L.D., Türkowsky, D., Wall, L.G., Erijman, L., 2015. Crop monoculture
rather than agriculture reduces the spatial turnover of soil bacterial communities at a regional
scale. Environ Microbiol 17, 678–688. doi:10.1111/1462-2920.12497
Foley, J.A., Ramankutty, N., Brauman, K.A., Cassidy, E.S., Gerber, J.S., Johnston, M., Mueller, N.D.,
O/’Connell, C., Ray, D.K., West, P.C., Balzer, C., Bennett, E.M., Carpenter, S.R., Hill, J.,
Monfreda, C., Polasky, S., Rockstrom, J., Sheehan, J., Siebert, S., Tilman, D., Zaks, D.P.M.,
2011. Solutions for a cultivated planet. Nature 478, 337–342. doi:10.1038/nature10452
Frey, S.D., Lee, J., Melillo, J.M., Six, J., 2013. The temperature response of soil microbial efficiency and
its feedback to climate. Nature Clim. Change 3, 395–398. doi:10.1038/nclimate1796

82

Hooper, D.U., Chapin, F.S., Ewel, J.J., Hector, A., Inchausti, P., Lavorel, S., Lawton, J.H., Lodge, D.M.,
Loreau, M., Naeem, S., Schmid, B., Setälä, H., Symstad, A.J., Vandermeer, J., Wardle, D.A.,
2005. Effects of Biodiversity on Ecosystem Functioning: A Consensus of Current Knowledge.
Ecological Monographs 75, 3–35.
Jangid, K., Williams, M.A., Franzluebbers, A.J., Sanderlin, J.S., Reeves, J.H., Jenkins, M.B., Endale,
D.M., Coleman, D.C., Whitman, W.B., 2008. Relative impacts of land-use, management intensity
and fertilization upon soil microbial community structure in agricultural systems. Soil Biology
and Biochemistry 40, 2843–2853. doi:10.1016/j.soilbio.2008.07.030
Jangid, K., Williams, M.A., Franzluebbers, A.J., Schmidt, T.M., Coleman, D.C., Whitman, W.B., 2011.
Land-use history has a stronger impact on soil microbial community composition than
aboveground vegetation and soil properties. Soil Biology and Biochemistry 43, 2184–2193.
doi:10.1016/j.soilbio.2011.06.022
Janssen, P.H., 2006. Identifying the Dominant Soil Bacterial Taxa in Libraries of 16S rRNA and 16S
rRNA Genes. Appl. Environ. Microbiol. 72, 1719–1728. doi:10.1128/AEM.72.3.1719-1728.2006
Jones, S.E., Lennon, J.T., 2010. Dormancy contributes to the maintenance of microbial diversity. PNAS
107, 5881–5886. doi:10.1073/pnas.0912765107
Kallenbach, C.M., Grandy, A.S., Frey, S.D., Diefendorf, A.F., 2015. Microbial physiology and necromass
regulate agricultural soil carbon accumulation. Soil Biology and Biochemistry 91, 279–290.
doi:10.1016/j.soilbio.2015.09.005
Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., Schloss, P.D., 2013. Development of a dualindex sequencing strategy and curation pipeline for analyzing amplicon sequence data on the
MiSeq Illumina sequencing platform. Applied and environmental microbiology 79, 5112–5120.
Lee, Z.M., Schmidt, T.M., 2014. Bacterial growth efficiency varies in soils under different land
management practices. Soil Biology and Biochemistry 69, 282–290.
doi:10.1016/j.soilbio.2013.11.012
Levine, U.Y., Teal, T.K., Robertson, G.P., Schmidt, T.M., 2011. Agriculture’s impact on microbial
diversity and associated fluxes of carbon dioxide and methane. ISME J 5, 1683–1691.
Matulich, K.L., Martiny, J.B.H., 2014. Microbial composition alters the response of litter decomposition
to environmental change. Ecology 96, 154–163. doi:10.1890/14-0357.1
Maul, J., Drinkwater, L., 2010. Short-term plant species impact on microbial community structure in soils
with long-term agricultural history. Plant Soil 330, 369–382. doi:10.1007/s11104-009-0211-y

83

McCune, B., Grace, J.B., Urban, D.L., 2002. Analysis of ecological communities. MjM software design
Gleneden Beach.
McDaniel, M.D., Grandy, A.S., 2016. Soil microbial biomass and function are altered by 12 years of crop
rotation. SOIL Discussions 1–39. doi:10.5194/soil-2016-39
McDaniel, M.D., Grandy, A.S., Tiemann, L.K., Weintraub, M.N., 2016. Eleven years of crop
diversification alters decomposition dynamics of litter mixtures incubated with soil. Ecosphere 7,
n/a-n/a. doi:10.1002/ecs2.1426
McDaniel, M.D., Tiemann, L.K., Grandy, A.S., 2013. Does agricultural crop diversity enhance soil
microbial biomass and organic matter dynamics? A meta-analysis. Ecological Applications 24,
560–570. doi:10.1890/13-0616.1
Navarro-Noya, Y.E., Gómez-Acata, S., Montoya-Ciriaco, N., Rojas-Valdez, A., Suárez-Arriaga, M.C.,
Valenzuela-Encinas, C., Jiménez-Bueno, N., Verhulst, N., Govaerts, B., Dendooven, L., 2013.
Relative impacts of tillage, residue management and crop-rotation on soil bacterial communities
in a semi-arid agroecosystem. Soil Biology and Biochemistry 65, 86–95.
doi:10.1016/j.soilbio.2013.05.009
Nelson, M.B., Berlemont, R., Martiny, A.C., Martiny, J.B.H., 2015. Nitrogen Cycling Potential of a
Grassland Litter Microbial Community. Appl. Environ. Microbiol. 81, 7012–7022.
doi:10.1128/AEM.02222-15
O’Brien, S.L., Gibbons, S.M., Owens, S.M., Hampton-Marcell, J., Johnston, E.R., Jastrow, J.D., Gilbert,
J.A., Meyer, F., Antonopoulos, D.A., 2016. Spatial scale drives patterns in soil bacterial diversity.
Environ Microbiol 18, 2039–2051. doi:10.1111/1462-2920.13231
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L.,
Solymos, P., Henry, M., Stevens, H., others, 2013. Vegan: Community Ecology Package.
Orgiazzi, A., Bardgett, R.D., Barrios, E., Behan-Pelletier, V., Briones, M.J.I., Chotte, J.-L., De Deyn,
G.B., Eggleton, P., Fierer, N., Fraser, T., Hedlund, K., Jeffery, S., Johnson, N.C., Jones, A.,
Kandeler, E., Kaneko, N., Lavelle, P., Lemanceau, P., Miko, L., Montanarella, L., Moreira,
F.M.S., Ramirez, K.S., Scheu, S., Singh, B.K., Six, J., van der Putten, W.H., Wall, D.H., 2016.
Global soil biodiversity atlas. Publications Office of the European Union, Luxembourg.
Powell, J.R., Welsh, A., Hallin, S., 2015. Microbial functional diversity enhances predictive models
linking environmental parameters to ecosystem properties. Ecology 96, 1985–1993.
doi:10.1890/14-1127.1

84

Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Peplies, J., Glöckner, F.O., 2013.
The SILVA ribosomal RNA gene database project: improved data processing and web-based
tools. Nucl. Acids Res. 41, D590–D596. doi:10.1093/nar/gks1219
R Core Team, 2012. R: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing, Vienna, Austria.
Ramirez, K.S., Lauber, C.L., Knight, R., Bradford, M.A., Fierer, N., 2010. Consistent effects of nitrogen
fertilization on soil bacterial communities in contrasting systems. Ecology 91, 3463–3470.
doi:10.1890/10-0426.1
Robertson, G.P., Hamilton, S.K., 2015. Long-term ecological research in agricultural landscapes at the
Kellogg Biological Station LTER site: conceptual and experimental framework, in: The Ecology
of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. Oxford
University Press, New York, New York, USA, pp. 1–32.
Roller, B.R., Schmidt, T.M., 2015. The physiology and ecological implications of efficient growth. ISME
J 9, 1481–1487. doi:10.1038/ismej.2014.235
Roller, B.R.K., Stoddard, S.F., Schmidt, T.M., 2016. Exploiting rRNA operon copy number to investigate
bacterial reproductive strategies. Nature Microbiology 1, 16160.
doi:10.1038/nmicrobiol.2016.160
Schlatter, D.C., Bakker, M.G., Bradeen, J.M., Kinkel, L.L., 2015. Plant community richness and
microbial interactions structure bacterial communities in soil. Ecology 96, 134–142.
doi:10.1890/13-1648.1
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A.,
Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Horn, D.J.V.,
Weber, C.F., 2009. Introducing mothur: Open-Source, Platform-Independent, CommunitySupported Software for Describing and Comparing Microbial Communities. Appl. Environ.
Microbiol. 75, 7537–7541. doi:10.1128/AEM.01541-09
Schmidt, T.M., Waldron, C., 2015. Microbial Diversity in Soil of Agricultural Landscapes and Its
Relation to Ecosystem Functions, in: The Ecology of Agricultural Landscapes: Long-Term
Research on the Path to Sustainability. Oxford University Press, USA.
Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W.S., Huttenhower, C., 2011.
Metagenomic biomarker discovery and explanation. Genome Biology 12, R60. doi:10.1186/gb2011-12-6-r60

85

Shange, R.S., Ankumah, R.O., Ibekwe, A.M., Zabawa, R., Dowd, S.E., 2012. Distinct Soil Bacterial
Communities Revealed under a Diversely Managed Agroecosystem. PLoS ONE 7, e40338.
doi:10.1371/journal.pone.0040338
Smith, R., Gross, K., Robertson, G., 2008. Effects of Crop Diversity on Agroecosystem Function: Crop
Yield Response. Ecosystems 11, 355-366–366.
Steinauer, K., Tilman, D., Wragg, P.D., Cesarz, S., Cowles, J.M., Pritsch, K., Reich, P.B., Weisser,
W.W., Eisenhauer, N., 2015. Plant diversity effects on soil microbial functions and enzymes are
stronger than warming in a grassland experiment. Ecology 96, 99–112. doi:10.1890/14-0088.1
Strickland, M.S., Lauber, C., Fierer, N., Bradford, M.A., 2009. Testing the functional significance of
microbial community composition. Ecology 90, 441–451. doi:10.1890/08-0296.1
Syswerda, S.P., Robertson, G.P., 2014. Ecosystem services along a management gradient in Michigan
(USA) cropping systems. Agriculture, Ecosystems & Environment 189, 28–35.
doi:10.1016/j.agee.2014.03.006
Tiemann, L.K., Grandy, A.S., Atkinson, E.E., Marin-Spiotta, E., McDaniel, M.D., 2015. Crop rotational
diversity enhances belowground communities and functions in an agroecosystem. Ecol Lett 18,
761–771. doi:10.1111/ele.12453
Venter, Z.S., Jacobs, K., Hawkins, H.-J., 2016. The impact of crop rotation on soil microbial diversity: A
meta-analysis. Pedobiologia.
Wardle, D.A., 2002. Communities and ecosystems: linking the aboveground and belowground
components. Princeton University Press.
West, P.C., Gerber, J.S., Engstrom, P.M., Mueller, N.D., Brauman, K.A., Carlson, K.M., Cassidy, E.S.,
Johnston, M., MacDonald, G.K., Ray, D.K., others, 2014. Leverage points for improving global
food security and the environment. Science 345, 325–328.
White, J.R., Nagarajan, N., Pop, M., 2009. Statistical Methods for Detecting Differentially Abundant
Features in Clinical Metagenomic Samples. PLOS Computational Biology 5, e1000352.
doi:10.1371/journal.pcbi.1000352
Yin, C., Jones, K.L., Peterson, D.E., Garrett, K.A., Hulbert, S.H., Paulitz, T.C., 2010. Members of soil
bacterial communities sensitive to tillage and crop rotation. Soil Biology and Biochemistry 42,
2111–2118. doi:10.1016/j.soilbio.2010.08.006

86

Zak, D.R., Holmes, W.E., White, D.C., Peacock, A.D., Tilman, D., 2003. Plant Diversity, Soil Microbial
Communities, and Ecosystem Function: Are There Any Links? Ecology 84, 2042–2050.
doi:10.1890/02-0433

87

CHAPTER 4: LEGUME COVER CROPS ALTER MICROBIAL CONTROLS ON SOIL
DENITRIFICATION RATES AND EFFICIENCY
ABSTRACT
Negative consequences of intensive inputs into agriculture, such as disruption of the global
nitrogen (N) cycle, have led to calls to improve ecosystem regulation of nutrients by increasing
diversity in cropping systems. Here we examine a long term row cropping experiment which has
received no external inputs except for biological N fixation, to see how increasing crop rotational
complexity, and in particular inclusion of cover crops, influences soil nitrous oxide (N2O)
emissions. Across a range of rotations from continuous monoculture, to rotations of annual
grains, to inclusion of overwintering cover crops, we measured N2O flux rates, soil N pools,
denitrification potential, and a gene for the nitrite reductase (nirK) pathway leading to N2O.
Mineral N values were highest in rotations with cover crops and in continuous soybean
monoculture. However potentially mineralizable N (PMN) was highest in rotations with cover
crops and an adjacent fallow system, ranging between 8.6 and 9.0 µg NH4+-N g-1 which was
approximately twice the concentration of all other rotations without cover crops. Soil mineral N
and PMN were not as strong predictors of N2O as the presence of a legume cover crop – these
rotations had the highest mean flux of 13.2 g N2O-N ha-1 day-1 in year one and 5.2 g N2O-N ha-1
day-1 in year two which was 3-4 times higher than all other rotations. Denitrification in rotations
with winter cover was also distinct. The ratio of N2O produced out of total denitrification (rN2O)
under cover crops was between 0.51-0.59, and significantly lower when compared to rotations
without cover which ranged from 0.7-0.83. In addition, rotations with cover crops had a
significantly greater abundance of nitrite reductase genes (nirK) derived from ammonia
oxidizing bacteria, suggesting greater potential for nitrifier denitrification. My results show that
neither high soil mineral N nor potential mineralization rates were sufficient to predict N2O flux,

88

but rather the presence of legume cover crops in rotation combined with a shift in the microbial
community were most closely related to rates of denitrification.

INTRODUCTION
The intensification of row crop agricultural production into only a few annual grain crop
species dependent on high inputs of mineral fertilizers has altered how soil and plants regulate
nutrient cycling (Drinkwater and Snapp, 2007; Vandermeer et al., 1998). The alteration of the
global nitrogen (N) cycle through agriculture, especially with anthropogenic additions of
industrially-fixed N now exceeding natural (e.g. microbially-fixed) sources of N into soils
(Galloway et al., 2004), has far-reaching downstream impacts and environmental consequences
(Robertson and Vitousek, 2009). Agriculture is the leading source of increasing rates of
atmospheric nitrous oxide (N2O), a potent greenhouse gas with 300 times the global warming
potential per molecule than carbon dioxide (IPCC, 2013). N2O production is largely a naturally
occurring process in soil, which has intensified due to land use change and N inputs to
agriculture (Shcherbak et al., 2014), highlighting the need to both understand and mitigate N2O
from agriculture (Robertson and Vitousek, 2009). One approach to tightening nutrient cycling in
soil is to restore plant diversity to the landscape and reduce intensive fertilizer inputs (Liebman
and Davis, 2000). However, in row crops, the effect of increasing plant diversity on N2O
emissions, while maintaining productive grain-based agroecosystems, is not well understood.
A broad range of microorganisms are responsible for soil N transformations and N2O
production largely through a series of oxidation-reduction reactions (Leininger et al., 2006;
Robertson and Groffman, 2015; Shoun et al., 1992). Nitrification is the critical pathway linking
mineral N transformations of ammonium (NH4+) to nitrite (NO2-) and then to nitrate (NO3-)

89

(Robertson and Groffman, 2015). In the first step, autotrophic ammonium oxidizing bacteria
(AOB) oxidize NH4+ as their principle source of energy, however under low oxygen conditions
or excess NO2-, AOB will reduce the latter to N2O (Kool et al., 2011). Denitrification is an
anaerobic process in which heterotrophic microbes respire C substrates via the serial reduction of
NO3- back to nitrite NO2-, nitric oxide (NO-), N2O and finally to N2 (Zumft and Cárdenas, 1979).
This process is crucial to the global N cycle as it returns reactive N back to the atmosphere
(Gruber and Galloway, 2008), though in the process denitrification also releases N2O, which is
exacerbated in agroecosystems (Gelfand et al., 2016; Robertson et al., 2000).
In nature, the relative contribution of nitrification and denitrification to N2O production is
difficult to ascertain, and while each is favored under different conditions these distinct processes
also share critical components. Unlike nitrification, denitrification is an anoxic process and
requires a C substrate, the lack of which can result in incomplete denitrification and higher
potential N2O flux (Groffman et al., 2006). In addition, denitrification potential, or the rate of
N2O production relative to N2, is distinct in agricultural soils in part due to the structure of the
microbial community and the enzymes expressed in the denitrification pathway (Cavigelli and
Robertson, 2000). The genes in this pathway, such as dissimilatory nitrite reductase (nirK), are
shared across many bacterial lineages (Jones et al., 2008), including AOBs that carry out
nitrification (Casciotti and Ward, 2001), thus offering one approach to distinguish between
processes which produce N2O in soil.
Numerous environmental and management factors influencing the N cycle can directly
and indirectly influence N2O emissions. Within agroecosystems N2O flux can be affected by
crop type, fertilization type and rate, tillage practice, and soil organic C (Gelfand et al., 2016;
Smith et al., 2011; Stehfest and Bouwman, 2006). For example, tillage leads to soil N

90

mineralization and a pulse in soil N2O (Grandy and Robertson, 2007). Plant communities have a
strong effect on N2O production, with perennial systems having far lower emissions than
agroecosystems (Gelfand et al., 2016; Oates et al., 2016). This follows findings in natural
ecosystems where plant diversity exerts a strong influence over the efficiency of soil nutrient
cycles (Tilman et al., 2001) and the N cycle in particular (Zak et al., 2003). And while plant
species diversity can increase soil fertility (Dybzinski et al., 2008), suggesting that increased
diversity in agriculture can offset high external inputs of mineral N (Blesh and Drinkwater,
2013), implementing diverse row crop systems that achieve high N use efficiency remains a
challenge (Robertson and Vitousek, 2009). Increasing the plant diversity of crop rotations (e.g.
by introducing legume cover crops) to enhance soil nutrient cycling and ecosystem processes
may also alter the many factors that control N2O emissions.
In this study we examine the effect of increasing rotational diversity in row crops on soil
N2O production in a long term cropping system that has received no external inputs (e.g.
fertilizer or pesticide) for over ten years. Plant composition ranges from continuous monoculture
to rotated annual grains with and without winter cover. Specifically we hypothesize that greater
abundance of legumes in rotation will increase nitrate availability and lead to greater N2O flux.
We expect this effect to be stronger in rotations with legume species, and smaller in rotations
that also included small grains, which would increase synchrony of mineral N uptake and reduce
N2O flux. In addition we predict that the long term effect of different crop rotations will alter the
composition of genes in the denitrification pathway, specifically nirK. Finally, we hypothesize
that more complex rotations, with greater frequency of legumes and small grains, will lead to
more complete denitrification (measured as DEA), due to greater coupling of C and N cycles,
and abundance of labile C and N forms.

91

METHODS
Site descriptions and experimental layout
We conducted our study at Michigan State University’s Kellogg Biological Station
(KBS) Long Term Ecological Research site in Hickory Corners, Michigan (42° 24′ N, 85° 24′
W, elevation 288 m). KBS lies in SW Michigan, in the eastern part of the U.S. Corn Belt. Prior
to 1988, the site had been conventionally managed for row crop agriculture (Robertson et al.,
1997). Soils developed on glacial outwash and are composed of Kalamazoo (Fine-loamy) and
Oshtemo (Coarse-loamy) mixed, mesic, Typic Hapludalfs (FAO soil order: Luvisols). The
climate is temperate, with approximately 1005 mm rainfall and an average snowfall of ~1.3 m.
Mean annual temperature is 10.1° C (Robertson and Hamilton, 2015).
This work was undertaken on the KBS Biodiversity Gradient Experiment
(http://lter.kbs.msu.edu/research/long-termexperiments/biodiversity-gradient/) which consists of
a gradient of rotational complexity of harvested annual grains including corn, soybean, and
wheat ranging from continuous monoculture (e.g., continuous corn, or continuous soybean) to
various rotations that also include an non-harvested overwintering cover crop of red clover
(Trifolium prantense L.), crimson clover (Trifolium incarnatum), or annual ryegrass (Lolium
multiflorum Lam.) either alone or in combinations. The study was established in 2000 and is
composed of plant species treatments from bare ground, 1, 2, 3,4, 6 and 10 species, in a
randomized complete block experiment with four replicates within 9 x 30 m treatment plots and
with every phase of each crop rotation represented every year. Since establishment, none of the
treatments have received synthetic chemical inputs (i.e. no N fertilizer or pest control agents) and
weeds are controlled mechanically.

92

For this study we selected seven treatments (Fig. 4.1): continuous corn (cC) and
continuous soy (cS), corn and soybean in rotation (CS), CS and wheat in rotation (CSW), CSW
with inclusion of clover varieties (CSW1cov), CSW with clovers and annual ryegrass
(CSW2cov), and a fallow plot tilled once each spring (Fspring) but otherwise allowed to return
to early secondary succession. Sampling occurred over five month periods during two growing
seasons, beginning in 2010 at planting, with most treatments in corn (except for cS and Fspring)
until after soybean harvest in 2011 (except for cC and Fspring).

Soil Sampling
Over each study year I sampled all treatments and blocks for soil C and N parameters,
and enzyme activity. I collected ten cores to 10 cm soil depth from five locations along a central
row, alternating cores from within and between crop rows. Cores were composited by plot and
placed on ice for transport to the lab. Within 24 hours of collection soils were sieved to < 2mm,
mixed, and weighed for soil moisture, enzymes, and mineral N analyses.

Soil carbon and nitrogen
A portion of air-dried soils was finely ground for determination of total SOC and total
soil organic nitrogen (SON) by dry combustion on a Costech ECS 4010 CHNSO Analyzer
(Costech Analytical Technologies, Valencia, CA). Air-dried soil was used to determine labile
soil C as permanganate oxidizable carbon (POXC) according to Weil et al. (2003). Duplicate 2.5
g samples of dried soil were mixed with buffered 0.02 M KMnO4 solution in 50ml conical tubes,
shaken at 120 rpm for two minutes and allowed to settle for eight minutes. From this reaction,
0.5 mL of supernatant were diluted with 49.5 mL of deionized water. The degree of oxidation
was measured colorimetrically at 550 nm on a Fisher Scientific Thermo Multiskan microplate

93

reader (Waltham, MA.) and standardized to a series of known KMnO4 standards. Water-filled
pore space (WFPS) was determined for air-died samples (Haney and Haney, 2010) and
potentially mineralizable carbon (PMC) was adapted from Franzluebbers et al. (2000) and Haney
(2008). Deionized water was added to 15 g of air-dried soil to 50% WFPS in canning jars with
lids fitted with airtight, rubber septa. Lids were placed loosely on the jars and soils were
incubated in the dark at 25° C. At 24-hours, jars were removed and soil respiration determined
using a LI-COR LI-820 infrared gas analyzer (LI-COR Biosciences, Lincoln, NE): jars were
sealed and a 0.5 mL of headspace gas was immediately removed and injected into the gas
analyzer, followed by three additional measurements over approximately 90 min. PMC was
calculated as CO2 flux per gram of incubated soil.
Soil mineral N was extracted from fresh soil within 24 hours of sampling. Triplicate
10.0 g subsamples of sieved soil were extracted with 1M KCl for one hour on a rotary shaker
(120 rpm). Samples were filtered and stored at -20 °C until analysis for ammonium (NH4+) and
nitrate (NO3-). Potentially mineralizable nitrogen (PMN) was assessed anaerobically according to
Drinkwater et al. (1996). Briefly, 10 mL of deionized water was added to triplicate 8 g fresh,
sieved soil in conical tubes. Dinitrogen gas was used to replace tube headspace air and bubbled
into the slurry for one minute prior to sealing with butyl rubber stoppers. Sealed tubes were
incubated at 25 °C for seven days. After seven days the stoppers were removed, buffer was
added to bring the slurry to 1M KCl, and samples were shaken, filtered, and stored in the same
manner as fresh samples. Concentrations of NH4+ and NO3- were determined colorimetrically on
a 96-well plate-reader. Soil NO3- was determined according to Doane and Horwáth (2003) and
read at 540 nm, and soil NH4+ concentration determined at 630 nm (Sinsabaugh et al., 2000).

94

After incubation and extraction, PMN was determined based on the concentration of NH4+
incubated soil minus NH4+ from initial soil extraction of the same soil.

Soil Enzymes
To analyze soil enzymes, soil pH was determined using 10 g fresh soil mixed with 20 mL
of deionized water, which was allowed to equilibrate for one hour with occasional stirring.
Following this determination, the pH of a 50 mM sodium acetate buffer was adjusted to the pH
of treatment soils. Approximately half of 100 mL of sodium acetate buffer was added to 1 g of
treatment soil, and homogenized vigorously in a blender for exactly one minute. The rest of the
buffer was used to rinse remaining soil contents from the blender into the slurry and then stored
at 4° C until analysis (within 1-2 hours). In soil slurries we measured enzymatic potentials for β1-4-N-acetyl-glucosaminidase (NAG) and tyrosine amino peptidase (TAP). At initiation of
enzyme assays, soil slurries were constantly stirred and the suspension was pipetted into 96-well
plates. Hydrolytic enzyme activity was monitored using substrate specific polymers containing
fluorescent labels (either 4-methylumbelliferone or methylcoumarin). Enzymes were incubated
in the dark at 15° C. At 4-6 hours, hydrolytic enzyme reactions were stopped by adding 10 µL,
1M sodium hydroxide, and plates were read on a Fisher Scientific Thermo Fluoroskan AscentTM
microplate fluorometer (Waltham, MA). Background absorbance-emission or fluorescence was
accounted for with controls for buffer, buffer + soil, buffer + substrate.

In situ nitrous oxide
At approximately 2-week intervals soil N2O fluxes were measured using static 25 cmdiameter chambers placed in situ in each treatment. The soil surface inside the chamber was
maintained free of living plant material. PVC chambers were inserted in the field mid-way

95

between rows and central furrows to a depth of 8 cm (Hoben et al., 2011). At each sampling
time, chamber height from soil surface was measured to account for headspace volume and then
chambers were sealed with an air-tight, O-ring PVC lid, fitted with a Vacutainer serum vial septa
(Becton-Dickinson, East Rutherford, NJ). Using a syringe, and mixing the headspace at each
sampling, 10 mL gas samples were removed via the lid septa at time zero and at three additional
20-minute intervals, and stored in 5.9 mL gas Exetainers (Labco, Ceredigion United Kingdom)
with headspace previously flushed. Samples were analyzed using gas chromatography (Hewlett
Packard 5890 Series II, Rolling Meadows, IL, USA) with gases separated on a Poropak Q
column (1.8 m, 80/100 mesh) at 80 °C. N2O was analyzed with a 63Ni electron capture detector
at 350 °C.

Denitrification enzyme assay (DEA)
We measured the potential activity of denitrification enzymes (Groffman et al., 1999)
with soil collected in November 2010. Five soil cores taken to 10 cm depth across each
treatment were pooled and stored on ice for transport. In the laboratory soils were sieved to 4
mm, and 25 g were added each of two 120 ml flasks. To these, 25 mL of a solution with 1 mM
KNO3 and 2 mM succinate was added and flasks were stoppered. To one flask, the headspace
was flushed completely with N2, and to the other approximately 10% of head space gas was
removed and replaced with acetylene. Flasks were incubated at 25 °C on a shaker at 200 rpm
and 3 mL gas samples (into vials previously flushed with N2) were taken at 15, 30, 45, and 60
min from each flask injected into a vial. Sample N2O was measured on an HP 5890II gas
chromatograph (Hewlett Packard, Rolling Meadows, Illinois) equipped with dual Poropak Q
columns set to 80 °C and a 63Ni electron capture detector (ECD). Incubations with acetylene
give the rate of total denitrification (N2O+N2 production) and incubations without acetylene give

96

net N2O production (N2O production - reduction). The molar ratio of N2O production
(N2O/(N2O+N2)) gives the efficiency of complete denitrification.

Nitrite reductase (nirK) PCR and quality filtering
A subsample of sieved soil collected above (for DEA) was frozen at -80°C and later
DNA was extracted and purified using a MO BIO PowerSoil® DNA Isolation kit (MO BIO,
Carlsbad, CA). The target nirK gene was selected following (Henry et al., 2004) and amplified
with PCR using primers nirK-q1F (5'-RTY GGC GGH CAY GGC GA-3') and barcoded nirKq1R (5'-GCC TCG ATC AGR TTG TGG TT -3') primers were modified as previously described
(Beszteri et al., 2010; Sogin et al., 2006). Each PCR reaction occurred in of 25 µL volumes
containing: 10 ng of purified DNA, 0.25 µM of each primer, 0.1 mM of each dNTP (Invitrogen,
Carlsbad, CA), 2.5 mM of MgCl2 (Invitrogen, Carlsbad, CA), and 2 U of Taq DNA polymerase
(Invitrogen, Carlsbad, CA), in 1x PCR Buffer (Invitrogen, Carlsbad, CA). Amplification
conditions for the PCR were as follows: 95ºC for 5 min, followed by 40 cycles of 94°C for 30 s,
60°C for 30 s, and 72°C for 30 s, and a final extension at 72°C for 7 min. PCR amplification
was performed using a PTC-200 Peltier Thermal Cycler (MJ Research, Waltham, MA). PCR
products were gel purified using the Wizard® SV Gel and PCR Clean-Up System (Promega,
Madison, WI) and sequenced on a Roche 454 Junior Titanium according to the manufactures
instructions. (454 Sequencing 2010 Manual (May)).
The nirK sequences were filtered in mothur v. 1.25.0 (Schloss et al., 2009) to remove
barcode sequences, screen out low quality sequences (qaverage=25) and trim primers from the
sequences. For this nirK region, aligned sequences have no gaps, so sequences not containing
the expected 126 base pairs were removed as likely due to sequencing error, followed by
clustering. Further community-level analysis of nirK sequence was performed in R (R Core

97

Team, 2016) using the VEGAN package (Oksanen et al., 2013). MEGA (Tamura et al., 2013)
was used to assess nirK sequence clustering against reference sequences of ammonia oxidizing
bacteria (AOB).

Statistical approach
Data for all soil analyses were examined for normality and heterogeneity of variance
assumptions following Zuur et al. (2010). Data deviating from normality were log-transformed
prior to analysis, and homogeneity of variance checked with Levine’s test. Analysis of variance
was calculated in R (The R Foundation for Statistical Consulting, Vienna, Austria) using the
lme4 package for linear, mixed-effect models (Bates et al., 2016), with time point nested within
block as a random effect and crop rotation as a fixed effect. On ten gas sampling dates, measures
were also taken for soil enzyme, mineral N, PMN and PMC. Along with soil moisture and
presence of legume cover crop these variables were individually related to N2O flux and then
used to generate mixed model comparisons using an iterative approach and assessing goodness
of fit with the Akaike Information Criterion (AIC). All results are reported as statistically
significant at α = 0.05, unless otherwise discussed.

RESULTS
Soil C and N
Among all treatments, mineral N was highest in CSW1cov, with a mean NH4+ of 2.5 µg
NH4+-N g-1 , which was significantly higher than all other treatments (P<0.05), as was NO3- with
a mean of 3.4 µg NO3--N g-1 (Table 4.2). Treatment cS had the second highest N values, with 2.2
µg NH4+-N g-1 (equal to or higher than all other treatments) and 2.7 µg NO3--N g-1 which was

98

nearly twice the NO3- of other non-covers. The two cover treatment, CSW2cov, also had higher
NO3- than non-covers (P<0.05) at 2.2 µg NO3--N g-1. The Fspring treatment had similar or lower
NH4+ and NO3- concentrations than the non-covers (except for cS). By contrast, the Fspring
plots and the cover cropped rotations had mean PMN between 8.6 and 9.0 µg NH4+-N g-1, which
was approximately twice the concentration of all other (non-cover) treatments (P<0.01). A
similar contrast occurred in soil enzyme activities for NAG and TAP, which had significantly
higher potentials in cover and fallow treatments (Table 4.2). Underlying these N values, the total
SON was significantly higher in CSW1cov, 0.91 g N kg soil-1, and CSW2cov, 0.93 g N kg soil-1,
than all other treatments with cC having the lowest SON (0.73 g N kg soil-1), (Supplemental
Table 4.1). Labile C as POXC was also significantly higher in CSW1cov 355 mg POX-C Kg-1
(P<0.05), followed by CSW2cov and Fspring. Similar to PMN, PMC was significantly higher in
both cover treatments and Fspring (P<0.01) when compared to all other non-cover rotations
(Table 4.2).

Nitrous oxide
In 2010, mean N2O production in the rotations with cover crops was 3-4 times greater in
magnitude than all other treatments (Table 4.3) with CSW2cov having the highest mean flux at
13.2 g N2O-N ha-1 day-1. Time points with the highest fluxes disproportionately influenced this
difference between treatments though median flux rates across cropped rotations still ranged
from 7.2 N2O-N ha-1 day-1 in CSW2cov to 2.3 N2O-N ha-1 day-1 in cC. Fspring had the lowest
median flux at 1.2 N2O-N ha-1 day-1 (Table 4.3). At least one of the cover cropped treatments
was higher in N2O flux at each measured time point except in November (Supplemental Table
4.5), with the strongest differences between treatments in the early part of the growing season

99

(Figure 4.1). Across all time points the Fspring treatment had N2O fluxes at or below the
magnitude of cropped treatments.
A similar pattern occurred in 2011, in which N2O fluxes from cover cropped treatments
diverged from all other rotations, especially with fluxes of larger magnitude in early spring
(Figure 4.1). Overall, the mean N2O flux was about a third as high in the second year as it was in
2010 across all treatments, with cover cropped treatments significantly higher than all other
treatments (P<0.05) and with highest mean flux in CSW1cov with 5.2 N2O-N ha-1 day-1 (Table
4.3). Similar flux patterns occurred among treatments as in the previous year, with cover
cropped rotations having 2-3 times greater magnitudes of flux as treatments without cover (Table
4.3). Fluxes were again higher under cover crops at every time point, with median values in
these rotations about double those of rotations without cover crops, and with the most significant
differences across all rotations occurring in the spring and early summer measurements
(Supplemental Table 4.5).

DEA
The denitrification potential and the relative rate of N2O production differed across
treatments based on results from DEA (Figure 4.2). In the anaerobic assay, N2O production had
high sample-to-sample variability and was not significantly different across treatments
(P=0.1033), although cover cropped treatments tended to have higher mean fluxes, from 197 200 ng N2O g-1 hr-1, compared to cropped treatments without cover, from 122 - 159 ng N2O g-1
hr-1. However, when N2 production was inhibited a clear and significant pattern emerged
(P=0.0042) with cover crop treatments denitrifying about twice as much as treatments without
cover, with CSW1cov and CSW2cov having the highest denitrification (N2O + N2) rates: 368

100

and 409 ng N2O g-1 hr-1 respectively. This meant that the relative rate of N2O production,
[ΔN2O/Δ (N2O + N2)], or rN2O was significantly lower in treatments with cover crops
(P=0.0135), with a ratio between 0.51 - 0.59 compared to treatments without a history of cover
which fell between 0.70 - 0.83 (Supplemental Table 4.2).

nirK PCR amplicons
Amplified nirK sequences differed significantly by rotation (Bray-Curtis, perMANOVA
constrained by block, p < 0.001), with nirK cover cropped rotations distinct from no-cover
rotations and from Fspring (Fig. 4.3). Clustering revealed two primary groups, one of which
contained sequences identified as AOBs. Although these were a small portion of the amplicons
(between 2 - 8 %), the percentage of sequences clustering with this group also differed by
treatment (P=0.012), with treatments with cover crops having a higher percentage than no-cover
treatments (Fig. 4.4).

DISCUSSION
Clear differences in N2O flux emerged across the cropping systems gradient over the
course of two years and appear to have been driven by changes in nutrient cycling and the soil
microbial community, which were induced largely by the presence of cover crops. I
hypothesized that treatments with a greater abundance of legumes in rotations would result in
higher soil NO3- and lead to increased N2O flux. While NO3- was significantly higher in
rotations with cover crops, available NO3- did not always lead to higher N2O. The presence of a
legume cover crop in rotation, however, was the best predictor of higher gas flux in part due to
higher potential mineralization of C and N. My hypothesis that cover crop rotations would shift

101

the phylogeny of nirK genes was supported and also suggested a change in pathway towards
nitrifier production of N2O under cover crops. Finally, I predicted that rotations with cover crops
would increase the efficiency of heterotrophic denitrification and DEA indicated that even
though in situ N2O flux under these rotations was higher, rotations without cover crops were
significantly less efficient at fully denitrifying. Along with influences from agronomic
management, higher N2O flux under rotations with cover crops appears to result from increased
nutrient cycling and altered microbial pathways as a result of including overwintering legumes in
rotation.
The presence of legumes in rotation increased mineral N, however while concentrations
of soil NH4+ and NO3- were as high or higher in cS compared to rotations with cover crops
(Table 4.1), cS had significantly lower N2O flux, suggesting that substrate availability alone did
not lead to the observed pattern in emissions. However, as contrasted to cS, the cover cropped
rotations had higher nutrient cycling potential with twice the value of PMN, PMC and enzyme
potentials associated with N mineralization (NAG+TAP), (Table 4.2), suggesting that
decomposition and release of NH4+ may be more indicative of available substrate than ephemeral
mineral N pools. However the fallow, Fspring, treatment also had PMN, PMC and enzyme
potentials equivalent to cover cropped treatments and consistently had the lowest N2O flux
(Table 4.3, Supplemental Table 4.4). This contrast in available N pools and N cycling potential
may be driven by the quality of plant detritus and degree of NH4+ mineralization versus
immobilization by microorganisms during decomposition (Robertson and Groffman, 2015).
Fspring had high mineralization potentials, but this treatment also had the highest soil C:N, 11.9,
while cS had the lowest at 9.9 (Supplemental Table 4.1). In addition, while all the treatments
were plowed in May of each year, the soil in cropped treatments continue to receive tillage

102

during the growing season for weed control, whereas Fspring was not be exposed to these
potential mineralization events (Grandy and Robertson, 2007) and instead successional plant
communities could scavenge available N. Thus, potentially greater soil N cycling, due in part to
more intense management, may more adequately indicate substrate availability in cover cropped
rotations that would lead to higher N2O flux.
Legume cover crops increase total soil N (Syswerda et al., 2011) and both cover cropped
rotations had significantly higher SON than all other treatments including Fspring (Supplemental
Table 4.1). While the soil C:N in cover cropped rotations was not different from rotations
without cover crops, spring plowing and incorporation of overwintering legumes introduces a
high quality plant tissue which can rapidly release N into soil (Parton et al., 2007). Among
individual variables, including available NO3-, the presence of legumes in rotation was the
strongest predictor of N2O flux (P=0.014), and when combined in a multivariate model, soil
moisture and PMN best fit flux patterns (Table 4.4). In spite of this finding, the highly stochastic
nature of soil N2O emissions (Nol et al., 2010) means that individual flux events may have
distinct underlying causes. In both years, peak flux events occurred in July (Figure 4.1). While
relatively higher soil moisture prevailed during these periods (favoring heterotrophic
denitrification), this is also a period of more intense management and tillage, approximately 6-8
weeks after cover crop incorporation. In this period the disproportionality higher emissions from
rotations with cover crops may be exacerbated by higher mineralization from cover crops
(Berthrong et al., 2013) stimulating both nitrification and denitrification. This confluence of
factors can lead to intense N2O flux events, or ‘hot moments’, (Molodovskaya et al., 2012),
which can dominate temporal analysis and present a challenge to modeling denitrification
(Groffman et al., 2009).

103

While short term environmental and management factors have an acute effect on soil
trace gas emissions (Gelfand et al., 2016), they essentially act on the activity of the underlying
microbial community (Cavigelli and Robertson, 2000). My data suggests that microbial
communities responsible for N2O flux are significantly different in rotations with cover crops.
Under anoxic conditions favoring heterotrophic denitrification, the relative rate of N2O to N2
production, rN2O, was between 16 - 39% higher in rotations without cover crops, suggesting that
in the field, incomplete denitrification could lead to high relative rates of N2O. This suggests
that denitrifiers in cover crop rotations are better adapted to coupling C respiration to complete
denitrification (as DEA is saturated with respect to labile C substrate). This may be a result of
greater availability of C, as POXC was higher in these rotations (Table 4.2), or better coupling of
available C with NO3- as PMC was significantly higher under cover crops and could supply more
substrate during ideal conditions for denitrification.
Functional differences in the denitrifier community in cover crop rotations was further
supported by the significant difference in the structure of nirK genes (Figure 4.3). While for all
soils the majority of these amplicons clustered with nirK from denitrifying bacteria, a portion
clustered with AOB lineages, and rotations with cover crops had between 2-10 times greater
relative abundance of AOB nirK (Figure 4.4). The diversity of genes underlying denitrification
pathways is broad and only partially described (Jones et al., 2008), such that the nirK genes
captured in this study represent only a fraction of the phylogenetic distribution of nitrite
reductases. However, along with a general shift in the denitrifier community, the increase in nirK
from AOB lineages suggests that nitrification may be an important source of N2O flux in
rotations with cover crops. Indeed, even in grassland systems fertilization can increase the
abundance of AOB leading to higher N2O from nitrification (Avrahami and Bohannan, 2009).

104

Increasingly, nitrification has been identified as a potent source of N2O in agricultural systems
(Zhu et al., 2013).
The structure of denitrifier communities broadly conforms to soil and land management
(Enwall et al., 2010), and their diversity influences ecosystem function (Hallin et al., 2009), such
that nir-type genes may be used in predictive models of denitrification (Powell et al., 2015).
While individual plant species can directly influence the structure of nirK genes and alter N2O
production (Abalos et al., 2014; Bremer et al., 2009), my study suggests that even in soil under
the same grain crop rotation, and without external inputs of mineral N, that cover crops alter the
structure and function of the denitrifier community.
A unique advantage to this study was the absence of external inputs or other differences
in management practices, such that the primary variable differing across all treatments was the
legacy of each specific crop rotation. The increase in N2O flux due to cover crops may be both
direct and indirect. Previously on this site, Tiemann et al. (2015) found that rotations with cover
crops had higher soil aggregation in spite of identical tillage regimes. Greater soil aggregation
alters microbial community structure as well as the oxygen environment in and outside of soil
aggregates (Paul et al., 2015), and has been found to influence soil N2O flux (Miller et al., 2009).
Soil structure may also interact with high substrate availability and soil moisture especially when
conditions align for high flux events.
Cover crops can improve field-scale N mass balances, particularly through reduced NO3leaching (Tonitto et al., 2006), and can immobilize NO3- that might otherwise available for
denitrification (McSwiney et al., 2010). A recent long term study of a corn-soybean rotation
with and without overwintering rye cover crop showed no difference in overall N2O flux, but
significant effects of reduced N2O in rye treatments when comparing paired flux events in

105

specific annual periods, especially in soybean (which was not fertilized as was the corn) (Parkin
et al., 2016). Apart from just N availability influenced directly by the rye, temporal changes in
N2O flux (Verma et al., 2006) may also be related to gene expression of denitrifiers in response
to seasonality (Németh et al., 2014) and management events (Wessen et al., 2011). In the current
study, the inclusion of rye cover in the most complex, CSW2cov, treatment did reduce NH4+ and
NO3- compared to CSW1cov with only a legume cover, even while PMN was the same (Table
4.2). However there was no difference in mean N2O flux (Table 4.3), and no clear trend in
paired flux events between rotations with cover, especially when considering peak ‘hot moment’
events. Still, with few time points in this study, especially in key seasonal periods, the effect of
reduced NO3- availability in treatments with rye may have some mitigating effect on N2O
emissions. Higher available N, likely through increased mineralization, along with incorporation
of legume residues may have an overriding influence on N2O (Mitchell et al., 2013) similar to
other low input row crop experiments - with cover cropped systems that showed increased N2O
emissions even outside of tillage or litter incorporation events (Peyrard et al., 2016).
While the scope of N2O measurements in this study did reveal consistent patterns across
treatments, and the potential for extremely high flux events under cover crops, the magnitude of
N2O rates were similar to those measured in conventionally managed and low-input row crop
systems at this experimental site (Gelfand et al., 2016; Oates et al., 2016). Interestingly adjacent,
conventionally managed row crop treatments at this site appeared to have the highest proportion
of AOB nirK when compared to the gradient in the current study (Supplemental 4.3). This may
indicate that differences in N use between these types of cropping systems may control potential
sources of N2O. While the presence of cover crops in this study led to more complete
denitrification than rotations without cover, knowledge of the microbial pathways leading to N2O

106

flux in rotations with cover crops would directly impact further strategies to mitigate N2O
emissions. Among numerous management approaches, greater mineralization and nitrification
might dictate greater emphasis on synchronizing N release and plant uptake (Grandy et al., 2012)
while increased denitrification might require careful timing of N immobilization (McSwiney et
al., 2010). Increasing N use efficiency in agroecosystems remains an overarching challenge to
reduce leaching and N2O emissions (Robertson and Vitousek, 2009; Snyder et al., 2014). This
study indicates that while cover crops increase efficiency in denitrification, they also pose a
unique set of challenges to reduce N2O emissions in low-input agroecosystems.

ACKNOWLEDGEMENTS
Support for this research was also provided by U.S. Department of Agriculture Soil Processes
Program, grant # 2009-65107-05961, and the National Science Foundation Long-term Ecological
Research Program (DEB 1027253) at the Kellogg Biological Station.

107

APPENDIX

108

Table 4.1 Crop rotations sampled on the KBS Biodiversity Gradient Experiment

ID
1
2
3
4
5
6

Treatment
cC
cS
CS
CSW
CSW1cov
CSW2cov

Rotation
corn
soy
corn + soy
corn + soy + wheat
corn + soy + wheat + clover
corn + soy + wheat + clover + rye

Species / yr.
1
1
1
1-2
2-3
3-4

Species / rot.
1
1
2
3
4
5

1
2
3
4
5

6

Month M J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D J F MAM J J A S O N D
C

S

W

red clover

109

crimson clover

rye

Table 4.2 Potentially mineralizable N (PMN), N-acetyl-glucosaminidase + tyrosine amino
peptidase (NAG+TAP), inorganic N (NH4+-N and NO3--N), potentially mineralizable C (PMC)
and permanganate oxidizable (POX) C on the biodiversity gradient treatments (see table 4.1 for
treatment acronyms) for the two study years. Values are means (standard error), n=4 replicate
blocks. Significance of treatment effect: * P ≤ 0.05, ** P = 0.001 to 0.01.

Rotation

PMN
µg NH4+-N g-1

cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

4.7(0.4)
3.8(0.3)
4.7(0.4)
5.1(0.4)
8.6(0.5)**
9.0(0.6)**
8.6(0.6)**

NAG+TAP
nmol g-1 h-1
31.8(2.0)
32.4(2.2)
31.0(2.0)
34.6(2.6)
45.7(3.2)*
51.7(3.7)**
46.8(4.0)*

NH4+-N

NO3--N

PMC

POXC

µg NH4+-N g-1

µg NO3--N g-1

µg CO2-C g-1

mg POX-C Kg-1

2.2(0.3)
1.6(0.2)
1.5(0.2)
2.2(0.3)
2.5(0.3)*
1.7(0.1)
1.9(0.2)

2.7(0.2)*
1.5(0.1)
1.7(0.2)
1.8(0.2)
3.4(0.4)**
2.2(0.2)*
1.1(0.2)

32.0(2.5)
29.4(2.3)
35.3(3.1)
34.7(3.2)
51.5(4.0)**
49.3(3.9)**
50.4(4.6)**

276(20)
263(16)
277(16)
278(14)
355(15)*
323(13)
301(19)

110

Table 4.3 In situ nitrous oxide fluxes in the biodiversity gradient treatments for the two study
years. Median, maximum (Max) and mean (standard error) for n=4 replicate blocks, followed by
letters denoting significant differences between treatments. See table 4.1 for treatment acronyms.

Year

Rotation

Median

2010

cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

4.5
2.3
3.1
4.3
3.6
7.2
1.2

2011

cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

1.5
1.0
1.1
0.8
2.3
2.6
1.7

111

Max
Mean (se)
g N2O-N ha-1 day-1
10.7
4.6 (0.8)ab
11.2
4.2 (0.8)ab
11.8
3.5 (0.6)a
26.5
4.4 (0.9)ab
41.1
10.5 (2.5)bc
58.6
13.2 (3.4)c
9.5
2.5 (0.6)a
3.0
10.5
7.0
7.7
34.0
29.4
4.1

1.3 (0.2)a
2.1 (0.7)a
1.5 (0.3)a
1.6 (0.4)a
5.2 (1.7)b
4.6 (1.2)b
2.0 (0.8)a

Table 4.4 Model reduction to best fit with N2O as response variable and predictor variables
presence of cover (0,1,2), moisture content, nitrate (NO3-), ammonium (NH4+), potentially
mineralizable N (PMN) and N-acetyl-glucosaminidase + tyrosine amino peptidase (NAG+TAP).
Random variables in the equation are treatment block and time point of flux measurement
(block:time point).
+

Model
N2O ~ cover + moisture + NO3 + NH4 + PMN + NAG+TAP + (1 | block:time point)

AIC
269.17

N2O ~ cover + moisture + NO3- + PMN + NAG+TAP + (1 | block:time point)

262.68

-

-

N2O ~ cover + moisture + NO3 + PMN + (1 | block:time point)

249.51

N2O ~ cover + moisture + PMN + (1 | block:time point)

244.18

112

θg (g H2O g soil -1)
θg (g H2O g soil -1)

--- θg

Figure 4.1 In situ N2O flux and gravimetric moisture (θg) for KBS biodiversity gradient for two
years. Time points are means with standard error for N2O, n=4, and means across all treatments
for moisture. Asterisks (*) show significant differences (P ≤ 0.05) by treatments at a specific
time point. See table 4.1 for treatment acronyms.

113

Figure 4.2 Denitrification enzyme assay (DEA) on the KBS biodiversity gradient. Bars
represent the mean rN2O or the relative rate of nitrous oxide flux per total denitrification
(including estimated N2) with standard error, n=4. See table 4.1 for treatment acronyms.

114

cS
cC

Figure 4.3 Non-metric multidimensional scaling (NMDS) plot of nitrite reductase (nirK) gene
amplicons from crop rotations, (n=3) on the biodiversity gradient experiment (see table 4.1 for
treatment acronym). Blue ellipse encircles treatments with cover crops. Amplicon count data
were transformed using Bray-Curtis dissimilarity matrix. Stress was 0.12349.

115

Figure 4.4 Proportion of nitrite reductase (nirK) gene amplicons from AOBs for crop rotations
on the biodiversity gradient experiment (n=3). See table 4.1 for treatment acronyms.

116

Supplemental Table 4.1 Soil organic C (SOC) and N (SON), and their ratio (C:N), in the
biodiversity gradient treatments. Values are means (standard error), n=4 replicate blocks.
Significance of treatment effect: * P ≤ 0.05. See table 4.1 for treatment acronym.

Rotation
cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

SOC
g C kg soil-1
7.93(0.96)
7.56(0.61)
7.95(0.96)
7.93(0.60)
9.30(0.35)*
9.38(0.38)*
10.42(1.35)

SON
g N kg soil-1
0.80(0.09)
0.73(0.06)
0.76(0.12)
0.77(0.07)
0.91(0.05)*
0.93(0.06)*
0.86(0.07)

117

C:N
9.9(0.13)
10.5(0.46)
10.8(0.58)
10.5(0.54)
10.3(0.45)
10.2(0.38)
11.9(0.81)

Supplemental Table 4.2 Denitrification enzyme potential assay (DEA) showing N2O flux, total
denitrification (N2O + N2) and relative rates of N2O production (rN2O) on rotations in the
biodiversity gradient experiment. Values are means (standard error), n=4 replicate blocks, with
overall treatment significance above values and letters denoting significant difference between
rotational treatments at P ≤ 0.05. See table 4.1 for treatment acronym.

Rotation
cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

N2O
ng N2O g-1-hr-1
(P=0.1033)
159 (19)
122 (28)
117 (9)
130 (23)
197 (39)
200 (16)
177 (25)

N2O + N2
ng N2O g-1-hr-1
(P=0.0042)
193 (8)ab
151 (34)a
164 (25)a
191 (32)ab
368 (110)bc
409 (30)c
167 (67)ab

rN2O
N2O/(N2O + N2)
(P=0.0135)
0.83 (0.1)b
0.81 (0.0)b
0.75 (0.1)ab
0.70 (0.1)ab
0.59 (0.1)ab
0.51 (0.1)a
1.82 (0.9) ¥

¥ Values for Fpring were left out of rN2O test for significance and treatment
comparisons.

118

Supplemental Table 4.3 Total number of nirK sequence reads, and within that pool the number
of sequences that clustered with known sequences of nirK from ammonia oxidizing bacteria
(AOBs). Treatments are from the biodiversity cropping experiment followed by a simultaneous
comparison with sequence from a nearby deciduous forest (DF) and conventional row crop
system (T1), for the years 2008 and 2009. See table 4.1 for treatment acronym.
Treatment
cS
cS
cS
cC
cC
cC
CS
CS
CS
CSW
CSW
CSW
CSW1cov
CSW1cov
CSW1cov
CSW2cov
CSW2cov
CSW2cov
Fspring
Fspring
Fspring
DF - 2008
DF - 2008
DF - 2009
DF - 2009
T1 - 2008
T1 - 2008
T1 - 2009
T1 - 2009

Field replicate
treatment #
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
1
2
1
2
1
2

Total # of nirK
sequence reads
440
2740
3938
1527
1682
1713
1244
1362
228
258
1647
241
623
919
2566
1548
611
513
2033
1804
1856
2991
1817
1934
1038
2121
883
786
2440

119

# of nirK sequence
clustering with AOB
10
103
152
36
67
10
22
51
4
1
36
4
49
112
105
39
41
19
18
19
38
28
2
7
0
431
235
67
286

Supplemental Table 4.4 In situ nitrous oxide fluxes in the biodiversity gradient treatments for individual time points over two years
with mean (standard error) for n=4 replicate blocks, followed by letters denoting significant differences between treatments. See table
4.1 for treatment acronym.

2010

g N2O-N ha-1 day-1

Rotation
07/13/10
cS
4.9 (1.9)a
cC
4.1 (0.5)a
CS
3.4 (0.5)a
CSW
5.6 (1.3)a
CSW1cov 17.5 (5.7)b
CSW2cov 20.3 (3.8)b
Fspring
3.2 (1.0)

07/21/10
10.7 (3.3)a
11.2 (3.1)ab
11.8 (3.9)ab
26.5 (18.9)ab
41.1 (10.3)ab
58.6 (11.6)b
3.3 (0.6)

07/28/10
7.9 (2.0)
7.0 (3.1)
5.0 (1.2)
9.5 (5.3)
13.7 (6.3)
12.1 (5.1)
1.3 (0.4)

08/04/10
1.3 (0.3)a
1.9 (0.5)ab
2.3 (0.6)ab
2.5 (0.8)ab
3.6 (0.4)b
3.2 (0.3)ab
1.1 (0.4)

120

08/31/10
1.2 (0.5)
1.5 (0.4)
1.6 (0.3)
1.5 (0.2)
2.3 (0.4)
2.2 (0.2)
1.2 (0.0)

09/09/10
6.4 (1.6)
8.4 (1.0)
5.6 (1.1)
8.4 (1.5)
10.4 (3.3)
11.0 (2.3)
9.5 (6.2)

09/23/10
4.1 (1.1)
2.3 (0.3)
3.1 (0.6)
4.3 (1.7)
3.1 (0.8)
7.2 (3.6)
1.8 (0.8)

10/07/10
n.a.
0.3 (0.1)
0.2 (0.1)
0.4 (0.1)
0.3 (0.1)
0.5 (0.1)
0.4 (0.0)

11/18/10
0.4 (0.3)
0.7 (0.3)
0.5 (0.4)
0.3 (0.3)
0.4 (0.2)
0.6 (0.2)
0.0 (0.3)

Supplemental Table 4.4 (cont’d)

2011
Rotation
cS
cC
CS
CSW
CSW1cov
CSW2cov
Fspring

g N2O-N ha-1 day-1
04/29/11
2.2 (0.3)ab
0.8 (0.2)b
1.4 (0.5)ab
0.7 (0.4)b
1.3 (0.5)b
3.6 (0.8)a
1.8 (1.3)

05/17/11
0.0 (0.2)
0.3 (0.2)
0.4 (0.2)
0.6 (0.2)
0.4 (0.1)
0.7 (0.2)
n.a

05/24/11
1.1 (0.4)
1.2 (0.4)
0.6 (0.3)
0.9 (0.3)
0.7 (0.3)
0.6 (0.3)
3.3 (1.3)

06/21/11
1.7 (0.2)ab
2.6 (0.7)ab
1.4 (0.3)ab
1.0 (0.2)a
3.6 (0.4)b
3.3 (1.0)ab
1.7 (1.6)

121

06/29/11
1.4 (0.2)ab
1.1 (0.4)a
1.2 (0.2)ab
1.6 (0.6)ab
3.9 (0.7)b
3.7 (1.4)ab
0.7 (0.2)

07/13/11
3.0 (2.2)a
10.5 (6.3)ab
7.0 (1.3)ab
7.7 (2.3)ab
34.0(14.6)b
29.4 (6.5)b
-1.2 (5.2)

Supplemental Table 4.4 (cont’d)

07/20/11
1.6 (0.9)
4.8 (2.5)
0.6 (1.2)
0.4 (1.0)
2.5 (0.5)
2.0 (0.9)
-2.4 (1.0)

08/04/11
2.0 (0.5)
1.0 (0.8)
2.6 (0.9)
3.9 (2.2)
9.9 (4.2)
5.5 (0.7)
1.8 (1.9)

08/12/11
0.4 (0.3)
0.7 (0.3)
1.0 (0.3)
0.6 (0.3)
1.2 (0.4)
0.8 (0.2)
3.2 (4.9)

08/17/11
0.7 (0.3)
0.7 (0.2)
0.5 (0.2)
0.6 (0.2)
2.0 (0.5)
1.5 (0.6)
4.1 (6.4)

122

08/31/11
1.0 (0.1)
0.9 (0.1)
0.3 (0.2)
0.3 (0.0)
1.3 (0.1)
1.3 (0.2)
7.1 (2.9)

09/23/11
0.5 (0.1)
0.4 (0.3)
0.4 (0.3)
1.6 (0.2)
1.5 (0.2)
1.6 (0.2)
2.3 (0.4)

Supplemental Figure 4.1 Potentially mineralizable N (PMN) on the biodiversity gradient
treatments for the two study years. All points are means across blocks (n=4) with standard errors
for a corn year (left) and following soybean year (right). See table 4.1 for treatment acronym.

123

Supplemental Figure 4.2 Permanganate oxidizable (POX) C on the biodiversity gradient
treatments for the two study years. All points are means across blocks (n=4) with standard errors
for a corn year (left) and following soybean year (right). See table 4.1 for treatment acronym.

124

Supplemental Figure 4.3 Proportion of nitrite reductase (nirK) gene amplicons from AOBs in
crop rotations on the biodiversity gradient experiment (n=3), (see table 4.1 for treatment
acronym) shown with treatments from the KBS Main Cropping Experiments (T1, conventional
management for 2008 and 2009) and Deciduous Forest (DF).

125

REFERENCES

126

REFERENCES

Abalos, D., De Deyn, G.B., Kuyper, T.W., van Groenigen, J.W., 2014. Plant species identity surpasses
species richness as a key driver of N2O emissions from grassland. Glob Change Biol 20, 265–
275. doi:10.1111/gcb.12350
Avrahami, S., Bohannan, B.M., 2009. N2O emission rates in a California meadow soil are influenced by
fertilizer level, soil moisture and the community structure of ammonia-oxidizing bacteria. Global
Change Biology 15, 643–655. doi:10.1111/j.1365-2486.2008.01727.x
Berthrong, S.T., Buckley, D.H., Drinkwater, L.E., 2013. Agricultural Management and Labile Carbon
Additions Affect Soil Microbial Community Structure and Interact with Carbon and Nitrogen
Cycling. Microb Ecol 1–13. doi:10.1007/s00248-013-0225-0
Beszteri, B., Temperton, B., Frickenhaus, S., Giovannoni, S.J., 2010. Average genome size: a potential
source of bias in comparative metagenomics. The ISME journal 4, 1075–1077.
Blesh, J., Drinkwater, L.E., 2013. The Impact of Nitrogen Source and Crop Rotation on Nitrogen Mass
Balances in the Mississippi River Basin. Ecological Applications. doi:10.1890/12-0132.1
Bremer, C., Braker, G., Matthies, D., Beierkuhnlein, C., Conrad, R., 2009. Plant presence and species
combination, but not diversity, influence denitrifier activity and the composition of nirK‐type
denitrifier communities in grassland soil. FEMS microbiology ecology 70, 377–387.
Casciotti, K.L., Ward, B.B., 2001. Dissimilatory Nitrite Reductase Genes from Autotrophic AmmoniaOxidizing Bacteria. Applied and Environmental Microbiology 67, 2213–2221.
Cavigelli, M.A., Robertson, G.P., 2000. The Functional Significance of Denitrifier Community
Composition in a Terrestrial Ecosystem. Ecology 81, 1402–1414. doi:10.1890/00129658(2000)081[1402:TFSODC]2.0.CO;2
Doane, T.A., Horwáth, W.R., 2003. Spectrophotometric Determination of Nitrate with a Single Reagent.
Analytical Letters 36, 2713–2722. doi:10.1081/AL-120024647
Drinkwater, L., Snapp, S., 2007. Understanding and managing the rhizosphere in agroecosystems.
Elsevier: Burlington, MA, USA.
Drinkwater, L.E., Cambardella, C.A., Reeder, J.D., Rice, C.W., 1996. Potentially mineralizable nitrogen
as an indicator of biologically active soil nitrogen, in: Methods for Assessing Soil Quality. Soil
Science Society of America, Madison, WI, pp. 217–229.

127

Dybzinski, R., Fargione, J.E., Zak, D.R., Fornara, D., Tilman, D., 2008. Soil fertility increases with plant
species diversity in a long-term biodiversity experiment. Oecologia 158, 85–93.
doi:10.1007/s00442-008-1123-x
Enwall, K., Throbäck, I.N., Stenberg, M., Söderström, M., Hallin, S., 2010. Soil Resources Influence
Spatial Patterns of Denitrifying Communities at Scales Compatible with Land Management.
Applied and Environmental Microbiology 76, 2243–2250.
Franzluebbers, A.J., Haney, R.L., Honeycutt, C.W., Schomberg, H.H., Hons, F.M., 2000. Flush of carbon
dioxide following rewetting of dried soil relates to active organic pools. Soil Science Society of
America Journal 64, 613–623.
Galloway, J.N., Dentener, F.J., Capone, D.G., Boyer, E.W., Howarth, R.W., Seitzinger, S.P., Asner, G.P.,
Cleveland, C.C., Green, P.A., Holland, E.A., Karl, D.M., Michaels, A.F., Porter, J.H., Townsend,
A.R., Vöosmarty, C.J., 2004. Nitrogen Cycles: Past, Present, and Future. Biogeochemistry 70,
153-226–226.
Gelfand, I., Shcherbak, I., Millar, N., Kravchenko, A.N., Robertson, G.P., 2016. Long-term nitrous oxide
fluxes in annual and perennial agricultural and unmanaged ecosystems in the upper Midwest
USA. Glob Change Biol 22, 3594–3607. doi:10.1111/gcb.13426
Grandy, A.S., Robertson, G.P., 2007. Land-Use Intensity Effects on Soil Organic Carbon Accumulation
Rates and Mechanisms. Ecosystems 10, 59–74. doi:10.1007/s10021-006-9010-y
Groffman, P.M., Altabet, M.A., Böhlke, J.K., Butterbach-Bahl, K., David, M.B., Firestone, M.K., Giblin,
A.E., Kana, T.M., Nielsen, L.P., Voytek, M.A., 2006. Methods for measuring denitrification:
diverse approaches to a difficult problem. Ecological Applications 16, 2091–2122.
Groffman, P.M., Butterbach-Bahl, K., Fulweiler, R.W., Gold, A.J., Morse, J.L., Stander, E.K., Tague, C.,
Tonitto, C., Vidon, P., 2009. Challenges to incorporating spatially and temporally explicit
phenomena (hotspots and hot moments) in denitrification models. Biogeochemistry 93, 49–77.
doi:10.1007/s10533-008-9277-5
Groffman, P.M., Holland, E.A., Myrold, D.., Robertson, G.P., Xiaoming, Z., 1999. Denitrification, in:
Standard Soil Methods for Long-Term Ecological Research. Oxford University Press, New York.
Hallin, S., Jones, C.M., Schloter, M., Philippot, L., 2009. Relationship between N-cycling communities
and ecosystem functioning in a 50-year-old fertilization experiment. ISME J 3, 597–605.
Haney, R., Brinton, W., Evans, E., 2008. Soil CO2 respiration: Comparison of chemical titration, CO2
IRGA analysis and the Solvita gel system. Renewable Agriculture and Food Systems 23, 171.

128

Haney, R.L., Haney, E.B., 2010. Simple and rapid laboratory method for rewetting dry soil for
incubations. Communications in soil science and plant analysis 41, 1493–1501.
Henry, S., Baudoin, E., Lopez-Gutierrez, J.C., Martin-Laurent, F., Brauman, A., Philippot, L., 2004.
Quantification of denitrifying bacteria in soils by nirK gene targeted real-time PCR. Journal of
Microbiological Methods 59, 327–335. doi:doi: DOI: 10.1016/j.mimet.2004.07.002
Hoben, J.P., Gehl, R.J., Millar, N., Grace, P.R., Robertson, G.P., 2011. Nonlinear nitrous oxide (N2O)
response to nitrogen fertilizer in on-farm corn crops of the US Midwest. Global Change Biology
17, 1140–1152. doi:10.1111/j.1365-2486.2010.02349.x
IPCC, 2013. , The Physical Science Basis. Summary for Policymakers. United Nations.
Jones, C.M., Stres, B., Rosenquist, M., Hallin, S., 2008. Phylogenetic Analysis of Nitrite, Nitric Oxide,
and Nitrous Oxide Respiratory Enzymes Reveal a Complex Evolutionary History for
Denitrification. Mol Biol Evol 25, 1955–1966. doi:10.1093/molbev/msn146
Kool, D.M., Van Groenigen, J.W., Wrage, N., 2011. Chapter Six - Source Determination of Nitrous
Oxide Based on Nitrogen and Oxygen Isotope Tracing: Dealing with Oxygen Exchange, in:
Stein, M.G.K. and L.Y. (Ed.), Methods in Enzymology, Research on Nitrification and Related
Processes, Part B. Academic Press, pp. 139–160. doi:10.1016/B978-0-12-386489-5.00006-3
Leininger, S., Urich, T., Schloter, M., Schwark, L., Qi, J., Nicol, G.W., Prosser, J.I., Schuster, S.C.,
Schleper, C., 2006. Archaea predominate among ammonia-oxidizing prokaryotes in soils. Nature
442, 806–809. doi:10.1038/nature04983
Liebman, Davis, 2000. Integration of soil, crop and weed management in low-external-input farming
systems. Weed Research 40, 27–47. doi:10.1046/j.1365-3180.2000.00164.x
McSwiney, C.P., Snapp, S.S., Gentry, L.E., 2010. Use of N immobilization to tighten the N cycle in
conventional agroecosystems. Ecological Applications 20, 648–662. doi:10.1890/09-0077.1
Miller, M.N., Zebarth, B.J., Dandie, C.E., Burton, D.L., Goyer, C., Trevors, J.T., 2009. Denitrifier
Community Dynamics in Soil Aggregates under Permanent Grassland and Arable Cropping
Systems. Soil Science Society of America Journal 73, 1843–1851. doi:10.2136/sssaj2008.0357
Molodovskaya, M., Singurindy, O., Richards, B.K., Warland, J., Johnson, M.S., Steenhuis, T.S., 2012.
Temporal Variability of Nitrous Oxide from Fertilized Croplands: Hot Moment Analysis. Soil
Science Society of America Journal 76, 1728–1740. doi:10.2136/sssaj2012.0039

129

Németh, D.D., Wagner-Riddle, C., Dunfield, K.E., 2014. Abundance and gene expression in nitrifier and
denitrifier communities associated with a field scale spring thaw N2O flux event. Soil Biology
and Biochemistry 73, 1–9. doi:10.1016/j.soilbio.2014.02.007
Nol, L., Heuvelink, G.B.M., Veldkamp, A., de Vries, W., Kros, J., 2010. Uncertainty propagation
analysis of an N2O emission model at the plot and landscape scale. Geoderma 159, 9–23.
doi:10.1016/j.geoderma.2010.06.009
Oates, L.G., Duncan, D.S., Gelfand, I., Millar, N., Robertson, G.P., Jackson, R.D., 2016. Nitrous oxide
emissions during establishment of eight alternative cellulosic bioenergy cropping systems in the
North Central United States. GCB Bioenergy 8, 539–549. doi:10.1111/gcbb.12268
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L.,
Solymos, P., Henry, M., Stevens, H., others, 2013. Vegan: Community Ecology Package.
Parkin, T.B., Kaspar, T.C., Jaynes, D.B., Moorman, T.B., 2016. Rye Cover Crop Effects on Direct and
Indirect Nitrous Oxide Emissions. Soil Science Society of America Journal 80, 1551–1559.
doi:10.2136/sssaj2016.04.0120
Parton, W., Silver, W.L., Burke, I.C., Grassens, L., Harmon, M.E., Currie, W.S., King, J.Y., Adair, E.C.,
Brandt, L.A., Hart, S.C., Fasth, B., 2007. Global-Scale Similarities in Nitrogen Release Patterns
During Long-Term Decomposition. Science 315, 361–364. doi:10.1126/science.1134853
Paul, E.A., Kravchenko, A., Grandy, A.S., Morris, S., 2015. Soil organic matter dynamics: Controls and
management for sustainable ecosystem functioning. Oxford University Press, New York, New
York, USA.
Peyrard, C., Mary, B., Perrin, P., Véricel, G., Gréhan, E., Justes, E., Léonard, J., 2016. N2O emissions of
low input cropping systems as affected by legume and cover crops use. Agriculture, Ecosystems
& Environment 224, 145–156. doi:10.1016/j.agee.2016.03.028
Powell, J.R., Welsh, A., Hallin, S., 2015. Microbial functional diversity enhances predictive models
linking environmental parameters to ecosystem properties. Ecology 96, 1985–1993.
doi:10.1890/14-1127.1
R Core Team, 2016. R: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing, Vienna, Austria.
Robertson, G.P., Groffman, P.M., 2015. Nitrogen Transformations, in: Soil Microbiology, Ecology and
Biochemistry (Fourth Edition). Academic Press, Boston, pp. 421–446.

130

Robertson, G.P., Hamilton, S.K., 2015. Long-term ecological research in agricultural landscapes at the
Kellogg Biological Station LTER site: conceptual and experimental framework, in: The Ecology
of Agricultural Landscapes: Long-Term Research on the Path to Sustainability. Oxford
University Press, New York, New York, USA, pp. 1–32.
Robertson, G.P., Paul, E.A., Harwood, R.R., 2000. Greenhouse Gases in Intensive Agriculture:
Contributions of Individual Gases to the Radiative Forcing of the Atmosphere. Science 289,
1922–1925. doi:10.1126/science.289.5486.1922
Robertson, P.G., Vitousek, P.M., 2009. Nitrogen in Agriculture: Balancing the Cost of an Essential
Resource. Annual Review of Environment and Resources 34, 97–125.
doi:10.1146/annurev.environ.032108.105046
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A.,
Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Horn, D.J.V.,
Weber, C.F., 2009. Introducing mothur: Open-Source, Platform-Independent, CommunitySupported Software for Describing and Comparing Microbial Communities. Appl. Environ.
Microbiol. 75, 7537–7541. doi:10.1128/AEM.01541-09
Shcherbak, I., Millar, N., Robertson, G.P., 2014. Global metaanalysis of the nonlinear response of soil
nitrous oxide (N2O) emissions to fertilizer nitrogen. PNAS 111, 9199–9204.
doi:10.1073/pnas.1322434111
Shoun, H., Kim, D.-H., Uchiyama, H., Sugiyama, J., 1992. Denitrification by fungi. FEMS Microbiology
Letters 94, 277–281.
Sinsabaugh, R.., Reynolds, H., Long, T.., 2000. Rapid assay for amidohydrolase (urease) activity in
environmental samples. Soil Biology and Biochemistry 32, 2095–2097. doi:10.1016/S00380717(00)00102-4
Smith, D.R., Hernandez-Ramirez, G., Armstrong, S.D., Bucholtz, D.L., Stott, D.E., 2011. Fertilizer and
Tillage Management Impacts on Non-Carbon-Dioxide Greenhouse Gas Emissions. Soil Science
Society of America Journal 75, 1070–1082. doi:10.2136/sssaj2009.0354
Sogin, M.L., Morrison, H.G., Huber, J.A., Welch, D.M., Huse, S.M., Neal, P.R., Arrieta, J.M., Herndl,
G.J., 2006. Microbial diversity in the deep sea and the underexplored “rare biosphere.” PNAS
103, 12115–12120. doi:10.1073/pnas.0605127103
Stehfest, E., Bouwman, L., 2006. N2O and NO emission from agricultural fields and soils under natural
vegetation: summarizing available measurement data and modeling of global annual emissions.
Nutr Cycl Agroecosyst 74, 207–228. doi:10.1007/s10705-006-9000-7

131

Syswerda, S.P., Corbin, A.T., Mokma, D.L., Kravchenko, A.N., Robertson, G.P., 2011. Agricultural
Management and Soil Carbon Storage in Surface vs. Deep Layers. Soil Science Society of
America Journal 75, 92–101. doi:10.2136/sssaj2009.0414
Tamura, K., Stecher, G., Peterson, D., Filipski, A., Kumar, S., 2013. MEGA6: molecular evolutionary
genetics analysis version 6.0. Molecular biology and evolution 30, 2725–2729.
Tiemann, L.K., Grandy, A.S., Atkinson, E.E., Marin-Spiotta, E., McDaniel, M.D., 2015. Crop rotational
diversity enhances belowground communities and functions in an agroecosystem. Ecol Lett 18,
761–771. doi:10.1111/ele.12453
Tilman, D., Reich, P.B., Knops, J., Wedin, D., Mielke, T., Lehman, C., 2001. Diversity and Productivity
in a Long-Term Grassland Experiment. Science 294, 843–845. doi:10.1126/science.1060391
Tonitto, C., David, M.B., Drinkwater, L.E., 2006. Replacing bare fallows with cover crops in fertilizerintensive cropping systems: A meta-analysis of crop yield and N dynamics. Agriculture,
Ecosystems & Environment 112, 58–72. doi:10.1016/j.agee.2005.07.003
Vandermeer, J., van Noordwijk, M., Anderson, J., Ong, C., Perfecto, I., 1998. Global change and multispecies agroecosystems: Concepts and issues. Agriculture, Ecosystems & Environment 67, 1–22.
doi:10.1016/S0167-8809(97)00150-3
Verma, A., Tyagi, L., Yadav, S., Singh, S.N., 2006. Temporal changes in N2O efflux from cropped and
fallow agricultural fields. Agriculture, Ecosystems & Environment 116, 209–215.
doi:10.1016/j.agee.2006.02.005
Weil, R.R., Islam, K.R., Stine, M.A., Gruver, J.B., Samson-Liebig, S.E., 2003. Estimating active carbon
for soil quality assessment: A simplified method for laboratory and field use. American Journal of
Alternative Agriculture 18, 3–17. doi:10.1079/AJAA200228
Wessen, E., Soderstrom, M., Stenberg, M., Bru, D., Hellman, M., Welsh, A., Thomsen, F., Klemedtson,
L., Philippot, L., Hallin, S., 2011. Spatial distribution of ammonia-oxidizing bacteria and archaea
across a 44-hectare farm related to ecosystem functioning. ISME J 5, 1213–1225.
Zak, D.R., Holmes, W.E., White, D.C., Peacock, A.D., Tilman, D., 2003. Plant Diversity, Soil Microbial
Communities, and Ecosystem Function: Are There Any Links? Ecology 84, 2042–2050.
doi:10.1890/02-0433
Zhu, X., Burger, M., Doane, T.A., Horwath, W.R., 2013. Ammonia oxidation pathways and nitrifier
denitrification are significant sources of N2O and NO under low oxygen availability. PNAS 110,
6328–6333. doi:10.1073/pnas.1219993110

132

Zumft, W.G., Cárdenas, J., 1979. The inorganic biochemistry of nitrogen bioenergetic processes.
Naturwissenschaften 66, 81–88. doi:10.1007/BF00373498
Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common statistical
problems. Methods in Ecology and Evolution 1, 3–14. doi:10.1111/j.2041-210X.2009.00001.x

133

CHAPTER 5: ALIGNING SOIL HEALTH TESTING AND FARMER KNOWLEDGE
FOR IMPROVED ON-FARM SOIL MANAGEMENT

ABSTRACT
The advent of soil health as a framework for management of row crops in the U.S.
Midwest has major implications for on-farm agronomic practices. A soil health approach posits
a new set of priorities for assessing soil quality compared to traditional forms of soil testing and
resultant management recommendations. Here I link farmer knowledge with results from soil
testing on their farms, including an emerging approach to test soil health. On grain farms in three
Michigan counties with contrasting soil types, I compare results from traditional field crops soil
tests with results from soil health tests. For testing, farmers were asked to select their “Best” and
“Worst” row crop fields, a “Choice” field of their own interest, and a non-row crop field, such as
a wood lot. I held interviews to discuss how soil test results aligned with farmers’ own
experience of individual fields, how soil testing informs their approach to soil management, and
how new approaches to soil testing might alter their practices. Results from traditional soil tests
showed no significant differences between comparisons of paired cropped fields for parameters
such as pH, phosphorus, potassium or nitrogen. In contrast, soil health test results revealed a
large range of differences between field types, in which parameters such as labile soil organic
matter and soil aggregation were significantly higher in ‘Best” compared to “Worst” fields. Soil
health test results also better aligned with how farmers described individual field performance,
such as how soils drained or ‘worked’ physically. While farmers discussed using traditional soil
tests primarily for informing fertilizer application rates and monitoring soil pH, soil health test
results elicited distinctly broader discussions of soil management considerations, including
tillage and crop rotation practices. The results showed that soil health tests can reveal important

134

differences across farmer fields and suggest both strong opportunities for and notable
deficiencies in aligning farmer field knowledge and soil testing to influence soil management
decisions that can achieve both socioeconomic and environmental objectives on farms.

INTRODUCTION
The environmental and social costs of industrial-scale agricultural production in the U.S.
have led to calls for a transformation to more ecologically-based approaches to management
(Drinkwater and Snapp, 2007; Robertson et al., 2014; Schipanski et al., 2016; Shennan, 2008).
Soil fertility management is a key dimension of agriculture that has far-reaching impacts for both
farm productivity and potential environmental degradation. For example, widespread use of
synthetic fertilizers is linked to greenhouse gas emissions and to more than 400 hypoxic zones
worldwide (Diaz and Rosenberg, 2008). The Great Lakes region, in particular, has experienced
recent severe water contamination and acute drinking water shortages primarily due to upstream
agricultural practices (Michalak et al., 2013). While a great deal of attention has focused on
opportunities to mitigate the downstream effects of agriculture by providing incentives for
farmers to practice alternate management strategies (Cherry et al., 2008), there remains a need to
understand how farmers think about and manage for soil fertility. Understanding this decisionmaking process is critical to inform how a broader range of practitioners could shift toward
ecologically-based soil management. Two key entry points directly tied to such a transformation
are: i) the quality of soil testing metrics used to assess and manage agricultural soils; and, ii) the
extent to which farmers rely on soil tests for management decisions, along with an understanding
of the broader frameworks that inform their soil management practices.
Promotion of ecosystem-based approaches to agricultural management has taken various
forms, especially in government programs designed to incentivize adoption of conservation

135

practices (NRCS-CEAP, 2011). However, altering landscape management to mitigate
downstream impacts from agriculture by coordinating with farmers to adapt to or confront
prevailing socio-ecological conditions poses considerable challenges (Allison and Hobbs, 2004;
Wolf and Allen, 1995), especially when soil fertility is so closely associated with crop
productivity. One potential leverage point for changing longstanding approaches to soil
management is the shift currently underway in soil testing—from a reductionist emphasis on
managing for individual inputs/deficits of key plant nutrients, towards a more holistic ecosystem
management approach, often termed soil health (Karlen et al., 2008).
Traditionally, soil testing has focused on the supply of plant macronutrients such as
nitrogen (N), phosphorus (P), and potassium (K), which Liebig’s law of the minimum –
formulated in the mid-nineteenth century – posits as key to limiting plant growth if insufficiently
supplied in soil. Along with manipulation of soil pH (to optimize nutrient availability to plants),
this framework of supplying plant nutrients via inputs to soil has largely guided soil testing and
fertility management up to the present (Havlin et al., 2005). Soil tests for field crops (e.g. corn)
return recommendations for altering soil pH, if needed, and for fertilizer application rates
(particularly for N) for achieving a target yield for a given crop.
Soil health (synonymous with soil quality) incorporates nutrient management into a
broader perspective defined as “the capacity of a soil to function, within ecosystem and land use
boundaries, to sustain productivity, maintain environmental quality, and promote plant and
animal health,” (Doran and Parkin, 1994). The multidimensional characteristics of soil health
include retaining and cycling nutrients, improved water infiltration and storage, suppression of
soil-borne pathogens, better stand establishment, and buffering against damage from drought
stress or heavy rain (Moebius-Clune et al., 2016), which are all dependent on agronomic

136

approaches such as tillage and crop rotation practices. Assessing soil health is most commonly
broken into three areas: chemical components such as soil pH and available nutrients, physical
components such as soil structure and water holding capacity, and biological components such as
the ability to alternately store or release soil carbon (C) and N. Most of the parameters in soil
health tests focus on economical or rapid measures of these soil ecosystem properties and
processes, and likewise, the recommendations that follow from soil health assessments are meant
to foster management practices that enhance these soil functions.
Traditional soil tests have generated decades of fine scale data, such as fertilizer rate
recommendations for specific crops in specific counties on specific soil types. In contrast, soil
health tests are still in the early stages of deployment and dissemination as information resources
for farmers. To support their effective implementation on a broad scale, they need to be
vigorously tested on farmer fields in a broad range of management and environmental contexts.
Even though many studies have demonstrated that soil health parameters are sensitive to changes
in soil management (Moebius-Clune et al., 2008), this may not ensure that soil health tests
capture soil quality differences on farm fields, or in distinct regions, in ways that correspond to
or are relevant to farmer experience. Many soil health testing parameters have proven effective
for assessing differences in soil quality in long-term experimental gradients on research sites
(Bhardwaj et al., 2011; Obade and Lal, 2016) but ultimately they must be able to reveal
differences on farmer fields on time-scales that can inform farm management decisions.
Compared to traditional soil tests that primarily recommend fertilizer and lime input rates, farmer
decisions based on soil health test output may extend into a broader suite of cropping system
practices, such as tillage, residue management, or crop rotation including use of cover crops.
Thus, it is critical that soil health tests provide accurate, field-specific data on which farmers can

137

rely for these decisions. Finally, even if soil health tests accurately capture soil biological,
physical, and chemical variability on farmer fields, ultimately how farmers conceptualize soil
quality will influence how they interpret and implement agronomic decisions based on test
results. While traditional soil tests directly link plant nutrient inputs to crop yield within a
growing season, addressing soil health problems – which may also severely impact crop
performance – may take careful planning over several seasons to produce a desired outcome. A
soil health management framework must therefore accurately reflect conditions on farmer fields,
effectively guide soil management decisions, and lead to positive agronomic performance for
farmers while minimizing impacts on surrounding ecosystems.
In this study I aimed to integrate technical and social dimensions to understand how to
adapt soil health testing to farms. Specifically our research addresses three overarching
objectives to inform soil health testing on Michigan farms: i) to evaluate how soil health tests
capture management variability on farmer fields, ii) to understand what information, including
soil testing, farmers consider when making soil management decisions, and iii) to assess how
farmers understand soil quality and respond to soil health test results in comparison to traditional
soil testing from their own fields. Using a targeted approach by asking MSU Extension staff to
identify farmers interested in soil testing, I conducted an in-depth, mixed-methods study in
partnership with 13 row crop farmers in three geographically distinct regions of Michigan.

METHODS
Experimental approach
MSU extension and staff from each region contacted farmers who would be willing to be
interviewed in exchange for soil testing on four distinct field types. I asked participant farmers

138

to identify four fields to sample for soil health testing based on their judgment of a ‘Best’ and
‘Worst’ field, an additional ‘Choice’ field, which they wished to test, and a non-row crop (NRC)
field, such as a woodlot or hay field. Through this approach I captured a wide range of field
variability as identified by growers on which to test and compare soil health parameters. It also
provided a participatory approach as a means to engage directly with farmers as decision-makers.
I collected soil samples from each of 52 fields across all farms, and measured managementsensitive parameters of soil health to examine how well these tests characterized soils on farm
fields, as compared to traditional field crop tests offered by Michigan State University (MSU). I
then conducted in-depth interviews with growers to evaluate how well soil health results
reflected the management histories for each field type, and to understand their own assessments
of soil quality and using soil testing, and how these inform their soil management decisions and
cropping practices. My approach presumes that managing for soil health on farms derives from
farmer knowledge of and practices on their own fields; from scientists’ capacity to accurately
assess and communicate soil health testing results; and, from feedbacks between these two types
of knowledge that may lead to emergence of a soil health management framework.

Field sampling
I sampled farms in three Michigan counties: Presque Isle (45°42’N, 83°81’W), Isabella
(43°60’N, 84°76’W), and Van Buren (42°21’N, 85°89’W), representing north, central and
southwestern regions, which have distinct climates and geography. I visited each farmer and
briefly discussed the history of the four differently managed fields they selected: ‘Best’, ‘Worst’
fields, NRC, and ‘Choice’. At each field I noted soil conditions and recorded crop management
observations, and followed the Cornell Soil Health Assessment approach (Moebius-Clune et al.,
2016) when taking soil samples. I selected five locations on each field representative of each

139

management designation, and used a shovel to made a small pit, a ‘slice’ of bulk soil from the
side of the pit, 15 cm depth, and approximately 4 cm thick and 9 cm wide. At each sampling
point in the field, two separate penetrometer (Soil Compaction Tester 6120, Spectrum
Technologies) readings were taken at 15 cm and 45 cm depths. All field samples were mixed
thoroughly in the field and a sub-sample (approximately 2 kg) was stored in a plastic bag, and
placed on ice until further processing. All samples were collected from May to June of 2014, on
thirteen farms for a total of 52 fields, and maintained at 4° C until processing.

Soil testing
As soon as possible after return from farmer fields, samples were sieved to < 8mm to
remove stones and a homogenized sub-sample was submitted to the MSU Plant, Soil and
Nutrient Laboratory (SPNL) for analysis of pH, total soil organic matter (SOM) by loss on
ignition, total Bray 1 phosphorus (P), potassium (K+), Calcium (Ca+2), Magnesium (Mg+) as well
as ammonium (NH4+) and nitrate (NO3-). We grouped these parameters into ‘chemical’ results
except for SOM. At the Kellogg Biological Station I carried out soil health analysis for soil
physical parameters (soil texture, aggregate stability, and available water capacity) and soil
biological factors (permanganate oxidizable carbon, potentially mineralizable carbon, potentially
mineralizable nitrogen, nitrification potential).
Soil texture was determined following Kettler et al. (2001) on soil sieved to < 2mm and
dried overnight at 60° C. A 14 g portion of this soil was placed into 50 ml Falcon tubes with 42
mL of 3% hexametaphosphate solution and tubes were placed on their side on a shaker at 120
rpm for two hours. Contents of each tube were poured through a 0.053mm sieve and thoroughly
washed with 600 mL deionized water, which was caught in a basin. The sand fraction on the
sieve was washed into a previously tarred drying can. Particles in the catch basin were

140

thoroughly re-suspended and allowed to settle form 4-6 hours, after which the clay particles in
suspension are decanted and settled silt particles are washed into another tarred drying can. All
cans were dried overnight at 105° C and contents weighed. Texture was calculated as: Sand % =
(oven dry sand mass / original sample mass) x 100%; Silt % = (oven dry silt mass / original
sample mass) x 100%; Clay % = 100 - (Sand % + Silt %).
Wet aggregate stability (AS) was determined following Moebius et al. (2007) from soil
dried to constant weight at 40° C (1-2 days in the oven) followed by isolation of aggregate size
fraction 0.25mm-2mm. Ten g of aggregates were spread evenly on a 0.25mm mesh, 125 mm
diameter sieve. The sieve was placed on a funnel containing a previously weighed filter paper
and the funnel with sieve was placed on a ring stand. Sieves were exposed to a rain simulator
(rate previously calibrated) for five minutes, after which any the material retained on the sieve
was thoroughly washed through and remaining particles (e.g. small stones) were washed off the
sieve surface into a drying can. The can and filter paper with slaked soil were dried for 1 day at
100° C oven, and AS was calculated as the percentage of soil retained on the sieve (difference
from what was not slaked onto the filter), and adjusting for the mass of un-sieved particles.
Available water capacity (AWC) was determined following Reynolds and Topp (2008),
with a portion of soil dried to 60° C was sieved to < 2 mm. Two 15 g portions were spread
evenly inside brass rings and placed on ceramic plates with known porosity, and saturated with
water. Plates were placed into high pressure chambers, one set to 10 kPa (to determine field
capacity) and the other set to 1500 kPa (permanent wilting point). After samples equilibrated
inside each chamber, soils were removed and weighed, dried at 105° C and then weighed again.
AWC was calculated as soil water loss between samples at 10 and 1500 kPa, and reported as g
water per g soil.

141

To determine labile carbon as permanganate oxidizable carbon (POXC), air-dried soil was
used according to Weil et al. (2003) . Duplicate 2.5 g samples of dried soil were mixed with
buffered 0.02 M KMnO4 solution in 50ml conical tubes, shaken at 120 rpm for two minutes and
allowed to settle for eight minutes. From this reaction, 0.5 mL of supernatant were diluted with
49.5 mL of deionized water. The degree of oxidation was measured colorimetrically at 550 nm
on a Fisher Scientific Thermo Multiskan microplate reader (Waltham, MA.) and standardized to
a series of known KMnO4 standards.
Air-dried soil was also used to determine potentially mineralizable carbon (PMC)
according to Franzluebbers et al. (2000). For each field, replicate ten g of soil were placed in
Mason jars and brought to 50% water-filled pore-space and incubated at 25o C. Following the
24-hour incubation, Mason jars were capped tightly and at time zero a CO2 reading was taken
immediately, by injecting 0.5 mL of headspace into a LI-COR LI-820 infrared gas absorption
analyzer (LI-COR Biosciences, Lincoln, NE). Three subsequent readings were taken over 90
minutes and a flux was calculated by regressing the change in CO2 against incubation period
(Robertson et al., 1999).
Potentially mineralizable N (PMN) was determined from fresh soil sieved to < 2mm and
incubated anaerobically following (Drinkwater et al., 1996). From duplicate 8 g soil samples,
NH4+ and NO3- were extracted with 1M KCl on a rotary shaker at 120 rpm for 1 hour. To two
additional 8 g replicates soil samples in conical tubes, 10 mL deionized water was added and
dinitrogen gas was used to replace tube headspace air and bubbled into the slurry for one minute
prior to sealing with butyl rubber stoppers. Sealed tubes were incubated at 25 °C for seven days.
After seven days the stoppers were removed, buffer was added to bring the slurry to 1M KCl,
and samples were shaken, filtered, and stored in the same manner as fresh samples.

142

Concentrations of NH4+ were determined colorimetrically at 630 nm (Sinsabaugh et al., 2000),
and NO3- at 540 nm following Doane and Horwath (2003). PMN was determined from the
concentration of NH4+ from incubated soil minus NH4+ from initial soil extraction.
Nitrification potential (NIT) was determined using fresh soil sieved < 2 mm and
following Norton and Stark (2011). Fifteen g soil was added to a solution with 1.0 mM NH4SO4
and 1mM phosphate (dilute mixture of 1.5 M KH2PO4 and 3.5 ml 0.2 M K2HPO4 in 1 L) in a
flask, loosely capped, and placed on a shake at 200 rpm. Over the next 24 hours, 1.5 mL
samples were taken from the slurry at four time points ( 2, 4, 18, 24 hours), placed in a
microfuge tube and spun down at 8000 x g for 8 minutes. Supernatant was decanted and stored
at -20° C for later analysis for NH4+ as described above. Potential was calculated by regressing
the change in NH4+ against incubation period time points.
Statistical approach
All soil data analyses were calculated in R (R Core Team, 2016). For all soil testing
parameters, paired t-tests were carried based on field type comparisons (e.g. Best compared to
Worst) across all farms.
Farmer interviews
After compiling all soil testing data, I held interviews, divided into three components,
with all participant growers. In the first component, I discussed the management history of each
field, including crop rotation, tillage, how the farmer had chosen specific management decisions
for each field, and why the farmer had assigned the given field to each category (i.e., ‘Best’ and
‘Worst’). In the second component, I discussed specific test results for all fields, grouping
results into physical, biological, and chemical categories. During the third component, I

143

facilitated an open discussion of soil test results and the factors that influenced management
decisions for specific fields. These open discussions were aimed at integrating soil test results
with farmer knowledge/experience to summarize how well the test metrics related to farmer
interpretation of soil management for each field type. All of the interviews were recorded, and
notes were written up within 24 hours of each interview. Recordings were transcribed and
analyzed for common themes related to each soil test parameter; different approaches to soil
management based on each of the field types; and, on the influence of soil testing on
management practices.
Using written notes and transcripts, the component interviews with each farmer were
analyzed based on three groupings. First, answers to core questions, which were asked of all
participants, were compiled and summarized into concise response summaries. Second, we
grouped farmer responses to the brief discussions of individual soil test results for each field type
by test result, and summarized findings for each soil test metric. Third was an analysis of open
discussions about testing interpretation and management for each grower on each field type.
These discussions also identified which results the farmers found most noteworthy/valuable, and
how they overlapped with their experience of management from each field type. For this third
section, farmers’ reactions to testing based on their experience of each field and its history of
management were noted and key quotes extracted.

RESULTS
Comparing soil tests parameters across field types
Comparing physical parameters across Best and Worst fields, both AS and AWC differed
significantly between paired fields on each farm (Table 5.2). Conversely, measures of surface
and sub-surface hardness did not differ between these field types. Surface and subsurface

144

compaction fluctuated with the status of the field on the day of sampling (Table 5.2). Different
fields had been tilled or prepared to differing degrees, or had variable amounts of residue or
cover which did not necessarily vary by the grower’s field type designations.
Three of the biological parameters PMN, POXC and NIT – did not show significant
differences between Best and Worst fields (Table 5.3). Percent SOM and PMC were both
significantly higher in the Best fields. None of the chemical data measured in the traditional
field crops test (pH, P, K+, Ca+2, Mg+, NH4+ and NO3-) showed any significant difference
between Best and Worst fields (Table 5.4).
The non-row crop (NRC) fields ranged greatly in their management and land-use
histories, from woodlot to hay field to buffer strip. However, all had been under some kind of
perennial cover for at least the previous five years (Table 5.5). In terms of the physical
parameters, only surface compaction was significantly higher on NRC (Table 5.2). Paired
comparisons of Best and NRC fields indicated that SOM, POXC and PMN were all significantly
higher on the Best field, and NIT was significantly lower on the NRC field (Table 5.3). Again,
none of the soil chemical parameters were different across these field type comparisons (Table
5.4). A similar set of results emerged when comparing Choice and NRC, in which OM, POXC,
PMN, and NIT were all more favorable for the latter field. The contrast between NRC and Worst
field was even greater, with the former having significantly higher AWC, OM, POXC, PMC and
PMN. The NRC fields were also higher in Ca+2 than the Worst field.

Farmer interviews on soil health test results
All farmers in the study conducted soil samples every 2-4 years for a given field (Table
5.6). Most of them use this information directly to determine their fertilizer and lime application
rates, rather than to provide indicators of soil quality. Two of the farmers also used soil tests

145

when considering controlled traffic (tractor passes) on fields, and two others used them to keep
track of excess P additions. In contrast, farmers primarily assessed soil quality as a physical
characteristic: how the soil ‘works’ (when planting and tilling), whether or not the surface crusts,
how a field drains, etc. Other indicators such as organic matter content, lack of soil-borne
disease, and the importance of longer crop rotations were also discussed by a minority of farmers
when asked to describe soil quality.
In describing how they make decisions about soil management practices, a majority of
farmer participants mentioned the role of social networks (e.g., family and neighbors), and the
importance of on-farm experimentation with different practices—either directly on their fields,
or observed on other’s fields. A minority of farmers also mentioned relying on information from
soil testing agencies, sales representatives, and on literature from extension agents, trade
magazines, or the World Wide Web as key knowledge resources.
Nearly all of the farmers designated the Best and Worst field types based on their
experience of yields from each field, as well as the contrasting predictability of management
outcomes between the two fields (Table 5.7). Best fields were more reliable, but also appeared
to receive more management attention (i.e., they were cared for more closely). Aside from yield,
farmers described Worst field characteristics in physical terms, as ‘working’ poorly or having
problems with too much or too little moisture. Finally, Choice fields were selected based on
yield concerns, but also due to more specific concerns, such as compaction or uncertainties over
prior management (Table 5.7).
Farmers responded differently to soil testing results depending on the parameter
measured for each field type. Those who responded to soil physical measures (e.g. AS, AWC)
did so with their explicit experience of a particular field, such as describing how the field

146

typically drains or ‘works’ (Table 5.9). For example, one grower responded to the contrasting
values for AWC on his Best and Worst field types by describing how the fields respond to rain.
He noted that in his Best field “ the water hits it and it absorbs it,” whereas for his Worst field,
“when water hits off this one it beats it, and then hits a table, and then just has to evaporate
off…then the sun bakes it and you got clay.”
For SOM, most farmers were aware of the SOM levels on each cropped field type, but
their interest in altering this value depended on the region. Referring to increasing SOM values
on his fields, one farmer in Isabella County – who farms a finer-textured soil – remarked, “I am
not going to live long enough to do that.” By contrast, in counties with coarser soils (Van Buren
and Presque Isle) farmer were more interested in how increase organic matter through changes in
management. A majority of farmers responded that they would like to be able to monitor shorterterm changes in soil C (such as PMC) if they could reliably do so. On most farms, C parameters
and physical characteristics showed clear contrasts across the fields of interest. Measures of
PMN and NIT did not, and farmers had fewer responses to these parameters. Finally, for soil
chemical parameters, farmers responded most strongly to values for pH and P across different
field types, as these are parameters they actively manage through soil testing.

DISCUSSION
Comparing soil testing approaches
Soil health tests reflected greater sensitivity to soil parameters across four different field
types than the traditional field crops test. Only total SOM, designated as a biological parameter
in this study, and assayed as part of the field crops test was significantly different between NRC
and all cropped fields (Best, Worst, Choice) as well as between Best and Worst. For the same
set of field comparisons, one or both soil health estimates of labile soil carbon, POXC and PMC,

147

also showed significant differences. By contrast, with the traditional field crops test, none of the
field comparisons revealed any significant difference in soil pH, NH4+ or NO3-. Neither were
other plant nutrients in the field crops test (P, K+, Mg+) significantly different between paired
field comparisons, except for Ca+2, which had a lower concentration in Worst and Choice fields.
Indeed, of the 52 fields measured for P, K+, Ca+2, and Mg+ (204 separate results), only ten results
rated below ‘Optimum’ concentration, as designated by the MSU Soil Testing Laboratory
(http://www.spnl.msu.edu/). The majority of these ratings were for K+, and included fields from
all four of the field types selected by participant growers.
Meanwhile, soil health test parameters captured differences in other managementsensitive indicators between field types. Soil aggregation and water holding capacity, positive
indicators of soil physical health, were higher in Best fields with AS and AWC significantly
greater compared to Worst fields. For indicators of soil N, mineralizable N and nitrification were
different in NRC fields, with PMN significantly higher in NRC fields when compared to all other
cropped field types, and NIT lower in NRC fields, both of which might be expected in an
uncropped field. However, among comparisons of cropped fields, soil health assays for labile
soil C and soil physical health showed clear differences between crop fields of interest to
farmers, especially when compared to traditional field crops tests.

How farmers make soil management decisions
The targeted approach of this survey meant that many of the farmer participants had been
identified by MSU Extension staff as careful and even exceptional managers for their region. In
terms of soil test metrics, pH was the only one that farmers actively managed on a longer time
scale. Farmers care about SOM, but indicated that they don’t generally manage for it. This is

148

likely because they receive the greatest financial incentives to intensify and manage for high
yields, using plant nutrient inputs, rather than to diversify crop rotations or adopt other ecological
practices that are likely to build SOM (even if that might ultimately feedback to increase
productivity with reduced input costs in the long term). By and large these farmers were
innovators and spoke in terms of having tried, or recognized the need to try, new approaches to
soil management. Most of them want to improve their cropping practices on specific fields and
were willing to make changes. That being understood, farmers knew that improving soil quality
can mean large changes in management, such as altered crop rotations, but have few guideposts
to point them in the right management direction.

Aligning soil quality with soil testing
Using farmer knowledge to sample a range of field types made it possible to assess how
well different soil testing parameters capture soil variability as characterized by farmers.
Traditional soil test parameters did not reflect the strong differences in productivity across these
field types, though farmers are already actively managing based on these testing parameters (e.g.
pH and P). Indeed, the farmers in this study were purposefully selected and agreed to participate
because of an expressed interest in soil testing. Soil quality, in contrast, is more intuitively
understood through farmers’ own experience and on-farm experimentation – they know it when
they see it and feel it. Farmers frequently described soil quality in physical terms for which soil
health parameters (AWC and AS) tended to overlap with their experience of managing specific
field types. Farmers also knew that managing to build organic matter improves soil quality and
expressed interest in knowing how soil health measures of labile soil C (POXC, PMC) could
give them greater management control over this resource.

149

Following a review of soil health results on their farm, we discussed how they approach
soil management across their fields to better understand how a soil health framework intersects
with farmer knowledge and practice. I identified three key themes that emerged from this
mixed-methods approach that combined knowledge from targeted soil testing with farmer
experience.

Emerging themes
I. Soil health metrics need to be solid and demonstrable.
Farmers are not sure how to address problems associated with lower soil health test
values. They believe the soil health numbers (e.g. the testing differences on contrasting field
types) but they need more specific guidance. For example, they want know how to link a
specific testing result to a change in practice, and to be able to trust and see that change.

II. Testers need to know how to use the test, and what its limits are.
Some structural constraints conflict with soil health testing. These may include crop
prices, or the time scales over which farmers are willing to consider management changes.
Educating farmers about soil health tests and their benefits will likely be insufficient to leverage
change if the recommendations are not likely to be adopted by farmers because of other
constraints. This could mean the need to develop targeted tests (e.g., perhaps starting with a full
assessment, with follow-up testing focusing on one or two critical metrics). Soil health tests may
be superior to traditional tests for informing more sustainable management, and, in the near-term
they are clearly important for on-going research on soil and cropping system management. In
the long-term, if changing socioeconomic conditions favor a greater number of sustainable
farming practices, these tests will be ideal tools for extension and technical assistance to inform

150

management goals on a broader range of farms. In either case, soil health tests are not useful for
farmers unless they target practices that farmers are willing to undertake, especially if the test
helps to motivate on-farm changes in management.

III. Soil health tests need to adapt to regional differences in soil and farm practices.
Different soil types and environmental conditions dictate both different management
styles and farmer responses to soil health metrics. For example, farmers on ‘heavier’ ground
care less about organic matter because it takes more effort to increase this resource. Instead, they
care more about soil structure (i.e., the aggregate stability test). Farmers on lighter, coarser soils
care more about how to improve their organic matter – they know they can change it from their
experience (not from following soil testing results) and from seeing the difference in improved
soil performance. Some test results may show poor numbers that are characteristic of regional
patterns. Similar to traditional testing, soil health tests need minimum data sets and databases to
strengthen their predictive ability and relevance to specific soil types and environmental
contexts.

Synthesis
Soil health testing revealed strong differences across farmer fields, particularly for values related
to soil organic matter and physical factors. Perhaps more importantly, these tests distinguished
key soil parameters between field types, which farmers had selected as either good or
problematic fields. The NRC field also proved useful in highlighting to farmers differences in
cropped fields, showing the greatest contrast in soil health measures of water holding capacity,
and pools of soil C and N. Further, the NRC field was the only field to show differences in

151

nitrification potential compared to other fields, highlighting the opportunity to further adapt
testing parameters to other management sensitive ecosystem processes. Michigan has a highly
diverse agricultural sector and to our knowledge, extensive soil health surveys have not been
conducted across different regions and agroecosystems. There is a critical need to establish
baseline surveys and build a database (which already exists for traditional soil tests), in order to
strengthen confidence in soil health testing results and recommendations.
Farmers rely on many different types of knowledge to make decisions about adjustments
to their soil management, including their own experience and competencies, and information
from outside sources such as crop advisers, extension agents, and soil testing agencies. In turn,
soil management decisions, are impacted by a complex set of social and ecological factors.
Adoption of soil management recommendations (such as might come from a soil health
assessment) may depend on the size and scale of the farm (Stuart and Gillon, 2013), the financial
capacity of the farmer (Singer et al., 2007), a farmer’s access to information and networks
(Baumgart-Getz et al., 2012), their trust in different sources of information (Wolf and Nowak,
1995), or on their willingness to engage with and seek out innovation (Blesh and Wolf, 2014).
Given the many factors that may influence a farmer’s decision about soil management, it is
critical that results from soil health tests intersect with farmers’ assessments of their own fields
so management recommendations can be aligned with farmer knowledge.
A framework for managing from soil health tests must not only reflect field conditions
but recognize that farmers have clear ideas about soil quality (Romig et al., 1995; Walter et al.,
1997), which inform how they manage fields and how they understand the effects of changing
practices. For example, my results suggest that farmers readily linked results from soil physical
measurements to their own observations in the field such as soil crusting, compaction, or field

152

drainage, and were more interested addressing these results through changing their practices.
Yet, even in this targeted sample of farmers, regional differences influenced soil testing results
and farmer interpretation. All farmers expressed interest in increasing organic matter, but the
weight placed on this value depended on the region. While traditional soil tests link plant
nutrient inputs directly to yield, findings from soil health tests, which reflect severe limits to crop
performance, such as poor soil aggregation or low labile organic C, may require a more complex
set of management decisions. Thus, changes in soil management may also depend on planning
and long-term horizons (Lambert et al., 2006), such as a farmer’s willingness to discount cost or
risk in the short-term in order to reach a longer term goal or benefit. Based on clear differences
in results across their fields, nearly all farmers in this study expressed interest in continuing soil
health testing, but wanted a better understanding of how specific management practices, under
their control, linked to positive testing outcomes and improved soil quality.

Conclusions
As private landowners, farmers have the greatest incentive to manage their fields for
near-term goals such as yield or profitability. Yet, a corollary to a focus on soil health
management is that agricultural landscapes are increasingly expected to supply a range of
ecosystem services beyond yield—including non-provisioning services (Dobbs and Pretty, 2004;
Robertson et al., 2014). Even though policy and market conditions often limit transitions toward
sustainability (Stuart and Gillon, 2013), a growing number of farmers are explicitly managing for
soil quality. However farmers lack soil testing information that may help guide management
toward soil quality. Results from this study suggest that soil health tests can detect differences in
a range of soil quality parameters across working farms. Since these parameters also intersect

153

with farmer knowledge of their fields, they can potentially inform changes in management. All
farmers in this study use information from traditional soil tests to inform critical and costly
management decisions such as fertilizer and lime applications. They also use other information
when deciding how to manage their soil, including information from networks of family and
neighbors, and from their own experimentation. When considering significant changes to soil
management that might follow from soil health test recommendations, these farmers expressed
concerns related to factors such as financial risks and availability of time, which have also been
identified in other farmer surveys (Baumgart-Getz et al., 2012). Using soil health testing in ways
that target important soil parameters and track progress from recommended management
practices, farmers may be more willing to adopt changes that have advantages to soil quality over
the long term (Chouinard et al., 2008). Likewise, knowledge from on-farm research and
understanding of farmer practice will inform the activities of soil researchers. Managed
ecosystems are dynamic, complex systems, and the need to integrate models of human behavior
and decisions with biophysical measures is critical to improve environmental outcomes (Antle et
al., 2001). Expanding soil health testing must include careful work with farmers in order to
integrate an ecosystem-based framework for managing soil.

AKNOWLEDGEMENTS
Support for farmer surveys and testing were supported by a the U.S. Department of
Agriculture’s Sustainable Agriculture Research and Education program, grant # GNC14-192,
and also by a summer fellowship from the Kellogg Biological Station’s Long-term Ecological
Research Program through the National Science Foundation (DEB 1027253).

154

APPENDIX

155

Table 5.1 Percent sand and soil texture class from 13 farms in three Michigan counties for four
field types: “Best Field” and “Worst Field” as characterized by farmers, a non-row crop field
(“NRC”) and a “Choice” field also selected by each farmer.

Best Field
County
Van Buren
Van Buren
Van Buren
Van Buren
Presque Isle
Presque Isle
Presque Isle
Presque Isle
Isabella
Isabella
Isabella
Isabella
Isabella

% Sand
81.6
64.8
85.3
85.0
87.9
81.5
73.6
70.0
41.6
62.5
66.1
66.3
86.1

Worst field

Texture class
Loamy sand
Sandy loam
Loamy sand
Loamy sand
Loamy sand
Loamy sand
Sandy loam
Sandy loam
Medium loam
Sandy loam
Sandy loam
Sandy loam
Loamy sand

156

% Sand
82.6
62.8
92.3
81.1
82.1
81.0
64.2
71.4
37.6
71.0
79.9
64.4
71.5

Texture class
Loamy sand
Sandy loam
Sand
Loamy sand
Loamy sand
Loamy sand
Sandy loam
Sandy loam
Medium loam
Sandy loam
Loamy sand
Sandy loam
Sandy loam

Table 5.1 (cont’d.)
NRC
County
Van Buren
Van Buren
Van Buren
Van Buren
Presque Isle
Presque Isle
Presque Isle
Presque Isle
Isabella
Isabella
Isabella
Isabella
Isabella

% Sand
83.2
65.2
85.8
85.9
84.6
82.8
81.6
76.1
60.4
71.4
77.9
91.0
44.9

Choice
Texture class
Loamy sand
Sandy Loam
Loamy sand
Loamy sand
Loamy sand
Loamy sand
Loamy sand
Sandy loam
Sandy loam
Sandy Loam
Loamy sand
Sand
Medium loam

157

% Sand
89.5
69.0
79.0
86.3
92.1
90.2
69.3
74.3
51.4
69.4
71.1
49.4
55.1

Texture class
Loamy sand
Sandy loam
Loamy sand
Loamy sand
Sand
Sand
Sandy loam
Sandy loam
Medium loam
Medium Loam
Sandy loam
Medium loam
Sandy loam

Table 5.2 Percent aggregate stability (AS), available water capacity (AWC), soil compaction measures as resistance at two soil depths
shown as means ± se for across Michigan farms (n=13). Field types are “Best Field” and “Worst Field” as characterized by farmers, a
non-row crop field (“NRC”) and a “Choice” field also selected by each farmer. Field comparisons are t-tests comparing paired field
types for each farm for each soil parameter. Significance of treatment effect: * P ≤ 0.05, ** P = 0.01 to 0.001, ***P<0.001.

Field
Type
Best
Worst
NRC
Choice

AS

AWC

Resistance at depth
0-15 cm
15-46 cm

%

g water g soil-1

psi

69.1 ± 4.0
58.7 ± 5.4
70.3 ± 5.0
59.8 ± 4.7

0.20 ± 0.02
0.15 ± 0.02
0.24 ± 0.02
0.19 ± 0.02

Field
comparison
Best vs. Worst
Best vs. NRC
Worst vs. NRC
Worst vs. Choice
Choice vs. NRC
Best vs. Choice

15.3 ± 3.6
20.3 ± 4.2
30.2 ± 7.5
25.1 ± 6.3

59.2 ± 3.1
59.0 ± 4.3
66.5 ± 6.9
61.2 ± 5.0

AS

AWC

Resistance
0-15 cm 15-46 cm

t-test

t-test

t-test

t-test

2.4*
0.1
-1.3
0.0
1.3
1.9

2.9*
-1.3
-4.2***
-1.1
1.9
1.7

-1.4
-2.3*
-1.4
-0.7
0.7
-2.1

0.0
-1.5
-0.7
-0.1
1.1
-1.0

158

Table 5.3 Percent organic matter (OM), permanganate oxidizable carbon (POXC), potentially mineralizable carbon (PMC) and
nitrogen (PMN) and nitrification potential (NIT) shown as means ± se for across Michigan farms (n=13). Field types are “Best Field”
and “Worst Field” as characterized by farmers, a non-row crop field (“NRC”) and a “Choice” field also selected by each farmer. Field
comparisons are t-tests comparing paired field types for each farm for each soil parameter. Significance of treatment effect: * P ≤ 0.05,
** P = 0.01 to 0.001, ***P<0.001.

Field
Type
Best
Worst
NRC
Choice

OM

POXC

PMC

PMN

mg POX-C kg

µg CO2-C g

µg NH4 -N g

mmol NO3--N kg-1 g-1

3.5 ± 0.3
3.1 ± 0.2
4.7 ± 0.6
3.2 ± 0.2

283 ± 28
287 ± 21
405 ± 26
295 ± 32

39.5 ± 3.4
29.2 ± 3.6
43.5 ± 4.6
30.3 ± 6.3

6.5 ± 0.6
6.0 ± 0.6
9.2 ± 0.7
6.6 ± 0.6

5.2 ± 0.3
4.8 ± 0.6
3.3 ± 0.4
4.8 ± 0.5

Best vs. Worst
Best vs. NRC
Worst vs. NRC
Worst vs. Choice
Choice vs. NRC
Best vs. Choice

-1

NIT

%

Field
comparison

-1

+

-1

OM

POXC

PMC

PMN

NIT

t-test

t-test

t-test

t-test

t-test

3.5**
-3.0*
-3.5***
0.1
3.0*
1.4

-0.2
-2.8*
-3.2***
-0.4
2.3*
-0.2

2.4*
-0.6
-3.1**
-0.1
2.0
1.6

0.6
-2.4*
-2.9*
-1.0
2.9*
-0.2

0.7
3.7**
2.1
0.2
-2.3*
1.0

159

Table 5.4 Soil pH, soil concentration Bray1 Phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), ammonium (NH4+) and
nitrate (NO3-) shown as means ± se for across Michigan farms (n=13). Field types are “Best Field” and “Worst Field” as
characterized by farmers, a non-row crop field (“NRC”) and a “Choice” field also selected by each farmer. Field comparisons are ttests comparing paired field types for each farm for each soil parameter. Significance of treatment effect: * P ≤ 0.05.

Field
Type
Best
Worst
NRC
Choice

pH
6.85 ± 0.18
6.85 ± 0.20
6.94 ± 0.19
6.88 ± 0.17

Field
comparison
Best vs. Worst
Best vs. NRC
Worst vs. NRC
Worst vs. Choice
Choice vs. NRC
Best vs. Choice

P
ppm
94.2 ± 18.6
82.9 ± 21.4
88.6 ± 16.9
76.3 ± 13.6

pH
t-test
0.3
-0.4
-0.9
-0.4
0.3
-0.1

P
t-test
-0.1
0.7
0.3
0.7
0.8
1.4

K
ppm
145.2 ± 14.0
138.5 ± 15.4
153.7 ± 26.9
142.9 ± 20.3

K
t-test
0.4
-0.1
-0.3
0.1
0.4
0.3

Ca
ppm
1378.6 ± 214.9
1228.5 ± 153.2
1658.3 ± 265.6
1194.4 ± 192.2

Ca
t-test
1.1
-1.1
-2.5*
0.1
2.0
2.9*

Mg
t-test
0.0
-1.2
-0.9
1.1
1.5
1.1

160

Mg
ppm
185.5 ± 24.4
194.3 ± 26.3
205.2 ± 33.1
173.8 ± 22.7

NH4+
t-test
-0.9
-1.7
0.4
1.0
1.3
-0.2

NO3t-test
0.7
0.6
0.1
1.8
1.7
1.6

NH4+
µg NH4+-N g-1
1.6 ± 0.1
3.2 ± 1.6
2.5 ± 0.5
1.7 ± 0.2

NO3µg NO3-N g-1
11.9 ± 2.4
10.9 ± 2.3
11.1 ± 1.6
7.5 ± 1.8

Table 5.5 Field observations from sampling on 4 field types on 13 farms across Michigan.
Farm
Number

Best

Worst

1

Best field is heavily tilled. Appears to
have been in soybeans the previous
year. Cloddy and crusted. Poor
structure but appears to have decent
organic matter. Formerly part of
swine operation? Poor aggregate
structure and largely unconsolidated.
No evidence of earthworm or other
biological activity. No weeds.

Planted and plowed. Beyond tree line
seen from best field. Structure is poor
and organic matter appears quite low.
Highly unconsolidated and loose –
little to no horizon.

2

‘High’ field. Was in corn with lots of
residue remaining. High level of
earth worm activity. Field has
received a lot of manure since it is
closest to the paddock. Planted in
soybeans already (no-till) and some
prills of fertilizer also visible. Soil is
dark. Nice structure. No evidence of
drainage problems. Looks in
excellent condition. Evidence of
mycorrhizae.

‘Dad/Bill’ field on the other side of
the highway. Planted in soybeans.
Field has been disked and appears to
have been in corn. Evidence of
earthworm activity, but looks like
poor organic matter and not very good
structure.

3

‘Dave Good – west of the white
‘Dad’ field. Soil is clayey but without house’. Lots of earth worm activity,
too much structure. Appears highly
and some crusting. Almost nothing
compacted and blocky. ‘Heavy.’
growing. Soil highly friable, but
Some oxidation – poor drainage and
decent structure and some blockiness
patchiness.
as peds. Coming out of corn with
large amount of residue remaining.

4

No-till and modified strip-till. Some
areas not tilled in 100 years. Flat and
recently planted to soy. Lots of corn
residue. Highly structured soil and
looks to be in excellent condition.
Low-lying but drainage appears good.

161

‘Smith/Low’, Compacted. Crusted.
Planted to soy last year with moderate
amount of residue remaining. Some
weediness ‘between’ rows. Some
gleying and possible drainage issues.

Table 5.5 (cont’d)

Farm
Number

Choice

Non-row crop

1

Across the road from the best field.
Clear drainage problems here –
evidence of outwash, rocky and
high level of mottling. Some
degree of variability across from
field too as it leads to trees.
Previously in corn. Crusted.

Grassy area need road and by
walnut trees. Highly
structured and deep horizon.
Some sandiness but generally
dark with high organic matter.
Covered primarily in orchard
grass. Highly moist.

2

‘Across the way’ On the other side
of a drainage ditch from the alfalfa
field. Field has more structure and
less of the drainage problems.
Slopey. Planted to corn. Lots of
earth worms.

‘Foxes-Hay’ is coming out of
pasture (alfalfa) which
appears to have been strayed.
Alfalfa looks poor and spotty.
Lots of earth worm activity.
Field appears to drain poorly
also – mottling with glaying
and oxidative patches.

3

CRP ground across the way from
main farm house and at the corner.
Deciduous forest. Sandy with but
high organic matter. Appears to be
poorly drained. Thick understory.
Highly structured.

Sandy and windswept. Rocky
at the surface. Field looks
‘burned’ but weedy – lots of
quackgrass. Soil very poorly
structured but with mottling –
drainage problems (streaky
oxidation)

4

Lots of earth worm activity.
Appears highly crusted. Planted to
no-till and appears to have been
kept in no-till for a while. Soil is
highly aggregated and wellstructured. Appears to drain well.

Grassy strip adjacent to high
yield field. In lush grass –
very deep horizon and high
organic matter. Full or roots
and well-structured.

162

Table 5.5 (cont’d)
Farm
Number

Best

Worst

5

Some crusting but rich-looking and
dark with lots of evidence of worm
castings. Sandy and does not hold
together well. Across and further down
the road from Dandelion buffer strip.
‘High-R17 and R15’. Would appear to
have received a lot of manure. Near
lots of old out-buildings.

Previously planted to Sorghum (this
is the filed that took an age to find –
last one of the day.) Crusty with lots
of weeds. ‘Kelly-Meyers’. Coming
out of sorghum. Moderate amount of
residue. Lots of weeds. Lots of
crusting. Field is highly sloped and
has some drainage problems at one
end.

6

Chris - Soybean, lots of corn litter.
Very hilly. Some weeds. Dark sandy
soil and wet (50% sand). No-till w/
Deep shank. 3-year rotation. Problem
with high alkalinity. A no-till field in
soy with clear evidence of corn stover.
Forested on south end and slight slope
toward north end, but somewhat hilly.
Some weeds, dark and sandy soil ~
50% sand. Field is on a three-year
rotation – use of deep shank tillage in
some years. Tends towards high
alkalinity, which has been problematic.

Dumfys - Corn, sandy, lots of rocks.
Tilled. A bit of litter. A little less
sandy. Some oxides. Sloped. Lots of
topography (variable). Field slopes
down the road steeply and then
plateaus about 50 meters in.
Construction equipment – clearing of
trees to make field bigger – trees
mostly on south (?) side – to the left
coming from the road. Some low
spots with drainage problems. Has
been in corn – sandy soil with lots of
rocks, tilled with a bit of litter left
and some evidence of oxidation and
poor drainage. High degree of
topographical variability.

7

Wheat (good) - Good soil structure. In
beans. Some moss on sand. Lots of
weeds. Lots of rocks not on surface but
deeper. Some good structure. Well
aggregated. Some oxidation. Towards
the ‘back’ of the land. In beans –
perhaps 10-days since planting. Some
moss on field and quite weedy. Quite
rocky, but more in sub-surface – small
glacial till. Good soil structure and
well-aggregated. Some evidence of
oxidation.

Alfalfa - Many earthworms. Mossy
cover on soil. Dark and decent
structure. Some spots more compact
than others. Towards the back of the
‘best’ field. Lots of weeds and moss
on surface. Soil is ‘dark’ with decent
structure. Compaction for
penetrometer seems highly variable.
Lots of evidence of earthworms.

163

Table 5.5 (cont’d)
Farm
Number

Choice

Non-row crop

5

Planted to wheat with a good stand.
Highly compacted 53.3 ±9.1 psi at
6 inches (69±8.3 at 18 inches).
Structure is poor – soil is a heavier
clay. Large field planted to red
wheat. Soil is much heavier –
much higher clay content? Appears
to be much more compacted. Crop
looks good. Appears well-drained.

Sloping buffer strip. Full of Dandelions.
Some water and drainage issues at lower
part. Decent structure. 36 ± 7 psi at 6
inches, 74±10 at 18. Across the road
‘High’ field. Tons of dandelions in spite
of being out of production. Some areas
must be highly disturbed. Other areas
have mixed grasses. Appears to have
drainage problems, but soil does have
dark color and decent structure.

6

Harvey’s - Crumbly soil (sandier
than Chris field). Corn and no litter.
Some spots are clotty and muddy.
80-day corn. Coming out of wheat.
Woods on right side as you enter
the field. Soil is ‘crumby’ but
sandier than field ‘A’. Extensive
tillage – almost no litter in evidence
– more soil clods and ‘muddy’ with
some evidence of poor drainage.
Field coming out of wheat and
going into 80-day corn.

Gregg's - Wheat crop looks good.
Extremely compacted. Plow pan, tough
dirt, very sandy and dry. No organic
matter. Some corn litter. No-till. Listed
as ‘unmanaged’ but really in wheat.
‘Extremely’ compacted soil – almost
surely a plow pan, although based on
corn litter it looks like the wheat was
planted with no-till drill. Soil extremely
sandy and dry with very little evidence
of organic matter. Very large field –
slopes toward the western edge down to
woods – hard to capture variability.

7

Wheat - Soil is well-aggregated but
evidence of a plow pan – highly
compact especially at sub-surface.
Was wheat going into hay. Pretty
good organic matter. Oxides. Off to
the side of the house/barn. Gently
sloping land. Pretty high organic
matter by visual inspection. How
does he use manure from his few
cows? Some evidence of oxidation
and poor drainage.

Hay - Very dry. Very compact. Mixture
of hay and orchard grass. Lots of
diversity in insects. Variable green color
in field – dark green strips/banding –
either variable drainage or older drain
tile in sub-surface. Highly productive for
grower. Decent mix of diversity – lots of
orchard grass.

164

Table 5.5 (cont’d)
Farm
Number

Best

8

Eva’s - Freshly tilled. Earthworms.
Good structure. Coming out of
wheat.

Lewaneowski - Surface looks cracked
and dry. Some weeds. Surface residue.
Crusty lots of stones. A little bit of clay
deeper past 3 cm. Came out of wheat.

9

STR-#5 - Slightly less compacted.
Blocky. A few earthworms.

STR-#3 - Very unsuccessful in
penetrating soil. Wouldn’t get below 6”.
Highly compact. Blocky. Small amounts
of oxidation. OK moisture.

10

Home - Sandy loam, uniform corn
and beans great yields clear rye
residue, no earthworms. Known as
‘Home’ field. Typically yields are
great for corn and beans. Sandy
loam. Clear evidence of a rye
cover crop. Little to no soil
structure and no evidence of earth
worm activity.

Vineyard - Hay field, sandy not good
growing soil lower part does well. “not
good growing soil’ according to owner.
‘Lower’ area does better in terms of
yield. Has been used for hay, but
evidence of a corn crop. Also some more
land being cleared. Slight rolling in field
with line of woods to the south.

11

Lyle - Mostly clay; soybeans; notill; lots of corn litter; fairly weedy;
different species of weeds; thicker
soil (clay texture), well structured,
plow pan at 10-15”

Mortimer - Even and not many hills;
some gravel, looks sandy; some corn
litter; not great aggregation; lots of
earthworms; some parts are dark and
clayish; less litter than Lyle

12

Carmichael - Sandy loam. 31 acres
of corn. Dark, more aggregates at
deeper regions. Hilly. Variable in
organic matter content.

Tribal - Some regions of nitrogen stress.
No-till, cover crop, corn, air seeded to
annual rye. Very sandy w/ earthworms.
Between 7 and 8 inches deep different
layer horizon top universally mixed.

13

Dad - Mostly plain not many hills,
soybean crop, lots of corn residue
(litter). Corn last year. Dark sand
holds together. Narrow spacing
between rows. No-till

Soy low - Doesn’t hold together very
well, no-till. Some grassy spots. Similar
to pivot. Lots of corn litter. Very sandy.

Worst

165

Table 5.5 (cont’d)
Farm
Number

Choice

8

Home - Soybean field. Hard, crusty
surface. Lots of residue. Darker
soil/good structure. Oxidation.
Good aggregation.

9

STR-#4 - More compacted within
row (can’t get below 6”). Doesn’t
go below middle of row. Slightly
sandy. Structure is good.
Earthworms.

Rye - Dry. Cracked at surface.
Compacted. Can’t penetrate below 5”!

10

Home Hay - 10 acres of hay, 4 year
old, east of house; row crop; had
been a pasture field. About 10
acres has been in alfalfa for 4 years.
This is east of the house and field
‘A’. Steep slope down to road and
then level for most of the rest of the
field – surrounded by woods to the
north and east. Soil appears heavily
compacted, especially at the
subsurface. It was recently
harvested – some areas not yet
bailed. Evidence of earthworms and
numerous grubs in areas where
sampled.

Krajewski - Had cattle there wet low
ground and swampy, soils black, slimy
and cold great soil, good aggregation,
dark soil (but sandy) pasture very
vegetated, great organic matter, lots of
flowers. This ground had cattle. It is low
and wet. High level of plant diversity
(lots of forbs in bloom) and seemingly
highly productive (mostly chest high
growth). Soils were black, ‘slimy,’ and
cold – muck-like. But excellent
aggregation in spite of sandiness. High
level of organic matter.

11

Home -No-till beans; cows grazing;
lots of corn litter; sandy, but good
structure; earthworms; weedy.

Bellis (hay field) - All hay and irrigated
mixture of alfalfa and some grass; great
soil structure.

12

Center pivot - Clay, sand, gravel
(wide range); variation in plant
growth (tall short plants); corn; lots
of hills; patchiness across field.
Plow pan at about 12” to 15.”

Vineyard/orchard - Within row less
compaction, between rows more
compaction.

13

Pivot - Narrow rows. Scattered
cropping. Lots of litter (corn). Dark
soil (sandy). Some grubs. Shaded
regions on the borders. No-till

Forest - Lots of vegetation. Lots of
shade. Thick brush. Dark rich soil.
Organic matter. No compaction. Moist.

Non-row crop

166

Buza - Been in pasture for years. Lots of
rocks. Thick grass.

Table 5.6 Responses from 13 farmer interviews regarding soil testing and information used for soil management.

Question
How often do you
soil sample?

Which soil test
parameters are you
most interested in
checking?

How do soil test
results inform your
soil management
decisions?

Soil Testing and Management

# of farmers

Test individual fields every 2-4 years.

13

Fertilizer recommendation

10

pH for lime recommendation

8

Organic matter content

7

Micronutrient concentration

3

Match up nutrient contents with soil types and yield maps

2

Target fertilizer application rates for yield goal

10

Use (or try to use) plant tissue tests to better adjust side dress

3

Dial fertilizer recommendations into variable rate equipment.

3

Relate test results to number of field passes and soil compaction

2

Avoid phosphorus excesses by adjusting crop rotation or application

2

167

Table 5.6 (cont’d)

Question

How do you assess or
understand soil quality?

Which sources of
information do you use to
inform soil management?

Soil Testing and Management

# of farmers

Physical characteristics (e.g. does not crust, 'works well')

9

How water acts (e.g. drainage, ease of getting equipment into field)

3

Organic matter content

3

Rotation type (e.g., soil under a long rotation has better quality)

3

Disease-free fields

2

Healthy soil biology

1

Chemical characteristics (e.g. pH, CEC)

1

Word of mouth (neighbors, family)

9

Experimentation with or inspection of own fields (or neighboring
fields)

8

Soil Testing

6

Reading/Literature (trade magazines, the Web, books)

6

Extension resources (including meetings or conferences)

4

Market prices

3

Sales representatives (fertilizer, seeds etc.)

3

168

Table 5.7 Responses from 13 farmer interviews regarding their selection of field types for soil sampling on their farms.

Question

Best

Why did you select this field

# of farmers

High yielding field

10

The soil ‘works’ well

5

Takes good care of the field

5

Reliable/less variable to manage

4

No apparent disease

4

Uses field to test new ideas/experiment

2

Low yielding field

9

Soil 'works' poorly (e.g. forms soil crust)

9

Field not ‘reliable’

4

Moisture problems

4

Compaction problems

3

History of poor management

3

Soil chemistry is ‘off’

2

Disease problems

1

Worst

169

Table 5.7 (cont’d)

Question

Why did you select this field

# of farmers

Physical problem (e.g. compaction, crusting)

7

Lower yield, ranging from ‘OK’ to ‘poor’

6

History of poor management (known or assumed)

4

Field highly variable

3

Curious about field, no explicit problems

2

Unsure what to do for management

2

Hay field

7

Field margin

3

CRP/wood lot

2

Buffer strip

1

Choice

Non-row crop

170

Table 5.8. Responses from 13 farmer interviews regarding their management of fields selected for soil sampling on their farms.

Management based on field type

Tillage

Best

Worst

Choice

No till

4

5

4

Chisel

4

5

2

Mix (e.g. chisel, subsoiler)

5

3

7

C-S-W

4

4

3

C-S

5

4

5

Other (e.g. W-S, hay)

3

4

5

Cover crop use

5

8

5

Currently

4

2

1

Past management

2

1

0

Crop/rotation

Manure

C=corn, S= soybean, W = Wheat

171

Table 5.9 Condensed responses from 13 farmer interviews regarding soil health test results from their farm.
Soil parameter (# of farmers who
commented on measure)

Physical

Soil aggregate stability
(AS) (6)
Available water capacity
(AWC) (7)

Soil Health metrics

Soil compaction

Biological,
Carbon

(6)

Total soil organic matter
(OM) (9)
Permangenate oxidizable
carbon (POXC) (2)
Potentially mineralizable
carbon (PMC) (10)

Biological,
Nitrogen

Potentially mineralizable
nitrogen (PMN) (4)
Nitrification

(3)

pH, phosphorus, potassium,
Traditional
calcium, magnesium,
field crops test
ammonium, nitrate (9)

Summarized feedback
Farmer who commented on AS and AWC felt the values in the soil test
corresponded to their experience across selected fields. Most had a ready
agronomic explanation for why a number might be low or high for a given
field. For soil compaction, however, there was much more confusion.
Some felt it did not reflect their experience of a field and some were
concerned and asked if they should alter their management based on
comparison between field types.
Most farmers were aware of total OM across different field types. In two
counties with coarsest soils there was more commentary over soil organic
matter/carbon. Those with 'heavier' soils were more likely to say they could
not easily change their OM value. A large number wanted to better
understand shorter term carbon (POXC and PMC) values, and commented
directly on how they could change practices to raise it in fields with lower
values to bring it closer to fields with higher faster cycling carbon pools.
Of the farmers that commented on PMN, all wanted to know how they could
increase this value in the field to extend nitrogen release. One readily
explained a high PMN value as due to recent history of including a perennial
legume (alfalfa) in the particular field. Two farmers understood nitrification
as a process occurring in their fields and three had used inhibitors.
In reviewing these parameters for all of their fields, farmers offered the most
feedback on values for pH and phosphorus, already having an expectation
for the range of these values for different fields. Some farmers also
explained why a specific field had a low or high value for a given parameter,
given their prior experience with testing or management of a field.

172

Table 5.10 Condensed responses from 13 farmer interviews regarding approach to soil management on their farms.

Themes

Summarized feedback

Quotes

In reflecting on testing results, nearly all
growers valued organic matter and wanted to
increase it, but they differed in terms of
approach to management.

Organic matter,
time and place

Farmers on finer soils, with generally higher
organic matter values, regardless of field types,
were less likely to manage for increasing
organic matter, even if they wanted to. Rather,
they focused on how their soil ‘worked’
(physically) as a metric of how to improve soil
quality.

Those on coarser or more easily degraded soils
place greater value on building organic matter
through management.

173

"It's not an overnight fix on that ground, we fixed
up a lot of ground that has not been into farming
over the years, and we tried to build that ground
up.”
"It's so hard to raise it."
"I am not going to live long enough to do that…"
"I mean it’s just slow, so slow going, my goodness
it takes forever to change a number.”
Referring to leaving more residue on the field "...I
was more impressed with the mellowness of the
dirt..., I don’t know that I gained a lot in yields but
the dirt works a whole lot better."
Referring to a how he improved a field: "when we
first started to farm it, it was like the desert out
there, that soil would just blow away...and that
goes back to the organic matter, and that takes time
to build that up, that organic matter. If you can get
that organic matter built up, then you can hold the
water out there."
"But if you got high organic matter, good soil CEC
test, everything, it's got a good holding capacity,
then your compaction isn't there. Like, say, just the
general physical characteristics of the soil are so
much better. "

Table 5.10 (cont’d)
Themes

Summarized feedback

Quotes
Stated soil management goal via increased cover cropping
and reduced tillage: "Keep it there, whatever I put on”.

Farmers in all counties referred to
altered crop rotation (including use
of cover crops) as the surest way to
improve soil quality

Changing crop
rotation to influence
measurable changes in
soil health

But altering rotation was
constrained, with some saying they
could not do it or faced constraints.
Reflecting on differences in soil
health metrics between different
field types, farmers wanted to be
able to use them if they had a better
sense of how to act on the numbers
and if this testing would give them
more control of their soil
management

Referring to using sorghum to improve degraded fields: ".. I
think that's one of the main things...is that root system is just
breaking things back up and allowing water to get through
the soil…it suppresses weeds, there are no weeds growing.
Referring to farmer practices on best fields he has seen: “,…
he rotated into a rye cover … he’d always put rye in between
[beans]…. he rotated oats into it he’d never grow a kernel of
corn, he’d seed with clover and that sort of thing, and that’s
soil health, that’s definitely reflected on his ground over
time"
“I'd like to get quaternary rotation if I can do it
profitably…but I have a hard time making money the year
corn was high. I would like the narrow rotation, no question
about it.”
“…I’d like to see more stable numbers that say: 'ya I’m
making progress, or no I’m not…"
Referring to mineralizable carbon: "So it's there, it just needs
to get mineralized. How do you trick it into mineralizing
that? What do you put on to trick it?"
“…If we can change that, if there’s a way to change it. That's
the big thing. For me, it's one thing to say, well you got a
problem there. What's the implementation of the change in
that and how do I change it? I want to change it."

174

REFERENCES

175

REFERENCES

Antle, J.M., Capalbo, S.M., Elliott, E.T., Hunt, H.W., Mooney, S., Paustian, K.H., 2001. Research Needs
for Understanding and Predicting the Behavior of Managed Ecosystems: Lessons from the Study
of Agroecosystems. Ecosystems 4, 723–735. doi:10.1007/s10021-001-0041-0
Baumgart-Getz, A., Prokopy, L.S., Floress, K., 2012. Why farmers adopt best management practice in the
United States: A meta-analysis of the adoption literature. J. Environ. Manage. 96, 17–25.
doi:10.1016/j.jenvman.2011.10.006
Bhardwaj, A.K., Jasrotia, P., Hamilton, S.K., Robertson, G.P., 2011. Ecological management of
intensively cropped agro-ecosystems improves soil quality with sustained productivity. Agric.
Ecosyst. Environ. 140, 419–429. doi:doi: 10.1016/j.agee.2011.01.005
Blesh, J., Wolf, S.A., 2014. Transitions to agroecological farming systems in the Mississippi River Basin:
toward an integrated socioecological analysis. Agric. Hum. Values 31, 621–635.
doi:10.1007/s10460-014-9517-3
Cherry, K.A., Shepherd, M., Withers, P.J.A., Mooney, S.J., 2008. Assessing the effectiveness of actions
to mitigate nutrient loss from agriculture: A review of methods. Sci. Total Environ. 406, 1–23.
doi:doi: DOI: 10.1016/j.scitotenv.2008.07.015
Chouinard, H.H., Paterson, T., Wandschneider, P.R., Ohler, A.M., 2008. Will farmers trade profits for
stewardship? Heterogeneous motivations for farm practice selection. Land Econ. 84, 66–82.
Diaz, R.J., Rosenberg, R., 2008. Spreading Dead Zones and Consequences for Marine Ecosystems.
Science 321, 926–929.
Doane, T.A., Horwáth, W.R., 2003. Spectrophotometric Determination of Nitrate with a Single Reagent.
Anal. Lett. 36, 2713–2722. doi:10.1081/AL-120024647
Dobbs, T.L., Pretty, J.N., 2004. Agri-Environmental Stewardship Schemes and “Multifunctionality.”
Appl. Econ. Perspect. Policy 26, 220–237. doi:10.1111/j.1467-9353.2004.00172.x
Drinkwater, L., Snapp, S., 2007. Understanding and managing the rhizosphere in agroecosystems.
Elsevier: Burlington, MA, USA.
Drinkwater, L.E., Cambardella, C.A., Reeder, J.D., Rice, C.W., 1996. Potentially mineralizable nitrogen
as an indicator of biologically active soil nitrogen, in: Methods for Assessing Soil Quality. Soil
Science Society of America, Madison, WI, pp. 217–229.
Franzluebbers, A.J., Haney, R.L., Honeycutt, C.W., Schomberg, H.H., Hons, F.M., 2000. Flush of carbon
dioxide following rewetting of dried soil relates to active organic pools. Soil Sci. Soc. Am. J. 64,
613–623.
Havlin, J.L., Beaton, J.D., Tisdale, S.L., Nelson, W.L., 2005. Soil fertility and fertilizers: An introduction
to nutrient management. Pearson Prentice Hall Upper Saddle River^ eNJ NJ.

176

Karlen, D.L., Andrews, S.S., Wienhold, B.J., Zobeck, T.M., 2008. Soil Quality Assessment: Past, Present
and Future.
Kettler, T.A., Doran, J.W., Gilbert, T.L., 2001. Simplified Method for Soil Particle-Size Determination to
Accompany Soil-Quality Analyses. Soil Sci. Soc. Am. J. 65, 849. doi:10.2136/sssaj2001.653849x
Lambert, D., Sullivan, P., Claassen, R., Foreman, L., 2006. Conservation-Compatible Practices and
Programs: Who Participates? (Economic Research Report No. 14). USDA Economic Research
Service.
Michalak, A.M., Anderson, E.J., Beletsky, D., Boland, S., Bosch, N.S., Bridgeman, T.B., Chaffin, J.D.,
Cho, K., Confesor, R., Daloğlu, I., DePinto, J.V., Evans, M.A., Fahnenstiel, G.L., He, L., Ho,
J.C., Jenkins, L., Johengen, T.H., Kuo, K.C., LaPorte, E., Liu, X., McWilliams, M.R., Moore,
M.R., Posselt, D.J., Richards, R.P., Scavia, D., Steiner, A.L., Verhamme, E., Wright, D.M.,
Zagorski, M.A., 2013. Record-setting algal bloom in Lake Erie caused by agricultural and
meteorological trends consistent with expected future conditions. Proc. Natl. Acad. Sci.
doi:10.1073/pnas.1216006110
Moebius, B.N., van Es, H.M., Schindelbeck, R.R., Idowu, O.J., Clune, D.J., Thies, J.E., 2007. Evaluation
of laboratory-measured soil properties as indicators of soil physical quality. Soil Sci. 172, 895–
912.
Moebius-Clune, B.N., Moebius-Clune, D.J., Gugino, B.K., Idowu, O.J., Schindelbeck, R.R., Ristow, A.J.,
van Es, H.M., Thies, J.E., Shayler, H.A., McBride, M.B., Wolfe, D.W., Abawi, G.S., 2016.
Comprehensive Assessment of Soil Health – The Cornell Framework Manual, Edition 3.0. ed.
Cornell University, Geneva, NY.
Moebius-Clune, B.N., van Es, H.M., Idowu, O.J., Schindelbeck, R.R., Moebius-Clune, D.J., Wolfe,
D.W., Abawi, G.S., Thies, J.E., Gugino, B.K., Lucey, R., 2008. Long-Term Effects of Harvesting
Maize Stover and Tillage on Soil Quality. Soil Sci. Soc. Am. J. 72, 960.
doi:10.2136/sssaj2007.0248
Norton, J.M., Stark, J.M., 2011. Regulation and measurement of nitrification in terrestrial systems.
Methods Enzym. 486, 343–368.
Obade, V. de P., Lal, R., 2016. A standardized soil quality index for diverse field conditions. Sci. Total
Environ. 541, 424–434. doi:10.1016/j.scitotenv.2015.09.096
R Core Team, 2016. R: A Language and Environment for Statistical Computing. R Foundation for
Statistical Computing, Vienna, Austria.
Reynolds, W.D., Topp, G.C., 2008. Soil water desorption and imbibition: tension and pressure
techniques. Soil Sampl. Methods Anal. 2nd Ed. Can. Soc. Soil Sci. Taylor Francis LLC Boca
Raton FL 981–997.
Robertson, G.P., Gross, K.L., Hamilton, S.K., Landis, D.A., Schmidt, T.M., Snapp, S.S., Swinton, S.M.,
2014. Farming for Ecosystem Services: An Ecological Approach to Production Agriculture.
BioScience 64, 404–415. doi:10.1093/biosci/biu037

177

Robertson, G.P., Wedin, D., Groffman, P.M., Blair, J.M., Holland, E.A., Nadelhoffer, K.J., Harris, D.,
1999. Soil carbon and nitrogen mineralization, nitrification, and soil respiration potentials, in:
Standard Soil Methods for Long-Term Ecological Research. Oxford University Press, New York,
pp. 258–71.
Romig, D.E., Garlynd, M.J., Harris, R.F., McSweeney, K., 1995. How farmers assess soil health and
quality. J. Soil Water Conserv. 50, 229–236.
Schipanski, M.E., MacDonald, G.K., Rosenzweig, S., Chappell, M.J., Bennett, E.M., Kerr, R.B., Blesh,
J., Crews, T., Drinkwater, L., Lundgren, J.G., others, 2016. Realizing Resilient Food Systems.
BioScience biw052.
Shennan, C., 2008. Biotic interactions, ecological knowledge and agriculture. Philos. Trans. R. Soc.
Lond. B Biol. Sci. 363, 717–739. doi:10.1098/rstb.2007.2180
Singer, J.W., Nusser, S.M., Alf, C.J., 2007. Are cover crops being used in the US corn belt? J. Soil Water
Conserv. 62, 353–358.
Sinsabaugh, R.., Reynolds, H., Long, T.., 2000. Rapid assay for amidohydrolase (urease) activity in
environmental samples. Soil Biol. Biochem. 32, 2095–2097. doi:10.1016/S0038-0717(00)001024
Stuart, D., Gillon, S., 2013. Scaling up to address new challenges to conservation on US farmland. Land
Use Policy, Themed Issue 1-Guest Editor Romy GreinerThemed Issue 2- Guest Editor Davide
Viaggi 31, 223–236. doi:10.1016/j.landusepol.2012.07.003
Walter, G., Wander, M., Bollero, G., 1997. A farmer-centered approach to developing information for soil
resource management: The Illinois Soil Quality Initiative. Am. J. Altern. Agric. 12, 64–72.
doi:10.1017/S0889189300007268
Weil, R.R., Islam, K.R., Stine, M.A., Gruver, J.B., Samson-Liebig, S.E., 2003. Estimating active carbon
for soil quality assessment: A simplified method for laboratory and field use. Am. J. Altern.
Agric. 18, 3–17. doi:10.1079/AJAA200228
Wolf, S.A., Nowak, P.J., 1995. Development of information intensive agrichemical management services
in wisconsin. Environ. Manage. 19, 371. doi:10.1007/BF02471979

178